Proceedings MVM2018 PDF

th
7 International Congress
Motor Vehicles & Motors 2018
ECOLOGY -
VEHICLE AND ROAD SAFETY
- EFFICIENCY
Proceedings
Motor Vehicles & Motors

Ministry of Education,
Science and Technological Development
October 4th - 5th, 2018

Kragujevac, Serbia
7th International Congress
ECOLOGY -
VEHICLE AND ROAD SAFETY
- EFFICIENCY
Proceedings

Kragujevac, Serbia
Publisher: Faculty of Engineering, University of Kragujevac
Serbia, 34000 Kragujevac, Street. Sestre Janjić broj 6
For Publisher: Prof. dr Dobrica Milosavljević - Dean
Editor: Dr Radivoje Pešić, prof.

Dr Jasna Glišović, prof.
Technical preparation: Dr Dragan Taranović, prof.

M.Sc. Nadica Stojanović, assistant
M.Sc. Ivan Grujić, assistant
Picture on the cover: Nemanja Lazarević
Print CD: Faculty of Engineering, University of Kragujevac, Kragujevac
ISBN 978-86-6335-055-7
Year of publication: 2018.
Number of copies printed: 200
CIP - Каталогизација у публикацији

Народна библиотека Србије, Београд
629.3(082)(0.034.2)
621.43(082)(0.034.2)
INTERNATIONAL Congress Motor Vehicles and Motors (7 ; 2018 ; Kragujevac)

Ecology - Vehicle and Road Safety - Efficiency [Електронски извор] :
proceedings / 7th International Congress Motor Vehicles & Motors 2018,
October 4th-5th, 2018 Kragujevac, Serbia ; [congress organizers Faculty of
Engineering, University of Kragujevac ... [et al.] ; editors Radivoje
Pešić, Jasna Glišović]. - Kragujevac : Faculty of Engineering, University,
2018 (Kragujevac : Faculty of Engineering, University). - 1 elektronski
optički disk (DVD) ; 12 cm
Sistemski zahtevi: Nisu navedeni. - Nasl. sa naslovne strane dokumenta. -

Tiraž 200. - Bibliografija uz svaki rad.
ISBN 978-86-6335-055-7
a) Моторна возила - Зборници b) Мотори са унутрашњим сагоревањем -
Зборници c) Електрична возила d) Хибридна електрична возила - Зборници
COBISS.SR-ID 268099340
Copyright © 2018 Faculty of Engineering, University of Kragujevac
Publishing of this book is supported by:
Ministry of Education, Science and Technological Development of the Republic of Serbia

SCIENTIFIC BOARD
President: Prof. Radivoje Pešić, Ph.D., UniKg, FE, Serbia
Secretaries: Assist. Nadica Stojanović, M.Sc. UniKg, FE, Serbia

Assist. Ivan Grujić, M.Sc. UniKg, FE, Serbia
Members: Prof. Mohamed Ali Emam, Ph.D., Helwan UniFE, Cairo, Egypt
Prof. Giovanni Belingardi, Ph.D., PoliTo, Italy
Prof. Murat Ciniviz, Ph.D., SelUni, FT, Turkey
Prof. Adrian Clenci Ph.D., University of Pitesti, Romania
Assist. Prof. Aleksandar Davinić, Ph.D., UniKg, FE, Serbia
Prof. Miroslav Demić, Ph.D., UniKG, FE, Serbia
Prof. Radan Durković, Ph.D., UniM, FME, Montenegro
Prof. Jovan Dorić, Ph.D., UniNS, FTS, Serbia
Prof. Ivan Filipović, Ph.D., UniSa, FME, BiH
Prof. Jasna Glišović, Ph.D., UniKg, FE, Serbia
Prof. Valentina Golubović Bugarski, Ph.D., UniBL, FME, RS, BiH
Prof. Dušan Gruden, Ph.D., TU Wien, Austria
Prof. Emil Hnatko, Ph.D., Croatian Academy of Engineering, Zagreb, Croatia
Prof. Rosen Hristov, Ph.D., TU – Varna, Bulgaria
Prof. Aleksandra Janković, Ph.D., UniKg, FE, Serbia
Prof. Hakan Kaleli, Ph.D., YTU, Istanbul, Turkey
Prof. Breda Kegl, Ph.D., UniMa, Slovenia
Prof. Božidar Krstić, Ph.D., UniKg, FE, Serbia
Prof. Dimitrios Koulocheris, Ph.D., NTU, Athens, Greece
Prof. Zbigniew Lozia, Ph.D., WUT, Warsaw, Poland
Prof. Jovanka Lukić, Ph.D., UniKg, FE, Serbia
Prof. Danijela Miloradović, Ph.D., UniKg, FE, Serbia
Prof. Slavko Muždeka, Ph.D., UniDef, MA, Serbia
Prof. Aleksandar Novikov, Ph.D., PFOSUni, Orel, Russia
Prof. Patrizio Nuccio, Ph.D., PoliTo, Italy
Prof. Snežana Petković, Ph.D., UniBL, FME, RS, BiH
Prof. Stojan Petrović, Ph.D., UniBg, FME, Serbia
Prof. Boran Pikula, Ph.D., UniSa, FME, BiH
Prof. Ralph Puetz, Ph.D., Landshut University, Germany
Prof. Dragoljub Radonjić, Ph.D., UniKg, FE, Serbia
Prof. Rajko Radonjić, Ph.D., UniKg, FE, Serbia
Assist. Prof. Dragan Ružić, Ph.D., UniNS, FTS, Serbia
Prof. Mihail Šatrov, Ph.D., MADI Moscow, Russia
Prof. Lazar Savin, Ph.D., UniNS, FA, Serbia
Prof. Hristo Stanchev, Ph.D., UniRu, FT, Ruse, Bulgaria
Prof. Dragoslava Stojiljkovic, Ph.D., UniBg, FME, Serbia
Prof. Dragan Taranović, Ph.D., UniKg, FE, Serbia
Academician Dušan Teodorović, Ph.D., UniBg, FTTE, Serbia
Prof. Miroljub Tomić, Ph.D., UniBg, FME, Serbia
Prof. Stevan Veinović, Ph.D., UniKg, FE, Serbia
ORGANIZATIONAL BOARD
President: Prof. Jasna Glišović, Ph.D., UniKg, FE, Serbia
Secretaries: Assist. Nadica Stojanović, M.Sc. UniKg, FE, Serbia

Assist. Ivan Grujić, M.Sc. UniKg, FE, Serbia
CONGRESS ORGANIZERS
 Faculty of Engineering, University of Kragujevac

 Department for Motor Vehicles and Motors, FE in Kragujevac
 International Journal “Mobility & Vehicle Mechanics”
 University of Kragujevac
CONGRESS PATRONS
 Government of the Republic of Serbia, Ministry of Education, Science and Technological

Development,
 Centre for technical proper function of vehicles, FME Kragujevac
 Faculty of Engineering, University of Kragujevac
PREDGOVOR
Tradicionalno organizovanje Međunarodnog Kongresa „Motorna Vozila i Motori - MVM“ je
postalo brend Fakulteta inženjerskih nauka Univerziteta u Kragujevcu. I ove godine nije izostao
veći broj radova iz zemlje i inostranstva, kao ni tradicionalna podrška Ministarstva za prosvetu,
nauku i tehnološki razvoj Republike Srbije, grada Kragujevca i naših prijatelja i saradnika iz
proizvodnih ili uslužnih delatnosti vezanih za automobile.
Moto Kongresa „MVM 2018“ je „Ekologija – Vozilo i bezbednost saobraćaja - Efikasnost“.

Radovi istraživača i naučnih radnika koji su publikovani na Kongresu „MVM 2018“ i ovoga puta
su pokazali značaj istraživanja u oblasti motornih vozila, motora i drumskog saobraćaja.
Motorna vozila su od svog nastanka privlačila veliku pažnju kako inženjera tako i poslovnih ljudi,
a u njihov razvoj su utkana naučna dostignuća iz mnogih oblasti. Ona imaju i korisnike širokog
spektra koji svojim zahtevima nameću i nove klase i nove kategorije automobila. U skladu sa
trendovima razvoja tehnike i tehnologije, kao i novim zahtevima korisnika, težište proizvođača
se pomera ka određenim vrstama ili klasama automobila, ali, automobil generalno, kao
aktuelnost, izazov i potreba, ne nestaje sa scene.
Jasno je da je automobil, kao multi-disciplinarni objekat razvoja i istraživanja sa jedne strane, i

kao potreba ili prestiž sa druge, meta za razvoj i ulaganja. Osim toga, automobil je sredstvo koje
sve brže spaja udaljene destinacije i koje je sa sve većim manevarskim sposobnostima u gradu,
pa kao objekat planiranja i regulisanja saobraćaja, zauzima, bez sumnje, najznačajnije mesto.
Stoga i postoji tendencija da se okviri prijavljenih tema radova sve više šire, da se osim
mehanike vozila i procesa u motoru, odgovarajući prostor dâ i oblastima kao što su ekologija,
alternativna goriva, regulisanje, autoelektronika, itd. Radovi sa rezultatima istraživanja u oblasti
elektro- i hibridnih vozila su sve prisutniji.
U Kragujevcu, 4. - 5. oktobar, 2018. godine Naučni i Organizacioni odbor

Kongresa „MVM 2018“
FOREWORD
Traditional organizing of International Congress "Motor Vehicles & Motors - MVM" has become
the brand of the Faculty of Engineering University of Kragujevac. This year also, there is a
considerable number of papers from the country and from abroad, as well as the traditional
support from the Ministry of Education, Science and Technological Development of the Republic
of Serbia, the city of Kragujevac and from our friends and associates from manufacturing or
servicing activities related to automobiles.
The motto of the Congress "MVM 2018" is "Ecology – Vehicle and Road Safety – Efficiency".
The papers of the researchers and the academics published at the Congress "MVM 2018", have
shown, this time also, the importance of the research in the field of motor vehicles, engines and
road traffic.
Since their occurrence, motor vehicles have attracted a considerable attention both from the
engineers and from the business people, while scientific achievements from various scientific
fields have been implemented in the process of vehicles development. Motor vehicles have
wide spectrum users, who impose new car classes and new car categories with their demands.
In accordance with trends in technical and technological developments, as well as new user
demands, the focus of the car manufacturers is displaced towards the specific car types or car
classes, but the car in general, as actuality, challenge and need, is still present on the "stage".
It is quite clear that the car, as the multidisciplinary object of development and research on the
one hand, and as the object of need or prestige on the other, is actually the target for
development and investment. Besides, the car is the means for faster and faster connection of
distant destinations, and with more and more improved mobility performances within the city, so
it has, without a doubt, the most significant place as the object of traffic planning and regulation.
Therefore, there is the tendency of broadening the papers topics from the vehicles mechanics
and the processes within the engine, in order to give appropriate space to the fields such as
ecology, alternative fuels, regulations, auto-electronics, etc. The scientific papers with the
research results from the field of electric and hybrid vehicles are all the more present.
In Kragujevac, October 4th - 5th, 2018 Scientific and Organizational Board of

the Congress “MVM 2018“
CONTENT
INTRODUCTORY LECTURES
WILL ELECTRIC MOTOR SUBSTITUTE INTERNAL

MVM2018-IL1 Dušan Gruden 3
COMBUSTION ENGINE?
ELECTRIC BUS VERSUS BUS WITH COMBUSTION
MVM2018-IL2 Ralph Pütz 13
ENGINE - AN ECOLOGICAL-ECONOMIC APPROACH
MVM2018-IL3 Miroljub Adžić FUEL CELLS AND ELECTRIC VEHICLES 25
TURBULENT DEVELOPMENT OF BODY CONCEPTS,
MVM2018-IL4 Josip Vlahović MATERIALS AND JOINING PROCESSES IN THE LAST 37
TWO DECADES
CURRENT STATE AND DEVELOPMENT TRENDS IN
MVM2018-IL5 Dobrivoje Ninković 47
THE FIELD OF LARGE DIESEL AND GAS ENGINES
DEVELOPMENT THROUGH SIMULATION OF A
MVM2018-IL6 Patrizio Nuccio TURBOCHARGED 2-STROKE G.D.I. ENGINE 65
FOCUSED ON A RANGE-EXTENDER APPLICATION
Giovanni Belingardi
OPTIMIZATION-BASED CONTROLLERS FOR HYBRID
MVM2018-IL7 Pier Giuseppe Anselma 71
ELECTRIC VEHICLES
Miroslav Demic
TRANSPORTATION SCIENCE: CHALLENGES AND
MVM2018-IL8 Dušan Teodorović 81
FUTURE RESEARCH
SECTION A
Power Train Technology
Sunny Narayan
Ivan Grujic ANALYSIS OF PRESSURE PLATE OF AN
MVM2018_001 95
Nadica Stojanovic AUTOMOTIVE CLUTCH ASSEMBLY
Vipul Gupta
Sreten Perić
EXPERIMENTAL RESEARCH OF THE
Bogdan Nedić
MVM2018_004 BIODEGRADABLE UNIVERSAL TRACTOR OIL BASED 101
Mile Stoiljković
ON THE VEGETABLE OILS
Radovan Radovanović
Slobodan Mišanović
Dragan Taranović THE IMPACT OF THE SYSTEM FOR HEATING AND
MVM2018_005 Jovanka Lukić AIR-CONDITIONING ON THE ENERGY EFFICIENCY 115
Radivoje Pešić OF THE ELECTRIC DRIVE BUS (E-BUS)
Jasna Glišović
Mikhail G. Shatrov
Leonid N. Golubkov ANALYSIS OF REALIZATION OF STEPPED FUEL
MVM2018_011 Andrey U. Dunin INJECTION WITHOUT CHANGING DESIGN OF 139
Pavel V. Dushkin COMMON RAIL INJECTOR
Andrey L. Yakovenko
Mikhail G. Shatrov
Vladimir V. Sinyavski EXPERIENCE OF DIESEL ENGINES CONVERSION
MVM2018_012 Ivan G. Shishlov FOR OPERATION ON NATURAL GAS OBTAINED IN 147
Andrey Y. Dunin MADI
Andrey V. Vakulenko
I
INVESTIGATION OF DIESEL ENGINE OPERATED
Hristo Stanchev
MVM2018_014 WITH DIESEL FUEL-BUTANOL BLENDS AND CETANE 155
Krasimir Markov
IMPROVER
Igor V. Alekseev METHOD AND SOME RESULTS OF INVESTIGATION
MVM2018_016 Mikhail G. Shatrov OF STRUCTURE-BORNE NOISE OF DIESEL ENGINE 165
Andrey L. Yakovenko AT ACCELERATION MODE
Branimir Milosavljević
Radivoje Pešić
STATISTICAL ANALYSIS OF INFLUENTIAL
Žarko Đorđević
MVM2018_017 PARAMETERS ON THE PROBABILITY THAT CARS 171
Nikola Kostić
MEET THE CO EMISSION TEST
Jelena Erić Obućina
Saša Babić
INFLUENCE OF DETERMINING THE PERIOD OF
Almir Blažević
IGNITION DELAY ON THE HEAT RELEASE
MVM2018_023 Dževad Bibić 183
CHARACTERISTIC OF DIESEL ENGINES FUELLED
Ivan Filipović
WITH DIESEL AND BIODIESEL
Ivan Blagojević
MVM2018_029 HYDROGEN AS A VEHICLE FUEL 193
Saša Mitić
Ivana Terzić
Vanja Šušteršič POSSIBILITIES FOR THE USE OF BIOFUELS IN
MVM2018_044 203
Aleksandar Nešović EUROPE AND IN SERBIA
Mladen Josijević
Ivan Grujić
Jovan Dorić
Nadica Stojanović NUMERICAL MODELING OF IC ENGINE
MVM2018_062 213
Aleksandar Davinić COMBUSTION PROCESS
Radivoje Pešić
Sunny Narayan
Nenad Raspopović
Jovan Dorić ANALYSIS OF HYBRID TURBOCHARGER IN MOTOR
MVM2018_063 219
Ivan Grujić VEHICLE IC ENGINE
Nebojša Nikolić
Saša Milojević
Radivoje Pešić
Aleksandar Davinić
INFLUENCE OF VARIABLE COMPRESSION RATIO ON
Dragan Taranović
MVM2018_064 EMISSION AND VIBE FUNCTION PARAMETERS OF 227
Snežana Petković
EXPERIMENTAL ENGINE
Emil Hnatko
Radmilo Stefanović
Stevan Veinović
Radivoje Pešić
EXPERIENCES RELATED TO THE INSPECTION OF
Saša Milojević
MVM2018_065 MOTOR VEHICLES RUNNING ON NATURAL GAS IN 245
Dragan Taranović
THE REPUBLIC OF SERBIA
Milan Stanojević
SECTION B
Vehicle Design and Manufacturing
THE ROLE OF INDUSTRIAL AND SERVICE ROBOTS

Isak Karabegović
MVM2018_002 IN FOURTH INDUSTRIAL REVOLUTION WITH FOCUS 257
Ermin Husak
ON CHINA
Nenad Kostić
AUTOMATED SELECTION OF ANALYTICALLY
Nenad Petrović
MVM2018_010 CALCULATED GEARBOX CONCEPTS ACCORDING 265
Vesna Marjanović
TO DIMENSIONAL CRITERIA
Nenad Marjanović
II
Saša Mitić END-OF-LIFE VEHICLE DISPOSAL AND IT’S
MVM2018_027 271
Ivan Blagojević INFLUENCE TO THE ENVIRONMENT
Dimitrios V. Koulocheris ALTERNATIVE DESIGN FOR A SEMI – TRAILER TANK
MVM2018_031 279
Clio G. Vossou VEHICLE
Ralph Pütz EXPLOITATION OF THE GOLF BALL EFFECT FOR
MVM2018_033 Matthias Kampfl FURTHER AERODYNAMIC OPTIMISATION OF RACE 293
Thomas Hechfellner CARS
Nadica Stojanovic
Jasna Glisovic
Ivan Grujic EXPERIMENTAL AND NUMERICAL MODAL ANALYSIS
MVM2018_034 313
Sunny Narayan OF BRAKE SQUEAL NOISE
Sasa Vasiljevic
Bojana Boskovic
Sandra Veličković
Blaža Stojanović
APPLICATION OF NANOCOMPOSITES IN THE
MVM2018_036 Lozica Ivanović 323
AUTOMOTIVE INDUSTRY
Slavica Miladinović
Saša Milojević
Aleksandar Poznić SECONDARY CONTRIBUTION TO THE ANALYSIS OF
MVM2018_040 333
Boris Stojić MAGNETORHEOLOGICAL BRAKE
Sonja Kostić THE INFLUENCE OF THE INTERNAL RADIAL
MVM2018_045 Zorica Đorđević CLEARANCE ON THE LOAD DISTRIBUTION ON THE 341
Milosav Đorđević ROLLER BODY OF A SINGLE RADIAL BALL BEARING
Aleksandar Radaković
Dragan Milosavljević
ELASTIC WAVES PROPAGATION IN COMPOSITE
MVM2018_046 Gordana Bogdanović 349
LAMINATES USED IN THE AUTOMOTIVE INDUSTRY
Dragan Čukanović
Petar Knežević
Dragan Čukanović THERMAL ANALYSIS OF FUNCTIONALLY GRADED
MVM2018_047 Aleksandar Radaković MATERIAL AS THE SPECIFIC CLASS OF ADVANCED 357
Gordana Bogdanović COMPOSITES USED IN AUTOMOTIVE INDUSTRY
AN ANALYTICAL MODEL OF AIR-CONDITIONED
MVM2018_056 Dragan Ružić 369
MOTOR VEHICLE CABIN
Boris Stojić
Dragan Ružić
LIGHT HYBRID-ELECTRIC VEHICLE AS
MVM2018_059 Zoran Ivanović 377
EDUCATIONAL AND RESEARCH PLATFORM
Evgenije Adžić
Nikola Ilanković
Saša Milojević EXPERIMENTAL DETERMINATION OF
Dragan Džunić TRIBOLOGICAL CHARACTERISTICS OF COMPOSITE
MVM2018_066 Dragan Taranović MATERIALS IN USE FOR PARTS IN ALUMINUM AIR 383
Radivoje Pešić COMPRESSOR
Slobodan Mitrović (PISTON AND CYLINDER)
Rodoljub Vujanac
DESIGN OF FORKLIFT SPEED LIMIT CONTROL
MVM2018_067 Nenad Miloradovic 393
SYSTEM FOR WAREHOUSES
Milan Djordjevic
SECTION C
Vehicle Dynamics and Intelligent Control Systems
Lazar Stojnić
DEVELOPMENT AND ANALYSIS OF A 4X4 TRUCK
MVM2018_037 Aleksandar Grkić 401
MULTIBODY SIMULATION MODEL
Slavko Muždeka
III
Boris Stojić EXAMINATION OF SENSITIVITY OF TRACTOR TIRE
MVM2018_042 Aleksandar Poznić FILTERING BEHAVIOR TO CONTACT LENGTH 411
Nebojša Nikolić CHANGE
OSCILLATIONS OF LIGHT WHEELED VEHICLE
Nina Živanović
MVM2018_051 UNDER THE INFLUENCE OF INTEGRATED 417
Gordana Bogdanović
WEAPON’S RECOIL FORCE
Satoru Furugori
AUTOMOTIVE SAFETY CONTROL SYSTEM BASED
MVM2018_052 Takeo Kato 427
ON TIMEAXIS DESIGN
Yoshiyuki Matsuoka
SECTION D
Driver/Vehicle Interface, Information and Assistance Systems
Valentina Golubović-
Bugarski VIBRATION COMFORT OF THE VEHICLE
MVM2018_003 Snežana Petković EXPRESSED BY SEAT EFFECTIVE AMPLITUDE 437
Željko Đurić TRANSMISSIBILITY
Goran Jotić
Janez Kopač TOUCHSCREEN CAR DASHBOARDS AS SERIOUS
MVM2018_015 447
Franci Pušavec DANGER FOR CAUSING TRAFFIC ACCIDENTS
Saša Babić
Jovanka Lukić
Nikola Kostić NOISE EMISSION WITH ASPECT OF VEHICLES
MVM2018_019 451
Žarko Đorđević FLEET STRUCTURE – THE CASE OF SERBIA
Jelena Obućina-Erić
Branimir Milosavljević
Jelena Trifunović
METHODOLOGY FOR DETERMINING THE AMOUNT
Aleksandar Manojlović
MVM2018_021 OF ECOLOGICAL FEE FOR THE USE OF 457
Snežana Kaplanović
PASSENGER CARS
Nebojša Bojović
Tijana Ivanišević
Dragan Taranović
ROLE OF VEHICLES ON ALTERNATIVE PROPULSION
MVM2018_035 Radivoje Pešić 465
SYSTEMS IN LOGISTICS AND TRANSPORT
Sreten Simović
Vedran Vukšić
Slavica Mačužić APPLICATION OF DIGITAL HUMAN MODEL IN ORDER
MVM2018_053 473
Jovanka Lukić TO DETERMINE THE HUMAN BODY DISCOMFORT
Slavica Mačužić THREE DIMESIONAL ANALYSES OF SEAT BELT AND
MVM2018_057 479
Igor Saveljić DRIVER IN CASE OF SUDDEN BRAKING
SECTION E
Road Traffic Safety
Slobodan Mišanović
Predrag Živanović
Dušan Savković TWO-YEAR SUCCESSFUL EXPLOITATION OF THE
MVM2018_006 487
Pavle Krstić ELECTRIC BUSES IN BELGRADE
Slobodan Ivanov
Slobodan Stević
Milja Simeunović
Zoran Papić THE ROLE OF APPLYING ELECTRIC VEHICLES IN
MVM2018_007 493
Milan Simeunović SUSTAINABLE URBAN MOBILITY
Pavle Pitka
IV
Goran Štetin
Zoran Papić ANALYSIS OF VEHICLE CRASH DEFORMATIONS
MVM2018_008 499
Nenad Saulić USING FE MODELS
Andrijana Jović
Vase Jordanoska COMPARATIVE ANALYSIS OF CAR FOLLOWING
MVM2018_013 Igor Gjurkov MODELS BASED ON DRIVING STRATEGIES USING 509
Darko Danev SIMULATION APPROACH
Jelena Eric Obucina
Nikola Kostic
SAFETY ASPECTS OF THE HYDRAULIC SERVO
MVM2018_020 Zarko Djordjevic 517
STEERING GEAR IN THE VEHICLE SYSTEM
Sasa Babic
Branimir Milosavljevic
Saša Vasiljević
Nadica Stojanović MEASUREMENT OF VEHICLE BRAKE COEFFICIENT
MVM2018_025 Ivan Grujić AS VEHICLE MAINTENANCE PARAMETER ON 523
Bojana Bošković TECHNICAL INSPECTION LINES
Milan Mladenović
Bojana Bošković
Milan Mladenović
Saša Vasiljević PERSONS WITH DISABILITIES AS PARTICIPANTS IN
MVM2018_026 531
Marina Teofilović THE TRAFFIC
Nadica Stojanović
Ivan Grujić
Clio G. Vossou A COMPUTATIONAL MODEL FOR THE
MVM2018_032 541
Dimitrios V. Koulocheris RECONSTRUCTION OF VEHICLE COLLISIONS
Žarko Đorđević
Nikola Kostić CONTRIBUTION TO THE RESEARCH OF
MVM2018_039 Jelena Erić Obućina PARAMETERS, EFFECTS AND METHODS OF ECO 553
Branimir Milosavljević DRIVING MEASUREMENTS
Saša Babić
Goran Pejić
Marko Bunjevac IMPACT OF INTRODUCTION OF LOW EMISSION
MVM2018_041 561
Matija Pečet ZONES IN THE CITY OF ZAGREB
Zoran Lulić
Kjosevski Stevan AN INDICATORS BASED APPROACH TOWARDS
MVM2018_058 Kochov Atanas DECISION MAKING AND POLICY MAKING 573
Kostikj Aleksandar REGARDING INTRODUCING ELECTRIC VEHICLES
INTEGRATED CONTROL OF DIVERGING DIAMOND
Aleksandar Jovanović
MVM2018_061 INTERCHANGE AND RAMP METERING: FUZZY 579
Katarina Kukić
LOGIC APPROACH
V
VI
Introductory
Lectures
1
2
International Congress
Kragujevac, Serbia
MVM2018-IL1
Dušan Gruden1
WILL ELECTRIC MOTOR SUBSTITUTE INTERNAL

COMBUSTION ENGINE?
ABSTRACT: The first transportation means driven by a DC electric motor was a six-passenger boat. Ground
vehicles with electric drive have also a longer history than vehicles with IC engines. At the beginning of the 20th
century, there were more cars driven by electric motors then those driven by IC engines. During the last 100 years,
there have been many attempts to replace the IC engine with supposedly better power units. Experiments with electric
vehicles were repeatedly performed. Around 1.1 million electric vehicles or about 0.12% of 947 million of passenger
vehicles is used in the world traffic today. In terms of CO2 emissions, electric cars have an environmental advantage
after 2 to 4 years of use, only if the electricity for their charging is created by regenerative sources: water, wind, solar
energy, high and low tides or nuclear energy (as in France). In China and Germany, where the largest part of the
electricity is produced in the coal or gas power plants, the CO2 emissions created by electric cars are significantly
higher than those created by cars with IC engines are. It is the opinion of many experts, that vehicles with combined
drive via an electric motor and a single, relatively small and optimized IC engine (so-called hybrid drive), present the
best solution passenger cars in the future, especially in terms of reduction of toxic emission in the cities, as well as
in reduction of fossil fuels consumption. The conclusion of this analysis of electric drives for motor vehicles, can only
state the same as it repeats for over half a century: "The piston IC engines will be substituted by the better powertrains
during the following 10 to 15 years and the existing oil reserves will be spent in the next 30 to 40 years!".
KEYWORDS: fuel cells, hybrid vehicles, electric vehicles, ecological fuels, IC engine
INTRODUCTION
In the world, and particularly in Europe, an intense discussion is being lead today about the need to forbid the use of
internal combustion engine (IC engine) because it, allegedly, has only negative consequences on human
environment. Representatives of the Green party in Germany are seeking a ban on the IC engine from the year 2030
in their program, while their counterparts from the Social-democratic party demand this ban from 2040.
French writer and philosopher Andre Marlo wrote: “Who wants to predict the future, must first go through the past”!
SHORT HISTORY OF ELECTRIC VEHICLES

The first transportation means driven by a DC electric motor was a six-passenger boat, patented by Morris Herman
von Jacobi in 1838, trying his own boat on the Neva in Saint Petersburg, Russia [1].
Ground vehicles with electric drive have also a longer history than vehicles with IC engines. As early as 1881, five
years before the first motor vehicles arrived, which are attributed to Daimler and Benz in Germany, an Englishman,
Gustav Truve, have built in the electric drive on a tricycle. In 1898, Gaston Šezelup-Labat laid the first car speed
record. With a car with electric drive, he achieved a speed of 63.12 km/h.
In 1890, the carriage maker from Vienna, Ludwig Lonner, gave an order to the young engineer Ferdinand Porsche
to develop a cab, which would be powered by the electric motor. At the world exhibition in Paris in 1900, Porsche
1
Dušan Gruden, Prof. Dr. techn. Dr.h.c., Honorary Professor of TU Wien, former director of Engine development
and Sector for ecology end energy at Porsche, Stuttgart, gruden@yudit.de
3
and Loner presented their patent: vehicle driven by electric motors built-in on its wheels, so that each wheel had its
own drive, Figure 1.
Figure 1 Loner-Porsche electric vehicle at the World exhibition in Paris in 1900
At the beginning of the 20th century, there were more cars driven by electric motors then those driven by IC engines.
In that time, the forecasts for the future of the cars with electric drives were significantly more optimistic than for cars
with IC engines. Back in 1910, over 30% of cars in the traffic in the US were electric cars. Accumulation (storage) of
electricity in these cars was carried out exclusively in so-called lead (Pb) accumulators – batteries. Their great weight,
for any larger energy capacity of battery, and the long time it takes to charge the battery, were the biggest
shortcomings of those cars. At that time, passenger vehicles with electric motors could develop speeds of around
100 km/h and had a range of about 70 to 100 km.
The rapid development of IC engines, especially after the invention of the electric starters for engines, where the lead
electric car battery as a power source for the start-up have been applied until today, has begun to suppress electrical
drives, first of all in passenger vehicles. The culmination of application of electric drives, especially in the field of
freight transport, was reached between 1920 and 1930.
To minimize the main disadvantage of electric vehicles - their small driving radius, so called hybrid drives or vehicle
drives with two engines - electric motor, consuming the electric power from lead batteries, and IC engine, which
drives a generator to produce electricity for charging the battery, were offered at the beginning of the 20th century,
Figure 2. The first hybrid car of that time was a Ford (1912).
Figure 2 The first hybrid vehicle
High production costs for these vehicles prevented their wider application.
The electricity, similar to the heat, has a feature that it has to be consumed the second when produced, because its
storage is not simple, but very difficult and complicated.
4
In vehicles which are directly connected to a source of electricity, like in rail transportation, trams and trolley buses,
electrical drives are established for decades, with all their advantages.
STATE OF THE ART IN DEVELOPMENT OF VEHICLE POWER UNITS

Since 1920s, until today, piston internal combustion engines (IC engines) have become the undisputed drive unit,
not only for passenger cars and freight vehicles, but also for a wide range of other different applications. The IC
engine owes this wide distribution to its characteristic that it uses one of the biggest and the cheapest energy sources
of the planet Earth, namely oil, with greater efficiency than other power units.
During the last 100 years, there have been many attempts to replace the IC engine with supposedly better power
units. Experiments with electric vehicles were repeatedly performed. From the 1980s, the attempts in Mendrizi
(Switzerland), Berlin (FR Germany), La Rošelu (France) and Amsterdam (the Netherlands) were well known. In the
early 1990s, extensive testing of electric cars were conducted on Reagan island in the Baltic Sea (FR Germany),
with always the same conclusions: main problems are the storage, accumulation of electric power in batteries, small
driving radius of vehicles and their high production costs. Although different batteries like NiCd (Nickel-Cadmium),
NaCl (Sodium chloride), NaS (Sodium-Sulfur) and NaNi (Sodium-Nickel) were tested at that time, the lead (Pb)
battery has always been the winner, despite all its faults. Its energy capacity is around 70 Wh/kg. It takes a lead
battery of 3,000 l of volume and 6.5 t of weight for the same amount of energy as in a reservoir for gasoline of 70 l
(gasoline 12,000 Wh/kg). The modern lithium-ion batteries (Li-Ion) have an energy capacity of approximately 100 to
250 Wh/kg (250 to 620 Wh/l), that is, in order to replace the energy amount in 70 l of gasoline, they need to have the
volume between 400 and 1,000 l and the weight of 0.4 to 2.0 t, Figure 3.
Figure 3 Energy needed for 500 km range of driving
Whether electric vehicles would succeed in this new attempt to break through to a large serial production and replace
piston IC engines depends on the further development of accumulator - battery, its production cost and energy
charging speed.
Until now, passenger vehicles with electric drives have been produced in relatively small numbers. Around 1.1 million
electric vehicles or about 0.12% of 947 million of passenger vehicles is used in the world traffic today. The share of
these vehicles in the traffic in the USA is about 0.76% and in Germany about 0.38%. These numbers suggest that,
until now, the use of the electric vehicles was limited to a small area of application, first in small distance traffic, in
the city traffic and as a second or a third vehicle in the household. In the future, vehicles with electric drives would
have to satisfy the demands, as complete family vehicles, able to cover larger distances.
The key problem of all electric cars is their driving radius. For fast covering of larger distance, not only a distance
travelled with one battery charge is important, but also the time it takes to charge the empty battery. Today's standard
battery charging is carried out with relatively low or medium-sized power, with the power supply voltage of 110 V or
220 V, and it lasts several hours, similarly to charging Li-ion batteries in mobile phones or laptops. These vehicles
should always be charged, when they are not used for a long time, for example at the parking lots or at homes, in
garages. To increase the speed of charging to approximate time needed to refuel the vehicles with the IC engines,
systems with high power and voltage of 800 V to 1,000 V are being developed, which causes a substantial increase
in investment for the design of infrastructure.
All consumers of energy in the electric vehicle are supplied with energy contained in the battery. Thus, the driving
radius of vehicles with electric drive is greatly influenced by the weather conditions in which the ride takes place (day,
night, rain, snow...). Vehicle’s air-conditioning, especially the heating of electric vehicles, can significantly reduce
their driving radius, Figure 4.
5
Figure 4 Driving radius of electric vehicles with air-conditioning and heating turned on
The typical battery of the electric vehicle consists of more than 100 cells. In Tesla S car, the battery consists of 16
modules with 7,104 Li-ion cells. It takes only a single cell not to work and the whole system has to be changed.
All batteries have certain temperature areas in which they can work optimally. These operating temperatures affect
the capacity of the battery (and hence the driving radius), as well as battery life. That is why the so-called thermal
management of the batteries is very important. For Li-ion batteries, the optimum operating temperatures lie between
20oC and 40oC.
It is often emphasized, that the ecological advantage of electric cars is that they do not emit toxic emissions during
their operation. That is true if only those emissions that arise on the path from the battery to the car's wheels are
considered (so-called “tank-to-wheel” analysis). However, if one looks at the overall ecological balance of electrical
cars, from its construction until the end of his lifetime (the "Life cycle assessment"), then the environmental image of
electric cars looks different.
While around 5 t of CO2 emissions arises during the production of one middle class motor vehicle, this emission
doubles during the production of an electric car, Figure 5 [4].
Figure 5 LCA (Life Cycle Assessment) of CO2 emission (ICE – Internal combustion engine, CNG – compressed
natural gas, E-Fuels – ecological fuels, BEV – electric vehicles)
In terms of CO2 emissions, electric cars have an environmental advantage after 2 to 4 years of use, only if the
electricity for their charging is created by regenerative sources: water, wind, solar energy, high and low tides or
nuclear energy (as in France). In China and Germany, where the largest part of the electricity is produced in the coal
or gas power plants, the CO2 emissions created by electric cars are significantly higher than those created by cars
with IC engines are, Figure 6.
6
Figure 6 CO2 emissions of electric car depending on type of electricity production
Today's electric cars with Li-ion batteries cannot meet the European regulations on the recycling of cars. At the end
of their lifetime, the passenger cars manufactured by 2015 must prove that 95% of their weight can be re-used or
recycled. Only 5% of the vehicle’s weight may be deposited on the landfill. With modern electric cars, the weight of
Li-ion batteries is approximately 25% to 30% of the vehicle’s weight. Until today, solutions for economic reasonable
recycling of these batteries have not been found.
Also in the area of public and freight transport, there are visible efforts to replace IC motors with electric drives. This
drive long ago dominates where there is a direct contact between the transport means and the electricity sources
(railways, trams, trolley buses). In some countries, the city transport also contains buses with electric drives, Figure
7. The batteries for these vehicles are charged at terminals over trolleys similar to those on trams or trolley buses.
Figure 7 Electric bus in Belgrade traffic
German industry has suspended work on these projects, because the weight of the batteries is to large, for slightly
larger distances that the vehicle would have to pass.
Hybrid vehicles
One of the biggest disadvantages of electric cars - their relatively small driving radius compared to the cars powered
by IC engines, has led to re-development of so-called hybrid cars. With these cars, the features and benefits of the
electric drives combine with the positive features of the IC engines. It is the opinion of many experts, that vehicles
with combined drive via an electric motor and a single, relatively small and optimized IC engine (so-called hybrid
drive), present the best solution passenger cars in the future, especially in terms of reduction of toxic emission in the
cities, as well as in reduction of fossil fuels consumption.
7
For years, the hybrid vehicles have become the subjects of huge series production. Out of all the different possible
combinations between the electric motors and the IC engines, it looks like a so-called “plug-in hybrid” is the most
attractive today. Statistically, one car travels a distance of 30 km to 50 km daily. The goal of the development of the
"plug-in” hybrids is to cover the daily distance in everyday traffic with electric drive, because the battery is
dimensioned in such a way, that the vehicle can cross this distance. The battery of these vehicles is charged over
the normal power supply network (plug-in), at home, or in the workplace. A relatively small IC engine is used in the
case when the battery runs out of power, in order to run the generator for the production of electric power and enable
the travel over the larger distance. About the half of the so-called electric vehicles in the world are "plug-in" hybrids.
How big an advantage the hybrid drive offers compared to pure electric or IC engine drive, can be shown on the
example of the Porsche 919 hybrid vehicle [8]. This vehicle have been winning the race on the famous 24 hours of
Le Mans race for three years in a row, Figure 8.
Figure 8 The winner of “The 24 hours of Le Mans” race in 2015, 2016 and 2017: Porsche 919 Hybrid
In the area of freight or commercial vehicles and buses, there are also extensive tests with hybrid drives (MAN, Volvo,
Iveco).
Electric drives in water traffic

A historical example of the first vehicle with electric drive from 1838 shows that the electric drive is not connected
just to the road vehicles on wheels. For more than 100 years, all submarines have electric (hybrid) drives. After
successful deployment of electrical drives on the road vehicles, there are more and more attempts to apply this drive
on vehicles in water transport. On many Alpine lakes in Austria, Slovenia, Switzerland and Germany, where it is
prohibited to use the boats and the ships with IC engines, vessels with electric drives and with heavy lead batteries
are used for decades.
New attempts to use electric drives for the ships are associated with efforts to reduce the toxic emissions from
exhaust gases of ship engines, especially in large ports, where the emission significantly exceeds the exhaust
emission of passenger vehicles. Application of two-stroke diesel engines and the bad quality of the relatively cheap
fuels for these engines (heavy oil fractions, mazut) lead to high emissions of sulphur dioxides (SO2), nitrogen oxides
(NOx) and particulate matter (PM) with these engines. The most attractive electrical drive is for ferries, where the
distance between the two ports is not large. The most famous ferry with electric drive is Norwegian ship "Ampera".
Its battery pack weighs 10 t [5]. Experts, however, agree that the road to all ships having electric drives is still very
long.
Electric drives in air traffic

The airplane industry is working intensively on the development of quieter, more economical and more
environmentally friendly planes. That is why the interest of the industry for the electric powered airplanes is getting
bigger and bigger [6]. The best chances are given to the hybrid drive, as the combination of the jet engine and the
electric engine. Turbine of the jet engine charges the battery through the generator, during a steady flight at the travel
height. At the landing or the take-off of the plane, additional power of the electric engine is used, which reduces
airplane noise. The biggest problem here is also the weight of the necessary batteries, because the energy capacity
of the fossil fuels (kerosene) is 60 to 100 times greater than the energy capacity of the best batteries today. It is not
8
expected for smaller planes with hybrid drives to take-off and fly in the next 15 years. Bigger airplanes are going to
take considerably more time. The hybrid aircrafts are expected to reduce the fuel consumption by 30% until 2035
and by 75% by 2050.
A VIEW INTO THE FUTURE

Less than 2% of today’s 1.38 billion vehicles (947 millions of passenger vehicles and 335 millions of commercial
vehicles) do not have gasoline or diesel engines, Figure 9. Most of the cars that do not use these two fuels have IC
engines that use natural gas (methane) or liquid petrol gas (propane) or so-called bio-fuels (ethanol, vegetable oils).
Although the number of the electric vehicles and vehicles with hybrid drives does not exceed 0.12% of the total
number of vehicles in the world today, these vehicles are mentioned more in public discussions than all others, as if
only they are expected to solve all the problems in the transport.
Figure 9 Modern electric cars (source: ADAC)
Whether the new attempt at introducing electric drives is to be successful, unfortunately, does not only depend on
the economic parameters that accompany this vehicle, but also on the political decisions of the governments. The
German Government has reached the decision that the number of registered passenger cars with electric drives
must be at least 1 million until 2020. Implementation of this decision would mean that, as of today, more than 20,000
electric or hybrid cars must be sold per month, which is far from the actual sales, which is 10 to 15 times lower. What
is interesting is that only about 15% of the registered electrical cars in Germany is in private ownership. The largest
part is found in larger companies (e.g. post offices) or in the shop windows of the shopping centres.
Fuel cells
The biggest drawback of electric vehicles - a small energy capacity of the batteries, has found the solution in part
through the so-called hybrid drive, where optimised IC engine is applied to drive electric generators. As the IC engine
also produces unwanted emissions of carbon monoxide (CO), carbon dioxide (CO2), nitrogen oxides (NOx), unburnt
hydrocarbons (HC) and particulate matter (PM), during its operation, there are proposals to replace the IC engine
with so-called fuel cell. A fuel cell, whose working principle was patented by Englishman William Robert Grove back
in 1839, operates on reverse electrolysis principle, Figure 10. While, in the process of electrolysis, water dissolves
into hydrogen (H2) and oxygen (O2) with the help of electricity, in the fuel cell, hydrogen combines with oxygen, so
the electric power, the heat and the water vapour (H2O) are produced. The electric power is used either directly to
drive the electric engine of the vehicle or to charge the batteries.
9
Figure 10 Energy from air and hydrogen: the principle of the fuel cell
Until now, hydrogen (H2) is the only fuel that allows the operation of fuel cells for cars. Although the forecasts for fuel
cells were very optimistic at the beginning of this century, this drive so far has not asserted itself on the world market.
The only market vehicle with the fuel cell is Toyota-Mirai, Figure 11.
Figure 11 Toyota Mirai with fuel cell
Many believe that, with the development of the battery for the accumulation of electricity, the fuel cells become
redundant, because, in order to obtain hydrogen (theoretically the best way is by electrolysis of water), a large amount
of electric power is needed. The biggest problems with hydrogen are with its storage, transport, or the absence of
entire infrastructure. In the fuel cell, the hydrogen is then used as a source for the production of electricity. The
question is why the power would not be directly used for charging the battery, instead through the hydrogen or the
fuel cell.
One of the leading companies in the field of fuel cells, the Daimler, has moved the focus of the research from the fuel
cells to the development of batteries for the accumulation of electrical energy [7].
It is often said that the fuel cells do not have a negative impact on the human environment, because they do not emit
other harmful substances beside water vapour (H2O). Today, the hydrogen, which is mainly used for industrial
purposes, is almost exclusively (96%) produced from fossil fuels (natural gas, petroleum, stone coal) with the
releasing of the large amounts of CO2. Carbon dioxide (CO2) is almost exclusively mentioned during the discussion
about the effects of greenhouse gases on the climate change. It is forgotten that the water vapour (H2O) has the
greatest share in the natural greenhouse effect and thus has a threefold greater influence on that effect than carbon
dioxide. There is no thought yet about the effect the water vapor will have on the environment if all the vehicles were
to transfer from fossil fuels to hydrogen.
e-(ecological) fuels
In parallel with the intense work on the development of electric drives for vehicles, the automotive industry is also
working on the further development of the IC engine and, together with the oil industry, also on the development of
10
new, ecologically better fuels. Above all, they include liquid fuels (alcohols, vegetable oils) that are obtained from
renewable sources or from the waste. These fuels have the advantage that they can immediately be applied on the
existing vehicles, as well as use the existing infrastructure. Hydrogen is also an ecological fuel, which is very good
fuel for IC engines and the engines that can successfully consume it were developed a long time ago.
CONCLUSIONS
Writing this paper in Detroit, one of the centres of the world automotive industry and observing the daily traffic of
thousands of cars with which the workers travel from their homes to work and back, and looking at the dimensions
of this continent, I wonder how long will it take for this traffic to cross to the electric drive. From the aspect of densely
populated European cities, the discussions about the electric cars have an entirely different dimension, than on this
continent. Probably, it is similar in Russia (Siberia is also Russian land), Brazil, India, Australia, China...
Well-built infrastructure for recharging the batteries is considered as one of the key issues for the success of the
electric cars, in addition to their price. The developing countries will not be able to build this infrastructure and will be
related to IC engines for a long time.
Wealthier people in the USA are buying the car, which bears the name of the famous scientist, Tesla, showing their
commitment to the conservation of nature. There is no mention of the fact that their private, heated swimming pools,
the air-conditions in their villas, numerous electrical and electronic devices with continuously turned on stand-by
lights, boats and ships, consume more energy and emit more CO2 than their cars with IC engines.
All of the world’s cars and commercial vehicles transport is responsible for less than 13% of the total anthropogenic
(i.e. of man created) CO2 emissions. The largest source of these emissions are electric power plants for electricity
generation, whose share ranges from 25% to 40% [10].
Behaviour of the countries - producers of oil has considerable effect on the success of the electric drive in the large
series production. Lowering the price of oil, whenever there was a danger of its replacement, has eliminated all the
efforts for the development of alternative powertrains and alternative fuels.
The conclusion of this analysis of electric drives for motor vehicles, can only state the same as it repeats for over half
a century: "The piston IC engines will be substituted by the better powertrains during the following 10 to 15 years and
the existing oil reserves will be spent in the next 30 to 40 years!"
REFERENCES
[1] “Die Geschichte der Elektromobilität”, Ferrytells Verlags und Beratungs GmbH, Vienna, 2017,
[2] Angermaier, A. Et al.: “E-Drive: Affordable-Further-Faster”, 38th International Vienna Motor Symposium, 2017,
[3] Winter, M., Friesen, A., Schappacher, F.M.: „Lithium-Ionen-Batterien: Stand der Technik in Performanz und
Sicherheit”, 38. Internationales Wiener Motorensymposium 2017, Vienna, April, 27th-28th, 2017,
[4] Scheider, W. H.: „Effiziente Verbrennungsmotoren und elektrische Antriebe der Zukunft: Der Beitrag von Mahle”,
38. Interantionales Wiener Motorensymposium 2017, Vienna, April, 27th-28th, 2017,
[5] Von Dierk, J.: “Elektrisch übers Wasser”, VDI-Nachrichten, April, 28th, 2017,
[6] “How electric aircraft could change aviation?”, SAE-Europe, 2017,
[7] “Brennstoffzelle für Daimler out”, VDI-Nachrichten, No. 17, April, 28th, 2017,
[8] Hatz, W.: “Porsche 919 Hybrid - Le Mans. The Future of the Sportcar”, 36th International Vienna Motor
Symposium, Vienna, 2016,
[9] ACEA: “Alternative Powertrain”, ACEA, Brussels, May, 4th, 2017,
[10] Gruden, D.: “Umweltschutz in der Automobilindustrie”. Vieweg+Teubner Verlag, Visbaden, 2008.
11
12
Kragujevac, Serbia
MVM2018-IL2
Ralph Pütz1
ELECTRIC BUS VERSUS BUS WITH COMBUSTION ENGINE

- AN ECOLOGICAL-ECONOMIC APPROACH
ABSTRACT: The automotive industry is facing a worldwide paradigm shift. With the threat of driving bans on
diesel vehicles in some European cities, raising the potential for further development of conventional powertrains and
the use of alternative fuels during the transformation phase on the road to future zero-emission mobility is
indispensable. In this paper, in a system-related and technology-open approach, various options for propulsion
systems of public transport buses are analysed and compared concerning ecological and economical aspects in a
systems approach on the time horizons “today” and “medium term”.
KEYWORDS: propulsion systems, buses, local and global emissions, sustainability, renewables
INTRODUCTION
Der politisch propagierte Paradigmenwechsel hin zu einer Elektromobilität im Straßenverkehr umfasst alle Bereiche
vom Pkw über leichte bis hin zu schweren Nutzfahrzeugen. Zu den Letzteren zählen auch Linienbusse, die im
Stadtverkehr aufgrund fest vorgegebener, reproduzierbarer Linien das Potenzial eines reduzierten Aufwands für die
erforderliche Energiezuführungs-Infrastruktur aufweisen und somit – neben Verteilerfahrzeugen – für die Einführung
der Elektromobilität prädestiniert scheinen. Die Entwicklung von Elektrobussen – mit der Ausnahme des (partiell)
oberleitungsgebundenen Trolley-(Hybrid-) Busses – steht jedoch noch am Anfang. Andererseits haben etablierte,
konventionelle Antriebe für Linienbusse – Diesel- und Erdgasantriebe – bereits einen derart hohen Reifegrad erreicht,
der bei allen Herstellern ein Nahe-Null-Emissionsniveau bei vertretbarer Wirtschaftlichkeit ermöglicht. Dies belegen
umfangreiche Messreihen des Instituts BELICON, die belastbar aussagen, dass die Emissionsreduzierung der Euro-
VI-Fahrzeuge einem „Quantensprung“ im Vergleich zur Vorgängerstufe Euro V/EEV entspricht. Dabei weisen einige
Vertreter der Euro-VI-Population bei den emittierten kritischen Komponenten Partikel (PM, Particulate Matter; PN,
Particulate Number) und Stickoxide (NOx; NO und NO2) je nach Umgebungsbedingungen sogar eine geringere
Konzentration auf als in der angesaugten Umgebungsluft. Folglich besteht hinsichtlich der Partikelimmissionen – und
hier insbesondere des als kanzerogen bewerteten Feinstaubs (PM10) – sowie hinsichtlich der Stickoxidimmissionen
– und hier insbesondere bezüglich des Reizgases NO2 – für Linienbusse mit Verbrennungsmotor ab der
Grenzwertstufe Euro VI bereits kein akuter Handlungsbedarf mehr.
Nach der Androhung der EU von Sanktionen gegen die deutsche Bundesregierung aufgrund signifikanter
Überschreitungen der Luftqualitätsgrenzwerte in einigen Kommunen und Ballungsräumen und den erfolgreichen
Klagen von Umweltverbänden gegen einzelne Kommunen kann für die Verkehrsunternehmen nach dem Urteil des
Bundesverwaltungsgerichts vom 27.02.2018 dort Handlungsbedarf vor Ort entstehen, wo die Immissionen an
Stickoxiden (gemessen wird das Reizgas NO2) überschritten werden. Im Fokus stehen aktuell – forciert durch die
vermeintliche „Dieselkrise“ aufgrund Betrugsvorwürfen gegen die Pkw-Industrie, die de facto jedoch lediglich eine
Krise der Pkw-Industrie ist – insbesondere die Stickoxidemissionen (NOx) aus Dieselmotoren. Straßenfahrzeuge
machen im Mittel etwa 40% der Stickoxidemissionen aus. Somit bestehen aktuell ernsthafte Bestrebungen der
Bunderegierung – insbesondere motiviert durch die Androhung von Klagen der EU und DUH wegen der
Überschreitung der Luftqualitätsgrenzwerte in einigen Kommunen –, die Plakettenzuordnung mit Bezug auf
Stickoxidemissionen zu ergänzen („hell- und dunkelblaue Plaketten“), was weiteren Handlungsbedarf – auch in
Bezug auf eine potenzielle Forderung von Nachrüstung mit SCR-Systemen bei älteren Dieselfahrzeugen –
1
Prof. Dr.-Ing. Ralph Pütz, Landshut University, BELICON Institute for Vehicle Research – Dept. Commercial
Vehicles, Am Lurzenhof 1, D-84036 Landshut, Germany, ralph.puetz@belicon-forschung.de
13
hervorrufen könnte. Aktuell herrscht daher in vielen deutschen Verkehrsunternehmen (und sicherlich auch
europäischen) eine Verunsicherung, welche Antriebstechnologie kurz- und mittelfristig für die Linienbus-
Fahrzeugneubeschaffung aus ökologisch-ökonomischen Erwägungen das Optimum bietet und welche Maßnahmen
zur Emissionsverbesserung bestehender Fahrzeugflotten zur Verfügung stehen. Jede Verbesserung der lokalen und
globalen Emissionen sowie hinsichtlich der Ressourcenschonung muss im liberalisierten Verkehrsmarkt auch
wirtschaftlich umsetzbar sein. Deshalb soll nachfolgend auf der Basis der Randbedingungen eines mittleren ÖPNV-
Verkehrsunternehmens, das sowohl Stadt- als auch Überlandverkehr betreibt, ein Vergleich der etablierten,
konventionellen verbrennungsmotorischen Antriebe für Linienbusse mit den alternativen Antriebstechnologien des
Spektrums Elektromobilität erfolgen und auch ein Ausblick auf die weitere, mittelfristige Entwicklung gewagt werden.
Das zugrunde gelegte, beispielhafte Unternehmen betreibt rund 200 Linienbusse mit einer mittleren
Zyklusgeschwindigkeit von ca. 22 km/h mit den Extrema 18 km/h auf Linien des Stadtverkehrs und bis hin zu 34
km/h im Regionalverkehr. Die mittlere jährliche Fahrstrecke je Bus beträgt rund 50.000 km.
Description of the System Approach and preliminary Remarks

Für eine umfassende ökologische und ökonomische Bewertung von Fahrzeugflotten müssen alle Stationen des
Lebenszyklus von Linienbusverkehrssystemen wie
 Fahrzeugproduktion (und ggf. Verwertung und Entsorgung),
 Kraftstoffbereitstellung (Well-to-Tank; WTT),
 Fahrbetrieb (Tank-to-Wheel; TTW) und
 Instandhaltung
in die Analyse einbezogen werden, da die isolierte Betrachtung nur des eigentlichen Fahrbetriebs zu völlig
falschen Schlüssen führen kann, da in allen o.g. Subsystemen Energie verbraucht wird und Emissionen entstehen.
Erst auf diese Weise können im engen Kontext einer „ganzheitlichen“ ökologischen Integrität zielführende
Lösungsansätze für Linienbusverkehrssysteme mit sowohl geringen lokalen als auch globalen Emissionen bei hoher
Energieeffizienz und reduzierten Geräuschen systembezogen identifiziert werden. So kann z.B. eine „Nullemission“
der Elektromobilität im Einsatz vor Ort durch ggf. höhere Emissionen in der Kraftstoffbereitstellung und
Fahrzeugproduktion zu signifikanten ökologischen Nachteilen führen. Bei der Analyse der Emissionen stehen als
lokal wirksame Hauptkriterien die Partikel- und Stickoxid-Emissionen und als global wirksame Hauptkriterien die
CO2-Emissionen bzw. das CO2-Äquivalent im Vordergrund, deren Wirkungen über die Bestimmung der externen
Kosten als ökologisches Profil einer Antriebstechnik zusammengefasst werden können.
Bei der ökonomischen Analyse erfolgt die Fokussierung auf die „Fahrzeugkosten“, die aus Kapitaldienst,
Kraftstoff/Energiekosten und Instandhaltungskosten gebildet werden, ergänzt durch die Kosten für die
Energiezuführungsinfrastruktur.
Als „konventionell“ können Linienbusse mit den etablierten Diesel- und Erdgasantrieben bezeichnet werden. Dabei
ist eine „Mild-Hybridisierung“ beim Dieselantrieb durch Einsatz von sogenannten Rekuperationsmodulen mit 48-V-
Bordnetz bereits heute als Serienstandard verfügbar. Linienbusse mit konventionellem, verbrennungsmotorischem
Antrieb machen heute immer noch rund 98% der deutschen Linienbusflotte aus (vgl. Tabelle 1).
Tabelle1 Antriebstechnologien in der deutschen Linienbusflotte 2017 (Stadt- und Überlandbusse)

(Quellen: VDV, KBA)
Antriebstechnologie Prozent
Diesel konventionell 94,85
Diesel-Hybrid 1,65
Erdgas (CNG) 2,91
Wasserstoff-Verbrennungsmotor (H2 ICE) 0,00
Brennstoffzellen (BZ)-Hybrid 0,04
Batterie 0,21
Trolley 0,29
Trolley-Hybrid 0,04
Summe 100,00
Trotz zahlreicher Pilotprojekte und Praxiserprobungen von Verkehrsunternehmen mit alternativen Antrieben, die in
der Regel national von den Ministerien BMVI, BMU und BMWI sowie international durch EU-Förderprojekte (z.B.
ZeEUS, Zero Emission Electric Urban Bus System und ELIPTIC, Electrification of Public Transport in Cities) finanziert
und medienwirksam präsentiert werden, ist der prozentuelle Anteil an alternativen Antrieben in der deutschen
Linienbusflotte noch vernachlässigbar gering. In der Summe erreichen Linienbusse mit Batterien und
Brennstoffzellenantrieb heute einen Anteil von deutlich weniger als ein Prozent der Linienbusflotte, obwohl
14
umfangreiche Marketingmaßnahmen einen anderen Anschein erwecken mögen. Mittelfristig wird auch nach
Bewertung der Nationalen Plattform Elektromobilität (NPE) und der „Batterie-Roadmap 2050“ weiterhin die Li-
Ionentechnik der 2. und 3. Generation eingesetzt werden. Erst die Post-Li-Ionen-Technologie, deren Einführung nach
2030/2035 erwartet wird, dürfte zu signifikanten Effizienzsteigerungen führen. Ob die Li-S, Li-Luft oder
Festkörperbatterietechnologie sich durchsetzen wird, ist heute nicht absehbar. Auch die zwischenzeitlich als
betrieblich serienreif zu bezeichnenden Busse der Brückentechnologie (Full-)Hybrid, die seit längerem im Einsatz
sind und bei den alternativen Antrieben dominieren, machen heute lediglich weniger als zwei Prozent der Linienbus-
Flotte aus. Der Autor sieht die optimistische Marktdurchdringung von Elektrofahrzeugen bei Neufahrzeugen im
Stadtbusbereich unter für die Elektromobilität günstigsten Annahmen wie in Abbildung 1 dargestellt. Dabei wird ein
Technologiefortschritt bei Batterien mit Migration auf die Post-Li-Ionen-Technik ab frühestens 2030 unterstellt.
Abbildung 1 Potenzielle Marktdurchdringung von Elektrobussen: Entwicklung der Neuzugänge

im Stadtbereich unter optimistischen Annahmen (Quelle: © BELICON)
Es muss konstatiert werden, dass der zu konventionellen Antrieben äquivalente Einsatz alternativer
Antriebstechnologien im deutschen ÖPNV, insbesondere von Batterie- und Brennstoffzellenbussen, heute noch sehr
begrenzt ist. Mit der Annäherung an eine betriebliche Serienreife dieser Technologien wird deren Einsatz zukünftig
weiter forciert werden. Kleinere und mittlere Verkehrsunternehmen unterhalb einer Flottengröße von 200 Bussen
dürften von einem sehr breiten Einsatz alternativer Antriebe mit heute noch i.d.R. signifikant reduzierten
Verfügbarkeiten des Gesamtsystems (inklusive der Energiezuführungs-Infrastruktur) überfordert sein, da mit
Ausnahmen der Unternehmen mit Straßenbahnen auch die Personal-Qualifikationsstruktur hier unzureichend ist. In
den kleineren und mittleren Unternehmen beschränkt sich der Einsatz alternativer Antriebe sinnvollerweise allenfalls
auf Einzelfahrzeuge oder Kleinstflotten, um Erfahrungen mit der Elektromobilität zu sammeln.
Zu den aktuell diskutierten, aussichtsreichen Konzepten der Elektromobilität zählen:
 Opportunity-Elektrobus (auch Opportunity Charger oder Gelegenheitslader, GL); verfügt über eine kleinere
Batterie und muss auf der Linie regelmäßig nachgeladen werden. Dies erfordert eine Schnellladung mit
hohen Strömen und Spannungen. Ladestationen sind zumindest an den Endhaltestelle der Linien
anzusetzen.
 Overnight-Elektrobus (auch Nachtlader, NL), verfügt über eine große Batterie und wird daher lediglich über
Nacht auf dem Betriebshof langsam und schonender nachgeladen.
 Brennstoffzellen-Hybridbus, BZ-Hybrid (konventionelle Brennstoffzellenbusse spielen keine Rolle mehr).
 Trolley-Hybridbus, auch „In-Motion-Charger“ oder Hybrid-Oberleitungsbus, HO (konventionelle
Trolleybusse werden zukünftig keine Rolle mehr spielen).
Es sei an dieser Stelle betont, dass in der Regel die Elektromobilität für jedes Verkehrsunternehmen und seine
Randbedingungen ein individuelles System darstellt, bei dem Fahrzeug- und Antriebstechnik, Batterietyp,
Energiezuführungsart/Ladetechnik und Betriebsmuster auf die jeweilige Linientopologie exakt abgestimmt sein
müssen.
Ein wichtiger umweltrelevanter Vorteil aller Elektrobuskonzepte sind die im Vergleich zu verbrennungsmotorischen
Antrieben reduzierten Geräuschemissionen. Praxismessungen des Instituts BELICON belegen, dass beim Verlassen
der Haltestelle Elektrobusse um rund 5 dB(A) reduzierte Schalldruckpegel gegenüber Dieselbussen aufweisen, siehe
Abbildung 2. Für eine quantifizierte Beurteilung der Umweltwirkungen aufgrund der Schallemissionen fehlt bislang
allerdings eine gesetzliche Grundlage anhand z.B. externer Kosten wie bei Schadstoffemissionen (siehe EU-
Beschaffungsrichtlinie 2009/33/EG). Dazu müssten jedoch nicht alleine die Schalldruckpegel, sondern diese auch
mit Bezug auf die jeweiligen Frequenzen, begutachtet werden. Deshalb unterbleibt nachfolgend eine Bewertung der
Geräuschemissionen.
15
Abbildung 2 Beispiel für Außengeräusche von Diesel (D-Bus)- und Batteriebus (E-Bus) bei
Haltestellenanfahrt (A-bewerteter Summenpegel) (Quelle: BELICON©)
Neben den ökonomischen Parametern auf der Basis der Fahrzeugkosten (Kapitaldienst, Kraftstoff- und
Instandhaltungskosten) sowie Infrastrukturkosten wird die Wirtschaftlichkeit auch durch betriebliche Parameter
beeinflusst. So weisen z.B. Overnight-Charger-Batteriebusse, die im Solobus Batteriekapazitäten von über
300 kWh besitzen, aufgrund der großen Batteriegewichte geringere Fahrgastkapazitäten als konventionelle Busse
mit Verbrennungsmotor auf. Dies bedeutet, dass zum Ersatz eines konventionellen Busses in der Morgenspitze zwei
Batteriebusse eingesetzt werden müssten, vgl. Abbildung 3. Des Weiteren wird im Winter und Sommer, wenn forciert
Heizenergie bzw. Energie zum Betrieb der Klimaanlage aus der limitierten Energiekapazität der Batterie geliefert
werden muss, die geforderte Tagesfahrstrecke von mehr als 300 km nicht erreicht, so dass aufgrund langer
Ladezeiten auch hier zwei Batteriebusse einen konventionellen Bus ersetzen müssen (vgl. Abbildung 4). Die
Kombination dieser Einflüsse, z.B. Morgenspitze im Winter, verstärkt die Problematik weiter signifikant. Die anderen
Optionen des Spektrums Elektromobilität weisen derart ausgeprägte Einschränkungen in der Fahrgastkapazität und
in den Ladezeiten (Energiezuführung) nicht auf. Beim Opportunity-Charger können sich jedoch bei Ausfall einer der
Ladestationen Beeinträchtigungen des gesamten Betriebsablaufs ergeben. Dies gilt auch bei Verspätungen und
Zugriff mehrerer Linien auf eine Ladestation.
Abbildung 3 Fahrgastkapazität von Solo-Linienbussen nach Antriebssystem
16
Abbildung 4 Reichweite von Solo-Linienbussen nach Antriebssystem (gefordert:
>300 km pro Tag)
Ecological Analysis
Die ökologische Gesamtbewertung erfolgt hier gemäß der sogenannten „EU-Beschaffungs“-Richtlinie 2009/33/EG
(Richtlinie zur Förderung sauberer und energieeffizienter Straßenfahrzeuge) anhand der Summierung der externen
Kosten aus den lokalen und globalen Emissionen aller Subsysteme (Kraftstoffvorkette, Fahrzeugproduktion,
Fahrbetrieb) und charakterisiert somit in der Summe die Umweltrelevanz der untersuchten Antriebsvarianten zum
jeweiligen Stand der Technik „heute“ (2017) und „mittelfristig“ (2027). Die Analyse bezieht sich exemplarisch auf
Solobusse.
Die Bereitstellung von Dieselkraftstoff erfolgt dabei auf Basis eines Mixes aus konventionellem Rohöl (heute
dominierend), unkonventionellem Rohöl sowie aus Kohle (Coal-to-Liquid) und Erdgas (Gas-to-Liquid). Zukünftig wird
die Bereitstellung von Dieselkraftstoff aus Ölsanden und Ölschiefer eine größere Bedeutung erlangen. Dies ist
nachfolgend im Mix für den Zeithorizont „mittelfristig“ (2027) berücksichtigt. Die Bereitstellung von Erdgas erfolgt auf
Basis eines Mixes aus unterschiedlichen Herkunftsländern. Dabei weist die Erdgasbereitstellung aus Russland und
Katar die höchsten Treibhausgasemissionen auf, jene aus den Niederlanden, Norwegen, Großbritannien und
Deutschland hingegen vergleichsweise geringe. Die Emission von globalen und lokalen Emissionen für die
Bereitstellung von Erdgas für Deutschland auf dem Zeithorizont „mittelfristig“ (2027) basieren nachfolgend auf dem
aktuellen EU-Mix sowie der Annahme, dass in diesem für 2027 ein Anteil von rund 11% an Erdgas aus noch größerer
Transportentfernung von den Erdgasfeldern (7.000 km) und 5% an Biogas mit einer 90,42%igen CO2-Reduzierung
(auf der Basis der Angaben der Bundesanstalt für Landwirtschaft und Ernährung, 2017) enthalten sind. Je nach
Genese verändern sich die Stickoxidemissionen des Bereitstellungspfades für Biogas gegenüber fossilem Erdgas in
einer enormen Bandbreite, so dass nachfolgend Worst Case-Werte Anwendung finden. Diesen steht die genannte,
hochsignifikante CO2-Reduzierung von 90,42 Prozent durch Biogas aus Abfällen entgegen.
Für die Bereitstellung von komprimiertem Wasserstoff (Compressed Gaseous H2, CGH2) wird ein Mix aus
Erdgasdampfreformierung (heute dominierend) und Elektrolyse mit erneuerbarem Strom angesetzt. In beiden Fällen
wird vorausgesetzt, dass der Wasserstoff vor Ort an der Tankstelle produziert wird, da eine zentrale, großtechnische
Elektrolyse-Infrastruktur nicht vorhanden ist und eine immense Investition erforderte, die nur die Bundespolitik durch
strategische Entscheidung mit nachfolgender massiver Förderung initiieren kann, was heute aber noch nicht
absehbar ist. Für den Zeithorizont „mittelfristig“ wird eine Wasserstoffherkunft zu 50 Prozent aus erneuerbarem Strom
angesetzt. Für die Zusammensetzung des deutschen Strom-Mixes im Jahr 2027 soll laut BMWi das Potenzial
Deutschlands zur Steigerung des Stromverbrauchs aus Solar- und Windenergie soweit verbessert werden, dass 40
bis 45 Prozent des in Deutschland verbrauchten Stroms aus erneuerbaren Energien produziert werden. Laut dem
Pariser Klimaschutzabkommen wurde von der EU insgesamt eine CO2-Reduzierung um 28 Prozent bis 2021
zugesagt. Deutschland will im Vergleich zum Jahr 1990 bis zum Jahr 2020 die CO2-Emissionen gar um 40 Prozent
senken. Die Erreichbarkeit dieses Ziels ist indes sehr fraglich, da bis zum Jahr 2015 lediglich eine Reduzierung um
27 Prozent zu verzeichnen war.
Zur Bestimmung der Umweltwirkungen der Busherstellung werden eventuelle Migrationen hin zu „Purpose-Design“-
Elektrofahrzeugen, die der Autor durchaus als zielführend erachtet, aufgrund von Unwägbarkeiten nicht
berücksichtigt, da hierfür eine grundlegende Änderung der Fertigungsstrukturen erforderlich wäre. Bei der
17
ökonomischen Betrachtung wird davon ausgegangen, dass die essentiellen Rohstoffe Lithium und Kobalt für ein
schnelles, weltweites Wachstum der Elektromobilität ausreichend vorhanden sind. Temporäre Engpässe an Lithium
und Kobalt bleiben wegen der in der Literatur sehr heterogenen Bewertungen hier unberücksichtigt.
Das Datengut für die Emissionen im realen Fahrbetrieb auf den Linien basiert nachfolgend auf den umfangreichen
PEMS-Messungen (Portable Emission Measurements) des Instituts BELICON an Solobussen für unterschiedliche
Linientopologien unter Einbeziehung des Energieverbrauchs für Heizung/Klimatisierung. Bei der Erdgasbus-Variante
wurde aufgrund heterogener Verbrauchswerte der aktuelle Benchmarksetter angesetzt. Oftmals wird für die
Versionen der Elektromobilität lediglich der Energieverbrauch für die Traktion genannt, der bei Solobussen für SORT-
2-Charakteristik in der Ebene bei etwa 1,0 kWh/km liegt. Der Heizenergiebedarf wird dabei i.d.R. unzulässigerweise
vernachlässigt. Die für die Batteriebusoptionen zugrunde gelegten Verbrauchswerte beruhen auf Praxismessungen
des Instituts BELICON zu allen Jahreszeiten und sind als Mittelwerte berücksichtigt worden. Für den Zeithorizont
„mittelfristig“ (2027) wurde für die verbrennungsmotorischen Antriebe mit Diesel und Erdgas eine Mild-Hybridisierung
als Serienstand mit einer durchschnittlichen weiteren Verbrauchsreduzierung um 10% angenommen, während für
die alternativen Antriebe des Spektrums Elektromobilität bereits für den Zeithorizont „heute“ nach den Erfahrungen
des Instituts BELICON eine mittlere Rekuperationsrate von 25% über das gesamte Einsatzjahr unterstellt wurde.
Die ökologische Gesamtbewertung für den Zeithorizont heute (2017) weist aus, dass für eine umfassende
Nachhaltigkeit – unter Berücksichtigung lokaler und globaler Emissionen sowie des Energieverbrauchs – bei den
verglichenen Varianten heute keine nennenswerten Verbesserungen durch alternative elektrische Antriebsvarianten
erreicht werden. Zeitnah bleiben deshalb Diesel- und Erdgastechnik auch mit fossilen Kraftstoffen eine veritable
Option im ökologisch nachhaltigen ÖPNV, siehe Abbildung 5. Insgesamt liegen die konventionellen Antriebe Diesel-
Euro VI und Erdgas-Euro VI für die mittlere Einsatzcharakteristik etwas hinter dem Opportunity-Charger
(Gelegenheitslader), aber vor den Optionen Overnight-Charger (Nachtlader) und Brennstoffzellen-Hybrid. Der
Einsatz von Full-Hybrid-Dieselbussen böte ebenso signifikante ökologische Vorteile wie die Verwendung von Biogas.
Der Autor merkt aufgrund der Messergebnisse seines Instituts BELICON jedoch an, dass die Energieeinsparungen
bei Hybridbussen eine enorme Schwankungsbreite über alle Hersteller von 5% bis 31% aufweisen. Beim heute
bereits möglichen Einsatz von 100% Biogas (CO2-Minderungspotenzial >90 Prozent) wäre die Erdgastechnik in
Verbindung mit Euro-VI-CNG-Antrieben heute die sauberste verfügbare Antriebstechnik.
Abbildung 5 Systembezogene Umweltrelevanz 2017 anhand externer Kosten gemäß 2009/33/EG für
Solobusse unter mittleren ÖPNV-Randbedingungen im gesamten Lebenszyklus (12 Jahre)
Weitere Optionen wie der Einsatz biogener synthetischer Dieselkraftstoffe (Biomass-to-Liquids; BtL) sowie von Strom
und Wasserstoff aus ausschließlich regenerativen Energien böten weiteres ökologisches Potenzial, das künftig aktiv
erschlossen werden sollte. Aufgrund der Energiewende mit Abschaltung der Atomkraftwerke und ersatzweiser
Nutzung von Braunkohle und Steinkohle hatte sich der Strom-Mix jedoch vorübergehend sogar ökologisch
verschlechtert. Biogene Kraftstoffe der 2. und 3. Generation sind für die Dieseltechnik aktuell (und wohl auch
mittelfristig) nur in geringen Mengen verfügbar. Zukünftige Entwicklungen werden sich vermehrt auch biogenem
Dieseltreibstoff aus Algen und der Herstellung von synthetischen Kraftstoffen aus regenerativ erzeugtem Strom
(Power-to-Liquids, PtL) widmen, was jedoch erst sehr langfristig Marktrelevanz erreichen und hier nicht berücksichtigt
wird. Analog ebenso unberücksichtigt bleibt bei Erdgasfahrzeugen die Option Power-to-Gas (PtG).
18
Die ökologische Bewertung ausschließlich der global wirksamen Emissionen für den Zeithorizont heute (2017)
zeigt eine Dominanz der globalen Emissionen mit bis zu über 80 Prozent am gesamten ökologischen Profil (externe
Kosten). Bezogen auf die globalen Emissionen wären heute Verbesserungen durch alternative Antriebsvarianten der
Elektromobilität gegenüber den konventionellen verbrennungsmotorischen Antrieben erreichbar, jedoch nicht
unbedingt gegenüber dem Diesel-Hybridbus. Die konkurrenzlos CO2-günstigste Option ist der Einsatz von Biogas
im CNG-Euro-VI-Antrieb, vgl. Abbildung 6.
Abbildung 6 Systembezogene globale Umweltrelevanz 2017 anhand externer Kosten gemäß 2009/33/EG für
Die ökologische Gesamtbewertung für den Zeithorizont mittelfristig (2027) der verglichenen Varianten unter
mittleren ÖPNV-Randbedingungen weist aus, dass die Umweltprofile der alternativen Technologien des Spektrums
Elektromobilität signifikante Vorteile gegenüber jenen der konventionellen Antriebe erreichen können, wenn sich die
hohen, politisch erwarteten Reduktionsziele (z.B. Anteil regenerativer Energien) auch tatsächlich einstellen werden,
vgl. Abbildung 7. Die ökologische Bewertung ausschließlich der global wirksamen Emissionen für den Zeithorizont
mittelfristig (2027) zeigt signifikante Verbesserungen durch alternative Antriebsvarianten des Spektrums
Elektromobilität gegenüber den konventionellen verbrennungsmotorischen Antrieben, auch gegenüber dem Diesel-
Hybridbus. Die Erdgastechnik mit Biogas böte bei den globalen Emissionen auch in 2027 weiterhin die beste
ökologische Option (siehe Abbildung 8).
Abbildung 7 Systembezogene Umweltrelevanz 2027 anhand externer Kosten gemäß 2009/33/EG für
19
Abbildung 8 Systembezogene globale Umweltrelevanz 2027 anhand externer Kosten gemäß 2009/33/EG für
Economical Analysis
Die ökonomische Analyse der derzeit diskutierten alternativen Antriebskonzepte der Elektromobilität im Vergleich
zum Euro-VI-Diesel- als Referenz unter den Randbedingungen eines mittleren ÖPNV-Betriebes erfolgt enemfalls auf
der Fahrzeug-Basis Solobus. Die Kosten je Bus-Kilometer setzen sich dabei zusammen aus:
 Fahrzeug-Kapitaldienst;
 Ersatzinvestitionen (z.B. für einen während der Lebenszeit des Busses notwendigen Batteriewechsel);
 Kraftstoffe bzw. Energie;
 Instandhaltung;
 Infrastruktur-Mehrkosten gegenüber Dieselbetrieb (hier nur exemplarisch darstellbar).
Es wird ein Linienbus in üblicher Vollausstattung als Basis angesetzt. Dabei werden für die Optionen des Spektrums
Elektromobilität während der betrieblichen Nutzungsdauer von 12 Jahren eine zweite Batterie bzw. eine zweite
Brennstoffzelle angesetzt, da aktuelle Erfahrungen hinsichtlich der Garantien der Elektrobushersteller darauf
hindeuten. Für Diesel-Full-Hybride werden heute mittlere Investitionsmehrkosten von 90.000 € angesetzt, für
Erdgasbusse 40.000 €. Für einen Opportunity-Charger mit einer 70-kWh-Batterie wird ein Mehrpreis von rund
170.000 €, für einen Overnight-Charger mit 350-kWh-Batterie ein Mehrpreis von rund 450.000 €, jeweils inklusive
zweiter Batterie, angesetzt. Für den Zeithorizont „mittelfristig“ (2027) werden jedoch erwartungsgemäß
hochsignifikante Investkosten-Senkungen bei allen alternativen Antrieben, insbesondere des Spektrums
Elektromobilität, angesetzt.
Da in einigen Bundesländern keine bzw. nur eine pauschale Busförderung existiert, soll nachfolgend die öffentliche
Förderung bei den konventionellen Antrieben außer Acht gelassen und lediglich für Elektrobusse eine
Anschubfinanzierung gemäß der Verlautbarung des Bundesumweltministeriums BMUB v. 26.02.2018 in Höhe von
80 Prozent der beihilfefähigen investiven Fahrzeug-Mehrkosten und 40 Prozent der beihilfefähigen
Infrastruktur(mehr)kosten angesetzt werden. Für die Entwicklung der Kraftstoffpreise wurden Abschätzungen gemäß
IFEU/BELICON, 2015, der Energiereferenzprognose von 2014 und A. Schulz, Diss. 2015 sowie auf der Basis
weiterer wissenschaftlicher Veröffentlichungen und Angaben der Energieunternehmen angesetzt. Demnach ist
mittelfristig ein signifikanter Anstieg der Kraftstoffpreise bei allen Energieträgern, mit der Ausnahme von Wasserstoff,
zu erwarten. Durch die Entkoppelung von Erdgas- und Dieselpreis und die Erschließung weiterer Erdgasfelder ist
der Preisanstieg beim Erdgas geringer. Die Instandhaltungskosten sind einerseits von den Personalkosten,
abhängig von den Personalkennzahlen und den durchschnittlichen Jahresgehalten (hier werden übliche Lohnkosten
von rund 50.000 €/a angenommen) sowie den Materialkosten abhängig.
Da fast alle Busverkehrsunternehmen über eine Diesel-Betankungsinfrastruktur auf den Betriebshöfen verfügen,
müssen bei einem Wechsel auf eine andere Antriebstechnik noch die Infrastrukturkosten für die Energiezuführung
kalkuliert werden. Die Infrastrukturkosten für alternative Antriebe sind neben der Flottengröße jedoch hochsignifikant
von den lokalen Gegebenheiten sowie dem Betriebsmuster und dem Liniennetz abhängig („Elektromobilität als
System“). Für einen Schnellladepunkt beim Opportunity-Charger (GL) werden z.B. Kosten in Höhe von 250.000 €
angenommen.
20
Die ökonomische Gesamtbewertung der reinen Fahrzeugkosten (ohne Infrastruktur) für den Zeithorizont 2017
(heute) bei einer Abschreibungsdauer von 9 Jahren ist in Abbildung 9 dargestellt. Daraus ist ersichtlich, dass heute
alle Elektrobustechnologien – trotz der angesetzten Förderung von 80 Prozent der Mehrkosten gegenüber einem
Dieselbus – noch hochsignifikant teurer als der konventionelle, hochsaubere Euro-VI-Dieselbus sind.
Abbildung 9 Relativer Kostenvergleich der Antriebstechnologien für Solobusse (Lebenszykluskosten ohne

Fahrerkosten) auf dem Zeithorizont „heute“ für mittleren ÖPNV – ohne Förderung (außer E-Mobilität!) und
ohne Infrastruktur
Bei einseitiger, massiver Förderung der Optionen der Elektromobilität könnte – ohne Berücksichtigung der
Energiezuführungsinfrastruktur – ein Opportunity-Charger zu etwa 14 Prozent Mehrkosten gegenüber einem Euro-
VI-Dieselbus beschafft werden. Alle anderen Optionen der Elektromobilität liegen dennoch weit jenseits dieser
Mehrkosten. Zum Brennstoffzellen-Hybridbus ist anzumerken, dass die heute verfügbaren Fahrzeuge des
Benchmarksetters noch die höchsten Kapitaldienst- und Instandhaltungsaufwendungen erfordern. Durch Reduktion
der Brennstoffzellenkosten sowie durch eine zukünftig höhere Lebensdauer der Stacks werden signifikante
Kostensenkungspotentiale gesehen, jedoch ist die Kostenentwicklung von Brennstoffzellenbussen weiterhin mit
hohen Unsicherheiten verbunden, die wesentlich auch von der Marktdurchdringung von BZ-Fahrzeugen in anderen
Märkten (z.B. Pkw) abhängt.
Zum Vergleich sollen nachfolgend die aktuellen betrieblichen Relationen ohne Förderung dargestellt werden – also
die tatsächlichen Marktkosten, vgl. Abbildung 10. Demnach erfordert die fahrzeugseitig kostengünstigste Option der
Elektromobilität, Opportunity-Charger, rund 50 Prozent fahrzeugbezogene Mehrkosten gegenüber einem Euro-VI-
Dieselbus.

Fahrerkosten) auf dem Zeithorizont „heute“ für mittleren ÖPNV – ohne Förderung und ohne Infrastruktur,
also echte Marktkosten
21
Bei zusätzlicher Berücksichtigung der Infrastrukturkosten ist beim Übergang auf die Elektromobiliät unter
üblichen Randbedingungen der Mehrzahl der ÖPNV-Betriebe mit gemischtem Stadt-, Vorort- und Überlandverkehr
– mit Fahrzeugförderung (80% der Mehrkosten) und Infrastrukturförderung (40% der Mehrkosten) bei den
Elektrobussen – heute mit signifikanten Gesamtmehrkosten von rund 22% (Opportunity Charger) und rund 35%
(Overnight Charger) zu rechnen. Auch der Erdgasbus (ohne zusätzliche Förderung) weist signifikante Mehrkosten
von 16% gegenüber einem Dieselbus auf, bei Verwendung von Biogas erreichen die Mehrkosten 18%.
Abgeschlagen liegt der (mit üblicherweise 60% der Mehrkosten geförderte) Brennstoffzellenhybridbus mit
Gesamtmehrkosten von rund 161% als teuerste Option. Demgegenüber sind die Infrastrukturkosten für z.B. eine
Erdgastankstelle nahezu unbedeutend. Etwaige zusätzliche Reservefahrzeuge bei der Elektromobilität (aufgrund
reduzierter Verfügbarkeit oder, im Falle des Overnight-Chargers, reduzierter Fahrgastkapazität) werden hier nicht
berücksichtigt. Nicht untersucht wurde hier die Option eines Wasserstoff-Verbrennungsmotorantriebs, der aktuell
auch nicht serienmäßig verfügbar ist.

Fahrerkosten) auf dem Zeithorizont „heute“ für mittleren ÖPNV – ohne Förderung (außer E-Mobilität!), aber
inklusive Infrastruktur (Infrastrukturförderung nur bei Elektrobussen!)
Ohne Berücksichtigung der heute üppigen, einseitigen Anschubfinanzierung des Bundesumweltministeriums für
Elektrobusse inklusive deren Infrastruktur ergäbe sich das in Abbildung 12 gezeigte, marktkonforme Bild. Damit lägen
die Optionen der Elektromobilität noch viel weiter hinter einer Wirtschaftlichkeit zurück.

Fahrerkosten) auf dem Zeithorizont „heute“ für mittleren ÖPNV – ohne jegliche Förderung,
inklusive Infrastruktur
Für die ökonomische Bewertung für den Zeithorizont „mittelfristig“ (2027) wird die prognostizierte Entwicklung der
Kraftstoffpreise angenommen, in den Investkosten für die Elektromobilität ebenso die prophezeite, signifikante
22
Kostendegression der Energiespeicher, jedoch weiterhin noch eine Ersatzbatterie über die 12 (bis 17)jäjhrige
betriebliche Einsatzdauer beim (ersten) Betreiber. Heute ist davon auszugehen, dass die Lebensdauer der Batterien
beim durchschnittlichen ÖPNV-Einsatz kaum mehr als ein „halbes Busleben“ (etwa 7 Jahre) erreicht. Für die
Brennstoffzellen (BZ)-Technik ist „mittelfristig“ ebenfalls eine signifikante Kostendegression angenommen.
Daraus folgt die in Abbildung 13 dargestellte ökonomische Gesamtbewertung für den Zeithorizont „mittelfristig“
(2027), wobei dann freilich bei allen Antriebsoptionen keine Förderung mehr angesetzt wird, da auch bei den
alternativen Konzepten der Elektromobilität (Batterie- und BZ-Technik) auf diesem Zeithorizont eine erreichte
Serienreife unterstellt wird. Außerdem wird dann die Energiezuführungsinfrastruktur als bereits vorhanden unterstellt.

Fahrerkosten) auf dem Zeithorizont „mittelfristig“ für mittleren ÖPNV – ohne Förderung und ohne Infrastruktur
Es ist ersichtlich, dass „mittelfristig“ (Jahr 2027) Elektrobusse – insbesondere wegen der anzunehmenden Steigerung
der Energiepreise für die fossilen Kraftstoffe Diesel und Erdgas sowie der Reduktion der Batterie- und BZ-Kosten
und trotz steigender Strompreise – sich wirtschaftlich an die konventionellen Antriebe annähern werden. Die
Wirtschaftlichkeit der etablierten, hochsauberen Konzepte mit Verbrennungsmotor wird jedoch auch mittelfristig von
den Optionen der Elektromobilität für die Gegebenheiten der meisten ÖPNV-Unternehmen nicht erreicht werden.
Weiter über dem Niveau der Diesel- und Diesel-Hybrid-Technik werden auch die Fahrzeugkosten der Erdgas-/
Biogastechnik liegen (+7% gegenüber dem Dieselantrieb). Mit der mittelfristig noch dominierenden, heutigen Li-
Ionen-Metalloxid-Batterietechnik wird ein Betrieb von Elektrobussen wie bei Bussen mit Verbrennungsmotor – also
Overnight Charger (GL) – aufgrund der hohen Investitionskosten bereits ohne Berücksichtigung der Investitionen
in die Infrastruktur nur mit hochsignifikanten Mehrkosten von rund 52% möglich sein. Der Opportunity Charger
(GL) wird – ohne Berücksichtigung der Infrastrukturkosten noch um rund 16% Mehrkosten gegenüber der
Dieseltechnik aufweisen.
Da die Weiterentwicklung der Elektromobilität im Pkw-Bereich als Treiber auch für die Kostendegression der
Elektromobilität bei Linienbussen angesehen wird, soll hier die Bewertung FEV Aachen (2017) für den Zeitverlauf
bis zum Jahr 2030 zitiert werden. Demnach werden im Jahr 2030 lediglich 20% der Neuzugänge im Bereich Pkw
und leichte Nutzfahrzeuge über keinen Verbrennungsmotor verfügen. Batteriefahrzeuge werden dann 19% und
Brennstoffzellenfahrzeuge lediglich 1% der Fahrzeugneuzugänge ausmachen. Allerdings sollen 91% der
Neuzugänge über einen elektrischen Antriebsstrang verfügen, was die zunehmende Bedeutung der (Plug-in)
Hybridisierung belegt. Den mit 51% größten Anteil an den Neuzugängen werden den Mild-Hybriden zugeschrieben.
Trost (2016) entwickelte in seiner Dissertation am Fraunhofer IWES eine Prognose für die Entwicklung des
Fahrzeugbestandes je Antriebstechnik im Zeitverlauf. Er prognostiziert für das Jahr 2030 einen Anteil an
Batteriefahrzeugen um rund 7%. Erst nach 2040, mit der erwarteten Einführung der Post-Li-Ionen-Technik (Li-Luft,
Li-S, Festkörperbatterien etc.) wird die Batterieelektromobilität demnach stärker an Bedeutung gewinnen.
CONCLUSION
Die ökologische und ökonomische Analyse der zurzeit eingesetzten oder im Test befindlichen alternativen
Antriebstechnologien für Linienbusse und deren erwartete Weiterentwicklung hat gezeigt, dass konventionelle „Near
Zero Emission“-Antriebe mit Diesel-Verbrennungsmotoren für die überschaubare Zukunft, d.h. den
23
Betrachtungshorizont „mittelfristig“, für die Bedingungen der meisten ÖPNV-Verkehrsunternehmen weiterhin die
erste Wahl für Linienbusse bleiben werden, denn sie ermöglichen es weiterhin sowohl wirtschaftlich als auch
umweltgerecht zu handeln. Alternative Antriebe des Spektrums Elektromobilität würden im Vergleich zu den
etablierten Verbrennungsmotor-Optionen heute (2017) systembezogen die Ökobilanz nicht (bzw. im Fall des
Opportunity Chargers nur unwesentlich) verbessern. Erst mittelfristig (2027) können die alternativen Optionen der
Elektromobilität das ökologische Niveau der etablierten Antriebsoptionen mit Verbrennungsmotor für die
Randbedingungen der meisten ÖPNV-Unternehmen unterbieten.
Ökonomisch betrachtet liegen alle alternativen Antriebe des Spektrums Elektromobilität heute (2017) – trotz der
massiven Förderung von 80 Prozent der Fahrzeugmehrkosten gegenüber einem Euro-VI-Dieselbus und trotz der
Förderung von 40 Prozent der Anschaffungskosten für die Energiezuführungsinfrastruktur – weitab jenseits eines
wirtschaftlichen Einsatzes. So wäre heute inklusive der aktuellen Fahrzeug- und Infrastrukturförderung des Bundes
mit jährlichen Mehrkosten (inklusive Infrastruktur) in der Größenordnung von rund 22 Prozent beim Opportunity
Charger (Gelegenheitslader) und 35 Prozent beim Overnight Charger zu rechnen. Etwaige Mehrkosten aufgrund
noch geringerer Verfügbarkeit oder reduzierter Fahrgastkapazität (Overnight-Charger) wurden dabei nicht
berücksichtigt.
Auch „mittelfristig“ (2027) werden die Mehrkosten der Optionen des Spektrums Elektromobilität – bei dann nicht mehr
unterstellter spezifischer Förderung („Anschubfinanzierung“) – noch nicht das Niveau der hochsauberen
Verbrennungsmotoroptionen erreicht haben und bei Berücksichtigung der Energiezuführungsinfrastruktur (dann
ohne Förderung) im Abschreibungszeitraum noch Mehrkosten zwischen rund 16 Prozent (Opportunity Charger) und
52 Prozent (Overnight Charger) gegenüber den etablierten Konzepten mit Verbrennungsmotor erfordern. Die
Mehrkosten für Brennstoffzellenhybride werden dann immer noch voraussichtlich bei rund 63 Prozent liegen. Für alle
Antriebsoptionen wurde auf dem Zeithorizont „mittelfristig“ das Vorhandensein einer entsprechenden Infrastruktur für
die Energiezuführung vorausgesetzt. Aufgrund der womöglich rasanten Entwicklung der Elektromobilität müssen
Einflüsse evtl. neuer Batterietechnologien, die aus heutiger Sicht zu diesem Zeitpunkt nicht als betrieblich serienreif
absehbar sind, zu gegebener Zeit eine Neubewertung erfordern.
Um das vorhandene Potenzial alternativer (elektromobiler) Busantriebe vollumfänglich zu entwickeln ist es wichtig,
dass einige – große – Verkehrsunternehmen weiter dazu bereit sind, solche Busse im Linienverkehrseinsatz zu
testen und so wertvolle Erfahrungen zu sammeln, die allen Verkehrsunternehmen und damit der gesamten ÖPNV-
Branche dienen. Kleinere und mittlere Verkehrsunternehmen sind von dieser Pionierrolle indes i.d.R. massiv
überfordert, da eine „kritische Größe“ oder Synergien durch das Vorhandensein einer Straßenbahn/Stadtbahn
vorhanden sein sollten bzw. müssen. Die Erprobung und Weiterentwicklung alternativer Antriebe des Spektrums
Elektromobilität ist darüber hinaus auch nur dann möglich, wenn kontinuierlich umfangreiche Mittel der öffentlichen
Hand über lange Zeit hierfür zur Verfügung gestellt werden oder die jeweilige Kommune hier äquivalent unterstützt.
Trotz der objektiven Fakten zu den tatsächlichen ökonomisch-ökologischen Realitäten der unterschiedlichen
Antriebsoptionen heute und mittelfristig ist allgemein ein massiver politischer Druck auf die Verkehrsunternehmen zu
erkennen, eine Elektromobilität im ÖPNV zu etablieren. Deshalb wählen aktuell viele Verkehrsunternehmen den
pragmatischen Weg, aufgrund der aktuell massiven Förderung zunächst einige wenige Elektrobusse sowie die
Energiezuführungsinfrastruktur zu beschaffen, so lange die Förderanreize bestehen, um einerseits den
Anforderungen der Politik Genüge zu tun und andererseits trotz der noch hochsignifikant reduzierten Verfügbarkeit
und Einschränkungen des Systems „Elektromobilität“ (Elektrobusse und Ladungsinfrastruktur) dann mit den
vorhandenen Dieselbussen den Betriebsablauf sicherzustellen. Vielfach werden diese Initiativen als
„Marketingmaßnahme“ verbucht, so dass der Autor der Politik zukünftig wieder zu einer
technikoffenen/technologieneutralen Förderung auf der Basis von Wirkvorschriften rät.
REFERENCES
[1] Pütz, R.: "Modell zur ökologischen und ökonomischen Analyse und strategischen Optimierung von
Linienbusflotten", Düsseldorf: Alba Fachverlag, 2010. [Model for the ecological and economic analysis and
strategic optimisation of public service bus fleets],
[2] Pütz, R.: "Einführung in die Linienbustechnik", Köln: Alba Fachverlag, 2012. [Introduction to public service bus
technology],
[3] Bergk, F., Lambrecht, U., Pütz, R., Landinger, H.: "Analyse der Lebenszykluskosten von Hybrid-
Oberleitungsbussen", Heidelberg, IFEU/BELICON, 2015 [Analysis of the lifecycle costs of hybrid overhead line
buses].
24
Kragujevac, Serbia
MVM2018-IL3
Miroljub Adžić1
FUEL CELLS AND ELECTRIC VEHICLES
ABSTRACT: The paper is concerned with fuel cell and hydrogen technologies and their application in electric
vehicles. Beside positive energy balance effects, these technologies can decrease pollutants and carbon dioxide
emissions. Fuel cells and hydrogen are quite mature technologies and already used in transportation systems. There
are still drawbacks that prevent wider usage of fuel cell and hydrogen: current costs of electrical vehicles, fuel cells
and hydrogen, hydrogen storage, very poor hydrogen infrastructure, lack of public awareness that hydrogen is a safe
technology, already developed and applied battery technologies and political issues. Regarding the applicability in
light vehicles, proton exchange membrane fuel cells together with hydrogen show good performance. It would not be
a surprise that in the future hydrogen might affect the geo-strategy concepts based on fossil fuels.
KEYWORDS: fuel cells, electric vehicles, hydrogen
INTRODUCTION
The main conclusion of Fuel Cell Seminar in 1992, which was supported by US Department of Energy, Commission
of European Communities and Japan Fuel Cell Development Center was: „Fuel cells represent a technology which
could potentially replace internal combustion engines in all areas of ground transportation in the 21st century“[1]. In
the meantime, although never fully supported, fuel cells have reached maturity, side by side with internal combustion
engines, when it comes to power performance and reliability. One should keep in mind that these two devices
substantially differ: fuel cells deliver electric, while engines deliver mechanical power. It appears that the long waited
good news for fuel cells happened in January 2017 at the World Economic Forum in Davos, when thirteen word
leading energy, transport and industry companies have launched a global initiative by establishing Hydrogen Council
Coalition with the ambition to “Accelerate their significant investment in the development and commercialization of
the hydrogen and fuel cell sectors as part of the future energy mix with appropriate policies and supporting schemes.”
The Coalition founders are Air Liquide, Alstom, Anglo American, BMW GROUP, Daimler, ENGIE, Honda, Hyundai
Motor, Kawasaki, Royal Dutch Shell, The Linde Group, Total and Toyota [2]. In one year, a number of other world
leading companies joined the Council.
FUEL CELLS
Fuel cells are electrochemical devices that are capable of directly converting chemical energy of a fuel into electrical
energy. Although the fuel cell principle was discovered almost 200 years ago and a practical device is relatively
simple, there are still problems not easy to solve and a room to improve the performance.
What are the differences between combustion of a fuel or its utilization in a fuel cell? A combustion is a process in
which fuel and oxidant are brought into state of a complex mixture of radicals, atoms and molecules that react and
recombine to combustion products. During the combustion process chemical energy of fuel converts into heat and
work. Combustion in real devices mainly happens in a turbulent flow regimes. If we speak about gaseous, premixed
combustion, depending on the turbulence intensity and scales, combustion may proceed as a distinctive laminar, two
dimensional, thin flame front, or as three dimensional process [3]. Unlike to combustion, fuel and oxidant do not meet
as such in fuel cells, Figure 1. They are physically separated by an electrolyte, a substance that must be an insulator
1
Miroljub Adžić, University of Belgrade Faculty of Mechanical Engineering, K. Marije 16, 11000 Beograd,
mikce2001@gmail.com
25
for electrons, good ion conductor and impermeable for gases. Hydrogen, the basic fuel for fuel cells, is ionized to a
proton and an electron: H → H+ + e-. The proton is transported through electrolyte to react with oxygen, the electron
is picked up by an electrode (anode) and transported into an electric circuit. By transporting ions and electrons electric
current is generated and electric work is being done. The energy balance of the process is on the account of chemical
energy of fuel which is converted into electrical work and to less extent into heat. There are different types of fuel
cells, classified on different grounds, but it is mostly common to be classified by the type of electrolyte. However, the
operating principle is similar for all types. For automotive use proton exchange membrane fuel cells (PEMFC) are
the best suited. Keeping in mind the aim of this paper, only hydrogen and PEMFC are considered. This will not affect
the general thermo and electrochemical analysis of fuel cell operation. A more detailed schematics of a proton
exchange membrane fuel cell is shown in Figure 2. The PEMFC electrolyte is a sulfonated tetrafluoroethylene based
fluoropolymer-copolymer, in the form of a thin plastic membrane. Between the electrode and the electrolyte
(membrane) there are two layers: a gas diffusion layer (GDL), which enables diffusion of gases and transport of
electrons, and a catalyst layer, a carbon supported platinum catalyst where main reactions take place. The gaseous
fuel (hydrogen) continuously flows over the anode while the oxidant (air) over to the cathode. The H2 molecules pass
through GDL and adsorb onto the catalyst layer where breaking of the H-H bond happens and atomic hydrogen
adsorbs onto the surface. In the presence of H2O the adsorbed hydrogen desorbs as proton (H+) while the liberated
electron (e-) is conducted to anode and further into the external electric circuit. The transport of protons H+ has to
overcome chemical and electrostatic field energy barriers, the sum of which is net energy barrier, as shown in Figure
3. Protons are transported through the electrolyte where at the cathode catalyst layer two protons react with oxygen
and with two electrons coming via the closed external electric circle, doing electric work, producing only H2O and
heat [4-6]. The overall reaction mechanism is complex, still not completely understood. The role of a platinum based
catalyst is crucial. It substantially lowers the activation energies of reactions and accelerates the rate controlling
steps.
Figure 1 Schematic of fuel cell
Figure 2 Schematic of the fuel cell process
26
Figure 3 Schematic of fuel cell membrane-cathode energy field
Fuel cells can be fuelled by any fuel, but presence of carbon, sulphur and some other components can pose very
serious problem depending on the fuel cell type. For that reason hydrogen is considered the ultimate fuel for fuel
cells. One should notice that contrary to combustion, which is a three dimensional process, a fuel cell operates as a
surface based process. This is one of the main reasons why combustion can deliver hundreds of megawatts of
thermal power per unit volume of combustor while fuel cells are for orders of magnitude less effective.
Thermo and Electrochemistry of a Fuel Cell

We consider a fuel cell a reversible isothermal open thermodynamic system, with steady flow, negligible changes of
potential and kinetic energy. By combining the first and the second law of thermodynamics we get
TdS = dU + pdV + δA = dH + δA, J/mol (1)
where T stands for absolute temperature, S entropy, U internal energy, V volume, A any other work except pdV and
H enthalpy. From this equation follows that the maximum available work a system can deliver, δA, is
δA = TdS - dH (2)
If we introduce the Gibbs free energy (available work) function ΔG term
dG = - δA = dH – TdS (3)
and if only the electric forces do the work, then the work done by a system is
ΔA = - ΔG = Eit (4)
where E is electric potential (voltage) delivered by a fuel cell, i is electric current flowing during period of time t. From
(4) the electric potential E is
E = - ΔG/it (5)
The equation (5) shows that a fuel cell potential depends only on change of Gibbs energy, which is a unique
characteristic of each fuel-oxidant system. In general, the potential is about 1 V.
Let us explain the operation of fuel cells fuelled by hydrogen. The overall chemical reaction is
H2 + 1/2 O2 = H2O (6)
The process starts with ionization of a hydrogen molecule
H2 = 2H+ + 2e- (7)
One mol of hydrogen (6.023 1023 molecules) is ionized to 2*6.023 1023 = 12.046 1023 protons and 12.046 1023
electrons. The charge of an electron is 1.602 10-19 C (coulombs), therefore the total charge q of free electrons is 192
970 C/mol. Keeping in mind that electric current is defined as a flow of electric charge q per time, i = q/t, we have
E = - ΔG/it = - ΔG/it = - ΔG/q (8)

27
As the change of Gibbs energy, ΔG for hydrogen-oxygen reaction (at T = 298 K, atmospheric pressure and water in
liquid state) is – 237 245 J/mole, the electric potential of a fuel cell is
E = - ΔG/q = 237 245/192 970 = 1.229, J/C or V
We should keep in mind that the value of calculated potential is theoretical one, at electrostatic conditions. Fuel cells
appear to deliver electric energy in an elegant and efficient way. In practice a fuel cell faces a number of complex
problems. With the onset of electric current the voltage decreases due to the onset of activation polarization (voltage
losses by overcoming the activation energy needed to transport ions through the established electro-chemical field
at the surfaces of electrodes), ohmic (Joule effect – heat generated by current flow) and mass transport losses
(consumption of reactants induces concentration gradients which affect the fuel cell voltage as the supply of reactants
is governed by Fick’s law of molecular diffusion). Typical voltage and power curves of hydrogen fuelled PEMFC are
shown in Figure 4 [7].
Figure 4 Typical voltage and power curves of an actual PEMFC fuel cell as a function of current density per
geometrical unit area of membrane
The decrease of voltage with current are actually losses due to the kinetic effects of polarization activation (needs
time to free the electrode surfaces from already occupied sites to take place for adsorption of new ions), ohmic (Joule
loses Ri2) and mass transfer losses (Fick’s law, Ddc/dx, decrease of concentration gradient).
The absolute (theoretical) efficiency of a fuel cell η is defined as the ratio between maximum available work ΔG and
the heat of combustion ΔH of a fuel
η = ΔG/ΔH (9)
For hydrogen-oxygen and liquid H2O, η = 237 245/285 830 = 0.83; for water in gaseous state, η = 0.945; for methane
and liquid water, η = 0.92. As one can see the absolute efficiency depends only on fuel used, not on fuel cell type. In
practice the efficiency is significantly lower. It is interesting to compare absolute (theoretical) efficiency of a fuel cell
hydrogen fuelled, and the Carnot cycle efficiency, Figure 5.
Figure 5 Comparison of fuel cell and Carnot cycle efficiency at lower temperature 298 K, [11]
28
Only at high temperatures over about 1000 K Carnot efficiency becomes superior to that of fuel cells. Keeping in
mind that PEMFC operates at temperatures of about 100 K, its superiority over efficiency of Carnot is evident.
The electrical efficiency ηe of a fuel cell is defined as the ratio of electrical power generated and the power of
combustion based on the upper heating value of fuel used
ηe = Ei/ΔH𝑚̇ (10)
where ΔH is the higher heat of combustion of fuel and ṁ is the fuel mass flow rate (kg/s). The electrical efficiency of
fuel cell stacks can reach 0.55 – 0.60 when hydrogen fueled.
As just shown, fuel cells are low voltage DC devices with the voltage output of about 1 Volt. For practical purposes,
in order to increase the voltage and power, individual fuel cells are stacked in series, stacks, Fig. 6. A stack compared
with unit cell is generally far more difficult to operate, because of more complicated mass, heat, liquid water and
electric current management resulting in the decrease of efficiency, reliability and life of fuel cells.
Figure 6 Ballard fuel cell stack
The theoretical effects of temperature and pressure on fuel cell voltage can be deduced from eq. 8 where E = - ΔG/q
and from δ (ΔG)/δp = ΔV, where ΔV is the volume change of particular reaction. One should keep in mind that with
the increase of temperature, ΔG increases (becomes less negative) but that affects the voltage of the order of
millivolts, while the change of volume, e.g. for hydrogen fuel cell, is large, therefore the effect of pressure on voltage
is evident. Nevertheless, in practice, these effects are much more complicated due to the effects of stack design,
size, operating temperature and pressure, heat and mass transfer management and others.
It is worth noting that an electric battery is also an electrochemical electric energy source, but contrary to a fuel cell
that can theoretically run for any period of time, as long as fuel and oxygen are supplied, a battery is just an energy
storage with limited capacity.
Fuel Cell Performance

The technology of PEMFCs is fully matured and in some elements better or in line with other systems like internal
combustion engines and batteries. It is a question of performance, cost, primary energy sources, infrastructure and
a policy to what extent PEMFC will be used in electric vehicles [7-14]. The brief visual comparison of some significant
performance, based on the status in 2015/2017 are given in Figures 7-10 and Table 1. (The performance of fuel
cells, motors and batteries, used in this paper, are the manufacturers’ published data, as given). The efficiency of
PEMFC (ratio between FC electric power and the chemical energy flow) including expectations in the future [5], and
the last achievements in thermal efficiency of the Toyota gasoline engine is shown in Figure 7. The supremacy of
fuel cell is impressive. Regarding mass to power ratio, life and startup time, Table 1 [5], as we can see, in next 10-
15 years the mass to power ratio of PEMFC is planned to decrease three times, the life time to increase almost three
times, and startup time to decrease for an order of magnitude. Even now the startup time is quite acceptable for
transportation vehicles, two minutes. Comparison of mass to power ratio of PEMFC, internal combustion engine,
electric drive and turboprop, as a function of power, is shown in Figures 8 and 9 [11].
29
Figure 7 Efficiency of PEMFC and Toyota gasoline engine (2018)
Table 1 Performance indicators of PEMFC

Mass / Power ratio, Lifetime, h Startup time, for a stack,
PEMFC fuel cell kg/kW seconds
Short term (5-10 years) 0.5-0.2 20000 120
Mid-term (10 to 15 years) 0.25 – 0.17 40000 60
Long term ( >>15 years) < 0.15 50000 10
Figure 8 Mass to power ratio of PEMFCs (Nedstack, Ballard), internal combustion engine (Mini Cooper S-BMW),
electric drive and turboprop engine (Pratt and Whitney 120)
As expected, the specific mass decreases with the power increase. It is worth noting that some of the Ballard fuel
cells are at least as good as internal combustion engines and electro drives, but the recent efforts of Toyota resulted
in very impressive results. Toyota managed the technological leap in a couple of years. The mass to power and
volumetric power ratios of the Toyota PEMFC are the same as modern internal combustion engines and electric
drives. In conclusion: the PEMFC mass-volume-power performance have reached the level of internal combustion
engines and electric drives.
30
Figure 9 Volume to power ratio of PEMFC (Nedstack, Ballard), internal combustion engine (Mini Cooper S-BMW),
electric drive and turbo-shaft engine (Pratt and Whitney 120)
The cost of PEMFC system primarily depends on number of units produced. Prediction of PEMFC cost in 2017 is
given in Figure 10 [5]. It is interesting that the effect of year of production is rather small. It is expected that in 2025
the cost will be about 10% lower. At the moment the Toyota fuel stack cost is about $ 11,000 for annual production
of 3,000 units. It is expected that with annual production of 30,000 the cost will drop to $ 8,000.
Figure 10 An 80 kW fuel cell power system for transportation prediction (2015), as a function of number of sold
units
HYDROGEN
A chemical energy source to run a fuel cell can be any fuel, but hydrogen is far the most appropriate. Some important
properties of hydrogen are given in Table 2 [5].
Table 2 Properties of hydrogen

Gaseous hydrogen
Density (kg/m3 at 273 K, 1.013 bar) 0.089
Lower heating value (MJ/kg) 120.0
Lower heating value (MJ/m3) 10.68
Stoichiometric air fuel ratio (m3/m3) 2.38
Liquid hydrogen
Density (kg/m3at 1.013 bar and boiling temp. 20.3 K) 70.9
Critical temperature (K) 33
Although the main constituent of space, free hydrogen is scarce in our environment and has to be produced. About
96% of world hydrogen production is via fossil fuels, about 50% from natural gas, 30% from oil and 18% from coal,
by different technologies. It should be stressed that the production of one kg of hydrogen from fossil fuels evolves
about 11 kg of CO2. New important technologies like the microorganisms and algae production of hydrogen from
biomass with decentralized production from biomass become available. Most of hydrogen is used as raw material in
fertilizer, chemical, oil, metal, food, pharmaceutical and glass industries. World consumption of hydrogen in 2016
31
was about 85 million tons, growing steadily. New developments are in the areas of innovative water electrolysers that
can provide low price hydrogen production based on polymer-membrane and solid oxide fuel cell technologies. The
attractive idea to get hydrogen from natural gas reforming and use in low temperature fuel cells like PEMFC has
serious drawback: platinum based electrode catalysts can be poisoned by impurities, such as carbon monoxide,
which is a natural constituent of the reformate. Besides, the very concept of fuel cells’ low impact on the environment
is jeopardized by high CO2 emission when using hydrogen produced from fossil fuels. Hydrogen for fuel cells ought
to be produced by lower impact systems, e.g., wind, solar or nuclear energy water electrolysis. At a glance, it does
not look reasonable to produce electric energy by any method, use it to get hydrogen by water electrolysis and use
it in a fuel cell to get electric energy again. The point is the energy accumulation effect of hydrogen storage that can
help solving intermittency, fast fluctuations and surplus of wind and solar energies which can disrupt the electric grid
balance and planning, as power plants cannot quickly adjust their production and balance total energy supply and
demand. We have recently witnessed unimaginable situation when German producers of electric power paid some
consumers to switch on and use electric utilities. And of course, the intentional production of hydrogen to be used in
fuel cells and other purposes, makes clear sense. One should also notice as well a very important fact that by this
approach the total water balance in nature is zero. The main problems when using hydrogen in fuel cells or other
energy production devices are low volumetric heat of combustion, rather poor storage properties, lack of
infrastructure, cost, public stereotype that hydrogen is dangerous and lack of internationally homogenized standards
and regulations.
Germany could be a good example of correlation between impressive increase of production of solar and wind based
electric energy and the support of electric vehicles, hydrogen and fuel cells, Figure 11[15]. The situation in Japan is
even more positive towards use of hydrogen.
Figure 11 Wind and solar electric energy production in Germany during last five years
Hydrogen Storage
There is a number of hydrogen storage methods. From compressed gas cylinders, low temperature high pressure
liquid, to adsorption over new exotic materials. For passenger vehicles the standard compressed hydrogen pressure
is 70 MPa, while 35 MPa is for light and heavy trucks. There have been many efforts on hydrogen storage
optimization, development of pressurized cylinders, valves, pressure regulators, pressure and temperature sensors
and also high efficiency hydrogen compressors. Beside classical systems, the innovative approach uses
electrochemical compressors with efficiency of 4kWh/kgH2 at 70 MPa. Table 3 presents some new storage systems
[5].
Table 3 New hydrogen storage systems

Materials Energy storage
Nanotubes physisorption (high pressure, low temperature) 13.2 MJ/kg
Nanotubes 6 MJ/l
Graphene chemisorption 9.6 MJ/kg
Polymer matrix 10.8 MJ/kg
Hollow Li20B60 cage 9.6 MJ/kg
The serious drawbacks of the new storage systems are as follows:

 Only laboratory proved
 Expensive
 High pressure, low temperature needed
 Low rates of sorption/desorption.
32
Hydrogen in a liquid state was abandoned, except for rocket propulsion, due to extreme conditions to fulfil very low
temperature of 20 K and high pressure. The energy storage of hydrogen storage systems capacities by mass and
volume are shown in Figure 12 [11].
Figure 12 Energy storage capacities by mass and volume of fuels, hydrogen and hydrogen storage systems
It can be seen that the oil based fuels make good compromise between volumetric and mass energy storage,
hydrogen has excellent mass but poor volumetric capacities. In general, hydrogen storage systems have low storage
capacities either by mass or volume.
About the hydrogen cost. Due to the political commitment and hydrogen technology production developments, the
hydrogen price shows constant decrease and it is expected that in the next decade the price will be close to the price
of liquid hydrocarbon fuels for transport, see Figure 13 [5]. The EU goal is the retail price at filing stations of 5.5 €/kg,
or 0.046 €/MJ, thus approaching the 95 RON gasoline retail price in Germany of 1.49 €/l or 0.044 €/MJ, in September
2018. The retail price of hydrogen in August 2018 in Japan was 0.064 €/MJ.
Figure 13 Centralized mass production of hydrogen, price prediction in USA for two standard storage pressures
The main obstacle for wider use of fuel cell vehicles is the lack of hydrogen infrastructure, number and distribution of
hydrogen filling stations. In June 2018 there were less than 300 hydrogen filling stations worldwide [5]: 93 in Japan,
44 Germany, 39 USA, 12 South Korea, 11 England, 10 Denmark, 9 Norway, 9 in France and less than that in other
countries. The good side is that hydrogen tank filling up is fast. It takes a couple of minutes.
ELECTRIC BATTERIES
Electric batteries is proven and common energy storage technology. They are easy to use and operate. The
drawbacks are low specific storage capaticies per mass and volume, price, and charging performance: either it takes
hours or high power charging systems have to be used. Besides, fast charging induces stresses and can lower the
discharge capacity of a battery. At the moment two types of batteries are of main interest for electric vehicles: nickel-
33
metal hydride (Ni-MH) and lithium-ion (Li-ion) batteries. The general performance indicators of electric batteries are
presented in Table 4 [5]. As one can see, the mass to power ratio of batteries is planned to decrease three to six
times, the lifetime is expected to reach over 1500 cycles in next 15 years.
Table 4 Performance indicators of batteries

Mass/energy Gasoline Hydrogen, kg/kWh Toyota Mirai, Lifetime, cycles
Batteries stored ratio, kg/kWh 700 bar cylinder
kg/kWh
Short term (5-10 years) 6-3 > 1500 cycles;
Mid-term (10 to 15 years) 3–2 0.08 0.53 capacity drop
Long term ( >>15 years) 2-1 < 30%
The capital cost for Li-ion battery pack per energy stored was about 400 US $/kWh, in 2015, with tendency to drop.
In 2017 it was about 250 $/kWh. Further drop is expected to slow down with the expected cots in 2030 of about 100
$/kWh, Figure 14 [5].
Figure 14 Electric battery cost
ELECTRIC VEHICLES
The concept of an electric vehicle is more than 200 years old. Electric vehicles have had ups and downs throughout
the history. The first serious success in modern era, 20 years ago, comes from Toyota Prius hybrid electric vehicle
equipped with nickel metal hydride battery. Later comes Tesla motors, and now there is practically no major car
company without an electric vehicle in its production program. Regarding the propulsion systems, there is a steady
progress in battery technology. The lithium-ion batteries have bloomed in last two decades. Electric motors and
components have also been significantly improved. Also, the mass, volume, power, reliability, life of fuel cell. The
electric propulsion interest spreads also to air traffic, (major manufacturers Boeing and Airbus), marine, and
transportation in general. The plug-in, plug-in hybrid, fuel cell vehicle technologies have been launched. Public
electric vehicle chargers have been installed worldwide. The electric vehicle cost decreased and the environment
has become much more supportive: free electric charging, free parking places, open yellow lanes, open access to
otherwise forbidden areas in city centers and others. Although the drawbacks are serious, mass to storage ratio, low
charging rates and cost, electric vehicles have become commercialized from all major and some specialized
manufacturers. The variety and combinations of propulsion systems, internal combustion engines, batteries, fuel
cells, controls, fuels, fuel storage, drive components, energy management, materials, improvements and
breakthroughs, costs and others is tremendous and very complex, making overall comparison of different types of
vehicles difficult and questionable. Having that in mind and the aim of this paper to deliver a limited overview of fuel
cell and hydrogen technologies and electric vehicles, this paper compares the cost of similar performance vehicles
with different power systems manufactured by the same manufacturer, in this case Toyota, as an indicator of the
technologies prospective, Figures 15-17 and Table 5.
34
Figure 15 Toyota Mirai and hydrogen filling station
Figure 16 Toyota Mirai main components
Figure 17 Toyota Mirai powertrain
The Toyota Mirai is a midsize luxury 4 door sedan car. The powertrain consists of 113 kW, 335 Nm electric motor,
PEMF fuel cell (370 cells) with maximum power output 114 kW, 0.50 kg/kW, 0.32 l/kW, two 70 Pa pressure hydrogen
cylinders with total volume of 122.4 l and 84 kg mass, hydrogen storage of 4.8 kg or 160 kWh, 1.6 kWh nickel-metal
hydride battery nominal voltage 244 V, capacity 6.5 Ah, the vehicle kerb weight is 1,850 kg, range about 500 km
(planned to reach 1000 km in 2025!), max speed 178 km/h, acceleration 0- 100 km/h in 9.6 seconds, refueling time
3-5 minutes. The battery assists fuel cell stack during acceleration and stores recovered energy during deceleration.
The Mirai fuel cell emission is water vapour, cost on the road $ 57,500.
Table 5 Cost of Toyota vehicles

Type Cost on the road $
Avensis - (internal combustion engine; gasoline) 30.200
Prius Hybrid - (battery and internal combustion engine; gasoline) 34.800
Prius Plug-in Hybrid - (battery and internal combustion engine; gasoline) 40.5001
Mirai - (fuel cell and battery; hydrogen) 57.5001
1
Total sales of Prius plug-in hybrid was about 128,000 and 5,300 Mirai vehicles in 2017
35
Compared with the similar performance Toyota classical and electrical cars, Table 5, Mirai is expensive. With rise of
production units, the cost is expected to decrease. An analysis shows that the cost of Mirai fuel cell of $ 11,000 could
drop to $ 8,000 with increased production. Mirai is a hydrogen fuelled environment’s best friend, which is an
expensive position. Mirai is not a prestige car like Tesla S, a must for rich people, a matter of prestige, but a product
of responsible national strategy policy, and vision of national future.
CONCLUSIONS
 Fuel cells and hydrogen are matured technologies of equal performance (power, mass, volume, reliability,
life) with internal combustion engines and electric drives, and already used for propulsion of vehicles.
 Fuel cells are superior to internal combustion engines regarding efficiency. When fuelled by hydrogen, fuel
cells have excellent pollution and CO2 emission performance.
 The main obstacles for wider usage are cost, hydrogen price and undeveloped hydrogen infrastructure.
 Breakthroughs in fuel cell and battery technologies are possible and expected which is not the case with
hydrogen storage technologies.
 The cases of Germany and particularly Japan, show signs of accelerated and more aggressive utilization of
hydrogen, meaning changes in strategy of national energy and geostrategic policies in the coming decades.
 Only time will tell what vehicle propulsion systems will prevail in the future.
REFERENCES
[1] The 1992 Fuel Cell Seminar, Tucson USA, 1992,
[2] Hydrogen Council, Available from: http://hydrogencouncil.com/, Accessed 25.06.2018.,
[3] Mahjoub, M. A., Milivojević, A., Adžić, V., Živković, M., Fotev, V, Adžić, M.: "Numerical Analysis of Lean Premixed
Combustor Fueled by Propane-Hydrogen Mixture", Thermal Science, Vol. 21, No. 6A, pp 2593-2602, 2017,
[4] Fuel Cell Handbook, EG and G Technical Services, Morgantown, DOE, 2014,
[5] US Dept. of Energy, Energy Efficiency and Renewable Energy, 2017,
[6] Internationa Energy Agency, Hydrogen Production and Storage, OECD, Paris, 2006,
[7] Adzic, M., Fotev, V., Miivojevic, A., Adzic, V., in the FP7, Efficient Use of Resources in Energy Converting
Applications – PEMFC project, EURECA 2012-2015,
[8] Alaswad, A., Baroutaji, A., Achour, H., Carton, J.: "Developments in fuel cell technologies in the transport
sector", International Journal of Hydrogen Energy, Vol. 41, Issue 37, 2016, pp 16499-16508,
[9] Pettersson, L., Westerholm, R.: "State of the art of multi-fuel reformers for fuel cell vehicles: problem identification
and research needs", International Journal of Hydrogen Energy, Vol. 26, Issue 3, 2001, pp 243-264,
[10] Wang, Y, Dai, K., You, H., Gao, M.: "Research on the design of hydrogen supply system of 70 MPa hydrogen
storage cylinder for vehicles", International Journal of Hydrogen Energy, Available online, September 2018,
[11] Adzic, M.: "Hydrogen and Fuel Cells for Aircraft Propulsion", Airbus Meeting, Munich, 2015,
[12] Alavi, F., Lee, E., de Wouw, N., de Schutter, B., Lukszo, Z.: "Fuel cell cars in a microgrid for synergies between
hydrogen and electricity networks" Applied Energy Journal, Vol. 192, 2017, pp 296-304,
[13] Trade car view, Available from: https://www.tradecarview.com/specifications/toyota/, Accessed 25.06.2018.,
[14] Simons, A., Bauer, C.: "The production and end-of-life processes for current and future proton exchange
membrane fuel cell (PEMFC)", Applied Energy Journal, Vol. 157, 2015, pp 884-896,
[15] Eurostat, https://ec.europa.eu/eurostat.
36
Kragujevac, Serbia
MVM2018-IL4
Josip Vlahović1
TURBULENT DEVELOPMENT OF BODY CONCEPTS,

MATERIALS AND JOINING PROCESSES IN THE LAST TWO
DECADES
ABSTRACT: Great changes of the car-body Production concepts in the last few years are the result of a rapid
development of the new materials, new Joining processes and a new body-concepts.
In the last two decades noticeable are big tendencies/aspirations (we can even say fights) for primacy of Aluminum.
That made it possible a many of new this competition resulted in development of forming methods and many joining
techniques which were developed only for aluminum. They are currently in production of about 30 Aluminum-cars.
On the other hand, today has the Aluminum every day more and more share in the new steel car-body-structures.
Following the model of the Aluminum structures development, there were also attempts to develop similar
technological methods for the steel as well. This attempt was not very successful. The real progress has been made
with a development of steel specific technologies.
It can be said that the advanced aluminum technologies developed in the 1990s, have forced the rapid development
of steel technologies as well primarily new Technologies for steel-sheet. It has resulted that the steel relieved, and
will be again a cost-effective car-body technology.
KEYWORDS: steel, aluminum, Mg, plastic, carbon, car-body-structure (body-in-white), hang-on parts, sheet metal
construction, mixed (hybrid) construction, SpaceFrame construction, monocoque, Tailoredblanks, Tailoredtubes,
Strips, warm and cold jojning processes, welding, brazing, joining, rivets, Clinch-spots, Flow Drill Screws, Tack
impact, FSW - Friction Stir Welding
INTRODUCTION
Development of lightweight bodywork, or to say development of passenger car-body-structures, was in the focus of
my work within the last 30 years. This can be seen in chart on the Figure 1. Here you can see, at a glance, what
happened since the production of the first series aluminum cars in 1903. After a long time of unchanged production
of high-volume cars with steel-sheet-metal car-bodies and rare models of aluminum-cars, in 1990s appeared many
new technologies, especially for aluminum, which have greatly influenced the way of further development of the car-
body-structure. This brought, in the long-standing unchanged steel-sheet-metal production, discomfort and has
opened many questions. It caused great changes of the car-body Production concepts. A rapid development of the
new materials and new Joining processes has especially supported the rapid development of the production of
aluminum cars. It is very interesting that these developments, which went in different conceptual directions, caused
big concern for the conventional steel vehicle bodies´ manufacturers.
It can be said that the advanced aluminum technologies, developed in the 1990s, have forced the further development
of steel technologies as well! Following the model of Aluminum development, the steel bodies manufacturers
attempted to develop similar hi-tech methods. This has brought no particular progress, because the technical
characteristics of the Aluminum and steel are very different. These improvements can be achieved only through
development of steel specific technologies. This challenge resulted in development of so many new appropriate
Technologies for steel-sheet cars-body. The results was that the steel vehicles production relieved and will be again
a cost-effective option for steel-car-body technologies.
1
Dr. Josip Vlahovic, Germany, Freiberg a/N
37
Figure 1 Car Body lightweight history
At the moment, there are in production about 30 Aluminum-cars-bodies. In addition to that the aluminum as a material
has every day more and more share in new steel and carbon car-body-structures.
It can be said too, there are less and less of the real 100% “Al cars”. The Aluminum-cars body-structure get steel
parts in critical body areas: A/B-pillar and body side structure. The fact is too, that the aluminum-cars-body are being
built more and more in hybrid construction method.
It is the fact too, that the so-called „steel cars“, are not all made of 100% steel anymore. Large number of new steel
vehicles, constantly contain more other materials, mainly aluminum, especially in the crash and the spring connection
structures.
One-material car-body technologies are slowly loosing dominance!
DEVELOPMENT STATUS OF CAR-BODY PRODUCTION - current status in

the year 2017
Total car production per year
Monitoring the production of passenger cars began in the 60's of the last century. In 1969, for example, 22,752 million
vehicles were produced. At that time, the total number of passenger vehicles on the road was about 168.7 million.
Ten years later the production has reached about 30 million, until 1999 when the production of passenger vehicles
reached the figure of 40 million. In the previous year, this number was over 81 million. Drastic growth has been seen
since 2005, when China produced 15.25 million vehicles, and in 2017, China produced as much as 24.8 mio with a
tendency of growth. The total number of vehicles running in 2015 totalled 947 million, and today this number exceeds
one billion. There are different sources stating that the total number of vehicles vary from 1.069 to 1.2 billion. It is well
known that from the beginning of the production of passenger cars between 1886 and 2014, some 2.552 billion were
produced. This means that 28.8% of them are still on the road.
38
What kind of technology we have used in the past?
Developer level techniques, the general economic conditions and the expectation of customers were determining the
factors of former mass-produced production of passenger vehicles. It was a simple, reliable and profitable production,
which can be generally and briefly described as:
 one construction method - steel-sheet construction
 one type of steel - steel-sheet with different wall thicknesses
 two methods of connection - Spot and MAG welding.
However, the new considerably stricter legal regulations, new environmental requirements and new more demanding
customer requirements, caused significant changes in the production of passenger cars.
What about the current technologies used in car body production today?
As noted above, today we have one complex production of passenger car bodywork in which various materials and
methodes have been applied, which result in lighter, safer and more reliable passenger vehicles.
These are the basic features of the modern construction of passenger vehicles-car-body:
Use of different Body concepts:
 conventional sheet metal construction,
 SpaceFrame (profile, cast, sheet)
 mixed construction: sheet-metal, SpaceFrame, monocoque
Use of different materials:
 Steel, Al, Mg, carbon and other plastic
Use of different joining technologies:
 different warm and cold joining methods
DEVELOPMENT OF CAR-BODY PRODUCTION TECHNOLOGIES OVERVIEW
Long time stable car body production

It can be said that since the beginning of the 50s of the last century, when the production of self-supporting car-body
began, the production was stable for almost 60 years with little conceptual and technological changes. It was based
on the basic principle of a self-supporting bodywork, made of steel sheet alone. It was the only construction method,
especially appreciated because it is a very fast, cheap, reliable and profitable method due to the extensive
automation.
Figure 2 presents the main milestones of technological development of conceptual engineering car-body.
These are important technical and technological moments which influenced the entire change in the production of
the car-body:
 suddenly many Al-cars in mass production
 space frame new construction method
 new different steel classes
 new different joining technologies
 new steel sheet construction methods.
Suddenly many Al-cars in mass production

Since 1903. until 1987. there was totally 55 different Al-cars in production. Basically, they were small symbolic series
until 1953 when Panhard Dyna Z1 in total production from 1953 to 1956 produced 39,460 exemplars.
Then there was a small break in the production and development of Al-cars, until 1987. That year, appeared Treser
a new unusual Al-car, which was not only in aluminium, but also applied a new concept of bodybuilding. It was the
first that implemented the Space Frame concept in a very small series. All that remained without much of attention
until 1984, when the mass production of the Audi A8 began, that also implemented the Space Frame concept (Figure
3 and Figure 4).
It was the first sign that something new was happening in the car-body production. From 1984 to 2000, 11 Al-cars
came into the series. It was only one of the emerging signals for steel car manufacturers that they also have to
change something in the long-standing “Long Time stable car body production”.
39
Figure 2 Development of car-body production technologies overview
Space frame as new car body construction method

With the Audi A8, began mass deployment of this new concept of construction of car-bodys. About this concept given
in the 90s a lot of talking, researching and experimenting occured. This Al-car-body construction concept is simply
described by the following definition:
„Space Frame supporting structures are thin-walled spatial structure, whose elements (straight, 2D or 3D curved)
have closed sections and in which the nodes of the elements consistently spatially interconnected“ [3].
It is clear that this principle of car-body design has “lightweight” as a leitmotif and as a result of years of work on this
topic the following definition was created:
„ Lightweight design is a balanced synergy between shape and mass in the available space“ [3].
Today, after almost a quarter of a century of applying this principle, there are visible changes and can be said that
slowly the Space Frame car-body concept further developed in the direction of "Multi material" design.
Figure 3 Number of Al-cars in production
40
Figure 4 Aluminum cars in production 1953 – 2017
New different steel classes

It is clear that after major changes in the development of the concept of car-body construction and application of
Aluminum as a big competitor to steel, there was a need to make improvements in the quality and the types of steel
used as well as in the field of construction method of car-body with steel. So, in the 1990's many new types of steel
have been developed, improved both in strength and in the way of forming process.
Figure 5 Development of High Strength Steels since 1975 [9]
41
New different joining technologies
In the 90's, the development of new "joining" technologies began to appear, especially the so-called "Cold joining
methods". Until then, dominating methods were "warm joining" as "spot welding", which were used in the serial
production of car-bodies since the 30’s of the last century. At that time the self-supporting bodywork method was
patented, based on "steel-sheet" components, as the only profitable, fast and cheap method.
Everything started with the development of the basic "Self-piercing riveting" technologies for Aluminum. Today there
are two types of this technology: "Semi-hollow and Grip Self-piercing riveting". In order to increase the effect of Self-
piercing riveting, since it is like the "Spot welding" method that results in "Discontinuous body structure", new
"Structural adhesive" methods has begun to intensify. So today we have great application of "Self-piercing riveting",
but with pre-glued "connecting flange". This results in a "Continuous body structure" which means that all "structural
parts" have a full length "closed cross sections". Such approach results in a body-structure with much greater
"torsional stiffness". This means that the body structure is much more stable, less vibrating, and ultimately
considerably lighter than the body-structure that does not have an adhesive "connecting flange".
At the same time, other "cold joining methods", the so-called " Different "Clinch-spots" methods, which are used as
auxiliary methods, to position the parts before, for example at "Structural adhesive" or before "Roller connection".
This method also allows the glued parts to hold and tighten and prevent interference and movement of parts and
vibrations.
Interestingly, the old simple "warm joining methods", the so-called "Resisting spot welding" for the traditional principle
of "steel / alu-sheet" body construction are improving as well. The same situation is with already known methods for
steel and aluminum joining, such as MIG, MAG and different Laser welding.
Other "cold joining methods" have also been developed, for example: Friction stir welding FSW (cold), Flow drills
screws, Tack impact, etc.
New steel sheet construction methods

In the late 90’s of the last century many new steel-sheet construction optimization methods have been developed
and perfected so that they can achieve the efficiency of new Al-SpaceFrame constructions. These methods enable
“steel-sheet construction” to be significantly stiffer and and lighter and thus remain the dominant "body construction"
method. A brief overview of these new methods is shown in Figures 6 and 7. All these methods are based on two
simple principles: Tailored & Patch of sheet metal parts.
Figure 6 New steel-sheet construction methods – Tailored Blanks [10]
42
Figure 7 New steel-sheet construction methods – Patchwork [11]
COMPARISON OF DIFFERENT MATERIALS AND JOINING TECHNOLOGIES

This is a theme that 30 years ago did not exist, because we have used only “one material and one technology”.
Today's situation provides a lot more creative possibilities for car-body design and that is why, this rendering as
information is very useful (Figure 8).
Figure 8 Comparison of number of different materials and joining technologies
If we look in the diagram, we can see significant oscillations on the left part (1996-2013) in the number of applied
materials and joining technologies. These oscillations are understandable in the process of searching for the
optimum.
It is evident that this number of applied materials, as well as the number of applied joining technologies, have
stabilized over time.
Talking about an average numbers, one could say that width a new modern car-body (forerunner) there are about
14 different materials and 9 different joining technologies
43
CONCLUSIONS - NEW VIEWPOINTS FOR THE FUTURE
All the facts indicate that the times when cars were made from a single steel sheet of various thicknesses have gone,
meaning a single material and a single technology is a matter of the past. Neither the method of a single material,
introduced with Aluminum, lasted for long. So, the all attempts to produce future cars in new technologies, but from
only one material, belong to the past.
The data has already shown that we do not have pure "Al-body" today, as well as all the less pure “St-body "except
in large-scale daily car manufacturing.
It is clear that new directions are being introduced and that they can be summarized to the following conclusions:
 Most car bodies will still remain made of steel as basic material, but with the introduction of modern
processing technology and new materials
 Medium and high class as well as premium cars will have in principle hybrid construction (As today
already partially exists) with a basic structure of steel or aluminum, with implementation of other materials
and made by introducing modern technologies.
 Exotic sports cars still have the basic structure (as it is to date) from carbon with the addition of other
materials and manufacturing technology.
In general, all new car body structure will be safer and more reliable. Today's body-structure consists of the supreme
structure and crash structure, which theoretically represents the ballast that every car must bare due to the risk of
collision. Maybe it's too early to say, but if the cars are electronically so reliably managed and secured against
collisions with other vehicles and facilities, it might be said that the body structure in the future may be significantly
more relaxed than as it is today, and this means another contribution in order to protect environment.
REFERENCES
[1] Jahreskatalog der AUTOMOBIL REVUE von 1968 bis 2015, HALLWAG Verlag, Bern CH,
[2] Enzyklopädie des Automobils, Weltbild Verlag GmbH Augsburg, Orbis Publishing Ltd, 1990. London,
[3] Vlahovic, J.: “Gestaltung von T-förmigen Knotenpunkten in dünnwandigen Rahmentragwerken”, Dissertation,
RWTH Aachen, 2004, ISBN 3-8322-4033-0,
[4] TRESER, Manufacturer brochures, 1990,
[5] Mayer, H., Audi AG: “Das Karosseriekonzept des neuen A8”, International Circle of Experts on Carbody
Construction, 7/8 Nov. 2002, Bad Nauheim,
[6] Koglin, K., Audi AG: “Neue Fertigungstechnologien für die A8-Karosserie”, International Circle of Experts on
Carbody Construction, 7/8 Nov. 2002, Bad Nauheim,
[7] Der fünfte 5er, Dullinger, M., Kempinger, K.: “Internationaler Rohbau-Expertenkreis”, 22.07.2003, BMW AG –
Dingolfing,
[8] Dullinger, M., Kempinger, K., Der fünfte 5er: “Internationaler Rohbau-Expertenkreis”, 22.07.2003, BMW AG –
Dingolfing,
[9] Bernhard, C.: Department für Metallurgie Montanuniversität Leoben (A), Vortrag an der HTL Waidhofen, 22.
Feb. 2008,
[10] Adam, ThyssenKruppDrauz, Evolution from first Tailored Blank to near net shape, Thyssen Tailored Tube,
Wolfsburg 26-27 Mai 2009,
[11] ArcelorMittal Patchwork 2010, EuroCarBody, Bad Nauheim, Oct. 2010,
[12] Aston Martin - the new Rapide Ian Minards, Andrew Syvret Aston Martin Lagonda, Ltd., GB, EuroCarBody 19-
21 October 2010,
[13] Manabu Ozawa, Atsushi Mikuni, Hideo Satake, EuroCarBody 2010, Global Car Body Benchmarking
Conference,
[14] Lexus the new LFA, Atsushi Mikuni, Manabu Ozawa, Hideo Satake Toyota Motor Corporation, JP, EuroCarBody
19-21 October 2010,
[15] Andreas Fidorra, Jürgen Baur: Audi-The new A8, EuroCarBody 2010, 12th Int. Car Body Benchmark
Conference, Bad Nauheim, 19/21 Oct. 2010,
[16] Manabu Ozawa, Atsushi Mikuni, Hideo Satake, EuroCarBody 2010, Global Car Body Benchmarking Conference
[17] BENTELER Automotive: Technologies & Products of PG Structures, EuroCarBody 2012, Bad Nauheim
[18] Fa. Variform, Official technical brochures, 14th Global EuroCarBody Award 2012, Bad Nauheim, Oct. 2012,
[19] Günther Ast, Michael Trabner, Daimler AG, EuroCarBody 2012, Global Car Body Benchmarking Conference
[20] The all new Range Rover L405, R. Lloyd, G. Movce, S.Black, M. Rickards, JLR, GB, EuroCarBody, 18. Oktober
2012,
[21] ThyssenKrupp, Tailored Strips - Brochures, 14th Global EuroCarBody Award 2012, Bad Nauheim, Oct. 2012,
[22] Günther Ast, Michael Trabner, Daimler AG: The new Mercedes-Benz SL, 14th Global EuroCarBody Award 2012,
Bad Nauheim, Oct. 2012,
44
[23] Lloyd, R., Movce, G., Black, S., Rickards M.: JLR, GB, The all new Range Rover L405, EuroCarBody, 18.
Oktober 2012,
[24] Shawn Tarr, Jared Cox, Honda, EuroCarBody 2016, 18th Global Car Body Benchmarking Conference,
[25] Shawn Tarr, Jared Cox, Honda, EuroCarBody 2016, 18th Global Car Body Benchmarking Conference,
[26] Presented by Petr Šimon & Jiří Nikodem EuroCarBody 2016, 18th Global Car Body Benchmarking Conference,
[27] Petr Šimon & Jiří Nikodem EuroCarBody 2016, 18th Global Car Body Benchmarking Conference,
[28] Tim Hämmerle, Dr. Dominik Hußmann, The newAudi A8, AUDI, EuroCarBody 2017, Bad Nauheim.
45
46
Kragujevac, Serbia
MVM2018-IL5
Dobrivoje Ninković1
CURRENT STATE AND DEVELOPMENT TRENDS IN THE

FIELD OF LARGE DIESEL AND GAS ENGINES
ABSTRACT: Large Diesel and gas engines with power higher than approximately 1 MW play a major role in the
sea and land transport of people and goods, but also in the production of electrical energy in remote areas, for
peak load balancing, and backup systems. While their total installed power and fuel consumption are vastly smaller
than with their road transport counterparts, the vehicles they propel (ships and trains) represent the most efficient
freight transport modes in existence. It is thus their importance for the world economy that has been driving their
development, resulting in these engines belonging to the group of the most efficient thermal machines available to
the mankind.
Presented in the paper is a survey of the current development state of the large Diesel and gas engines in terms of
their thermodynamic performance, emission levels and the means for achieving them, and fuels used. Due to its
indispensability for achieving the high performance of these engines, turbocharging is given a particular treatment
in the paper.
Based on the study of the contemporary literature, an attempt is made to identify main development trends in this
area. With regard to the latter, special attention is devoted to the so-called dual-fuel engines (Diesel and gas),
which appear to have enjoyed a rapid development in the last decade.
KEYWORDS: large engines, diesel engines, gas engines, dual-fuel engines
INTRODUCTION
The year 2017 marked the 125th anniversary of the Rudolf Diesel’s patent on the theory and the realisation method
for combustion engines (in original: “Arbeitsverfahren und Ausführungsart für Verbrennungskraftmaschinen”), a
patent formulated rather broadly, but which in essence dealt with an engine with constant temperature and slow
combustion. The thermodynamic goals that Diesel set for his engine concept in 1892 were bold, but at the same
time held a powerful prophecy for the future: maximum cylinder pressure of 253 bar and an efficiency of 58%; and
that at the time when the dominating prime mover – the steam engine – ran with an efficiency of mere 7%.
The patent brought neither a fortune nor a personal happiness to his inventor, but it can be seen as the starting
point for the rise of an engine branch without which the modern large-scale goods and passenger transport would
have been unthinkable, namely the large engine division.
It definitely took time for this engine branch to set foot and gain momentum, for it wasn’t until 1912 that a four-
stroke Diesel engine was used as the main propulsion set on a high-sea ship, the MV Selandia [1]. However, it was
not a sign of change in the sea transport, for in the Second World War two thirds of the shipping was still driven by
steam, and there were only two capital warships powered exclusively by Diesel engines. The situation started to
change with the first application of the turbocharging to a two-stroke Diesel engine in 1946 [2], partly due to the
“downsizing” of this engine kind brought about by the turbocharging, for turbocharged four-stroke engines were in
use on ships since 1927.
It is now a recognized fact that shipping is the most efficient form of bulk transport, without which the globalization
of the world trade would hardly have been possible. Given that the carrying capacity of modern container ships
already exceeds 21000 TEU (Twenty-foot Equivalent Units) containers, and that they are almost invariably
1
Dobrivoje Ninković, Dr. Tech. Sci., Ruchwiesenstr. 28, 8404 Winterthur, Switzerland, dninkovic@bluewin.ch
47
propelled by a single low-speed two-stroke engine (75570 kW in the case of OOCL Hong Kong, currently the
world’s largest container ship), the importance of the large engines can hardly be overstated.
The breakthrough in rail transport came about in 1939 with the introduction of the EMD FT, the two-stroke Diesel-
electric locomotive in the U.S.A. [3]. The Second World War slowed the proliferation of the Diesel engine in this
transport sector too, because it was forbidden to build Diesel engines in the war-time U.S.A. on account of the
steam engine being much better known and requiring less precious materials in the production.
Diesel engines have also been widely used to generate electricity, an example of which is the Diesel locomotive
whereby the engine is used to generate electricity to be supplied to the electric motors driving the wheels. On
ships, Diesel generators are used to supply electricity for lighting, ventilation, powering the ballast water pumps for
balancing the ship, but also for auxiliary propellers, i.e. thrusters, used for enhancing the ship’s maneuverability,
especially in ports.
Another important application field of large engines is the stationary electricity generation, either as the so-called
“island solutions” in areas without connection to an electric grid system, or for supporting a national grid system at
meeting the peak or a sudden load demand. A typical power plant of this kind may consist of one to twenty 10 to 20
megawatt units ready to start at short notice, which is of an ever increasing value in connection with the often
unpredictable availability of the renewable electrical energy sources, such as e.g. the wind turbines [13].
Lest the impression be given here that the Diesel engines are the only large engines in existence, the gas-burning
engines actually predate the Diesel ones by some 30 years, as gas was available for lighting, and thus the so-
called atmospheric gas engines appeared. However these engines were very inefficient and the development
impetus for gas engines came about with the four-stroke patent of Otto in 1876.
Large gas engines experienced an ever accelerating development rate in the last 20 to 30 years, constituting
nowadays a mature segment of the large engine family. At powers of up to 50 MW, low-speed, two-stroke gas
engines are more efficient than the corresponding gas turbines; and their small environmental impact makes them
also attractive for being used as prime movers in the transportation as well, competing ever more with the four- and
two-stroke Diesel engines on ships. As a matter of fact, the former clearly-cut difference between the gas and
Diesel engines has been progressively blurring in the past decade or so due to the introduction of the dual-fuel
engines that are capable of burning both fuels, switching between them according to the situation at hand.
Given that the slow-running two-stroke Diesel engines are at least equal if not better in terms of the fuel efficiency
than the gas turbine installations, being thus the most efficient thermal machines known to man, and the
importance of the large engines in general to the large-scale transportation of goods and people, it is of interest to
review their current state and development trends, borrowing thus the motto “Quo Vadis Large Engines” from a
recent Guest Commentary in MTZ Industrial [4].
MAIN ENGINE TECHNOLOGY DRIVERS

The engine technology has in the past been driven by the need to continuously fulfill the standard criteria of the
general product development, namely:
 Efficiency, i.e. the operating costs
 Price, i.e. the production price at the manufacturer’s side, and the purchase price at the buyer’s side
 Reliability and safety
 Ease of maintenance (also a part of the operating costs)
The state and its agencies may have intervened insofar as to define regulations concerning the reliability and
safety of the products, which has traditionally been the concern of the insurance business as well, leaving the rest
to the negotiations between the manufacturers and buyers.
The situation changed sharply as the concern arose for the environment pollution and its consequences upon the
human and animal health, which in the case of large ship engines may be dated to the year 1997 as the IMO
(International Maritime Organization) published the MARPOL (Maritime Pollution) Annex VI [5]. This agreement
was intended to come into effects on 1.1.2000, subject to being ratified by enough countries to reach 50% of the
world ship tonnage. There were several previous declarations to the subject of the environment protection, but it
was the first time that concrete emission values were to be applied along with the measures and the timetable for
their implementation and enforcement. A process was thus set up that would become a major development driver
in the engine business for the decades to come, as the regulations were to be updated, i.e. become ever stricter,
on a regular basis [18].
48
The evolution of the IMO limits on the nitric oxides (NOx) emissions from large engines is presented in Figure 1
below [6]. It is valid for marine Diesel engines with power larger than 130 kW, but not for vessels operating in
national waters, where local legislation applies. It is seen that the limits were lowered in stages (so-called Tiers),
first mildly, and then drastically in the so-called Emission Control Areas (ECA). These are to be defined by the
individual countries. It is also apprent that the slow-running engines (< 250 rpm) were allowed higher NOx values in
the beginning, but the Tier III practically eliminated the difference between the two- and four-stroke machines.
The sulphur limits were also scheduled to be reduced with time at the global and ECA levels, but since the SOx
emissions are primarily determined by the sulphur content in the fuel, they have repercussions upon the engine
only insofar as regards the exhaust after-treatment.
The soot (black carbon) and particle emissions are currently not regulated, and it is apparently not likely that the
discussions within the IMO are to be soon concluded.
Figure 1 Evolution of the Nox legislation for large maritime engines
The emission limits for rail engines came about at approximately the same time, in the form of the EU Directive
97/68/EC, which was adopted by the Int. Union of Railways (UIC) as the Code 624V, shown in Figure 2 below. The
former was superseded in 2016 by the all-encompassing NRMM (Non-Road Mobile Machinery) legislation (EU
2016/1628), which covers practically everything from the small agricultural machinery to the inland waterway
vessels and electricity generating sets [7]. The relevant emission limits for the rail transport are shown in Figure 3
below.
Figure 2 The UIC Code 624V emission legislation for railway propulsion engines
Similar legislation for the rail transport exists in other countries, with the U.S.A. NOx limits of 1.74 g/kWh for their
Tier 4 (2015+) locomotive engines being more stringent than the Stage V (2021+) EU limits of 4 g/kWh [6].
The above mentioned exhaust gas limits must be met without sacrificing the fuel efficiency. However, the latter
determines the CO2 emissions; and as a Greenhouse Gas (GHG), CO2 is also the subject of regulatory measures.
Although the global shipping is currently responsible for only 2.5% of the worldwide CO2 emissions [9], IMO has
initiated a program for reducing the total annual GHG emissions by 50% relative to the 2008 level, i.e. the CO2
emissions by 40% by the year 2030. To this end, IMO has introduced mandatory technical (EEDI) and operational
(SEEMP) measures for the energy efficiency of ships [10].
The term EEDI (Energy Efficiency Design Index) is a comprehensive efficiency parameter that takes into account
not only the main engine effciency, but also all other GHG emissions, the ship's capacity, the so-called „hotel load“
(crew facilities), and speed. Referring to Figure 4 below, there is a reference EEDI line for each ship comissioned
49
after January 1st of 2013 as a function of its DWT (Deadweight Tonnage), and a set of required lines for the new or
retrofitted ship generations as a parameter, reducing by 10% every five years [9].
Figure 3 The EU 2016/1628 emission legislation for railway propulsion engines
Ship owners can improve the respective EEDI figures of their ships by various measures, such as making changes
to the engine (installing better injectors, turbochargers, etc.), and by optimizing the entire energy flow of the vessel.
Engine manufacturers are thus stimulated not only to improve their products, but also to offer various devices for
enhancing the performance of the existing engines in service.
Figure 4 The required IMO EEDI lines for the existing and future ships [8]
As regards meeting the above goals set by the respective legislation bodies, Schlemmer-Kelling [5] writes that the
potential for increasing the efficiency at high engine loads by improving the combustion process, the gas exchange,
and the mechanical efficiency has already been exhausted. The only remaining measures would be increasing the
firing pressure, and opting for a two-stage turbocharging with an intercooler. In his opinion, it is the part-load where
larger potential for improvement exists in terms of better air supply to the cylinders, and optimized fuel injection.
It is at this place necessary to mention one technology that has been and remains instrumental to achieving the
improvements required for meeting the exhaust legislation, namely the supercharging, which in a vast majority of
cases is realized by means of turbochargers. There is now hardly a large engine without turbocharging, and new
theoretical IC cycle variants such as Miller and Atkinson would have been impossible to realize without this
technology. It has been developing in parallel with the engines, exchanging the development incentives in both
50
directions. This is also true of the newest T/C products, namely the two-stage turbochargers that have been offered
as mature products by all major T/C manufacturers in the last years, e.g. [11][12].
The fact that the GHG emissions from many energy sectors in Europe have been falling, and those from transport
have increased over the last decades, and are expected to increase significantly to 2050 [14], has become an
additional technology driver in the large engine sector. The GHG emissions of fuels with a smaller carbon content
than the standard marine fuel oil (MFO) and the heavy fuel oil (HFO), such as .e.g. the Liquid Natural Gas (LNG),
have a favourable impact on the emissions [15], and the use of LNG is thus expected to rise [16], but presupposes
the availability of the corresponding engines [17].
SLOW-RUNNING TWO-STROKE ENGINES
Current development state

The two-stroke, low-speed large engine industry has undergone a concentration process, with the result that there
are now only two major companies developing but not necessarily manufacturing the largest engines, these being
MAN Energy Solutions (formerly MAN Diesel & Turbo) of Augsburg, Germany, and a company now called WinGD
(Winterthur Gas & Diesel), which started in 2015 as a joint venture of Wärtsilä of Finland with the China State
Shipbuilding Corporation (CSSC), but is from 2016 owned solely by CSSC.
The business model of both companies consists of developing state-of-the-art Diesel engines, which are then built
in license almost exclusively in the Far East. MAN also manufactures turbochargers, whereas WinGD do not,
relying mostly on ABB Turbo Systems as the turbocharger supplier. The most powerful engines of both companies
are in the same range (Wärtsilä: 80 MW, MAN: 87 MW), with Wärtsilä having the largest one in use with the power
of 80080 kW on board the container ship Emma Maersk.
Kobe Diesel, having recently acquired the large engine program of Mitsubishi Heavy Industries (MHI), is the third
company developing and manufacturing this class of engines (about 2% of the world market), their range currently
extending up to about 35 MW units.
Before the financial crisis of 2008, the container ships were built for speed, and it was expected that the future ship
classes with the carrying capacity of the order of 20000 TEU would need main propulsion power of over 100 MW.
However, this trend has not materialized because the shipping companies realized that significant fuel savings are
possible by reducing the cruising speed somewhat. For example, reducing the ship speed from 27 to 22 knots, i.e.
by 19%, lowers the nominal engine power to 42%, saving thus 58% of the hourly fuel consumption [19]. So now the
21000 TEU OOCL Hong Kong with the main propulsion power of 75 MW has a smaller engine than the 16000 TEU
Emma Maersk.
Main characteristics of modern two-stroke, main propulsion ship Diesel engines are:
 Long-stroke crosshead design
 Cylinder diameter of up to about 100 cm, stroke-to-bore ratio about 4 and higher (in newer engines)
 Hydraulically activated exhaust valve in the cylinder cover, electronic valve control
 Single-stage turbocharging, mechanical blower for the low power regime
 Electronically controlled fuel injection, common rail (MAN and WinGD), mechanical (MHI)
 Up to three injection nozzles per cylinder in order to achieve fuel injection flexibility
 HFO and MFO as the main fuel types, switchable on demand
Three general technology trends are discernible in the low-speed two-stroke large engine segment:
 Comprehensive engine optimization aimed at improving the fuel efficiency, reliability, maintainability, etc.
achieving thus low life-cycle costs
 Changes with regard to meeting the emission legislation, esp. the current Tier III norm
 System optimization (ship, engine, propulsion, operating procedures) for improving the EEDI figure
The above trends are not isolated from each other – they rather intersect and interact in the attempt to meet the
changed economic criteria of the customers, namely the Total Cost of Ownership (TCO) that now dominates the
decision making process instead of the previously most important criterion of the First Cost, i.e. the purchase price.
Discussing the numerous measures and solutions that have been studied and realized in order to arrive at the
current optimization level of this engine class is beyond the scope and limits of this paper. Therefore, only a short
list will be given of the most important improvements and optimization areas [21], [22], [23], [24], [25]:
 Fuel injection: electronic control, common-rail in different variants, pumps and injectors
 Exhaust valve electronic control
 Auto-tuning in the engine control
51
 Mechanical improvements of the engine structure
 Minimization of the mechanical (friction) losses in the critical areas, e.g. in the cylinder ring packing and
the crosshead guide shoe bearing
 Cylinder lubrication system
 Cylinder cooling
 Turbocharging, e.g. variable turbine area, turbocharger cut-out at low loads, etc.
As an example, shown in Figure 5 below is a diagram showing a possible fuel oil cost savings for a 1 g/kWh
reduction of the lubrication oil consumption with optimized piston rings [25].
Figure 5 Possible savings by reducing the lubrication oil consumption by 1 g/kWh [25]
In looking at the optimization possibilities in the entire ship propulsion system, one of the obvious elements is the
ship’s propeller, and its feature of increasing the efficiency with the diameter. Using larger propellers necessitates
lower engine rotation speeds, and in order to keep the engine power at the level required by the ship, the stroke-to-
bore (S/B) ratio must be increased. This was the approach that led MAN to develop the G-Type engine, with S/B
ratio of up to five [20]. According to MAN, CO2 reductions of up to 7% are possible as a part of the propulsion
package. Referring to Figure 6 below, overall efficiency increase of 4-5% should be possible with the new engine
and a larger propeller.
With the the introduction of the RT-flex mid-size engines (48 to 68 cm bore) in the mid-nineties, Wärtsilä also
increased the S/B ratio to over 4.0, and this trend continued with the current Generation X-engines. The latter
engines have an S/B ratio between 4.1 and 4.5, with the largest model W-X92 (92 cm bore) still at 3.8 for the lack
of manufacturing capacity at the crankshaft factory [21]. The authors point out the compromises that need to be
made when realizing the long-stroke engines, both from the standpoints of the manufacturing, and of the engine
room limitations on the ship.
Figure 6 Efficiency gain by using a larger propeller with a G-Type engine [20]
52
As regards meeting the emission legislation, particularly the NOx pollution, already at the introduction of the Tier II
norm the engine designers were confronted with the so-called “Diesel dilemma”. This refers to the fact that there is
no simple way towards simultaneously lowering the fuel consumption and reducing the emissions, for the mecha-
nisms of their generation within the cylinder are different and lead in part to opposite effects. Referring to Figure 7
below, reducing the NOx emissions is accompanied by a fuel consumption increase, which MAN in 2010 quantified
to 6 g/kWh for their mechanically controlled engines, and 4 g/kWh for the electronically controlled ones [26].
Figure 7 The “Diesel dilemma” [18]
MAN subsequently reduced the fuel consumption penalty for their Tier II engines to 1-2 g/kWh by a combination of
measures [26]:
 increased scavenging air pressure
 reduced compression ratio (two-stroke Miller timing)
 increased maximum combustion pressure
 adjustments of the compression volume, and other engine design changes
Wärtsilä introduced a new generation of two-stroke, Tier II conformable Diesel engines in 2011 [22]. Beside the
already quoted optimization measures, the new RT-flex engines featured full electronic control of the exhaust
valves and the common-rail injection. As mentioned above, these engines are now offered by WinGD.
Compliance with the Tier II limits adds one more parameter to the procedure for matching the engine speed and
power with the nominal operating point of the vessel’s propeller. This means that there are limits to the lowest
feasible BMEP (Brake Mean Eff. Pressure) level at lower engine ratings in the engine operation map, calling thus
for stronger internal NOx reduction measures, which in turn may have adverse effects on the BSFC (Brake Specific
Fuel Consumption). These measures include the reduction of the scavenging air pressure, use of extended “Miller
timing” (late exhaust valve closure in conjunction with increased scavenge pressure), and reduction of the peak
cylinder pressure – the ultimate measure because of its adverse effects on BSFC. The availability of electronic
controls on the engine greatly facilitates the application of these measures to the engine [27].
Today, all commercially available low-speed two-stroke engines comply with the IMO Tier II legislation without any
additional measures to the ones mentioned previously. Reducing the NOx emissions by additional 75% for
compliance with the Tier III values stipulated for the ECA requires much more complex technologies [27], most of
which are accompanied by considerable financial expenditures. Since the HFO remains the principal fuel in this
field [5] and its high sulphur content will be tolerated only until 2020 (<5000 ppm in non-ECA afterwards), and that
because of the high capital costs the oil refineries are unlikely to modernize their HFO plants for lower sulphur
content, it is the engine manufacturers who will have to solve the SOx emission problem.
Several technologies for meeting the Tier III criteria have been already investigated and experimentally tested, and
some of them have been also commercially available for a number of years, e.g. [28]:
 Selective Catalytic Reactor (SCR) with urea injection for reducing the NOx content; injection of ammonia
gas (NH3) has been introduced as an alternative recently
 Exhaust gas scrubbing and/or washing in order to reduce the SOx content
 Dual-fuel engines (HFO at high sea, gas in the ECA)
 EGR (Exhaust Gas Recirculation)
Starting with the SCR technology, there has been experience with it on the low-speed two-stroke engines previous
to the establishment of the IMO Tier II norm, but it was limited to stationary power plants, e.g. [28]. In this particular
case, the engine manufacturer (MAN) quotes both the power plants equipped originally with the exhaust gas
treatment, and the retrofitted ones. Given the power range of the plants mentioned (40 to 50 MW), the SCR and the
53
auxiliary equipment may be quite large. While in these cases there was enough space to accommodate whatever
equipment was necessary, an application of this technology on ships is confronted with a number of restricting
factors. Retrofitting an existing ship engine must take into account factors like tight space in the engine room, ship
balance, engine room deformation with various cargo arrangements, having enough space for routine engine
maintenance, possibilities for a relocation of the existing auxiliary components of the engine, etc. [29].
Referring to Figure 8 below left, the SCR from the reference above is seen at the RHS of the engine. While this
installation is not a true retrofit (the system was previously tested on the test-bed) and the engine room was
redesigned to accept thus modified engine, it conveys the idea of the space requirements. The RHS picture shows
a different SCR arrangement on an engine of comparable size to the previous one [30].
Figure 8 The SCR systems on two similar, middle-power two-stroke engines [29], [30]
Being installed between the exhaust receiver and the T/C turbine inlet, the SCR solutions shown above operate at
high exhaust pressure and temperature. However, in both cases there is a single turbocharger on the engine,
which greatly facilitates adding the SCR, but limits the solution to relatively low power engines. Large engines have
typically several turbochargers, typically one for each three to four cylinders, and it is questionable whether this
concept is applicable to such engines without major modifications. An SCR system after the turbochargers, i.e. at
the funnel inlet is a more attractive solution for the high-power engines, and one such system was recently
developed and commercially produced [31].
Dual-fuel engines have experienced very intensive development in the last years on the account of being capable
of full Tier III compliance without the need for the exhaust after-treatment measures. The approach consists of
using the HFO on the high sea, and gas (typically LNG) in the ECA. The two fuels differ significantly, but the engine
remains essentially a Diesel, with modifications for the gas combustion. Two concepts have been established, the
low-pressure one by Wärtsilä [32] and the gas injection (GI) of MAN [33], recently extended to the liquid gas
injection (LGI), methanol, and ethane gas, i.e. practically to all gas fuels [34]. A comparison of the two concepts is
presented in [37].
Referring to Figure 9 below, shown at the left is the low-pressure dual-fuel engine concept of Wärtsilä in the gas
fuel mode. The gas is injected at a low pressure (about 16 bar) in the middle of the compression stroke. The
compressed air-gas mixture is ignited at the end of the compression stroke by injecting the pilot fuel oil (MDO) into
the prechambers in the cylinder cover. The mixture has a high air excess ratio, creating thus the conditions for a
lean combustion. However, since the air-gas mixture is compressed to a high pressure, there may be a tendency
towards knocking combustion under certain conditions, so the engine must be monitored by the control system. In
addition, the compressioin ratio of the engine is determined by the gas fuel operation, which means that the engine
is derated in the Diesel fuel mode.
The MAN dual-fuel engine is based on high pressure gas injection directly into the cylinder, and subsequent
ignition of the mixture by a pilot fuel flame. The combustion process is the same as in the same engine operating
with Diesel fuel injection, such that there is no knock limit, as well as no limit on the compression rate and BMEP.
The engine thus achieves the same thermal efficiency as when operating with HFO or MDO.
54
The fuel gas is supplied to the engine at a pressure between 200 and 315 bar [33]; the pressure is dependent upon
the engine load. If the gas is available in the gaseous state, it is supplied by a suitable compressor, whereas LNG
must be supplied by a cryogenic pump and a vapouriser (LNG is kept at -162 °C in the tank). Since the gas fuel is
potentially explosive if it leaks and mixes with the surrounding air, proper high-pressure sealing of the gas system
must be guaranteed at all times.
Figure 9 The Wärtsilä [32] and MAN [33] dual-fuel engine concepts
If a dual-fuel engine is for any reason not an attractive option for the ship owner, the Tier III limits can be met by
burning MDO and employing the SCR and/or EGR for the exhaust gas treatment. The latter is only possible with a
very low sulphur content in the fuel oil, for sulphur is responsible for the major part of the PM (Particulate Matter) in
the exhaust gas; alternatively, an SOx scrubber/washer can be used.
EGR solutions have been tested by MAN and their licensees, and found to meet the Tier III criteria. In comparison
with the SCR, the necessary installation space is obviously much smaller (Figure 10), but the air supply to the
engine, i.e. the turbocharging solution, must be redesigned [35], and a number of other non-trivial conditions met
[36].
Figure 10 The MAN EGR concept (left) vs. the SCR solution for the same engine (right) [36]
55
Future R & D directions
According to the German engine consulting firm FEV, for price reasons the HFO remains the fuel of choice in the
next decade, and this prediction has repercussions on the future research activities in the low-speed two-stroke
engine field. While LNG is a very attractive fuel on an LNG tanker, using it as main fuel on an e.g. container ship is
hardly a practicable idea, for LNG must be kept at -162 °C in the fuel tank and the inevitable boil-off must be
continuously recompressed, which requires a special compressor. Therefore, there will be a legislation-driven need
for attractive exhaust gas after-treatment solutions [5].
The other factor influencing the future research is expected to be the change to the TCO concept at the customer’s
side, which means that the engine is only a part of the much larger picture. However, since the TCO approach
tends towards lower ship speeds, it leads sometimes to surprising results. Referring to Figure 11 below, a TCO-
based optimal ship speed for a 1000 TEU container feeder would be 14 knots (prices as of 2015), roughly the value
achieved by the sailing ships of the 19th century [37]. Because of the thus prolonged transport times, the ship
operator will all the more be interested in having maximum efficiency of the entire business unit, including the
engine.
Figure 11 TCO-based optimal ship speed for a 1000 TEU container feeder [37]
The future development directions, as predicted by the FEV [5], are shown in Figure 12 below. Some of the
solutions that should be important in the next decade are already available, such as common-rail injection, and the
two-stage turbocharging [39]. The latter was tested a decade ago on an engine test bed and found unattractive, but
should now become interesting on account of offering a possible 2 to 7 g/kWh reduction of the fuel combustion.
Figure 12 FEV road map for the necessary R & D activities in the two-stroke low-speed sector [5]
The above road map does not include a recent addition to the turbocharging technology that holds a considerable
potential for a BSFC reduction in the slow-steaming ship mode, especially at the low end of the engine load (below
40%). It is based on adding an electric motor/generator (M/G) to the turbocharger, and in the solution of Mitsubishi
[58] the electrical machine is an intrinsic part of the latter, sharing its shaft. Referring to Fig 13 (left) below, the M/G
is located at the compressor inlet, in the middle of the filter/silencer. At high engine loads, the excess turbocharger
56
power is used to drive the generator, reducing thus the demand on the main electricity generator; in a case
reported in the paper, the generator consumed about 13% of the T/C turbine power [58].
At low engine loads, when the T/C turbine does not become enough energy in order to supply enough scavenging
air to the engine, it is customary to use an electrically driven blower for this purpose. However, the electrical
machine within the T/C can also be used in the motor mode, providing thus additional energy to the compressor,
possibly obviating the need for the auxiliary blower. In the case reported in [59], generating the scavenging
pressure by this method consumed 30% less electricity as compared to the auxiliary blower, reducing simul-
taneously the total fuel consumption by 1.3 to 6.7%, depending on the main engine operating point [58]. Obviously,
this approach has a solid potential for the future, but the motor/generator price is currently rather high.
Figure 13 The “Electric-Assist” T/C of Mitsubishi [58], and the available scavenging pressure [59]
Regarding the EEDI as a technology driver, an interesting combination of the low-pressure EGR and the OCR
(Organic Rankine Cycle) as a possibility for low-temperature waste heat recovery has recently been reported in the
literature [60]. As a matter of fact, not only the EGR system can be used as the OCR heat source, but practically
any low-temperature energy source can be harnessed for this purpose. Savings of the order of 6 to 7% can be
expected from this heat recovery method.
FOUR-STROKE LARGE ENGINES

Traditionally, engines in this group are differentiated according to their speed (medium and high speed) and the
fuel they burn (Diesel or gas). While the speed criterion is still valid (<1000 rpm: medium speed), the fuel criterion is
not sharp anymore, for the proportion of dual-fuel engines appears to be rising in this segment too [40], [41].
With the exception of the air and road transports, large four-stroke engines are encountered on ships, boats, and
submarines (main propulsion, auxiliaries), in the rail transport (locomotives and railcars), in the stationary power
plants (electricity generation), on the oil drilling platforms, and in nonroad mobile machinery (excavators, etc.).
Current development state

The developments in this engine class have also been influenced by the exhaust gas legislation, especially in the
maritime sector, but also by fuel price fluctuations and customer requirments. For example, a TCO projection for a
10 MW DF ship engine calculated in 2014 would give a three-year ROI (Return on Investment) for the gas
operation, but due to the falling HFO prices, the same calculation made in 2016 would have incurred heavy losses
[42]. This is where the flexibility of the DF engines could play a prominent role, provided that the vessel has
provisions for storing both fuel types. Practically all modern ships with the four-stroke main propulsion are equipped
with DF engines.
Modern four-stroke engine portfolios are optimized to satisfy not only the exhaust gas norms, but also the ever-
present striving towards maximum efficiency across the entire load range, low manufacturing costs, high reliability,
and low maintenance costs. Treating an engine in the development process as a part of the entire object (ship,
power plant, etc) requires solutions facilitating the system integration, both at the engine level itself, and at its
„working place“.
57
Turbocharging has apparently played a decisive role in bringing about both increased effciencies and lowered
emissions of polluting gases. The effects of higher charging pressure have been known for more than twenty years,
but the impetus for developing such solutions came from the emission legislation, e.g. [43]. At the 2007 CIMAC
Conference, Wärtsilä presented the first test-bed results for their Type 20 engine turbocharged with an ABB two-
stage T/C delivering the air at a pressure ratio of 9.1, which enabled the use of the Miller timing [44]. Valuable
results were obtained as regards the NOx reduction, and further research and development activities indicated.
Theoretical studies of the Miller cycle indicate that the maximum cylinder pressure must not be increased in
comparison with the Diesel cycle in order to obtain an efficiency gain [43]. Referring to Figure 14 below left, it is
seen that the efficiency gain increases with the reduction in the maximum cylinder pressure, such that e.g. pmax of
190 bar with the Miller process results in the same efficiency as with the Diesel pmax of 270 bar, lowering the NOx
generation at the same time. In making these calculations, it is important to use the ideal gas hypothesis
(temperature-dependent specific heats), because the perfect gas hypothesis leads to incorrect results [43].
In order to achieve the above results, a large pressure drop over the engine is needed, calling thus for a
turbocharging efficiency of about 70%, which in turn means a T/C compressor ratio of the order of ten (Figure 14,
right). Such pressure ratios are impossible to realize with a single stage compressor if a useful performance map is
to be simultaneously obtained, hence the reason for a two-stage compression. Together with the intercooling,
which renders the overall compression process approximately isothermal, i.e. energy-efficient, favourable
conditions are created for the realisation of the Miller process.
Figure 14 Theoretical Miller and Diesel efficiencies (left) and the required turbocharging efficiency (right) [43]
In 2015, Wärtsilä announced its “31” medium speed, four-stroke engine with the world’s highest efficiency in its
class (BSFC of 165 g/kWh with MDO), available in the Diesel, dual-fuel, and spark-ignited gas versions [45]. In the
current Product Specification sheet (2017), the BSFC is quoted at 170 g/kWh for somewhat heavier distillates with
the sulphur content of up to 2% [46]. An SCR is needed for compliance with the Tier III when operating with these
fuel oils. In order to secure acceptable performance at part loads, the inlet valve drive has variable closure timing.
The turbocharging technology has also made advances since the introduction of the first two-stage models. For
example, the second generation of the two-stage turbochargers of ABB Turbo Systems (Baden, Switzerland) has a
potential for achieving the equivalent turbocharging efficiency of up to 80% at overall pressure ratios of 12 [47],
giving the engine designer possibilities for further performance improvements (Figure 15, left). Looking at the low-
pressure stage turbine efficiency of almost 90% (Figure 15, right, one has a feeling of the physical limits being
about to be reached (turbine efficiency of over 90% has been attained on the test stand). Generally, the benefit of
the two-stage turbocharging is a BSFC reduction of 3 to 5%, depending on the application [49].
Figure 15 Turbocharging efficiency (left) and low-pressure turbine efficiency of the ABB two-stage T/C [47]
58
However, designing a two-stage turbocharged engine solution requires solving a number of problems in connection
with the mounting of the T/C onto the engine, dealing with the noise and the changed vibration patterns and levels,
securing serviceability of the entire system, etc. For example, going from a single-stage to a two-stage T/C
increases the mass to be held by the engine by 1.5 to 1.7 times, making a new structural analysis of the entire
assembly necessary [48]. Fundamental research is sometimes necessary in order to achieve the engine
performance goals under the new conditions; and the literature abounds with papers on the subject of combustion,
heat transfer, and gas flow optimization through the valves, e.g. [40].
Thermodynamical efficiency of the large four-stroke engines appears to have been rising at a rate of 0.25% per
year (Figure 16, left) for the Diesel and 0.3% for the gas fuel (Figure 16, right) in the last two decades (note that the
ordinate in the LHS plot was in the original paper incorrectly labelled, and has been corrected by the present author
in the image below). Both trends appear to have continued to the present with the already mentioned Wärtsilä 31
engine (the efficiency figure has not been published, but with the BSFC of 165 g/kWh it is likely to be about 50%),
and the newest Jenbacher J920 gas engine genset with an electrical efficiency in excess of 50% has the engine
efficiency of more than 51% [52].
Figure 16 Diesel (left) and gas engine (right) efficiency trends in the past decades [50], [51], resp.
The performance figures of the new J920 gas engine came about through a comprehensive system-based analysis
optimization of the already efficient first version of the engine (electrical efficiency of 48.7%, [53]). Referring to the
comparison of the respective contributions of the fundamental thermodynamical and mechanical processes in the
first and second engine versions, presented in Figure 17 below, it is seen that the performance improvement came
about through reductions of the losses due to incomplete combustion, gas exchange, and friction [52].
Figure 17 Reduction of losses in the Jenbacher J920 gas engine design process [52]
Summarizing, modern four-stroke large engines are capable of meeting the current exhaust legislation while
keeping or improving the engine efficiency, and hold promise for the future. The current development state has
been attained by intensive R & D efforts which, given the importance and the omnipresence of these engines, are
going to continue.
59
Nowadays, the four-stroke engine family is characterised by a combination of features, such as:
 Two-stage turbocharging, making the Miller and/or Atkinson cycles possible
 Flexibility with regard to the fuel to use
 Compliance with the IMO Tier III and other equivalent exhaust legislation norms without external exhaust
gas treatment when burning gas fuel
 Availability of a combination of intra-engine, e.g. EGR, and external measures for compliance with the
norms in the case of liquid fuel oils
 Variable valve timing
 Improvements in the fuel supply to the cylinders (common-rail injection, optimized gas supply)
 Comprehensive electronic control of the entire engine, capable of securing optimal results in the entire
operation map; and individual cylinder control has already been reported upon
 Prolonged maintenance intervals of the order of a year’s continuous service
 Modular design, facilitating quick repairs but also supporting future upgrades and retrofits
It is the opinion of the author that the above represents a solid basis for, as formulated in [54], a “future-proof”
engine.
Future R & D directions

It is to be expected that current fine-tuning efforts will be continued and extended, especially as regards the part-
load operation. For example, contemporary DF engines must revert to Diesel operation when the load drops below
50% either because of instable operation or the increased emissions. It appears that this problem can be alleviated
by selectively switching individual cylinders off [56].
Transient operation is known to give rise to smoke emission (limited by some countries in the ECA), and it has
been shown that a better control of the turbocharging can eliminate this problem [55].
Individual cylinder pressure sensors are regularly used in low-speed two-stroke engines, and the trend is expected
to carry over to the four-stroke segment. For example, Caterpillar have developed an in-cylinder pressure module
(ICPM) that not only measures the maximum pressure, but also evaluates several combustion-relevant parameters,
enabling thus the balancing of cylinders in operation [41]. Such efforts then almost inevitably lead to a closed-loop
cylinder control, which would be instrumental in further reducing the engine emissions.
The Mitsubishi hybrid turbocharger mentioned above within the low-speed two-stroke context has also a fuel saving
potential in the four-stroke segment. Tested on a rather small engine (660 kW), it demonstrated a net fuel oil
consumption reduction of between 1.9 and 2.8% [58]. It is therefore clear that further R & D efforts are to be expec-
ted with this approach, reducing the price being one of the targets.
Variable valve timing remains essential for achieving high efficiency and low emissions under varying operating
conditions, but it cannot solve the derating problem in low-pressure dual-fuel engines. Essentially, a straightforward
solution would be an engine with a variable compression ratio, which has been known in general for a long time. A
conference with this topic was held in 2017 [57], showing the current activities at several institutions. One of the
concepts is based on an eccentric bushing, integrated in the upper connecting rod eye, which can change the
effective connecting rod length in two fixed positions [42]. The VCR idea and theoretical BSFC reduction obtainable
with this technology (up to 10 g/kWh with Diesel engines, and up to 8 g/kWh for the gas ones) are shown in Figure
18 below. Intensification of the research is to be expected in this domain.
Figure 18 The FEV variable compression ratio proposition and the expected effects [57]
60
CONCLUSIONS
The legislation limiting the emission of hazardous gases has led to unprecedented R & D efforts in the field of large
engines, aimed at meeting the emission limits. But it is also their importance for the world economy that has been
driving their development, resulting in these engines belonging to the group of the most efficient thermal machines
available to the mankind. Due to the work done in the last two decades, all engines offered in the market today
comply with the exhaust gas regulations; while in many cases there has also been a considerable progress in
improving their efficiency even further.
In the low-speed, two-stroke engine field, compliance with the legislation was brought about mostly by external
measures, such as after-treatment of the exhaust gas, and switching to a different fuel in the ECA’s. The latter led
to the introduction of the dual-fuel engines and their further development into multi-fuel engines. System integration
plays a key role in reducing the GHG emissions, in that not only the engine efficiency is of importance, but rather
the efficiency of the entire object containing the engine (ship, power plant). For example, this led to the change in
the stroke/bore ratio of the ship’s main propulsion engine from about 3.5 to up to five in order to enable using large,
i.e. more efficient, propellers.
In the case of the four-stroke engines, one technology has played a major role in making possible intra-engine NOx
reduction, and this is the turbocharging. Having been known for a long time for making out 10% of the engine price,
but bringing about 75% of its power, in its new, two-stage version it provided the high charging pressure required
for employing the Miller/Atkinson cycle. Alone or combined with other measures, such as variable valve timing,
EGR, SCR, urea or ammonia injection etc. two-stage turbocharging secures full compliance of the current four-
stroke engines with the exhaust gas legislation.
Two-stage turbocharging has also led to these engines attaining thermal efficiency of almost 50% with a potential
for further increases, especially with gas engines. However, obtaining high efficiency over a broad range of
operating conditions, and/or with different fuels, will require further measures and new concepts. One such solution
that has been known for a long time, but for cost reasons never utilized in commercial engines, may be the variable
compression ratio (VCR).
The general trend towards digitalization has been slowly taking grasp of the large engine area. In addition to the
local computer control of the engines themselves, which has already become indispensable, optimization of the
vehicles (ships, locomotives) and plants (electricity generation) containing large engines, seen as integrated
systems is becoming possible. Large amounts of data are already being transferred from ships en route to their
control centres, making applications such as e.g. remote monitoring and on-line optimization possible.
Summarizing, it can be stated that large Diesel and gas engines will continue to provide the global mass transport
with a reliable propulsion power that complies with the environment protection legislation while simultaneously
offering ever increasing efficiency and fuel flexibility.
ACKNOWLEDGMENTS
The author is indebted to Dr Ioannis Vlaskos, formerly of ABB Turbo Systems Ltd. and Ricardo plc. for proofreading
the manuscript and providing comments and suggestions, the inclusion of which improved the paper. However, all
errors and oversights remain the sole responsibility of the author.
REFERENCES
[1] Maritime propulsion, Available from: http://articles.maritimepropulsion.com/article/Memorable-2012-100th-
Anniversary-of-MV-SELANDIA20845.aspx, Accessed 25.08.2018.,
[2] Ninkovic, D.: "Large Engine Turbocharging Simulation: The Role, State of the Art, and the Development
Trends”, Proceedings of the International Congress Motor Vehicles and Motors 2008, Paper MVM2008IL03,
Kragujevac, Serbia, 2008,
[3] Wikipedia, Available from: https://en.wikipedia.org/wiki/EMD_FT, Accessed 25.08.2018.,
[4] Vlaskos, I.: “Quo Vadis Large Engines", MTZ Industrial, March 2015,
[5] Schlemmer-Kelling, U.: “Agenda 2030 – Mega Trends in the Large-Bore Marine Engine Business”, Proc. 12th
International MTZ Conf. on Heavy-Duty Engines, 2017,
[6] "Exhaust Emission Legislation – Diesel and Gas Engines", VDMA, January 2017, Frankfurt,
[7] "A new EU Regulation for engines in non-road mobile machinery (NRMM) ", European Commission, GRPE
Session, June 2016, Available from: https://www.unece.org/fileadmin/DAM/trans/doc/2015/wp29grpe/GRPE-
73-06.pdf, Accessed 19.08.2018.,
61
[8] Humans At Sea; Available from: https://www.humansatsea.com/2016/10/08/top-10-biggest-container-ships-in-
the-world/, Accessed 20.08.2018.,
[9] "EEDI – Energy Efficiency Design Index", MAN Diesel & Turbo; Available from: https://marine.mandieselturbo.
com/docs/default-source/shopwaredocumentsarchive/eedi.pdf?sfvrsn=4, Accessed 22.08.2018.,
[10] "Adoption of the Initial IMO Strategy on Reduction of GHG Emissions From Ships and Existing IMO Activity
Related to Reducing GHG Emissions in the Shipping Sector", IMO, April 2018.,
[11] ABB; Available from: https://new.abb.com/turbocharging/technologies/power2, Accessed 22.08.2018.,
[12] MAN Energy Solutions; Available from: https://turbocharger.man-es.com/products/tcx, Accessed 25.08.2018.,
[13] "Hybrid Power – Engines Supporting Wind Power”, MAN Diesel & Turbo, 2009, Available from:
https://powerplants.mandieselturbo.com/docs/librariesprovider7/brochures/hybrid-power-engines-supporting-
wind-power.pdf?sfvrsn=6, Accessed 19.08.2018.,
[14] Hartman, S. Kofod, M.: "Cost Effective Reduction of Life Cycle GHG Emissions in the Heavy-Duty Sector –
GHG Marginal Abatement Cost Curves for Global Shipping Including Fuel Shift to LNG”, Proc. 12th interna-
tional MTZ Conf. on Heavy-Duty Engines, 2014,
[15] "Costs and Benefits of LNG as Ship Fuel for Container Vessels – Key Results From a GL and MAN Joint
Study", Germanischer Lloyd, 2011,
[16] "Global Marine Fuel Trends 2030", Lloyd’s Register, 2014,
[17] Schlemmer-Kelling, U.: "Welcher Kraftstoff ist für den maritimen Einsatz in der Zukunft sinnvoll? ", Proc. 9th
[18] Götze, H.-J.: "Treiber für die Entwicklung von Großdieselmotoren: Internationale Abgasgesetzgebung",
Tagungsband 2. Rostocker Grossmotorentagung "Die Zukunft von Großmotoren II", 2012,
[19] Wiesmann, A.: "Slow Steaming – A Viable Long-Term Option", Wärtsilä Technical Journal, 02.2010.,
[20] “G-Type Engine: Revolutionary Ultra-Long Stroke", MAN, 2012,
[21] Brücki, A. et al.: "WinGD Generation-X Low-Speed Engines",Tagungsband 4. Rostocker Grossmotorentagung
"Die Zukunft von Großmotoren IV", 2016,
[22] Frigge, P. et al.: "New Two-Stroke Diesel Engines from Wärtsilä", MTZ, Vol. 72, 11/2011, pp 10-17,
[23] Clausen, N.B.: "Marine Diesel Engines – How Efficient Can a Two-Stroke Engine Be? ", https://www.ship-
efficiency.org/onTEAM/pdf/Clausen.pdf, Accessed 22.08.2018.,
[24] Kaiser, A.C. et al.: "Common Rail Pumps and Injectors for Large Diesel and Dual-Fuel Two-Stroke Engines”,
MTZ Industrial, Sept. 2015, pp 14-19,
[25] Mittler, R.: "Two-Stroke Piston Rings for Reduced Oil Feed and Optimised Distribution", MTZ Industrial, Sept.
2015, pp 27-31,
[26] MAN Global Corporate Website, Available from: https://www.corporate.man.eu/en/press-and-
media/presscenter/MAN-B-and-W-Engines-Set-Sights-on-Tier-II-Limits-58502.html, Accessed 25.08.2018.,
[27] Weisser, G. Hattar, Chr.: "IMO Tier III – Entiwicklung von Lösungen für die Wärtsilä Zweitakt-Grossdiesel-
motoren", Tagungsband 2. Rostocker Grossmotorentagung "Die Zukunft von Großmotoren II", 2012,
[28] "Two-Stroke Low-Speed Diesel Engines for Independent Power Producers and Captive Power Plants“, MAN
Diesel, 2009,
[29] Fujibayashi, T. et al.: "Development of Marine SCR System for Large Two-stroke Diesel Engines Complying
with IMO NOx Tier III", Proc. 2013 CIMAC Conference, Shanghai,
[30] "First Wärtsilä two-stroke engine with Tier III compliant high pressure SCR produced in China introduced",
Press release of Wärtsilä Corp., March 16th 2015, Available from: https://www.wartsila.com/media/news/16-03-
2015-first-wartsila-two-stroke-engine-with-tier-iii-compliant-high-pressure-scr-produced-in-china-introduced,
Accessed 25.08.2018.,
[31] Ryu, C. et al.: "The World’s First Commercialized Low Pressure SCR System on 2-Stroke Engine, DelNOx
System", Proc. 2016 CIMAC Congress, Helsinki, Paper No. 305
[32] Ott, M. et al.: "Niederdruck-Dual-Fuel Technologie für langsam laufende Schiffsmotoren", MTZ extra Heavy-
Duty-, On- und Off-Highway-Motoren, Nov. 2015, pp 34-38,
[33] Juliussen, L.R. et al.: "The MAN ME-GI Engine: From Initial System Considerations to Implementation and
Performance Optimization", Proc. 2013 CIMAC Conference, Shanghai, Paper No. 424,
[34] "At the Forefront of Green Engine Technology: The MAN B&W ME-GI and ME-LGI Dual Fuel Concepts",
Available from: https://marine.mandieselturbo.com/docs/librariesprovider6/test/at-the-forefront-of-green-
engine-technology.pdf?sfvrsn=13, Accessed 19.08.2018.,
[35] Shirai, T. et al.: "Testing an Integrated EGR System on a Marine Two-Stroke Diesel Engine", MTZ Industrial,
April 2014, pp 46-51,
[36] Garshasebi, A.: "ETB - Electrical Turbo Blower - MAN’s EGR Blower Series”, Proc. 2016 CIMAC Congress,
Helsinki, Paper No. 105,
[37] Strödecke, D.: “Two-stroke Dual-Fuel Technology Evaluation", MTZ Industrial, March 2016, pp 38-44,
[38] Schlemmer-Kelling, U.: "Efficiency Optimized Future Propulsion Concepts", MTZ Industrial, March 2015, pp
26-33,
[39] Ryser, R.: "Power2 - Is 2-Stage Turbocharging Interesting for 2-Stroke Engines? ", Proc. 2016 CIMAC Con-
gress, Helsinki, Paper No. 281,
[40] Menage, A. et al.: “The New Dual Fuel Engine 35/44 DF from MAN Diesel & Turbo SE”, Proc. 2013 CIMAC
Conference, Shanghai, Paper No. 291,
62
[41] Ritscher, B. Greve, M.: "Caterpillar M46 Dual Fuel Engine with New Cylinder Pressure Based Control
Strategies", Proc. 2013 CIMAC Conference, Shanghai, Paper No. 411,
[42] Schlemmer-Kelling, U. et al.: "How Legislation Bodies and Customers Influence Maritime Development", MTZ
Industrial, Sept. 2017, pp 15-21,
[43] Codan, E. Mathey, Chr. "Emissions – A new challenge for turbocharging", Proc. 2007 CIMAC Congress,
Vienna, Paper No. 245,
[44] Wik, Chr. Hallback, B.: "Utilization of 2-Stage Turbo Charging as an Emission Reduction Mean on a Wartsila 4-
Stroke Medium-Speed Diesel Engine", Proc. 2007 CIMAC Congress, Vienna, Paper No. 101,
[45] "Wärtsilä launches the new Wärtsilä 31 engine: a breakthrough in efficiency", Wärtsilä Press Release of 2nd
June 2015, Available from: https://www.wartsila.com/media/news/02-06-2015-wartsila-launches-the-new-
wartsila-31-engine-a-breakthrough-in-efficiency, Accessed 23.08.2018.,
[46] Wärtsilä 31 Product Specifications 2017, Available from: https://cdn.wartsila.com/docs/default-source/product-
files/engines/ms-engine/brochure-o-e-w31.pdf?sfvrsn=9, Accessed 23.08.2018.,
[47] Behr, T. et al.: "Second Generation Two-Stage Turbocharging For Large Four-Strokes", MTZ Industrial, April
2014, pp 26-33,
[48] Vaarasto, M.: "Experience of 2-Stage Turbocharged Engines", Proc. 2016 CIMAC Congress, Helsinki, Paper
No. 191,
[49] Cover Story Interview: “Marine Engines Need to Operate at Optimum Efficiency While Complying With Local
Laws", MTZ Industrial, Oct. 2013, pp 24-26,
[50] Poensgen, Chr.: “Development of the MAN Diesel & Turbo Engine Portfolio and Future Challenges", Proc. 12th
[51] Lang, J. et al.: “Development of a New Generation of GE’s Jenbacher Type 6 Gas Engines”, Proc. 11th Interna-
tional MTZ Conf. on Heavy-Duty Engines, 2016,
[52] Birgel, A. et al.: "GE’s J920 Gas Engine – 10.4 MW Power and More Than 50 % Electrical Efficiency", MTZ
Industrial, Sept. 2017, pp 40-46,
[53] Trapp, Chr. et al.: "GE’s All New J920 Gas Engine - A Smart Accretion of Two-Stage Turbocharging, Ultra
Lean Combustion Concept and Intelligent Controls", Proc. 2013 CIMAC Congress, Shanghai, Paper No. 289.
[54] Åstrand, U.: "Modular Medium Speed Four-stroke Engine", MTZ Industrial, Aug. 2016, pp 14-21,
[55] Toda, K. et al.: "Improvement Technology for Non-visible Smoke during Ship Transient Operation", MTZ
Industrial, March 2018, pp 14-21,
[56] Conrad, J. et al.: "Investigation of the Cylinder Cut-out for Medium-speed Dual-fuel Engines", MTZ Industrial,
March 2018, pp 54-63,
[57] International FEV Conference on Variable Compression Ratio, 2017, Available from:
https://www.fev.com/fileadmin/user_upload/Programm_VCR_Web.pdf, Accessed 25.08.2018.
[58] Shiraishi, K. et al.: "Energy Savings Through Electric-Assist Turbocharger for Marine Diesel Engines",
Mitsubishi Heavy Industries Technical Review, Vol. 52, No. 1, March 2015, pp 36-41,
[59] Shiraishi, K and Krishnan, V.: "Electro-Assist Turbo for Marine Turbocharged Diesel Engines“, Proc. of ASME
Turbo Expo 2014, paper No. GT2014-25667,
[60] Lion, S. et al.: "Overview of the Activities on Heavy Duty Diesel Engines Waste Heat Recovery with Organic
Rankine Cycles (ORC) in the Frame of the ECO-MATTE EU FP7 Project", Energy Procedia, Vol. 129, Sept.
2017, pp 786-793.
63
64
Kragujevac, Serbia
MVM2018-IL6
Patrizio Nuccio1
DEVELOPMENT THROUGH SIMULATION OF A

TURBOCHARGED 2-STROKE G.D.I. ENGINE FOCUSED ON
A RANGE-EXTENDER APPLICATION
ABSTRACT: In the last years, an originally designed 2-stroke prototype engine, which is equipped with an
electronically controlled gasoline direct injection (G.D.I.) apparatus was tested and these performance were
compared to those obtained with a commercial crankcase-scavenged 2-stroke engine. Very encouraging results
were obtained as far as fuel consumption and unburned hydrocarbons (HC) in the exhaust gas are concerned.
For example, a reduction of about 37% in specific fuel consumption was found, along with a reduction of over 90%
in the HC specific emissions in comparison of a commercial crankcase-scavenged 2-stroke engine using an indirect
injection fuel feeding system (the specific values were evaluated with reference to indicated mean effective pressure,
owing to the different friction mean effective pressure of the two engines).
High reductions were possible because the injection timing causes all the injected gasoline to remain in the
combustion chamber, so taking part in the combustion process. In other words, fuel does not participate to the
scavenging process, so it does not flow to the exhaust pipe without burning.
Moreover, a force-feed lubrication system has been employed, like those usually exploited on mass-produced 4-
stroke engines thanks a presence of an external pump. In fact, adding oil to gasoline is no more necessary, as
gasoline does not pass through the crankcase volume.
These first results also provided suggestions for engine future modifications. In fact, further improvements have been
implemented in a new design: the aim of this paper his therefore to present the main features of this 2-stroke engine
which mainly consist of the adoption of two inlet valves and two exhaust valves on the cylinder head; of a turbo
compressor for scavenging and supercharging process as well, together with a small volumetric impeller for start
transitory.
Thanks to the limited size, weight and high power output at low speed, this engine should be particularly interesting
as range extender application in electric vehicles.
In support of the previous experimental experience, a numerical simulation was also carried out by means of a one-
dimensional model that furnished encouraging results, for example, a brake mean effective pressure of about 8 bar
and a brake-specific fuel consumption of 270 g/kWh at 3000 rpm.
KEYWORDS: 2-stroke engine, engine power, fuel consumption, gasoline direct injection, electric vehicle
1
Patrizio Nuccio, Professor, Energy Department, Politecnico di Torino, Italy, patrizio.nuccio@polito.it
65
PRESENTATION
66
67
68
69
70
Kragujevac, Serbia
MVM2018-IL7
Giovanni Belingardi1
Pier Giuseppe Anselma2
Miroslav Demic3
OPTIMIZATION-BASED CONTROLLERS FOR HYBRID

ELECTRIC VEHICLES
ABSTRACT: Hybrid electric vehicles (HEVs) are more and more of interest in the present vehicle market, because
of the relevant reduction both in the fuel consumption and in the CO2 emission, in particular for the urban utilization.
Different architectures are possible and have been considered for the vehicle hybridization. The most convenient
architecture is depending on the application itself. This paper deals with a design methodology based on an
optimization scheme to make the selection among the consistent number of alternatives. Optimization-based control
strategies play a central role in the design process of HEVs. In early design phases they allow the comparison of
different HEV powertrain architectures, thus supporting the selection of an appropriate topology. Furthermore, they
lay the foundations for the development of real-time optimal energy management strategies to be implemented in the
HEV on-board control unit. Fuel economy and design efficiency can overall be enhanced in this way. This paper aims
at providing a comprehensive review of different optimization-based energy management strategies for HEVs. An
analysis of strength and drawbacks of each considered strategy is carried out based on different evaluation criteria
such as global optimality, computational cost and uniformity of the powertrain operation. Finally, simulation results
for a HEV powertrain from the industrial state-of-art validate the conceptual and methodological comments related
to the analyzed controllers.
KEYWORDS: control optimization, fuel economy, hybrid electric vehicle, optimal design, powertrain modelling
INTRODUCTION
Regulations to reduce motor vehicle CO2 emissions and fuel consumption have been enforced worldwide recently
[1]. This stringent trend is expected to be confirmed and even made more severe over the next decades. Vehicle
producers have been subsequently pushed to employ new technologies, including the use of fully electrified
powertrains. In this framework, hybrid electric vehicles (HEVs) are establishing as one of the most promising solutions
to satisfy customer requirements and CO2 emission regulations at the same time. A HEV adds an additional power
source (e.g. battery, ultra-capacitor, etc.) and one or multiple actuators (i.e. electric machines) to the conventional
powertrain. The additional power devices help to improve system efficiency and fuel economy by engine right-sizing,
load levelling, regenerative braking and pure electric mode.
To design a HEV, the engineer typically first selects one configuration to focus on. The vehicle design parameters
(e.g. motor size, battery size, planetary gear sizes, etc.) and control strategy then need to be determined. Obviously,
to achieve the near-optimal overall performance for the selected configuration, an iterative process needs to be
executed. However, the problem for this approach is that even with this optimal performance, it is not known whether
the selected configuration offers the best solution among all possible configurations. To achieve this goal, the exact
same process, beginning from the selection of another configuration and then iteratively approaching the optimal
performance, has to be repeated. Moreover, only when the optimal performance is gained for each configuration,
then the comparison between them is a sensible task. With numerous options for the configuration design variations,
1
Giovanni Belingardi, Politecnico di Torino – DIMEAS - Italy
2
Pier Giuseppe Anselma, Politecnico di Torino – DIMEAS - Italy
3
Miroslav Demic, Serbian Academy of Engineering Science
71
such an iterative process can only be developed by means of a systematic method with many underlying techniques,
including the automated model generation and simulation with optimal design and optimal control techniques. The
many different possible configurations and additional propulsion components bring new challenge and research
opportunities to vehicle designers.
As several control problems reported in the literature, the control of HEVs can have a two level hierarchical
architecture: the lower level control, and the supervisory control. For the lower level control, each subsystem (e.g.
engine, electric motors, battery) is equipped with actuators, sensors and a control system to regulate its behavior, in
response to the supervisory control commands. The design of lower level controllers can be separated from the
supervisory controller, and it will not be considered in this paper. The supervisory control of the HEVs determines the
operating mode and power levels of all power devices to balance design objectives such as drivability, fuel economy
and battery health. The supervisory level control and its use in assessing the optimality of design candidates are the
focus of this paper.
This paper is organized as follows: the different architectures for an HEV and its design process are presented. The
optimization-based control strategies for HEVs are subsequently illustrated and analyzed. A case study for the
assessment of these strategies is then presented. Conclusions are finally given.
THE HEV DESIGN PROCESS
HEV powertrain architectures

To date three main categories are identified for a HEV architecture: series, parallel and power-split. The
correspondent block diagrams are illustrated in Figure 1.
Figure 1 Different HEV architectures
Series HEV
In a series HEV, the propulsion is provided by one traction motor or more than one motors; the conventional engine
drives the electricity generator, and they are both decoupled from the drivetrain. Since there is no mechanical
coupling between the engine and vehicle drive axle, the engine speed and power are not rigidly constrained by the
vehicle speed and road load, which enables the engine to operate always at high efficiency. In addition, because the
traction motor usually can provide enough traction torque, transmission, in particular a gearbox, may not be needed.
Despite simple and easy to control, the series hybrid vehicle powertrain suffers from high energy conversion losses:
100 % of the engine output must be converted into electrical power and some of it is further converted into
electrochemical form and stored in the battery. The low efficiency is more pronounced when the vehicle is running in
highway cruise or steady state because of double energy conversion (mechanical-electric-mechanical), while the
series configuration is good for very transient drive cycles.
Parallel HEV
In a parallel HEV, both the electric Motor/Generator (MG) and the Internal Combustion Engine (ICE) can contribute
to the propulsion directly. In other words, the engine torque and the electric motor torque are additive. When the MG
is relatively small, it can only start/stop the engine, provide some regenerative power features, and drive the vehicle
in limited circumstances; when the MG is large, it can drive the vehicle by itself or simultaneously with the engine.
The MG can be used to shift the engine operating points to a higher-efficiency area by acting as a generator when
the power demand is low or as a motor at high power demand.
The efficiency of parallel hybrid vehicles can be very high on highways since the engine can directly drive the vehicle
near its sweet spot and energy circulation between the mechanical energy and electric energy can be significantly
reduced, while it reveals to be less convenient in transient driving cycles.
72
Power-split HEV
In a typical power-split HEV, an engine and two electric machines are connected to make a so-called “power-split
device”, represented by one or multiple planetary gear (PG) sets, through the carrier, the sun gear, and the ring gear.
The lever diagram can be used to represent the 2 degrees of freedom dynamics of a PG, as illustrated in Figure 2.
Figure 2 Planetary gear set and the lever diagram
The early power-split transmission appeared in the late 1960s and early 1970s, when such power-split mechanisms
were used in lawn tractors. Although other early studies on power-split hybrid vehicles followed, at authors’ best
knowledge, in the market there was no passenger power-split hybrid vehicle until the Toyota Motor Corporation
introduced the Prius, the first mass-production HEV in the world, in Japan in 1997 [2]. This hybrid powertrain system,
called the Toyota Hybrid System (THS), is the framework and the foundation of all Toyota hybrid vehicles, as well as
hybrid vehicles from several other companies, including the Ford Fusion Hybrid and the General Motor Allison Hybrid
system [3].
Power-split hybrid vehicles are efficient in city driving conditions as a result of the pure electric drive function.
However, due to energy circulation from the generator to the motor, the power-split vehicles may have higher energy
losses than parallel HEVs in highway driving. This problem for single-mode power split hybrids can be avoided by
adding clutch connections between different PG nodes, thus realizing multi-mode hybrid designs. Engagements and
disengagements of clutch connections can enable different operating modes, thus improving the overall efficiency
and performance of the system. Each operating mode typically best suits a specific case of vehicle operation (e.g.
launching, accelerating, cruising at high speed, regenerative braking). Power-split architectures are the most
successful and represent a large portion of the current population of HEVs. This paper particularly focuses on the
power-split HEV powertrains with 2 PGs. A lever diagram for an example of this kind of system, taken from the
industrial state-of-the-art, can be seen in Figure 3 [4]. The correspondent PG and final drive gear ratios are reported
in the figure as well.
Figure 3 Example of a power split HEV
Modelling method for a power-split HEV

Many of today’s power-split hybrid vehicles use two MGs to complement the ICE. In this paper, we only consider the
case that each PG set is connected with two powertrain components, since having three powertrain components on
the same PG will lead to very limited design flexibility. If we consider only the cases in which the engine and output
shaft are on different PGs and each is complemented by a MG (inspired by Prius and GM Volt Gen 2), there are 144
possible configurations for components location. Three examples of configurations are reported in Figure 4.
73
Figure 4 Examples of different components location for a 2 PG HEV
𝑁𝑐𝑙𝑢𝑡𝑐ℎ , the total number of possible clutches, results to be 16. This value is found by using equation (1) and
considering the number of PG sets (Np) equal to 2. A correspondent graphical interpretation can be seen in Figure
5.
𝑁𝑐𝑙𝑢𝑡𝑐ℎ = (𝐶2𝑁𝑝 − 2𝑁𝑝 ) + 3𝑁𝑝 − 1 (1)
The first parenthesized term ( C2Np ) represents the possible clutches added between each pair of nodes and
corresponds to 15 for a double PG set. The second parenthesized term represents the redundant clutches to be
eliminated. Since only one clutch could lock a planetary gear set by connecting any two nodes out of three in that
planetary gear set, the other two possible connections are redundant (displayed in red in Figure 5). 3Np represents
the grounding clutches: all the gear nodes could be grounded except the one attached to output, eliminated by the
last term.
Figure 5 All possible clutch locations of a double PG HEV
Setting the maximum number of clutches to 3, this leads to generate 7280 different designs for each location of
components. Therefore, the total number of candidate transmission designs (144x7280) for each set of input
parameters is more than one million [5].
A methodology to model the modes of multi-mode HEVs was proposed in [6]. The dynamics of any specific mode is
described by the characteristic matrix [A*], as shown in equation (2). This 4x4 characteristic matrix [A*] governs the
relationship between the angular acceleration of powertrain components and their corresponding torques. The
detailed derivations have been described by X. Zhang et al. in [6].
𝜔̇ 𝐼𝐶𝐸 𝑇𝐼𝐶𝐸 𝑎11 𝑎12 𝑎13 𝑎14 𝑇𝐼𝐶𝐸

𝜔̇ 𝑂𝑈𝑇 𝑇 𝑎21 𝑎22 𝑎23 𝑎24 𝑇𝑂𝑈𝑇
[ ] = [𝐴∗ ] [ 𝑂𝑈𝑇 ] = [𝑎 𝑎32 𝑎33 𝑎34 ] [ ] (2)
𝜔̇ 𝑀𝐺1 𝑇𝑀𝐺1 31 𝑇𝑀𝐺1
𝜔̇ 𝑀𝐺2 𝑇𝑀𝐺2 𝑎41 𝑎42 𝑎43 𝑎44 𝑇𝑀𝐺2
Engine model
The ICE can be taken into account through its experimental furl flow map. This fuel map is the result of the adopted
electronic fuel injection system's setting for regulating the air/fuel mix and is created by engineers during the
construction and tests of an engine. This lookup table represents a summary of the engine’s entire operating regime
with respect to torque and speed. For supervisory control studies and fast prototype design, the engine transient
dynamics due to spark-timing and fuel injection are ignored.
74
Electric motors model
The two motor/generators assisting the thermal engine in the powertrain model are direct current machines. In our
transmission model, the size and performances of the electric machines is not a design parameter, but rather an
optimization parameter. This means that a model of the losses in the electric machines is not required, but rather the
electric machines are already existing and manufactured. Their implementation in the transmission model goes
through their correspondent loss map, which can be empirically obtained from measurement tests on the real
machines. For every possible combination of torque and speed, the loss map returns the measured lost power of the
electric machine.
Battery model
Based on the dynamic characteristics and working principles of the battery, the equivalent circuit model was
developed by using resistors, capacitors and voltage sources to form a circuit network. In this paper we use the Rint
model, which is the most popular battery model for studying hybrid powertrain design due to its capacity to simulate
battery dynamics while having a simple approach. The Rint model implements an ideal voltage source VOC to define
the battery open-circuit voltage, together with an internal resistance RIN. The variation in time of the battery State-of-
Charge (SOC), SOC , can thus be expressed as
√𝑉𝑂𝐶 2 − 4 ∙ 𝑅𝐼𝑁 ∙ 𝑃𝑏𝑎𝑡𝑡 − 𝑉𝑂𝐶

̇ =
𝑆𝑂𝐶 (3)
2 ∙ 𝑅𝐼𝑁 ∙ 𝑄𝑏𝑎𝑡𝑡
where Pbatt and Qbatt represent the power requested from the battery and its capacity, respectively. 𝑃𝑏𝑎𝑡𝑡 can be
evaluated from the power requested or provided by the MGs, adding the correspondent values of lost power from
the experimental maps.
OPTIMIZATION-BASED HEV CONTROL

In order to assess the millions of HEV powertrain configurations above generated, a proper control strategy needs
implementation. In early vehicle design phases, the powertrain operation is optimized offline. In other words, the
trajectory of the vehicle speed is known a priori and it is determined by standard duty cycles (e.g. the New European
Driving Cycle (NEDC), the Urban Dynamometer Driving Schedule (UDDS)). This contrasts with actual real driving
conditions, where the powertrain control unit does not have input information concerning the future operating
conditions. In offline HEV control, optimization-based control approaches can be adopted to estimate the fuel
consumption of each analysed powertrain design and consequently identify the optimal one.
In general, three main approaches can be identified related to offline optimization of multimode power split HEVs:
1. The Pontryagin’s Minimum Principle (PMP)
2. Dynamic Programming (DP)
3. The Power-weighted Efficiency Analysis for Rapid Sizing (PEARS)
All these approaches aim at analysing the HEV performance in Charge-Sustaining (CS) mode, having the battery
SOC with equal values at the beginning and at the ending of the optimization time horizon.
The Pontryagin’s Minimum Principle

The PMP is a general case of the Euler-Lagrange equation in the calculus of variation. It optimizes a single operating
trajectory for the HEV powertrain, thus achieving local optimal solutions without guaranteeing global optimality [7].
The algorithm is dived into two steps: an inner-loop problem solved at each time point of the considered drive cycle,
and a time-horizon control problem.
Inner-loop problem
Before solving the control problem for the overall considered drive cycle, an inner-loop optimization process is
performed to obtain the family of the best ICE operating points correspondent to specific values of output torque and
speed. Firstly, 𝑃𝑏𝑎𝑡𝑡 and the fuel consumption 𝑚̇𝑓𝑢𝑒𝑙 are evaluated as a function of the ICE torque and speed values
(𝑇𝐼𝐶𝐸 and 𝜔𝐼𝐶𝐸 , respectively) in (4).
𝑃𝑏𝑎𝑡𝑡 = ℎ(𝜔𝐼𝐶𝐸 , 𝑇𝐼𝐶𝐸 ) (4)
75
The inner-loop optimal problem to minimize the fuel consumption subject to a defined battery power is then defined
as
min [ 𝑚̇𝑓𝑢𝑒𝑙 = 𝐿 (𝜔𝐼𝐶𝐸 , 𝑇𝐼𝐶𝐸 ) ]

(5)
Subject to ℎ(𝜔𝐼𝐶𝐸 , 𝑇𝐼𝐶𝐸 ) − 𝑃𝑏𝑎𝑡𝑡 = 0
The optimized control variables (i.e. TICE*

and ICE
*
) can thus be defined by choosing a point of minimal fuel
consumption on each feasible ICE operating line as per Pbatt . Figure 6 illustrates all the possible operating points for
the inner-loop optimization problem, discriminating between pure electric and hybrid points whether the ICE is used
or not. The optimal operating points are represented in Figure 6 by the lower edge of the blue points cloud.
Figure 6 Instantaneous fuel consumption points for specific required output torque and speed values
In the time-horizon optimal control, this process reduces the bi-dimensional control variable ( TICE and ICE ) to the
single dimension of Pbatt only. The fuel consumption rate mfuel can thus be determined from Pbatt , which could be
decided by an on-board supervisory algorithm. The optimal ICE operating points in Figure 6 can be interpreted as a
sort of Pareto frontier with a clear physical interpretation: in the best operating conditions, less battery power is
needed when more fuel is consumed and viceversa.
Time-horizon control problem
From the assistance of the inner-loop optimal solutions, only Pbatt is the control variable that decides all the operating
points in the time-horizon plane of the optimal control problem. This variable sets the fuel consumption rate and the
ICE operating point, which subsequently fixes all the other system variables such as the speed and torque of the
MGs and the transmission status. The time-horizon control problem can thus be defined as
𝑡_𝑒𝑛𝑑
min 𝐽 = ∫ 𝑔(𝑃𝑏𝑎𝑡𝑡 , 𝑡) 𝑑𝑡
𝑡_0
̇ = f(SOC, 𝑃𝑏𝑎𝑡𝑡 ) (6)

Subject to 𝑆𝑂𝐶
𝑆𝑂𝐶(𝑡0 ) = 𝑆𝑂𝐶(𝑡𝑒𝑛𝑑 )
*
where g is the best fuel consumption rate according to Figure 6. The optimal control variable Pbatt in the PMP is
obtained minimizing a performance measure, which is defined as a Hamiltonian H. Its mathematical formulation is
illustrated in (7)
̇
H = g(𝑃𝑏𝑎𝑡𝑡 , 𝑡) + λ ∙ 𝑆𝑂𝐶 (7)
where λ is a constant co-state variable that can be tuned to achieve the optimal control objective. Recent studies
illustrated different methods to properly tune this parameter, obtaining fuel economy results consistently comparable
with the global optimum [7, 8]. However, accuracy may be questionable when the operating conditions change.
Moreover, tuning the equivalence factor may result computationally inefficient when dealing with component sizing
in the HEV powertrain design procedure.
76
Dynamic Programming
The DP approach is by far the currently most applied approach for HEV control. The concept of DP was proposed
by Richard Bellman in the 1940s and refined by Bellman himself in 1954 [9]. This global optimization method was
firstly introduced to the HEV problem by H. Mosbech in the 1980s [10]. However, because it was constrained by the
computation power available at that time, this approach did not draw much attention until the later work by Brahma
et al. in 2000 [11]. Since then, this topic has been studied extensively and was extended to power-split HEVs by Liu
in 2006 [12].
The DP approach guarantees global optimality through an exhaustive search of all control and state grids. Its process
is implemented backward from the final drive cycle time point to the initial one by searching for the optimal trajectory
among the discretized grid points, as illustrated in Figure 7. Particularly, the Bellman’s principle of optimality states
that the optimal policy can be obtained if a single-stage sub-problem involving only the last stage is solved first, then
the sub-problem involving the last two stages, last three stages, etc. until the entire problem is solved step by step.
Figure 7 Dynamic programming process
For the HEV control problem, this signifies the minimization of the cost function J defined in equation (8) over the
considered time horizon.
𝑁−1
𝐽 = ∑(𝑚̇𝑓𝑢𝑒𝑙𝑘 + 𝛼 ∙ 𝛥𝑆𝑂𝐶 2 )
𝑘=0
(8)
𝑆𝑂𝐶𝑘 − 𝑆𝑂𝐶𝑡𝑎𝑟𝑔𝑒𝑡 𝑆𝑂𝐶𝑘 < 𝑆𝑂𝐶𝑡𝑎𝑟𝑔𝑒𝑡
𝛥𝑆𝑂𝐶 = {
0 𝑆𝑂𝐶𝑘 ≥ 𝑆𝑂𝐶𝑡𝑎𝑟𝑔𝑒𝑡
SOCt arg et is the desired value of battery SOC, while 𝛼 represents an operating factor. DP is demonstrated to achieve
global optimality under a wide range of operating conditions, but its major drawback refers to the computational power
needed for exhaustively searching through all the possible solutions and tuning α.
Power-weighted Efficiency Analysis for Rapid Sizing

The Power-weighted Efficiency Analysis for Rapid Sizing (PEARS) has been introduced by Zhang et al. as a rapid
near-optimal control strategy for HEVs [13]. In the PEARS algorithm, mode overall efficiency values are retained as
the weighting factor for selecting hybrid or electric powertrain operation. Beforehand, speed and torque of power
components are swept to determine the optimal combination in terms of mode efficiency at each driving cycle point.
The overall mode efficiency values to maximize (  EV and HEV ) are illustrated in (9) for pure electric and hybrid modes,
respectively.
𝑙𝑜𝑠𝑠
𝑃𝐸𝑉
𝜂𝐸𝑉 = 1 − 𝑖𝑛
𝑃𝐸𝑉
𝑃𝐼𝐶𝐸_1 𝜂𝐺 𝜂𝑏𝑎𝑡𝑡 /(𝜂𝐼𝐶𝐸_𝑚𝑎𝑥 𝜂𝐺_𝑚𝑎𝑥 ) 𝑃𝐼𝐶𝐸_2 𝜂𝐺 𝜂𝑀 /(𝜂𝐼𝐶𝐸_𝑚𝑎𝑥 𝜂𝐺_𝑚𝑎𝑥 𝜂𝑀_𝑚𝑎𝑥 )
𝜂𝐻𝐸𝑉 = +
𝑃𝑓𝑢𝑒𝑙 + 𝜇𝑃𝑏𝑎𝑡𝑡 𝑃𝑓𝑢𝑒𝑙 + 𝜇𝑃𝑏𝑎𝑡𝑡 (9)
𝑃𝐼𝐶𝐸_3
+ 𝜇𝑃𝑏𝑎𝑡𝑡 𝜂𝑏𝑎𝑡𝑡 𝜂𝑀 /𝜂𝑀_𝑚𝑎𝑥
𝜂𝐼𝐶𝐸_𝑚𝑎𝑥
+
𝑃𝑓𝑢𝑒𝑙 + 𝜇𝑃𝑏𝑎𝑡𝑡
77
loss in
For the electric modes, PEV includes both battery loss and electric drive loss, and PEV refers to the power flowing
into the system. For the hybrid modes, all the possible power flows are illustrated in Figure 8. PICE _1 is the engine
power from the engine through the generator to the battery, PICE _ 2 is the engine power that flow from the engine
through the generator to the motor, PICE _ 3 is the engine power that flows directly to the final drive.
PICE _1  PICE _ 3  PICE _ 3 is the total engine power, Pbatt is the battery power and μ is a flag for battery assist. Pfuel is the
rate of fuel energy injected; subscripts G and M refer to generator (when the power is negative) and motor (when the
power is positive or zero). ICE _ max , G _ max , and M_ max are the highest efficiency of the engine, the generator, and
the motor, respectively.
Figure 8 Power flow in the hybrid modes
Once the entire driving cycle is analyzed to extract the optimal power split for each operating mode at each time step,
the powertrain is initially set to operate in electric modes only (the most efficient one according to speed and torque
output). Subsequently, a recursive process starts that aims at replacing electric with hybrid operation in the driving
cycle points where the smallest ranges between hybrid and electric mode efficiencies are observed. This iterative
procedure is conducted until charge-balance is realized and the battery State-of-Charge (SoC) exhibits equal values
at the beginning and at the end of the driving cycle. The mode-shifting schedule and the resulting fuel consumption
can be evaluated in this way. Details regarding the operation of the algorithm can be found in [13].
The PEARS algorithm was demonstrated to be able to obtain results similar to Dynamic Programming (DP), while
being over 10000 times faster Error! Reference source not found.. The issue with implementing a PEARS
technique is that it generates an unrealistic mode-shifting schedule. To overcome this drawback, the authors of the
algorithm tried to combine PEARS with DP: more uniform mode-shifting schedules were obtained, however
computational cost was increased at the same time Error! Reference source not found.. Anselma et al. [14]
detected and analyzed the problematic points of the PEARS algorithm. Subsequently, a solution to minimize mode-
shifting events was proposed without excessive increase of the computational cost.
CASE STUDY
After having presented the majorly employed optimization-based controllers for HEVs, this session presents a case
study to assess the advantages and drawbacks of each control strategy. The analysed vehicle data are reported in
Table 1, while the powertrain layout corresponds to the one of Figure 3.
Table 1 Vehicle parameters

Component Parameters
188 kW @ 5800 rpm
Engine
320 Nm @ 4400 rpm
PMG1max [kW] 60
PMG 2max [kW] 85
Final Drive Ratio 3.59
R1 : S1 3.15
R2 : S2 1.59
Vehicle Mass [kg] 2248
78
All the three control strategies presented above were implemented and simulated in MATLAB© software. In general,
a quasi-static approach was adopted considering the time step equal to 1 second. The UDDS driving cycle profile
was retained as a good representative of the urban driving conditions. The correspondently obtained results for the
fuel consumption trends and battery SOC trajectories can be seen in Figure 9. Table 2 reports the calculated total
fuel consumption values and the correspondent computational time considering a desktop computer with Intel Core
i7-8700 (3.2 GHz) and 32 GB of RAM.
Table 2 Simulation results

Control strategy UDDS fuel consumption [g] Computational time [minutes]
DP 258 5
PMP 261 4
PEARS 284 2
Figure 9 Simulation results for the control strategies in UDDS
Simulation results confirm the capability of reaching global optimal fuel economy employing a DP approach. On the
other hand, PMP and PEARS establish themselves as near-optimal control strategies for HEVs, with the
correspondent fuel economy predicted values increased by 1.16 % and 11.1% respectively. In this particular case
study, the performance achieved by the PEARS algorithm in fuel economy is lightly lower than its major competitors.
This is due to the PEARS especially designed for HEV powertrains with several operating modes, where it is able to
realize consistent fuel economy with DP [13]. On the other hand, the HEV powertrain considered in Figure 3 can
operate only one electric mode and one hybrid mode, thus compromising the flexibility and freedom to operate of
PEARS. However, the PEARS achieves interesting results in terms of computational efficiency being the most rapid
control strategy to determine the powertrain operating schedule. Indeed, DP requires consistently increased
computational effort to evaluate the global optimal solution for the HEV control problem, while PMP represents a
trade-off between the other two approaches in this sense.
CONCLUSIONS
In this paper, references are provided concerning the different powertrain architectures for HEVs. A modelling and
optimization technique for power split HEV powertrains is illustrated, which include analytical formulations for the
transmission and experimental tables for the power components.
The three mainly adopted optimization-based control techniques for HEVs (i.e. PMP, DP and PEARS) are
subsequently presented and their mathematical formulations are analysed. These strategies are then simulated
considering the operation of a HEV powertrain design from the industrial state-of-art in the urban driving cycle.
Results show that the DP is effectively capable of achieving global optimal performance in terms of fuel economy.
On their behalf, PMP and PEARS demonstrate near-optimal fuel economy results while increasing the computational
efficiency.
79
The control strategy to adopt in the HEV design process depends on the peculiar application. The PEARS algorithm
reveals particularly efficient for optimizing power split HEVs with multiple operating modes, due to its ease of
implementation and flexibility in the operation. Moreover, its computational rapidness represents a consistent
advantage when analysing millions of different HEV powertrain configurations. On the other hand, employment of
DP and PMP may be suggested for the cases of HEV powertrains with a limited number of different operating modes,
since they can efficiently predict optimal fuel economy values.
The optimization of a HEV architecture may demonstrate significantly improved when the accuracy of the vehicle
model is increased. As an example, micro and macro road profiles may replace the standard driving cycles to simulate
real-world driving conditions [15]. Moreover, including the position and the characteristics of the powertrain may be
accounted as well to reduce the negative effects of vibrations [16]. Finally, the optimization reliability can be
consistently enhanced when the physical perception and the behaviour of the driver are included in the model [17]
18].
REFERENCES
[1] Regulation (EU) No 333/2014 of the European Parliament and of the Council of 11 March 2014 amending
Regulation (EC) No 443/2009 to define the modalities for reaching the 2020 target to reduce CO2 emissions
from new passenger cars, April 2014,
[2] Liu, J., Peng, H., Filipi, Z.S.: “Modeling and Analysis of the Toyota Hybrid System”, Proceedings of the 2005
IEEE/ASME International Conference on Advanced Intelligent Mechatronics, 24-28 July, 2005 Monterey, USA,
[3] Arata, J., Leamy, M., Meisel, J., Cunefare, K., Taylor, D.: "Backward-Looking Simulation of the Toyota Prius and
General Motors Two-Mode Power-Split HEV Powertrains", SAE Int. J. Engines, Vol. 4, No. 1, pp 1281-1297,
2011,
[4] Pittel, M., Martin, D.: “eFlite Dedicated Hybrid Transmission for Chrysler Pacifica”, SAE Technical Paper 2018-
01-0396, 2018,
[5] Anselma, P.G., Huo, Y., Roeleveld, J., Emadi, A., Belingardi, G.: “Rapid optimal design of a multimode power
split hybrid electric vehicle transmission”, Proceedings of the Institution of Mechanical Engineers, Part D: Journal
of Automobile Engineering, online Jan. 2018,
[6] Zhang, X., Li, S., Peng, H., Sun, J.: “Efficient Exhaustive Search of Power Split Hybrid Powertrains with Multiple
Planetary Gears and Clutches”, Journal of Dynamic Systems Measurement and Control, Vol. 137, No. 12, 2015,
[7] Kim N., Cha, S., Peng, H., "Optimal Control of Hybrid Electric Vehicles Based on Pontryagin's Minimum
Principle", in IEEE Transactions on Control Systems Technology, Vol. 19, No. 5, Sept. 2011, pp. 1279-1287,
[8] Kim, N., Rousseau, A.: “Sufficient conditions of optimal control based on Pontryagin’s minimum principle for use
in hybrid electric vehicles”, Proceedings of the Institution of Mechanical Engineers, Part D: Journal of
Automobile Engineering, Vol 226, Issue 9, 2012, pp 1160 – 1170,
[9] Bellman, R.: “The Theory of Dynamic Programming”, Bulletin of American Mathematical Society, 1954,
[10] Mosbech, H.: "Optimal control of hybrid vehicle", Proceedings of the Int. Symp. Automotive Technology
Automation (ISATA '80), Vol 2, 1980, pp 303-320,
[11] Brahma A., Guezennec, Y., Rizzoni, G.: "Optimal Energy Management in Series Hybrid Electric Vehicles," in
American Control Conference, 2000, Chicago, IL,
[12] Liu, J. Peng, H.: "Control optimization for a power-split hybrid vehicle," in American Control Conference, 2006,
Minneapolis, MN,
[13] Zhang, X., Peng, H., Sun, J.: “A near-optimal power management strategy for rapid component sizing of power
split hybrid vehicles with multiple operating modes”, in American Control Conference (ACC), 2013, Washington,
DC, USA, pp 5972-5977,
[14] Anselma, P.G., Huo, Y., Amin, E., Roeleveld, J., Emadi, A., Belingardi, G.: “Mode-shifting Minimization in a
Power Management Strategy for Rapid Component Sizing of Multimode Power Split Hybrid Vehicles,” SAE
Technical Paper 2018-01-1018, 2018,
[15] Demic, M., Diligenski, D., Demic, I.: “New design of a road profiler”, International Journal of Automotive
Technology, Vol. 11, No. 6, 2010, pp 801-808,
[16] Demić, M.: “Contribution to the optimization of the position and the characteristics of passenger car powertrain
mounts”, International Journal of Vehicle Design, Vol. 11, No. 1, 1990, pp 87-103.
[17] Demić, M.: “Optimization of Vehicles Elasto-Damping Elements Characteristics from the Aspect of Ride Comfort,
Vehicle System Dynamics”, International Journal of Vehicle Mechanics and Mobility, Vol. 23, No. 1, 1994, pp
351-377,
[18] Demic, M., Lukic, J.: “Investigation of the transmission of fore and aft vibration through the human body”, Applied
Ergonomics, Vol. 40, No. 4, 2009, pp 622-629.
80
Kragujevac, Serbia
MVM2018-IL8
Dušan Teodorović1
TRANSPORTATION SCIENCE: CHALLENGES AND FUTURE

RESEARCH
ABSTRACT: The need for transportation evolves from a variety of man’s activities. People take a trip motivated
by business, professional, cultural, or private reasons. Thousands of vehicles and passengers that travel from one
place to another generate flows of cars on highways, bicycles on streets, peoples in bus and metro stations,
pedestrians on crossings, and aircraft on airport’s taxiways and runways. Passenger and freight flows are the
consequences of spatial interaction among different regions.
Study, planning and design of transportation systems are the heart of the transportation engineering. Planners and
traffic engineers are continually confronted with the question how to plan, design and maintain high-quality
transportation system and livable cities? Transportation engineering deals with planning, design, operations, control,
management, maintenance and rehabilitation of transportation systems, services and components. Inside
transportation engineering, technology, mathematics, physics, computer science, social sciences, and cultural
heritage meet. Transportation planning and traffic control techniques have high influence on transportation system
performances (level of service, capacity, safety, reliability, resource consumption, environment, economics, etc.).
Transportation and traffic systems are, in essence, different from other technical systems. Their performances very
much depend on the users’ behavior. Studying of the human decision-making mechanism is one of the crucial factors
in the transportation planning process, as well as in developing suitable real-time traffic control. There are continuous
construction, and expansion of traffic networks. Previous to the construction of a new bridge, road expansion, or
development of a toll road, it is essential to consider how the potential users of the facility will react. Good
understanding of the human decision-making mechanism is of a high importance, since it has been shown that
building additional roads, in some cases, does not, without doubt, create decrease in total travel time in the
transportation network.
A number of the existing transportation systems are characterized by outstanding performances. For example, the
annual average delay of Shinkansen trains in Japan is only 0.9 minutes per operational train (including delays due
to unmanageable causes, such as natural disasters). On the other hand, a variety of transportation systems in a lot
of countries are inefficient, and not enough reliable. Simultaneously, they are huge consumers of energy and great
polluters. A number of the scheduled flights are cancelled on a daily basis. Traffic incidents occur on highways and
many links in a traffic networks are completely congested. Traffic engineers and operators must be also able to find
practical, “good” solutions for the complex transportation problems initiated by random events.
A number of transportation networks are extremely big. They are characterized by complex relationships between
specific nodes and links, and they are frequently congested. Accordingly, it is not simple to observe and study them,
and to find suitable solutions for traffic problems. Most often, it is not possible to control the whole network in a
centralized way.
Transportation science and transportation engineering suggest a range of techniques related to transportation
modeling, transportation planning, and traffic control. These techniques should be used for predicting travel and
freight demand, planning new transportation networks, and developing traffic control strategies. The array of
1
Dušan Teodorović, University of Belgrade, Faculty of Transport and Traffic Engineering, Vojvode Stepe 305,
11000 Belgrade, Serbia, Serbian Academy of Sciences and Arts, Knez Mihailova 35, 11000 Belgrade, Serbia
81
engineering concepts and methods should be used to make future transportation systems safer, more cost-effective,
and more “greener”.
We highlight in this lecture challenges and future research in the areas of transportation planning, traveler behavior,
transportation networks, public transit, traffic control, vehicle routing and scheduling, and logistics.
KEYWORDS: transportation, traffic, research
PRESENTATION
82
83
84
85
86
87
88
89
90
CONCLUSION
In modern world, passengers and goods move over the land, under the land, over the seas and oceans and, over
the sky. The development of technology and transportation systems enabled modern world to become a "global
village". Simultaneously, mankind has achieved significant economic and cultural development. Modern
transportation systems have made possible the development of political, economic, cultural, touristic, and sporting
relations among people. Sustainable and efficient transportation is one of the most important factors for the survival
and progress of modern civilization.
Modern, large-scale transportation networks, their complexity, high level of congestion, and high transportation costs
call for comprehensive approach to transportation planning and traffic control problems. Transportation planners and
engineers should plan, design and maintain transportation systems in a way to interact in a socially desirable way
with other systems. They should always consider possible consequences of the transport policy, transportation
planning and/or traffic control on various areas of the society (land use, regional planning, air pollution, safety,
agriculture, food distribution). Wrong decisions in the transportation arena are environmentally damaging and highly
costly.
91
REFERENCES
[1] Beckmann, M. J., McGuire, C. B., Winsten, C. B.: "Studies in the Economics of Transportation", Yale University
Press, New Haven, Connecticut, 1956,
[2] Blondé, B.: "At the cradle of the transport revolution? Paved roads, traffic flows and economic development in
eighteenth-century Brabant”, The Journal of Transport History, Vol. 31, 2010, pp 89-110,
[3] Boyce, D. E., Mahmassani, H. S., Nagurney, A.: "A Retrospective on Beckmann, McGuire, and Winsten’s
Studies in the Economics of Transportation", Papers in Regional Science, Vol. 84, No. 1, 2005, pp 85-103,
[4] Braess, D., "Uber ein Paradox der Verkerhsplannung", Unternehmenstorchung, Vol. 12, 1968, pp 258-268,
[5] Clark, J.: "Ships, Clocks, and Stars: The Quest for Longitude: National Maritime Museum Greenwich, London“,
The Journal of Transport History, Vol. 36, 2015, pp 124-126,
[6] Fitzpatrick, S.M., Callaghan, R.: "Magellan’s Crossing of the Pacific: Using Computer Simulations to Examine
Oceanographic Effects on One of the World’s Greatest Voyages", The Journal of Pacific History, Vol. 43, No. 2,
2008,
[7] Gazis, D., “The Origins of Traffic Theory”, Operations Research, 50, 50th Anniversary Issue, 2002,pp 69-77,
[8] Harvey, E.: "Pavage grants and urban street paving in medieval England, 1249–1462”, The Journal of Transport
History, Vol. 31, 2010, pp 151-163,
[9] Janić, M.: "Advanced Transport Systems: Analysis, Modelling and Evaluation of Performances", Springer,
Springer-Verlag, 2014, London, UK,
[10] Matthews, K., Jr.: "The embattled driver in ancient Rome", Expedition, Vol. 2, 1960, pp 22-27,
[11] Nichols Busch, T.: "Connecting an empire Eighteenth-century Russian roads, from Peter to Catherine", The
Journal of Transport History, Vol. 29, 2008, pp 240-257,
[12] Purcher, J., Buehler, R., Bassett, D.R., Dannenberg, A.L.: "Walking and Cycling to Health: A Comparative
Analysis of City, State, and International Data", American Journal of Public Health, Vol. 100, 2010, pp 1986–
1992,
[13] Takagi, K.: "Expansion of High Speed Rail Services: Development of High-Speed Railways in China", Japan
Railway & Transport Review, Vol. 57, 2011, pp 36-41,
[14] Teodorović, D, (Editor), "The Routledge Handbook of Transportation", 2016, New York,
[15] Teodorović, D., Janić, M.: "Transportation Engineering: Theory, Practice and Modeling", Elsevier, 2016, New
York,
[16] Vickrey, W.: "Congestion theory and transport investment”, American Economic Review (Papers and
Proceedings), Vol. 59, 1969, pp 251-261,
[17] Vuchic, V.: "Urban Transit: Operations, Planning, and Economics", John Wiley & Sons, 2005, New Jersey,
Hoboken, USA,
[18] Vuchic, V.: "Urban Transit Systems and Technology", John Wiley & Sons, Inc., Hoboken, 2007, New Jersey,
USA.
92
SECTION
Power Train Technology
 SI Engines
 CI Engines
 Alternative Fuel Engines
 Hybrid & Electrical Power
Systems
 Fuel Cell Systems
 After treatment Systems
 Transmission Systems
93
94
Kragujevac, Serbia
MVM2018-001
Sunny Narayan1
Ivan Grujic2
Nadica Stojanovic3
Vipul Gupta4
ANALYSIS OF PRESSURE PLATE OF AN AUTOMOTIVE

CLUTCH ASSEMBLY
ABSTRACT: The project presents design of an automotive clutch assembly using finite element method for
selected car model, which is Toyota KUN 25. Pressure plate of a single plate clutch was modelled and ANSYS
software was used for structural analysis. Structural analysis was done to obtain equivalent Von-Misses stress and
total deformation.
KEYWORDS: automotive systems, design, modelling
INTRODUCTION
A clutch (Figure 1) is a machine member used to connect driving shaft to the driven shaft so that the driven shaft
may be started or stopped at will, without stopping the driving [1]. There are three states in which clutch can be
found. These states are disengaged, clutch slipping and engaged.
Figure 1 Clutch and flywheel [1]
1
Sunny Narayan, Qassim University, Faculty of engineering, Saudi Arabia, rarekv@gmail.com
2
Ivan Grujic, M.Sc., assistant, University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić STR., 34000
Kragujevac, Serbia, ivan.grujic@kg.ac.rs
3
Nadica Stojanovic, M.Sc., assistant, University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić STR., 34000
Kragujevac, Serbia, nadica.stojanovic@kg.ac.rs
4
Vipul Gupta, Indus University, Faculty of Engineering, India
95
In an automobile, a gearbox is required to change the speed and torque of the vehicle. If we change a gear, when
the engine is engaged with gearbox or when the gears are in running position then it can cause wear and tear of
gears.
To overcome this problem a clutch is used between engine and gearbox. Clutch is the first element of power train.
The main function of clutch is to engage and disengage the engine to transmission, when the driver need or during
shifting of gear. When the clutch is in engaged position, the power flows from the engine to the wheels and when it
is in disengage position, the power is not transmitted [2].
The two main types of clutch are: positive clutch and friction clutch. Positive clutches are used when positive drive
is required. The simplest type of a positive clutch is a dog or claw clutch. A friction clutch has its principal
application in the transmission of power of shafts and machines which must be started and stopped frequently. The
force of friction is used to start the driven shaft from rest and gradually brings it up to the proper speed without
excessive slipping of the friction surfaces. In automobiles, friction clutch is used to connect the engine to the drive
shaft. The primary aim of this work is to design a rigid drive clutch system that meets multiple objectives such as
Structural strength.
Gradual engagement clutches like the friction clutches are widely used in automotive application for the
transmission of torque from the flywheel to the transmission [1]. The three major components of a clutch system
are the clutch disc, the flywheel and the pressure plate. Flywheel is directly connected to the engine's crankshaft
and hence rotates at the engine RPM. Bolted to the clutch flywheel is the second major component: the clutch
pressure plate. The spring-loaded pressure plate has two jobs: to hold the clutch assembly together and to release
tension that allows the assembly to rotate freely. Between the flywheel and the pressure plate is the clutch disc.
The clutch disc has friction surfaces similar to a brake pad on both sides that make or break contact with the metal
flywheel and pressure plate surfaces, allowing for smooth engagement and disengagement [2].
When the clutch begins to engage, the contact pressure between the contact surfaces will increase to a maximum
value at the end of the slipping period and will continue to stay steady during the full engagement period. During
the slipping period, large amount of heat energy is generated at the contact surfaces, which gets converted to
thermal energy by first law of thermodynamics. The heat generated is dissipated by conduction between the clutch
components and convection to the environment. Another loading condition is the pressure contact between the
contact surfaces that occurs due to the axial force applied the diaphragm spring. In addition to the above output
responses, this work also considers the vibration characteristics of the clutch plate during the full engagement
period. The engine and the transmission components experience dynamically varying loads during normal
operation. This will cause vibrations and hence, one must design the clutch system so as to avoid resonance with
the transmission and engine components [3-12].
METHODOLOGY
The aim of calculations was to find optimum thickness of pressure plate based on properties of material used.
The pressure plate is subjected to a pressure force that exerted towards the clutch plate by the spring. And the
thickness of the pressure plate can be calculating as following:
 The maximum bending stress,  max :
P  d02
 max  K1 (1)
4t 2
 The deflection, Ymax is estimated by this equation:
P r4
Ymax  K 2 (2)
16  E  t 3
Where,
t – thickness of pressure plate
d 0  260 mm – clutch outer diameter
r  102.5 mm – mean radius
d0
K1 and K 2 are constants. In this case from  1.7333 so using this value from table, K1  0.99 and K 2  0.43
di
P  0.4 N/mm2 from properties
d i – clutch inner diameter
96
The maximum bending stress in the pressure plate is given as:
400,000  0.262
 max  0.99
4t 2
6692.4
 max 
t2
The selected material is cast iron alloy (Code - ASTM A220) with properties:
 Tensile strength,   350 MPa and
 Yield strength,  y  215 MPa.
From the factor of safety equation we find the maximum bending stress:
y
 max  (3)
Fs
Where, factor of safety is 3.
215
 max   71.667 MPa
3
Then substituting this in the maximum bending stress, we get the thickness of the plate:
6692.4
71.667  106 
t2
t  9.66 mm
This is the optimum thickness of cast iron material that is structurally suitable to make pressure plate. Using these
dimensional values, deformations and stress concentrations were analyzed using FEA method discussed in next
section.
FEA ANALYSIS
The finite element analysis is the most widely accepted computational tool in engineering analysis. Through solid
modelling, the component is described to the computer and this description affords sufficient geometric data for
construction of mesh for finite element modelling.
The model of the pressure plate was meshed. The boundary conditions are defined as follows:
 Moment = 343 Nm;
 Pressure 1 = 0.2245 MPa and
 Pressure 2 = 0.2745 MPa.
Pressure 1 acts on the back surface of the pressure plate; it is the reaction pressure from the clutch plate, and
pressure 2 acts on the front surface of the pressure plate facing the diaphragm spring. All these values were taken
from calculations based on results of Maximum pressure theory.
The Figure 2 shows the distribution of equivalent von-Misses stress over the entire pressure plate. The Figure 3
shows the distribution of total deformation over the entire pressure plate.
97
Figure 2 The equivalent Von-Misses stress plot for the pressure plate
Figure 3 Total deformation plot for the pressure plate
CONCLUSIONS
In this work pressure plate of an automotive clutch assembly has been designed using and simulatio was done
using ANSYS. Among those different lining materials cast iron friction material was selected for analysis. The heat
between friction disk and flywheel, which is the main reason for clutch wear can be minimized by selecting suitable
material. A good contact pressure also reduces wear during slippage time.
Furthur comparisons can be made by using different materials for pressure plate and the best material can be
selected that is structurally suitable.
REFERENCES
[1] Narayan, S.: "A review of diesel engine acoustics", FME Transactions, Vol.42, No.2, 2014, pp 150-154,
[2] Bo, Z., Yi, C., Youbai, X., Kun, Z.: "Dynamics and lubrication analyses of a planar multibody system with
multiple lubricated joints", Proceedings of the Institution of Mechanical Engineers, Part J: Journal of
Engineering Tribology, Vol. 232, No. 3, 2017, pp 326-346,
[3] Fang, C., Xianghui, M., Youbai, X.: "A piston tribodynamic model with deterministic consideration of skirt
surface grooves", Tribology International, Vol. 110, 2017, pp 232-251,
[4] Bo, Z., Shen, F.: "A numerical coupling model for a multibody system with multiple lubricated clearance joints",
2017 International Conference on Mechanical, Aeronautical and Automotive Engineering (ICMAA 2017), Vol.
108, 2017,
[5] Cavalli, M., Lavacchielli, G., Tonelli, R., Nicoletto, G., Riva, E.: "Comparison of analytical and multibody
dynamic approaches in the study of a V6 engine piston", Proceedings of the Institution of Mechanical
Engineers, Part K: Journal of Multi-body Dynamics, Vol. 231, 2017, pp 420-438,
[6] Buzzoni, M., Mucchi, E., Dalpiaz, G.: "A CWT-based methodology for piston slap experimental
characterization", Mechanical Systems and Signal Processing, Vol. 86, 2017, pp 16-28,
[7] Meng, F. M., Yang, T.: "Preliminary study on mechanism of cavitation in lubricant of textured sliding bearing",
Proceedings of the Institution of Mechanical Engineers, Part J: Journal of Engineering Tribology, Vol. 227, No.
7, 2013, pp 695–708,
98
[8] Narayan, S.: "Noise Optimization in Diesel Engines", Journal of Engineering Science and Technology Review,
Vol. 7, No. 1, 2014, pp 37 – 40,
[9] Narayan, S.: "Wavelet Analysis of Diesel Engine Noise", Journal of Engineering and Applied Sciences, No. 8,
2013, pp 255-259,
[10] Narayan, S.: "Time-frequency analysis of diesel engine Noise", ActaTehnicaCorviniensis – Bulletin of
Engineering, Vol. 7, No. 3, 2014,
[11] Narayan, S.: "Pressure Waveform Modelling in Diesel engines", ActaMechanicaSlovaca, Vol. 19, No. 1, 2015,
pp 48-51,
[12] Narayan, S.: "Correlation between in cylinder pressure and noise emissions from engines", Journal of Kones
powertrain and Transport, Vol. 22, No. 1, 2015, pp 243-254.
99
100
Kragujevac, Serbia
MVM2018-004
Sreten Perić1
Bogdan Nedić2
Mile Stoiljković3
Radovan Radovanović4
EXPERIMENTAL RESEARCH OF THE BIODEGRADABLE

UNIVERSAL TRACTOR OIL BASED ON THE VEGETABLE
OILS
ABSTRACT: Vegetable oils are potential substitutes for petroleum-based oils. Not only they are environmentally
friendly, renewable and less toxic, but also they have excellent lubricating properties. Despite their safety,
vegetable oils suffer from several drawbacks, for instance in terms of thermal and oxidative stability. The
combination of additives in vegetable oils is the best option to overcome these problems. However, the additives
are often combination of many substances which may be toxic. This paper presents the features of the oils based
on rapeseed oil, sunflower, soya and mixture of rapes oil with mineral oil is, after corresponding testing, compared
with the features of the commercially available mineral oil based UTTO. According to the obtained results for
physico-chemical properties of various oil samples, it may be concluded that almost all the investigated properties
of biodegradable universal tractor oils, satisfy the John Deere and Massey Ferguson specifications, and some
characteristics are even better when compared to the properties of universal mineral based oil. The exceptions are
the oxidation stability and low temperature characteristics. The properties of vegetable oils were improved by the
corresponding additives, specially designed for these oils.
KEYWORDS: biodegradable universal tractor oil, oil analysis, physico-chemical characteristics, tribological
characteristics
INTRODUCTION
Due to the aforementioned potential dangers, during some last twenty years the ecologically acceptable oils are
more and more used. Ecologically acceptable oils are the oils which in contact with the environment produce the
minimum of harmful effects [1-7]. The conditions for the ecological acceptance are biodegradability and no toxicity
of lubricants. Besides, ecologically acceptable lubes are produced from the renewable sources (vegetable oil) thus
reducing the dependence of mineral oils. The interest for the ecologically acceptable oils is growing every year,
especially in the field where the application of mineral oils can produce harm in the environment, like the oil losses
of motor saws, agricultural, forester and civil engineering machines, railroad and tramway crossings, outboard
motors for the motor boats, oils for mold release in civil engineering, steel ropes and chains and the other flow
lubrications where the total quantity of lubricant remains in the environment. The lubricant producer’s strategies are
geared toward the production of ecologically acceptable lubricants. For many producers the market of the future is
the market of biodegradable lubricants.
1
Perić Sreten, Ph.D., University of Defence, Military Academy, Pavla Jurišića Šturma Str. No. 33, 11000 Belgrade,
Serbia, sretenperic@yahoo.com
2
Nedić Bogdan, University of Kragujevac, Faculty of Engineering, 6 Sestre Janjic Street, 34 000 Kragujevac,
Serbia, nedic@kg.ac.rs
3
Stoiljković Mile, Ph.D., NIS GAZPROMNEFT, Serbia, mile.stojilkovic@nis.eu
4
Radovanović Radovan, Ph.D., The Academy of Criminalistic and Police Studies, 196 Cara Dušana Street, 11080
Belgrade-Zemun, Serbia, radovan.radovanovic@kpa.edu.rs
101
Vegetable oils are base oils for production of ecologically acceptable lubricants. In fact, they are triglycerides made
up of complex mixtures of fatty acids having different chain lengths and number of double bonds. Triglycerides are
very quickly biodegraded and have excellent lubricity properties. The advantages of vegetable oils versus mineral
oils are in the following characteristics - no toxicity, biodegradability, good lubricity, high flash point, high IV and low
volatility. The disadvantages of vegetable oils versus mineral oils are low oxidation stability, low fluidity on low
temperatures, low hydraulic stability and the price which is 1.5 to 2 times higher than the price of mineral oils [8-13].
Lubricants based on vegetable oils are suitable only for low temperatures not higher than 80 °C.
In the field of the application of ecologically acceptable lubricants, attention should be given to the technical
requirements including protection from the wear, corrosion protection, high load carrying properties, heat transfer
capability and fulfillment of all the requirements imposed by the producer of the mechanical system [14, 15].
THE ANALYTICAL COMPOSITION OF THE BIODEGRADABLE UNIVERSAL

TRACTOR OIL BASED ON THE VEGETABLE OILS
The features of oils based on rapeseed oil, sunflower oil, soybean oil and a mixture of rapeseed oil with mineral oil
were, after the corresponding testing, compared with the features of the commercially available mineral-based
universal tractor oil, UTTO (Table 1).
Table 1 Oil samples

Sample number Oil name Oil code
1. Rapeseed oil without additives RE
2. Soybean oil without additives SO
3. Sunflower oil without additives SU
4. Rapeseed oil with additives REA
5. Soybean oil with additives SOA
6. Sunflower oil with additives SUA
7. Rapeseed oil + additives + 10% SN150 REAM10
8. Rapeseed oil + additives + 20% SN150 REAM20
9. Mineral UTTO MIN
The ASTM D 4951 and ASTM D 4927 AAS methods have been applied to obtain the elemental composition of the
additives used in the test oils, as shown in Table 2 and Figure 1.
Table 2 The elemental composition of the additives used in the test oils
P Ca Zn S
% m/m
REA 0.07 0.08 0.13 0.19
SOA 0.06 0.07 0.14 0.19
SUA 0.06 0.07 0.14 0.19
REAM10 0.08 0.15 0.14 0.23
REAM20 0.09 0.18 0.15 0.28
MIN 0.11 0.34 0.15 0.54
0.6
0.5
0.4
% m/m
0.3
0.2
0.1
0
REA SOA SUA REAM10 REAM20 MIN
P Ca Zn S
Figure 1 The elemental composition of the additives for the test oils
102
It can be seen from the Figure 1 that the elemental composition of additives is quite similar for all formulated
vegetable based UTTO: REA, SOA and SUA. The reference mineral UTTO oil (MIN) contains a significantly higher
level of phosphorous, calcium and sulfur, than the ester-based oils. Zinc and phosphorous are parts of AW/EP
additive package, while calcium is a typical element in a detergent additive. Samples REAM10 and REAM20
contain increased levels of phosphorus, sulfur and calcium, as they contain proportionally higher percentage of
mineral oil.
Vegetable oils are generally less additivated than mineral oils, because they posses good lubricating properties
due to their polar nature. This makes them good solvents for sludge and dirt, which would otherwise deposit on
metal surfaces [16-19]. Because of these properties, it is possible to reduce the amount of friction modifiers,
antiwear additive package, and dispersants, when formulating biodegradable universal tractor oils.
PHYSICAL AND CHEMICAL PROPERTIES OF THE BIODEGRADABLE

UNIVERSAL TRACTOR OIL
The physical and chemical properties of the vegetable oils were examined in accordance with standard methods
(Table 3).
Table 3 Laboratory test methods

Method No. Physical and chemical properties Test method
1. Density, kg/m3 ASTM D 1298
2. Kinematic viscosity at 40 °C, mm2/s ASTM D 445
3. Kinematic viscosity at 100 °C, mm2/s ASTM D 445
4. Viscosity Index ASTM D 2270
5. Pour point, °C ASTM D 97 or ISO 3016
6. Flash point, °C ISO 2592, ASTM D 92
7. Foaming, ml/ml 24 °C; 94 °C; 24 °C ASTM D 892
8. Deaeration, minutes DIN 51381
9. Oxidation stability, minutes ASTM D 2272
10. Corrosion on copper, 3 hours at 121 °C ASTM D130
11. P content, % ASTM D 4927
12. S content, % ASTM D 2622
13. Ca content, % ASTM D 4628
14. Zn content, % ASTM D 4628
15. Wear, (1h; 65 °C; 40 kg and 1500 rpm), mm ASTM D 4172
16. 4-ball EP test - scuffing, kg ASTM D 2783
The results of experimental testing of physico-chemical properties are presented in Table 4 [20]. Experimental work
was carried out in accordance with the manufacturer specifications and proper standards, by using the necessary
testing equipment.
Table 4 Physico-chemical properties of oils

Physico-chemical Measuri
Methods RE SO SU REA SOA SUA REAM10 REAM20 MIN
properties ng unit
Density at 15 °C kg/m³ ASTM D1298 916 918 920 918 921 922 912 907 877
Kinematic
mm²/s ASTM D 445 34.85 32.73 35.12 42.35 37.70 38.64 42.03 41.00 70.56
viscosity at 40 °C
Kinematic
mm²/s ASTM D 445 7.9 7.82 7.93 9.49 9.17 9.27 9.23 9.18 10.05
viscosity at 100 °C
Viscosity Index ASTM D2270 210 224 209 218 227 226 211 201 126
Flash point °C ASTM D 92 322 326 328 254 260 250 248 246 234
Pour point °C ASTM D 97 -8 -13 -11 -23 -25 -24 -26 -27 -36
Foaming,
I sequnce, 24 °C 0/0 0/0 0/0 25/0 15/0 30/0 25/0 20/0 5/0
ml ASTM D 892
II sequence, 94 °C 0/0 0/0 0/0 20/0 20/0 20/0 20/0 20/0 20/0
III sequence 24 °C 0/0 0/0 0/0 20/0 20/0 10/0 10/0 10/0 5/0
Deaeration minutes DIN 51381 6 5 6 5 7 1
Oxidation stability,
minutes ASTM D2272 13 8 10 109 60 70 120 149 214
RBOT
103
Table 4 (continued)
Wear, (1h; 75 °C; ASTM D
mm 0.66 0.69 0.68 0.38 0.39 0.37 0.39 0.39 0.36
40kg and 1200 rpm) 4172
4-ball EP test - ASTM D
kg 200 200 200 200 160 100
scuffing 2783
Protection against
ASTM D 665 pass pass pass pass pass pass pass pass pass
corrosion, Test B
Corrosion on
copper, 3 hours at ASTM D130 1A 1A 1A 1A 1A 1A 1A 1A 1A
121°C
Kinematic viscosity
Most tractor lubricants possess kinematic viscosity between 9 and 11 mm2/s at 100 °C. This viscosity is found to
provide sufficient thickness to promote good protection for the transmission system and anti-squawk performance,
yet still to be a suitable viscosity for the hydraulic system.
Kinematic viscosity
Viscosity at 40 C [mm2 /s]
80 70.56
60
42.35 42.03 41
34.85 37.7 38.64
40 32.73 35.12
20
Figure 2 Kinematic viscosity at 40°C, ASTM D 445
As can be seen from Table 4, viscosity of vegetable oils produced from oil seeds falls between 32.7 and 42.3
mm2/s at 40 °C, and between 7.8 and 9.4 mm2/s at 100 °C.
Kinematic viscosity
12
Viscosity at 100 C [mm2 /s]
10.05
9.49 9.17 9.27 9.23 9.18
10
7.9 7.82 7.93
8
6
4
2
0
Figure 3 Kinematic viscosity at 100°C, ASTM D 445
Vegetable oils have a much higher viscosity index (VI > 200) compared with mineral oils, which enables a reliable
tractor operation at higher temperature changes. As shown by Table 4, the viscosity of all vegetable oil based
UTTO, reveals them to be thicker at 100 °C, and with significantly higher viscosity index as compared with the
mineral based UTTO, labeled MIN.
104
Viscosity index
250 224 218 227 226
210 209 211 201
200
Viscosity index
150 126
100
50
Figure 4 Viscosity index, ASTM D2270
Oxidation stability
Most vegetable oils are triglycerides constituting a complex mixture of fatty acids with different chain length and
instauration content [18-24]. The alcohol component (glycerin) is the same in all vegetable oils. The fatty acid
components are plant-specific and therefore variable. The fatty acids differ in chain length and number of double
bonds. From the fatty acid composition of the oils, it is observed that chain length C18 is dominating (Table 5).
Main fatty acids with double bonds are linolenic, linoleic and oleic. The oxygen absorption rate is 800:100:1
respectively, therefore less double bonds in a carbon chain result in better oxidation stability [15]. Generally the
oxidation stability of vegetable based oils decreases with the increased level of instauration.
The content of polyunsaturated fatty acids (C18:2) is rather high for soybean (SO) and sunflower (SU) oil. Under
thermal conditions, the double bonds in polyunsaturated fatty acids polymerize much faster than monounsaturated
(C18:1 and C22:1) or saturated (C16:0 and C18:0) fatty acids. Unfortunately, the saturation of fatty acid
degenerates the low temperature behavior or pour point of the oil.
Iodine value characterizes particular oil on the base of unsaturated fatty acids. Oils with high iodine values are
more problematic for oxidation processes, however values fewer than 100 are not recommended since such oils
are more problematic for changing the characteristics at lower temperatures.
Table 5 Fatty acid content and iodine value of vegetable base oil
Physico-chemical properties Unit Rapeseed oil Soybean oil Sunflower oil
Iodine value gI2/100g 118.41 126.2 131.2
Fatty acid content %
C14:0 (Myristic acid) 0.06 0.05 0.04
C16:0 (Palmitic acid) 6.58 10.24 6.35
C16:1 (Palmitoleic acid) 0.36 0.15 0.13
C18:0 (Stearic acid) 2.88 5.24 5.35
C18:1 (Oleic acid) 53.10 29.33 27.13
C18:2 (Linoleic acid) 28.72 47.95 58.53
C18:3 (Linolenic acid) 6.54 5.35 0.16
C20:0 (Arachidic acid) 0.41 0.52 0.41
C20:1 (Eicosenoic acid) 0.73 0.29 0.20
C22:0 (Behenic acid) 0.28 0.65 1.31
C22:1 (Erucic acid) 0.17
C24:0 (Lignoceric acid) 0.10 0.20 0.35
The oxidation stability of oil samples was examined by the ASTM D2272 test (RBOT - Rotating Bomb Oxidation
Test). As concerning the neat vegetable oils (without additives), rapeseed oil (RE) shows better oxidation stability
as compared to the soybean (SO) and sunflower (SU) oil. The reason is a high content of oleic acid and a low
iodine value present in rapeseed oil. The oxidation stability of vegetable oils without additives is very low.
The improvement of the oxidation stability of vegetable oils was accomplished by adding the antioxidant additives
and the mineral base oil. The concentration of the additives was the same for all vegetable oils. The oxidation
stability values determined by RBOT method for neat vegetable oils, vegetable oils with additives and vegetable
oils with mineral base oil, were compared with the values for commercial UTTO oil, as it is shown in Figure 5.
105
The best oxidation stability shows sample REAM20 (149 minutes). The oxidation stability is improved by adding the
antioxidant additive: for rapeseed oil, the improvement is more than eightfold. The stability is further increased by
adding a mineral oil, but the biodegradability is reduced.
A low oxidation stability of newly formulated biodegradable oils limits their use for the production of motor oils and
other oils that require high oxidation stability. Therefore, such oils may be used in the fields where the high
oxidation stability is not required (flow lubricating oils, hydraulic oils, universal tractor oils), or in the agriculture and
forestry, where a low toxicity and biodegradability of lubricants is mandatory [25-31].
Oxidation stability
250 214
200
149
150 120
minutes
109
100 60 70
50 13 8 10
0
Figure 5 Oxidation stability, ASTM D2272 (RBOT)
Flash point
Flash point is important in transport and storage due to risk of fire. Vegetable oils have higher flash point values in
comparison with mineral oils (Figure 6).
Flash point for vegetable oils is higher than 300 °C. By adding a package of additives and mineral oil according to
the formulation, the flash point is reduced, but it is far above the allowed values according to the specifications of
tractor manufacturers (Massey Ferguson CMS M1141, Massey Ferguson CMS M1143; John Deere J20C: Flash
point ≥ 200 °C).
Flash point
400
322 326 328
Temperature [ C]
300 254 260 250 248 246 234
200
100
Figure 6 Flash point, ASTM D 92
Pour point
The flowability of vegetable oils at low temperatures is extremely low, which limits their use at low operating
temperatures (Figure 7). Vegetable oils form crystal structures at low temperatures, by agglomeration of
triglycerides, wherein the oil flowability is reduced. In order to improve the low-temperature characteristics,
vegetable oils are added additives labeled as pour point depressants (PPD). The function of these additives is to
prevent the crystallization of the triglyceride molecules at low temperatures and their further grouping. The optimal
concentration of additive PPD of 1% in the final formulation of vegetable oils, significantly improves their low
temperature properties.
106
Pour point
0
-5
-10 -8
Temperature [ C]
-11
-15 -13
-20
-25 -23 -24
-25 -26
-30 -27
-35
-36
-40
Figure 7 Pour point, ASTM D 97
Oil foaming
Foam affects the lubrication process very adversely, and also it accelerates the oil oxidation. Antifoaming additives
have a role to prevent the occurrence of foam, by preventing the entry of air into the oil, which causes the foam. It
is necessary to add this type of additives into the vegetable oils.
The tests of oil foaming have led to the interesting results. The neat vegetable oils free of additives have shown a
good resistance to the appearance of foam (0/0; 0/0; 0/0) (Table 6). A mixture of vegetable and mineral oils gives
also good results related to the appearance of foam, which means that the addition of mineral oil does not affect
the foaming. However, after the addition of additives, there was an intense foaming (Table 6). Further laboratory
analysis of oil with each individual additive, led to the conclusion that the foam occurrence is caused by additives
named pour point depressants (PPD) and additives for protection against extreme pressures and shock loads (EP).
On the contrary, the additives that protect against wear (AW) and oxidation (AO) did not initiate the appearance of
foam.
Table 6 The laboratory analysis results of oil foaming, without antifoaming agents
Foaming Unit RE SO SU REA SOA SUA REAM10 REAM20
I seq., 24 ºC 0/0 0/0 0/0 350/140 380/120 440/210 500/280 600/290
II seq., 94 ºC ml 0/0 0/0 0/0 290/0 270/0 400/0 680/0 350/0
III seq., 24 ºC 0/0 0/0 0/0 390/190 350/130 300/110 180/60 280/25
To solve the problem of vegetable oil foaming, the antifoaming agents were added. The application of conventional
antifoaming additives did not give the required results. Better results were obtained with 30 ppm of the additive
based on silicone (dimethylsiloxane), which improved the sequence II, or with the second additive
(alkylmethacrylate) in an amount of 80 ppm, which improved the sequences I and III. However, the results obtained
(180/0; 50/0; 20/0) were not within the acceptable limits, according to the manufacturer's specifications. Only after
the addition of 60 ppm of the substance based on silicone (dimethylsiloxane), and 160 ppm of alkylmethacrylate,
the satisfying results were obtained. The oil foaming data are presented in Table 7.
Table 7 The laboratory analysis results of oil foaming, with antifoaming agents
Foaming Unit REA SOA SUA REAM10 REAM20 MIN
I seq., 24 ºC 25/0 15/0 30/0 25/0 20/0 5/0
II seq., 94 ºC ml 20/0 20/0 20/0 20/0 20/0 20/0
III seq., 24 ºC 20/0 20/0 10/0 10/0 10/0 5/0
Rust and corrosion

Rust is a product of reactions between a metal and corrosive environment. Corrosion is an electrochemical
process, usually caused by the action of acidic chemical compounds on the metallic surface. Corrosion inhibitors
(anti-corrosion additives) protect metal parts that are exposed to the rust and corrosion [32]. The basic mechanism
of action of anti-corrosion additives is based on the formation of a protective layer on the metal surface, by physical
and chemical adsorption [33, 34]. Rust and corrosion inhibitors are commonly sulfonates, which are surface-active
substances and provide good results in vegetable oils. Antioxidant additives are multifunctional and they can also
protect metal against corrosion [20].
107
All of the oil samples showed good results in terms of protection against rust, according to the test B, ASTM D 665
method (the result was “passes”) and corrosion according to the method ASTM D130 (the result was “1A”), as
shown in Table 4.
Abrasion and high pressures resistance

Tests of the abrasion and high pressures resistance were carried out on the device with four balls ("four ball" test),
and the results are shown in Table 8. The wear resistance testing was carried out according to the method ASTM
D 4172.
Table 8 The wear resistance test parameters and results

Characteristics Unit RE SO SU REA SOA SUA REAM10 REAM20 MIN
Wear mm 0.66 0.69 0.68 0.38 0.39 0.37 0.39 0.39 0.39
4-ball EP test kg 140 140 140 200 200 200 200 160 100
The test conditions were as follows: the pressing force is 392 N, the top ball is rotated at 1200 rpm for 60 min., and
the temperature of the test lubricant is regulated at 75 °C. The limiting value of the scar diameter worn on the ball is
maximum 0.4 mm for UTTO oil, according to the Massey Ferguson MF 1135 specification. The vegetable oil
samples without additives (RE, SO, SU) did not pass the standard test, because the wear intensity was 70% higher
than allowed. The wear parameters for other samples are presented in Figure 8, and it is seen that they are quite
uniform, and within the allowed limits (< 40mm).
0.69 0.68
0.66
0.7
0.6
Diameter Wear [mm]
0.5
0.38 0.39 0.37 0.39 0.39 0.39
0.4
0.3
0.2
0.1
0
Figure 8 Wear resistance of vegetable oils and their mixtures
The measurement of extreme-pressure properties was performed according to the ASTM D 2783 (Four-ball
method). The load is steadily increased until welding occurs, and the welding value of load is recorded as a
maximum load which can be carried out by lubricant. The results in Table 6 indicate that vegetable based UTTO
oils possess higher load values comparing to mineral oils, which means that they can better withstand extreme
pressures and suddenly applied stress. The laboratory tests by using four-ball method have revealed that
vegetable oils free of EP additives, possess even better results than mineral oils with the additives, as it is seen in
Figure 9.
Extreme pressure resistance

200 200 200 200
200
160
140 140 140
150
Load [kg]
100
100
50
Figure 9 Extreme pressure resistance of vegetable oils and their mixtures
108
TRIBOLOGICAL PROPERTIES OF THE BIODEGRADABLE UNIVERSAL
TRACTOR OIL
Most tribological properties are determined using defined standard tests, as shown in Table 9. For these purposes,
some of the standard commercial tribometers such as Falex, Timken, Plint, Koehler, SRV, etc. can be used.
Table 9 Lubricant tribological characteristics tests

Standard Contact geometry
Wear Preventive Characteristics
ASTM D2670 Pin and Vee Block
ASTM D4172 Four-Ball
Extreme Pressure Properties
ASTM D2782 Block on Ring
ASTM D2783 Four Balls
ASTM D3233 Pin and Vee Block
IP 240 Block on Ring
Friction characteristics
ASTM D 5183 Four-Ball
Tribological behavior of lubricant in devices with real mechanical elements
ASTM D5182 FZG – Gear Test Rig
The measuring system applied to determine tribological properties of the tribomechanical system elements is
shown in Figure 10. This new measurement equipment for tribological studies have been developed and produced
over the past few years at Faculty of Engineering, University of Kragujevac [35, 36]. It is composed of:
 a TPD-93 Tribometer for normal force, frictional force, and frictional coefficient measurements;
 thermopairs for temperature measurements of oil and elements in contact;
 a PQ-2000 particle quantifier;
 a microscope for wear parameter measurements (length, width, and depth of worn zone);
 Talysurf 6, a computerized measuring device for surface topography and wear parameter measurements.
Figure 10 Measuring system applied for determination of tribological properties
The results obtained during the tests contain information about the friction coefficient, friction force, width and depth
of the wear track, wear shape of the block's contact surface, change in the friction coefficient and temperature
during the time of contact, block's and disc's surface topography parameters before and after the testing, etc. To
investigate the dynamics of the contact process, the force and friction coefficient signals were recorded
continuously for later processing.
Table 10 Tribological characteristics of oils

Friction coefficient μ Boundary lubrication
Oil name Fn=30N Fn=100N Fn=300N Block's wear width Block's wear depth
t=15 min t=15 min t=30 min (mm) (μm)
RE 0.058 0.065 0.062 1.117 8.9145
SO 0.092 0.095 0.098 1.278 11.6702
109
Table 10 (continued)
SU 0.063 0.069 0.069 1.175 9.8645
REA 0.053 0.071 0.077 1.460 15.2325
SOA 0.098 0.098 0.091 1.513 16.3588
SUA 0.093 0.084 0.080 1.345 12.9264
REAM10 0.087 0.084 0.079 1.455 15.1281
REAM20 0.094 0.090 0.088 1.502 16.1217
MIN 0.076 0.098 0.099 1.585 17.9537
The friction coefficients measured under various loads and in different periods are presented in Figure 11. In the
first phase, the coefficient was measured under the force of 30N during 15 mins, then the force increased to 100 N
for 15 mins, and in the final stage the force was 300 N during 30 mins.
When low force is applied (30 N), the lowest friction coefficient is recorded for rapeseed oil with additives
(biodegradable UTTO oil based on rapeseed oil), while the highest coefficient is for the soybean oil with additives
(biodegradable UTTO oil based on soybean oil).
Under the middle load values (100 N), the rapeseed oil without additives has the lowest coefficient, whereas the
additive free soybean oil and mineral UTTO oil show the highest coefficient.
Finally, at the highest applied force (300 N) during 30 mins experiment, the additive free rapeseed oil again shows
the lowest coefficient, and the highest value is again recorded for mineral UTTO oil.
It may be concluded from the diagram in Figure 11 that the friction coefficient significantly depends on the oil type,
and that the rapeseed oil, which possesses the highest percent of monounsaturated fatty acids (oleic acid, C18:1)
and the lowest iodine number (118.41 gI2/100g), shows the lowest coefficient.
Friction coefficient µ
0.1
Friction coefficient µ
0.08
0.06
0.04
0.02
Fn=30N t=15 min Fn=100N t=15 min Fn=300N t=30 min

Figure 11 Friction coefficients measured under various loads and examination times
Figure 12 illustrates the block's wear width measured after the examination of different oil samples. The highest
width value was observed for the mineral UTTO oil sample (MIN UTTO, 1.585 mm), and the lowest for the
rapeseed additive free oil sample (RE, 1.117 mm). The low block's wear width values for the RE, SO and SU
samples indicate good lubrication properties of vegetable oils, i.e. they prove that the triglyceride structure of these
oils is sufficiently effective against wear even at the highest pressures.
Wear width
1.585
1.46 1.513 1.455 1.502
1.6
1.345
1.4 1.278
1.117 1.175
Wear width [mm]
1.2
1
0.8
0.6
0.4
0.2
0
Figure 12 Block's wear width for different oil samples

110
CONCLUSIONS
According to the obtained results for physico-chemical properties of various oil samples, it may be concluded that
almost all the investigated properties of biodegradable universal tractor oils, satisfy the John Deere and Massey
Ferguson specifications, and some characteristics are even better when compared to the properties of universal
mineral based oil. The exceptions are the oxidation stability and low temperature characteristics. The properties of
vegetable oils were improved by the corresponding additives, specially designed for these oils.
 The vegetable oils show considerably higher viscosity index (VI > 200) than mineral oils, allowing a reliable
tractor operation at wider temperature changes.
 Flash point is higher for vegetable oils as compared to the mineral.
 Low temperature fluidity of vegetable oils is far from satisfactory, thus limiting their use at low
temperatures. However, PPD additive lowers pour point for these oils to -15 0C or even -23 0C, and these
values satisfy most standards.
 All vegetable oil samples show good protection against rust and corrosion.
 Some of the additives used (PPD and EP) increase foaming above the allowed limits, but after the addition
of antifoaming agent, good results are obtained.
 Oxidation stability of vegetable oils without additives is very low. For instance, the result from RBOT test for
additive free soybean oil is only 8 mins. The improvement of this characteristic of vegetable oils has been
accomplished by adding the antioxidant additive and mineral based oil. The mineral oil addition enhances
the oxidation stability, but lowers the biodegradability of vegetable oils. The best oxidation stability was
found for the rapeseed oil sample (REA=109 mins), which was the expected result.
 The wear resistance tests were performed on “four ball” device. The results from this examination were
quite uniform for all samples and within the standard limits (< 40mm).
 The extreme pressure (EP) resistance was tested by using “four ball” device. The vegetable UTTO oils
show much higher ability to withstand extreme pressures as well as shock loads, in comparison to mineral
oils. The laboratory tests give evidence that even vegetable oils without EP additives show better results
than mineral oils with the additives.
The tribological testing consisted of measuring the change in the tribological properties of the elements which were
in contact with the oil sample. The results obtained give information on the friction coefficient, friction force, width
and depth of the wear track and wear shape of the block's contact surface, the change in the friction coefficient and
temperature during the contact, block's and disc's surface topography parameters before and after the testing. The
obtained results about the friction coefficient enabled identifying certain friction-connected tribological processes,
while the data on block’s wear width and depth enabled the explanation of the wear process.
The change in the friction coefficient for the rapeseed and sunflower oil samples is very similar, while soybean oil
shows higher coefficient. Nevertheless, these three oils without additives possess lower friction coefficient than
mineral based oil. It gives rise to the assumption that vegetable oils, due to the triglyceride structure, possess
excellent lubricating ability in all load phases. As a matter of fact, esters of naturally occurring fatty acids are even
used as additives for the improvement of lubricating properties.
The friction coefficient of the vegetable oils with additives is higher when compared to these oils without additives,
but it is still lower than the coefficient value for the reference mineral UTTO oil, particularly for high loads. The
friction coefficient increases proportionally by adding the mineral oil, indicating that the lubricating property
deteriorates.
The friction coefficient was measured under various loads and in different time periods. In the first phase, the
coefficient was measured under the force of 30N during 15 mins, then the force increased to 100N for 15 mins, and
in the final stage the force was 300N during 30 mins. When low force is applied (30N), the lowest friction coefficient
is recorded for rapeseed oil with additives (biodegradable UTTO oil based on rapeseed oil), while the highest
coefficient is for the soybean oil with additives (biodegradable UTTO oil based on soybean oil). Under the middle
load values (100N), the rapeseed oil without additives has the lowest coefficient, whereas the additive free soybean
oil and mineral UTTO oil show the highest coefficient. Finally, at the highest applied force (300N) during 30 mins
experiment, the additive free rapeseed oil shows again the lowest coefficient, and the highest value is again
recorded for mineral UTTO oil. The friction coefficient significantly depends on the oil type, so that the rapeseed oil,
which possesses the highest percent of monounsaturated fatty acids (oleic acid, C18:1) and the lowest iodine
number (118.41 gI2/100g), shows the lowest coefficient.
After the examination of various oil samples, the block's wear width was measured. The highest width value was
observed for the mineral UTTO oil sample, and the lowest for the rapeseed additive free oil sample. The low block's
wear width values for the RE, SO and SU samples indicate good lubrication properties of vegetable oils, i.e. they
give evidence that the triglyceride structure of these oils is sufficiently effective against wear even at the highest
pressures.
111
The block’s and disk’s surface roughness was measured by using Talysurf device. The surface peaks were clearly
observed on the fresh block and disk surface. However, after the contact of coupled elements, where various oil
samples were used as a tribological element, the peaks were deformed and torn. After the surface topography
examination, it may be concluded that the smallest wear track was recorded for the additive-free rapeseed oil,
while the widest track was observed when the mineral UTTO oil was used as a tribological element.
After the analysis of the results obtained in the present work, one may suggest possible directions for future
investigation:
 By using the methodology developed in the present article, the experimental analysis should be expanded
to other lubricants, but the following rule should be obeyed: an ecological lubricant must be developed,
which will satisfy all technical, exploitation, and other requests defined by current standards and the
application conditions.
 The development of ecological consciousness brings the interest for ecologically acceptable lubricants. In
the areas where the environmental pollution is a real threat, it is necessary to accelerate the usage of
ecological lubricants with the assistance of state institutions. One way to achieve this goal is to adapt the
appropriate laws related to this subject, which may bring long-term benefit for the community.
 The main goal in the future investigation should be the improvement in the oxidation stability and low
temperature properties of vegetable oils. This could be achieved by using new technologies for the
production of vegetable oils and by the development of new additives.
REFERENCES
[1] Becker, R., Knorr, A.: "An Evaluation of Antioxidants for Vegetable Oils at Elevated Temperatures", Lubrication
Science, Vol. 8, 1996, pp 95-117,
[2] Adhvaryu, A., Biresaw, G., Sharma, B.K., Erhan, S.: "Friction behavior of some seeds oils: Biobased lubricant
application”, Industrial and Engineering Chemistry Research, Vol. 45, 2006, pp 3735-3740,
[3] Adhvaryu, A., Erhan, S.Z., Perez, J.M.: "Tribological studies of thermally and chemically modified vegetable
oils for use as environmentally friendly lubricants”, Wear, Vol. 257, 2004, pp 357-367,
[4] Adhvaryu, A., Erhan, S.Z.: "Epoxidized soybean oil as a potential source of high-temperature lubricants”,
Industrial Crops and Products, Vol. 15, 2002, pp 247-254,
[5] Arnsek, A., Vizintin, J.: “Scuffing load capacity of rapeseed-based oils", Lubrication Engineers, Vol. 55, 1999,
pp 11-18,
[6] Arnsek, A., Vizintin, J.: "Lubrication properties of rapeseed-based oils", Lubrication Science, Vol. 16, 2000, pp
281-296,
[7] Arnsek, A., Vizintin, J.: "Pitting resistance of rapeseed-based oils", Tribology 2000-Plus, 12th International
Colloquium Esslingen, Vol. 11-13, 2000, pp 143-148,
[8] Аsadauskas, S., Еrhan, SZ.: "Depression of Pour Points of Vegetable Oils by Blending with Diluents Used for
Biodegradable Lubricants", Journal of the American Oil Chemists' Society, Issue 73, 1996, pp 313-316,
[9] Asadauskas, S., Perez, J.M., Duda, J.L.: "Oxidative stability and antiwear properties of high oleic vegetable
oils", Lubrication Engineers, Vol. 52, 1996, pp 877-882,
[10] Bartz, W.J.: "Lubricants and the Environment", Tribology International, Vol. 31, pp 1-3, 1998,
[11] Becker, R., Knorr, A.: "An evaluation of antioxidants for vegetable oils at elevated temperatures", Lubrication
Science, Vol. 8, 1996, pp 95-117,
[12] Defrang, M.: "Loss lubricating - use of vegetable oil and derivatives in concrete industry", Proceedings of
CTVO-Workshop on Lubricants and Hydraulic Fluids, 17th February 1999, Eibar, Spain.
[13] Erhan, S.Z., Asadauskas, S.: "Lubricant basestocks from vegetable oils", Industrial Crops and Products, Vol.
11, 2000, pp 277-282,
[14] Gapinski, R.E., Joseph, I.E., Layzell, B.D.: "A vegetable oil based tractor lubricant", International Off-Highway
& Power plant Congress & Exposition 1994, September 12-14 1994, Milwaukee, Wisconsin,
[15] Ohkawa, S.A., Konishi, H., Hatano, K., Tanaka, K., Iwamura, M.: "Oxidation and corrosion characteristics of
vegetable base biodegradable hydraulic oils", SAE Technical Paper 951038, 1995, pp 55-63,
[16] Ravasio, N., Zaccheria, F., Gargano, M., Recchia, S., Fusi, A. Poli, N., Psaro, R.: "Environmental friendly
lubricants through selective hydrogenation of rapeseed oil over supported copper catalysts", Applied Catalysis
A: General, Vol. 233, 2001, pp 1-6,
[17] Robertson, A.J., Randles, S.J.: "The use of laboratory techniques to simulate biodegradation of lubricants in
the environment", JUGOMA Professional Publications, Vol. 205, 1990,
[18] Krzan, B., Vizintin, J.: "Use and development of biodegradable oils, in tribology of mechanical systems; a guide
to present and future technologies", ASME Press: New York, 2004,
[19] Chen, B., Zhang, N., Liang, K., Fang, J.: "Enhanced biodegradability, lubricity and corrosiveness of lubricating
oil by oleic acid diethanolamide phosphate", Tribology in Industry, Vol. 34, 2012, pp 152-157,
[20] Stojilkovic, M. PhD Thesis: "Research application of ecological oils in the tribological systems", Faculty of
Mechanical Engineering, University of Eastern Sarajevo, 2014,
112
[21] Kodali, D.R.: "High performance ester lubricants from natural oils", Industrial Lubrication and Tribology, Vol.
54, 2002, pp 165-170,
[22] Gapinski, R.E., Kernizan, C.F., Joseph, I.E.: "Improved gear performance through new tractor hydraulic fluid
technology", Tribology 2000-Plus, 12th International Colloquium, Vol. 3, 2000, pp 2269-2276,
[23] Kassfeldt, E., Goran, D.: "Environmentally adapted hydraulic oils", Wear, Vol. 207, 1997, pp 41–45,
[24] Ulm, L., Vržina, J. Schiesl, V., Šateva, M.: “Biodegradability of water-resistent greases", International
Conference on Tribology "Slotrib '98", 1998, pp 104-112,
[25] Lal, K., Carrick, V.: "Performance testing of lubricants based on high oleic vegetable oils", Journal of Synthetic
Lubrication, Vol. 11, 1994, pp 189–206,
[26] Rhodes, B.N., Mammel, P., Landis, P., Erikson, F.L.: "Water rejection of vegetable oil base stocks for
tractor/hydraulic fluids", SAE Technical Paper, 1995, pp 1-4,
[27] Sraj, R., Vizintin, J.: "Ecologically Acceptable hydraulic fluids based on rapeseed oils”, Proceedings of the
International Conference on Tribology, Gozd Martuljek, Slovenia, 1996, pp 110-112,
[28] Stachowiak, G.W., Batchelor, A.W.: “Engineering Tribology, Third edition", Elsevier, Amsterdam, 2005,
[29] Stempfel, E.M.: "Practical experience with highly biodegradable lubricants, especially hydraulic oils and
lubricating greases", NLGI Spokesman, Vol. 62, 1998, pp 8-23,
[30] Battersby, N.S.: "The biodegradability and microbial toxicity testing of lubricants + some recommendations",
Shell research and technology centre, Thornton, UK, 1999,
[31] Willing, A.: "Lubricants based on renewable resources - an environmentally compatible alternative to mineral
oil products", Chemosphere, Vol. 43, 2000,pp 89-98,
[32] Willson, B.: “Lubricants and functional fluids from renewable sources", Industrial Lubrication and Tribology,
Vol. 50, 1998, pp 6-15,
[33] Bučko, M., Rogan, J., Stevanović, S.I., Perić-Grujić, A., Bajat, J.B.: "Initial corrosion protection of Zn–Mn alloys
electrodeposited from alkaline solution", Corrosion Science, Vol 53, 2011, pp 2861-2871,
[34] Bučko, M., Lačnjevac, U., Bajat, J.B.: "The influence of substituted aromatic aldehydes on Zn-Mn alloy
electrodeposition", Journal of the Serbian Chemical Society, Vol. 78, 2013, pp 1-22,
[35] Nedic, B., Peric, S.: "Monitoring oil for lubrication of tribomechanical engine assemblies", Journal of the Balkan
tribological association, Vol. 16, 2010, pp 242-257,
[36] Peric, S., Nedic, B., Trifkovic, D., Vuruna, M.: "An experimental study of the tribological characteristics of
engine and gear transmission oils", Strojniški vestnik – Journal of Mechanical Engineering, Vol. 59, 2013, pp
443-450.
113
114
Kragujevac, Serbia
MVM2018-005
Slobodan Mišanović1
Dragan Taranović2
Jovanka Lukić3
Radivoje Pešić4
Jasna Glišović5
THE IMPACT OF THE SYSTEM FOR HEATING AND AIR-

CONDITIONING ON THE ENERGY EFFICIENCY OF THE
ELECTRIC DRIVE BUS (E-BUS)
ABSTRACT: The electric drive buses in the field of providing thermal comfort for passengers and drivers during
the winter and summer seasons must meet the necessary requirements to make the journey comfortable. There are
different systems for heating and air-conditioning that are applied to the electric drive buses are usually conditioned
by the geographical and climatic characteristics of the city where they are used. The heating and air-conditioning
systems that use only electricity and the impact of these systems on the energy efficiency of the e-bus on the example
of exploitation in Belgrade, on line EKO1 in different periods during the year is analysed in this paper.
KEYWORDS: e-bus, thermal comfort, energy efficiency
INTRODUCTION
Vehicles used in the urban passenger transport system must provide a satisfactory thermal comfort for passengers
in terms of heating the passenger compartment during the winter period, also in terms of ventilation, humidity, and
air conditioning during the summer and transition periods (spring-autumn). According to the recommendations VDV
(Verband Deutscher Verkehrsunternehmen e. V.) as shown in Figure 1, define the minimum and maximum values
of the passenger compartment temperature in the buses to be used in public transportation (Tin), which must be
provided by the use of a heating system or air conditioning in the function of outside temperature (Tout), dashed line).
The recommended values for the E-bus are shown in red [1].
From the standpoint of thermal comfort for all types of vehicles used in public transport, an ideal range of temperature
within the passenger compartment is between 18 °C and 22 °C, which is represented by the yellow line. In the case
of an electrically powered bus, the air conditioning system is exclusively electrically powered, in contrast to the
heating system that can be:
 Fully Electric heating system,
 "Webasto" system that uses diesel fuel,
 A combination of a pure electric heating system and a heat pump system that utilizes "waste heat" on
inverters of drive motors and other electrical devices on the vehicle.
1
Slobodan Mišanović, Project manager, City Public Transport Company “Belgrade“, Kneginje Ljubice 29, 11000
Belgrade,PhD candidate, University of Kragujevac, Faculty of Engineering, slobodan.misanovic@gsp.co.rs
2
Dragan Taranović, PhD, associate prof., University of Kragujevac, Faculty of Engineering, tara@kg.ac.rs
3
Jovanka Lukić, PhD, full profesor, University of Kragujevac, Faculty of Engineering, lukicj@kg.ac.rs
4
Radivoje Pešić, PhD, full profesor, University of Kragujevac, Faculty of Engineering, pesicr@kg.ac.rs
5
Jasna Glišović, PhD, associate profesor, University of Kragujevac, Faculty of Engineering, jaca@kg.ac.rs
115
Outside temperature [˚C]
Figure 1 Range of the temperature passenger compartment in vehicles for public transport
Which system will be applied depends on the climatic area where the buses on electric power use. In Figure 2
presents the different climates in the case of 15 countries: Austria, Belgium, Czech Republic, Brazil, Estonia,
England, France, Germany, Italy, Mexico, Portugal, Turkey, Serbia and Switzerland [2].
Figure 2 Types of climate on the example of 15 countries
In geographic areas with cold continental or sharp continental climate, the Webasto heating system is applied, while
in other cases the electric heating system is more represented.
CONSUMPTION OF ELECTRIC ENERGY IN E-BUS

The choice and proper dimensioning of the heating and air conditioning system for electrically powered buses is a
very important issue, which is important when it comes to the energy efficiency of vehicles, especially in extreme
conditions of winter and summer exploitation. In this regard, in addition to the consumption of electricity for mobile
vehicles, the operation of auxiliary systems (compressor, control pump, signaling), determined by the driving cycle
and load of the vehicle, the electric energy consumption of heating and air conditioning system, must be taken into
account, shown in Figure 3.
116
Air -Cond./Heating
Traction,auxiliary
devices
Figure 3 Total electricity consumption of E-bus
The operation of the heating and air conditioning system and the impact on the electricity consumption can be
connected to the outside temperature. In Figure 4, the ambient temperature range of -15 °C to + 15 °C, implies the
use of a heating system. At ambient temperatures of + 15 ˚C to + 24 ˚C, ventilation of the air is ensured by the micro-
conditions of the passenger compartment. Ventilation can be achieved by "natural" flow through the side windows or
roof lufter or using a fan powered by an electric motor. In cases of using a fan, there is an additional requirement for
electricity consumption from the electrical energy storage system. At outdoor temperatures exceeding + 25 ˚C, the
air conditioner is used.
Heating Ventilation (yes/no) Cooling
Outside Temperature [˚C]
Figure 4 Using the heating and air conditioning system in the function of external temperature
In real terms of exploitation, the total consumption of E-bus, depending on the outside temperature, or the use of the
heating or air conditioning system, can be expressed as:
Ec=Ectr (1)
Ec=Ectr+Ech (2)
Ec=Ectr+Ecac (3)
Ec=Ectr+Ecw (4)
Ec - total electricity consumed E-bus [kWh]

Ectr - consumed electricity on the movement E-bus [kWh]
Ech - consumed energy of the heating system [kWh]
Ecac - consumed energy of the air conditioning system [kWh]
Ecw - consumed energy of the ventilation system [kWh]
Improper selection of heating and air-conditioning or improper use can significantly increase energy consumption
and as a result, have an unplanned discharge of the battery (supercapacitors), and therefore reduced the autonomy
of the vehicle.
AIR-CONDITIONING AND HEATING SYSTEM IN E-BUS "HIGER", WHICH IS

IN EXPLOITATION IN BELGRADE
From 1st September 2016, five E-bus, model KlQ6 by manufacturers ''Higer'', are in regular exploitation in Belgrade
on the line EKO 1 (Vukov monument-Belleville). The electrically powered bus is equipped with the air conditioner of
117
the manufacturer ''Thermo King'' and fully meets all the criteria of cooling and ventilation of the passenger
compartment in the vehicle as a whole during its exploitation.
The power of the air conditioners in the cooling mode is max. 24 kW. The air conditioner is designed to allow the air
recirculation, as well as the treatment of the freshly inserted vehicle. The distribution of treated air through the
passenger compartment of the vehicle is carried out with special channels. The outdoor unit is equipped with
interchangeable air filters. In the vehicle roof, there is ''lufter'' for ventilation of the vehicle, which is activated by the
electric command from the instrument table of the driver. The digital control panel of the air conditioning system is
shown in Figure 5.
1. ON / OFF button 2. Mode selection button (OUT) 3. Temperature button (blue) (MINUS button) 4. Temperature button (red) (PLUS button)
5.Button for speed of ventilation (INPUT) 6. Button for selecting fresh air 7. AC control icons 8. Temperature icons 9. Display 10. Buster pump
icons 11. Icons for mode ventilation 12. Fresh mode icons 13. Red alarm 14. Yellow alarm
Figure 5 Control panel air conditioning [3]
The heating system in the vehicle is designed as an electric, type manufacturer "Spheros". Heating system power is
max. 30 kW. There are special temperature regulators for the air and driver's cab. The system consists of 9 heaters,
which are heated up to 75 ˚C. The heated work fluid then passes through the tube on the radiators in the passenger
compartment (Figure 6) and the driver's cab. For the driver's windshield warming is used "Defroster", particular or
together with the "Spheros" system [3].
Figure 6 Position radiators for heating
In the winter period of operation, when the temperature of the interior space falls below 18 ˚C, the system
automatically switches on the thermostat and reinsures the passenger compartment until it reaches 22 ˚C when the
heaters are switched off. The regulation of the heating system is multistage, which allows optimum engagement of
the heaters depending on the outdoor temperature. The heating system complies with UN ECE 122 or 2001/56/EC
regulations.
RESEARCH OF THE EFFICIENCY OF THE HEATING AND AIR

CONDITIONING SYSTEM, ON THE ENERGY EFFICIENCY OF E-BUS ON THE
LINE EKO 1 IN BELGRADE
The energy consumption of the heating and air conditioning system depends on several influential factors: outside
temperature, internal temperature, number of passengers in the vehicle, number of a door for entry/exit of
passengers, duration of entry/exit of passengers, thermal insulation of the vehicle, duration of travel.
118
This impact can also be observed depending on the period of the year, whether it is a winter, summer or transitional
period. Under conditions of exploitation in Belgrade, these periods can be observed by months, as presented in Table
1.
Table 1 Typical periods of exploitation in Belgrade

Month / Period of exploitation Winter Summer Transition
Janauary x
February x
March x
April x
1 st ½ May x
2 nd ½ May x
June x
July x
August x
1 st ½ September x
2 nd ½ September x
1 st ½ October x
2 nd ½ October x
November x
December x
Belgrade has a moderate continental climate. The average annual temperature during 2016.year was 13.5 °C, while
the hottest month was July, with an average temperature of 24.4 °C, and the coldest January with an average
temperature of 2.4 °C. [4]. Belgrade has an average of about 25 days a year with temperatures above 30 °C In the
last few years, during the summer months, frequent temperatures have been observed from + 35 °C to + 38 °C, and
in the winter period from -10 °C to -15 °C.
In Figure 7, the values of solar radiation energy for Belgrade during the year expressed in Wh·m-2 are presented for
one day [5].
Figure 7 Daily energy potential of the sun for the city of Belgrade, measured over a year
Since air conditioners in vehicles are most intensively used during July and August of the month, the radiation energy
values are about 6000 Wh·m-2 for one day. During the summer period, the average day lasts for 16 hours, so in the
hottest period of time, the intensity of the sun's radiation is the highest and reaches the intensity of 1000 Wh·m-2. To
a similar result of solar radiation intensity referring to the mid-European climatic conditions when the sun reaches the
maximum angle during the day, the radiation energy is 1064 Wh·m-2 (of which 800 Wh·m-2 is direct radiation and 264
Wh·m-2 of diffuse radiation) [6].
The operation of the air conditioning and heating system during the summer and winter period is related to the number
of passengers in the vehicle, since the human body releases a certain amount of heat as a result of chemical
processes in the human body (metabolism) and physical activity. Metabolism measurement unit is labeled Met (1 Met
= 58.15 W·m-2) and refers to the body surface unit [6]. The body surface is calculated according to the following
formula [7, 8]:
Asp=0.202 mb0,425 hb0,725 (5)

119
Asp - the surface of the human body [m2]
mb - body weight [kg]
hb - body height [m]
For an average passenger weight of 68 kg and a height of 1.75 m, while sitting, during a journey, corresponding to a
value of about 1 Met, the passenger will give a heat of approximately 105 W.
For the efficiency of the air conditioning system and the heating of the buses on electric drive, thermal insulation of
the vehicle and proper use of the side windows is also significant, since the transfer of heat between the vehicle as
a system and the environment is transmitted through convection and conduction.
WINTER PERIOD OF EXPLOITATION

In the following example will be presented the results of measuring the electric energy consumption of the heating
system and the temperature of the interior space for the E-bus (garage number 2103, Figure 8), which was conducted
in the Belgrade, 2nd March 2018 [9]. The outside air temperature was -1 ˚C. A measurement was performed according
to the following plan:
- Installation of measuring equipment in the E-bus garage number 2103, which was performed in the enclosure of
the depot "Dorćol". The measured temperature in the depot was 11.5 ˚C.
- Measurement of electricity consumption of heating and temperature systems in the passenger compartment on
depot ''Dorćol'' - Terminal of the Vuk monument, without passengers, without opening the door.
- Measurement of the electrical energy consumption of the heating system, passenger compartment temperatures,
energy consumption and recovery at the super condenser in real exploitation, direction "A" (Vuk's monument-Belvil).
Measuring equipment was used "Fluke 189", "Fluke 289" and current clamps, and measured the following sizes:
Voltage on the supercapacitor, discharge/charging current of the supercapacitor, ''Spheros'' system heating energy
consumption and heaters windscreen "Defroster", the interior temperature of the passenger compartment. The
measuring device for recording the ambient temperature of the interior of the vehicle was placed in the middle part
of the vehicle at a height of 1.5 m. (Figures 9, 10, 11)
Figure 8 Е-bus ''Higer'' Figure 9 Measurement of the current

consumption of the system for heating the
passenger compartment
Figure 10 Measuring the temperature of the Figure 11 Measurement of voltage and current
passenger compartment on a supercondenser
120
After mounting the measuring equipment at the depot, started with measuring the electricity consumption of the
heating system '' Spheros '' and '' Defroster '' and the temperature in the passenger compartment from the depot
''Dorćol'' - terminal ''Vuk monument''. The aim was to determine the change in the interior temperature of the vehicle
and the consumption of electricity on the section from the depot "Dorcol" to the "Vuk monument'' terminus. In addition
to the driver, there were also two persons who were measuring. All the while driving, the doors were closed. The E-
bus started from the depot "Dorćol" at 9:43:45 hours. The initial temperature in the passenger compartment was 11.9
˚C. The outdoor temperature was -1˚C. The heating system was in maximum power mode. The measured values of
the voltage and current on the heating system are shown in Figure 12.
Figure 12 Values voltage, power and current of the heating system in the period 09:43:45-09:46:50 am
The change of temperature in the passenger compartment, the consumed electric energy of the heating system, the
mean power of the system and the operating time are presented in Table 2 [9].
Table 2 The results of the measurements, the change in temperature of the passenger compartment, consumed
energy and power of the heating system in the period 09:43:45-09:46:42 am.
Inside Consumpted energy Average Power Energy –
Period Temperature ’’Spheros+Defroster’’ ’’Spheros+Defroster’’ cumulants
[hh:mm:ss] [˚C] [kWh] [kWh] [kWh]
09:43:45 11.9 0 0 0
09:45:27 13.8 0.461 16.1 0.461
09:45:52 16.9 0.1141 16.4 0.5751
09:46:22 18.8 0.1344 16.1 0.7095
09:46:42 19.4 0.0876 15.7 0.7971
From Table 2 and Figure 13, it can be seen that for 178 seconds were spent 0.7971 kWh of electricity, and the
temperature of the passenger compartment has been increased from 11.9 °C to 19.4 °C.
Figure 13 Changing the passenger compartment temperature, the energy consumed and the power of the heating
system at the depot '' Dorćol '' – ''Vuk's monument''
121
Upon arrival at the terminus''Vuk Monument'', vehicle is complemented electricity on the charger with the occasional
inclusion of a heating system to maintain the temperature reached the passenger compartment After that, E-bus was
started at 10:23 hours, in the regular traffic on the line EKO 1, the direction ''A'' (Vukov monument - Belleville) where
were done measuring the electricity consumption of the heating system, the temperature in the passenger
compartment, and the consumed and recovery energy on the supercapacitor in real exploitation.
Table 3 shows the results of measuring the consumed electricity at measuring points [9].
Table 3 Results of measuring the consumption of electricity in the period 10:23:05 - 10:52:40 am, direction ''A''
Energy consumption from supercapacitor [kWh] 13.165
Energy consumption of heating system ’’Spheros+Defroster’’ [kWh] 3.569
Recovered energy in supercapacitor [kWh] -2.482
Exchange energy in supercapacitor [kWh] 10.63
Recovery [ %] 18.9
Lenght od line [km] 7.47
-1
Average of consumption E-bus [kWh·km ] 1.43
Average of consumption E-bus [kWh·km-1]* 1.51
* with estimated losses of 5% in the filling phase
Table 4 presents results of the engaged power of the heating system ''Spheros'' and the system ’’Defroster’’ with the
period of operation and the temperature of the passenger compartment [9].
Table 4 Engaged power of the heating system ''Spheros'' and the system ’’Defroster’’
in period 10:23:05-10:52:40 am, direction ''A''
''Defroster Temperature
Start End Duration ''Defroster'' ''Spheros'' +Spheros'' passenger compartment
[hhmmss] [hhmmss] [hhmmss] [kW] [kW] [kW] [˚C ]
10:23:05 10:25:42 0:02:37 1.937 20.9 -18.3
10:25:43 10:31:40 0:05:57 1.920 12.96 14.88 18.3 - 23.9
10:31:41 10:36:04 0:04:23 1.895 23.9 - 18.6
10:36:05 10:40:00 0:03:55 1.884 11.99 13.88 18.6 - 23.1
10:40:01 10:49:23 0:09:22 1.873 23.1 - 18.8
10:49:24 10:52:40 0:03:16 1.868 10.37 12.24 18.8 - 20.3
The total driving time was 29.5 minutes. The average ambient temperature of the passenger compartment was 19.5
˚C. During the whole time, "Defroster" was switched ''on'' and its average power was 1.89 kW. The ’’Spheros’’ heating
system was engaged of 13 minutes and 8 seconds, or 44.5% of the driving time with an average power of 11.77 kW.
The maximum number of passengers in the vehicle was 28 passengers. Figure 14 gives the current values: ambient
temperature of the passenger compartment, power of the heating system, the cumulative energy consumption of the
heating system and the number of passengers registered in the vehicle during the measurement. During the
measurements, the passenger compartment temperature ranged from 18.1 °C to 23.9 °C, which fully meets the set
requirements regarding the temperature of the passenger compartment.
Figure 14 Graphical display of the results of measurements in the period 10:23:05 - 10:52:40 am, direction ''A''
122
If the heating system was switched ''off '' during operation, the following measurement results would be obtained, as
shown in Table 5.
Тable 5 The results of the measurement of electricity consumption E-bus

with switched ''off'' heating system 10:23:05-10:52:40 am, direction ''A''
Energy consumption of heating system ’’Spheros+Defroster’’ [kWh] 0
Recovered energy in supercapacitor [kWh] -2.482
Exchange energy in supercapacitor [kWh] 7.114
Recovery [%] 25.9
Lenght od line [km] 7,47
Average of consуmption E-bus [kWh·km-1] 0.95
Average of consуmption E-bus [kWh·km-1]* 1.01
Analysis of the obtained results is shown that when an ambient temperature of -1 °C, the effect of the operation of
the heating system to the energy efficiency of the E-bus increased by 49.5%, compared with the heating system,
when is not in operation.
In the following example are presented results of the measurements of the electricity consumption of the heating
system "Spheros", which was done on 22nd December 2016, on the line EKO 1, direction "A" (Vuk monument - Belvil)
in the period from 15:27:36 to 16:03:22 hours. The outdoor air temperature was -6 ˚C. It was a strong southeast wind,
which made the real outside temperature of -10 °C. The measurement was carried out in the same manner as in the
example given above [10]. Table 6 presents the results of measurement:
Тable 6 Results of measurement of electricity consumption from 22/12/2016, direction ''A''

Energy consumption of heating system Spheros [kWh] 5.97
Lenght od line [km] 7.47
Average of consumption E-bus [kWh·km-1] 1.848
Average of consumption E-bus [kWh·km-1]* 1.940
Results of the engaged power of the heating system ''Spheros'' with the period of operation and the temperature of
the passenger compartment are presented in Table 7 [10].
Тable 7 Engaged power of the heating system ''Spheros'' and the temperature of the passenger compartment
from 12/22/2016, direction ''A''
Power Temperature
Start End Duration ''Spheros'' passenger
[hh:mm:ss] [hh:mm:ss] [hh:mm:ss] [kW] compartment
[˚C]
15:27:36 15:39:19 0:11:43 12.3 17.2-18.4
15:39:20 15:40:02 0:00:42 0.12 18.4-18.5
15:40:03 15:44:37 0:04:34 10.61 18.5-19.5
15:44:38 15:45:22 0:00:44 0.12 19.5-19.6
15:45:23 15:53:48 0:08:25 10.13 19.6-20.6
15:53:49 15:54:32 0:00:43 0.11 20.6-20.7
15:54:33 16:03:22 0:08:49 8.76 20.7-21.6
Total driving time was 35.5 minutes. The average temperature of the passenger compartment was 19.4 ˚C. The
heating system ’’Spheros’’ was engaged of 33 minutes and 31 seconds, or 93.9% of the driving time with an average
power of 9.95 kW. The maximum number of passengers in the vehicle was 43 passengers. Figure 15 gives the
current values: ambient temperature of the passenger compartment, power of the heating system, the cumulative
energy consumption of the heating system and the number of passengers registered in the vehicle during the
measurement. During the measurements, the temperature of the passenger compartment ranged from 17.2 ˚C to
21.6 ˚C, which fully meets the set requirements regarding the ambient temperature of the passenger compartment.
123
Figure 15 Graphic display of measurement results from 22/12/2016, direction ''A''
If was assumed that the heating system during the operation was switched ’’off’’, the following measurement results
would be obtained as shown in Table 8.
Тable 8 The results of the measurement of electricity consumption E-bus with

switched ’’off ’’ heating system from 22/12/2016, direction ''A''
Energy consуmption of heating system ’’Spheros’’ [kWh] 0.0
Lenght of line [km] 7.47
Average of consumption E-bus [kWh·km-1] 1.05
-1
Average of consumption E-bus [kWh·km ]* 1.10
The analysis of the obtained results shows, at an ambient temperature of -10 °C, the effect of the operation of the
heating system to the energy efficiency of the E-bus is extremely bad, and the total consumption of the E-bus is
increased by 75.2% compared when the system for the heating not in operation. From the measurements, it can be
concluded the operation of the heating system increases with decreasing outdoor temperature.
In the first measurement, the engagement was 44.5% at -1 °C. When the outside temperature is -10 °C, engagement
is 93.9%. The mean power of the ’’Spheros’’ heating system measured on 22/12/12, had a slightly lower value of
9.95 kW, which was the result of a larger number of passengers in the vehicle that give more heat to the passenger
compartment and longer time travel, which results in a reduction in the DC voltage of the E-bus, especially at the end
of the trip, which affects the reduction of the power of the heating system. To the results of the total consumption of
electricity E-bus, which includes the consumed energy for movement, an operation of auxiliary devices, and the
energy of the heating system can be reached by analyzing the degree of discharge of the battery/ supercapacitor
(SOC- state of charge) of each direction at different outside temperatures. During of January and February 2018,
recorded the changes in SOC for the different dates in which the outside temperature was in the range of -11 °C to
+ 10 °C [11]. Analysis of a sample of 91 measurements overall consumption of E-bus (with calculated losses in the
filling phase) expressed in kWh·km-1, were in the range of 1.05 to 2.1 kWh·km-1. Distribution of total consumption is
presented in Figure 16.
Figure 16 Distribution of total electricity consumption in the period January-February 2018 on the line EKO 1,
direction ''A''
124
The total electricity consumption of E-bus expressed in kWh·km-1, in the function of outside temperature for the
observed period is shown in Figure 17.
Figure 17 Total electricity consumption of E-bus expressed in kWh·km-1, in the function of outside temperature
(winter period), direction ''A''
These results relate to the real conditions of exploitation covering periods of peak load, out peak load, the first/last
departures, characterized by the different load of E-bus and different durations of driving time, caused by traffic
conditions. Between the consumption of electric energy of E-bus, expressed in kWh·km-1 and the outside temperature
there is a strong correlation (R2 =0.983), which described a quadratic function,
y =0.0009x2-0.0417x+1.4664
Based on the representative sample (Xi), a statistical analysis of the results of electricity consumption of the E-bus
was performed using the mean value (Xs) and the standard deviation (Sd) [12].
Xs = 𝑋1+𝑋2+𝑋3….+𝑋𝑛
𝑛
(6)
∑𝑛𝑖=1(𝑋𝑖 − 𝑋𝑠).2
𝑆𝑑 = √ (7)
𝑛
The mean value of the treated samples was 1.52 kWh·km-1, the standard deviation of 0.263 kWh·km-1. Considering
that the distribution of sample corresponded to the Normal distribution, can be considered that the expected interval
of electricity consumption of E-bus, based on the estimates of the mean value(μ) of the basic set with a probability
95% confidence, the sample n≥ 30, represented by the following equation [12] :
𝑠𝑑 𝑠𝑑
P (Xs - 1.96 <<Xs+1.96 )=0.95 (8)
√𝑛 √𝑛
In the concrete case P(1.465<μ<1.574) = 0.95, it can be considered that the average value of electric consumption
of electric buses will be between 1.465 kWh·km-1 and 1.574 kWh·km-1 .
If the energy consumption is analyzed only for the ’'Spheros’’ heating system in function of the difference between
the temperature of the passenger compartment Tin and the external temperature Tout, based on the measured values
125
(Tables 4 and 7) and the trend value of the consumption of the E-bus, square function in Figure 17, is represented
in Table 9.
If the value for outside temperature +4 ˚C, put in the trend function, the total energy consumption of E-bus is 1.313
kWh·km-1. If is assumed that energy only for the movement of the vehicle 0.97 kWh·km-1, the energy for the operation
of the heating system was obtained as a difference between these two values and multiplied by 7.47 km, which the
length of the route in the direction "A''.
Тable 9 Results of the measurement of electricity consumption of the heating system ''Spheros'' as a function of
the difference temperature between the passenger compartment and the outside temperature
Tout Tin T ’’Spheros’’- consumed energy
[˚C] [˚C] [˚C] [kwh]
9 19.4 10.4 1.371
4 19.4 15.4 2.499
-1 19.4 20.4 3.56
-10 19.4 29.4 5.97
The dependence of the electrical energy consumption of the heating system and the temperature difference Т can
be described by the quadratic function y=0,0022x2+0.1536x-0.444 with extremely high correlation (R2=0.999) as
shown in Figure 18.
Figure 18 Dependence of the electrical energy consumption of the heating system

and the temperature difference Т
If the resulting correlation applied for a range of outside temperatures, that occur in the winter, the estimated energy
consumption of the ’’Spheros’’ heating system is presented in Table 10.
Тable 10 Results of measurement and calculation of electrical energy consumption of the heating system in
function of the difference in internal and outside temperature T
’’Spheros’’ ’’Spheros’’
consumed energy consumed energy
Tout Тin T measured calculation
[˚C] [˚C] [˚C] [kwh] [kwh]
12 19.4 7.4 0.797 0.814732
11 19.4 8.4 1.004392
10 19.4 9.4 1.198452
9 19.4 10.4 1.371 1.396912
8 19.4 11.4 1.599772
7 19.4 12.4 1.807032
6 19.4 13.4 2.018692
5 19.4 14.4 2.234752
4 19.4 15.4 2.491 2.455212
3 19.4 16.4 2.680072
2 19.4 17.4 2.909332
1 19.4 18.4 3.142992
126
Тable 10 (continued)
0 19.4 19.4 3.381052
-1 19.4 20.4 3.56 3.623512
-2 19.4 21.4 3.870372
-3 19.4 22.4 4.121632
-4 19.4 23.4 4.377292
-5 19.4 24.4 4.637352
-6 19.4 25.4 4.901812
-7 19.4 26.4 5.170672
-8 19.4 27.4 5.443932
-9 19.4 28.4 5.721592
-10 19.4 29.4 5.97 6.003652
If the average power of the heating system at a certain outside temperature is known, as well as the engagement of
the system, the consumed energy of heating systems can be represented by the following equation:
Ech=Phs·td ·Ku (9)
Ech - consumed energy of the heating system [kWh]

Phs - average power of heating system [kW]
td - driving time [h]
Ku - coefficient of system engagement
The values of the average power of the heating system, the outside temperatures, the coefficient of system
engagement (Ku) in relation to the driving time and the consumed energy on the line EКO 1, are presented in Table
11.
Тable 11 Average power of the heating system, outside temperatures, coefficient of system engagement
and the consumed energy on the line EКO 1
Heating system- Driving Energy consumed of
Tout Тin T
Average Power time Ku heating system
[˚C] [˚C] [˚C]
[kW] [h] [kWh]
9 19.4 10.4 10 0.54 0.25 1.35
4 19.4 15.4 12 0.54 0.37 2.3976
-1 19.4 20.4 13.66 0.54 0.45 3.31938
-10 19.4 29.4 11.5 0.54 0.93 5.7753
Displayed value of the average power of the heating system, outside temperature coefficient engagement system in
relation to driving time and energy consumed of heating system is described how the strong correlation, as shown in
Figure 19.
Figure 19 The dependence of energy consumption, power and level of engagement of the heating system
as a function of outside temperature
127
During the winter period the total electricity consumption of the E-bus on the EKO 1 line, direction "A" (Vuk's
monument-Belvil) was observed from the aspect of consumption to drive (motion, engine losses, transmission,
auxiliary device), losses during charging and consumption of the heating system, results are presented in Table 12.
Table 12 The distribution of the total consumption of E-bus during the winter season, direction ''A''
(traction, heating, loss in the filling phase)
Consumed Heating system- Consumption
Outside energy on the consumped energy Energy losses TOTAL TOTAL Heating
Temperature drive -calculation during charging CONSUMPTION CONSUMPTION sistem
[˚C] [kWh] [kWh] [kWh] [kWh] [kWh·km-1] [ %
-10 7.246 6.004 0.662 13.912 1.862 43.2
-9 7.246 5.722 0.648 13.616 1.823 42.0
-8 7.246 5.444 0.634 13.324 1.784 40.9
-7 7.246 5.171 0.621 13.038 1.745 39.7
-6 7.246 4.902 0.607 12.755 1.708 38.4
-5 7.246 4.637 0.594 12.478 1.670 37.2
-4 7.246 4.377 0.581 12.204 1.634 35.9
-3 7.246 4.122 0.568 11.936 1.598 34.5
-2 7.246 3.870 0.556 11.672 1.563 33.2
-1 7.246 3.624 0.543 11.413 1.528 31.7
0 7.246 3.381 0.531 11.158 1.494 30.3
1 7.246 3.143 0.519 10.908 1.460 28.8
2 7.246 2.909 0.508 10.663 1.427 27.3
3 7.246 2.680 0.496 10.422 1.395 25.7
4 7.246 2.455 0.485 10.186 1.364 24.1
5 7.246 2.235 0.474 9.955 1.333 22.4
6 7.246 2.019 0.463 9.728 1.302 20.8
7 7.246 1.807 0.453 9.506 1.273 19.0
8 7.246 1.600 0.442 9.288 1.243 17.2
9 7.246 1.397 0.432 9.075 1.215 15.4
10 7.246 1.198 0.422 8.867 1.187 13.5
11 7.246 1.004 0.413 8.663 1.160 11.6
12 7.246 0.815 0.403 8.464 1.133 9.6
In the winter season, at extremely low outside temperatures, the proportion of the heating system in the overall
consumption of E-bus amounts to about 43%. For external temperatures around 0 °C, the heating system is around
30%. With the increase in outside temperature is above 0 °C, the heating system has a lower impact on the total
power consumption E-bus. At +10 °C the share of heating is about 13.5%.
SUMMER PERIOD OF EXPLOITATION

During the summer period, there is the intense use of air conditioners in the vehicle to ensure a satisfactory
temperature of the passenger compartment. The effect of the operation of the air conditioner and the impact on the
electricity consumption of the E-bus, have been seen from the results of measurements performed on 11th August
2017 [13] at an outside temperature of 38 ˚C. Measurement of the temperature of the passenger compartment was
made by the measuring device "Fluke 189", which was placed on the central part of the vehicle. The measuring probe
was set at a height of 1.5 m (Figure 21). The device was recorded the current temperature values every 15 seconds.
Measurement of energy consumption of air conditioners ’’Thermo King’’, was done by other two instruments ’’Fluke
289’’ and power cleats (Figure 22) that had been placed at the incoming cable in the air conditioner in order to
measure the instantaneous current (A) driven by the operation of the air conditioner, which was in the "automatic"
mode. The second measuring device measures the voltage (V) of the DC-bus with which the air conditioner is
connected.
The survey included the electricity consumption of air conditioning systems:

 when there are no passengers in the vehicle
 with passengers in regular exploitation on the line EKO 1
The effects of the air conditioning system - without passengers

Preparing for the measurement (Figure 20) on the E-bus (garage number 2104) was performed indoors depot for
vehicle maintenance.
128
Figure 20 Е-bus in depot Figure 21 Position of the measuring Figure 22 Position devices ’’Fluke
device ’’Fluke189’’ for measuring the 289’’ to measure current air
temperature of the passenger conditioners
compartment
After mounting the measuring equipment, the current temperature in the passenger compartment was 31.9 ˚C. All
doors were closed and the air conditioner was put into in the "automatic" mode of operation. The measured values
of voltage and current on the air conditioner are shown in Figure 23 [13].
Figure 23 The values of the current and voltage of the air conditioners in the period 11:48:29 - 11:49:57
Changing the temperature in the passenger compartment, the electrical energy consumption of the air conditioning
system, the mean power of the system and the time are presented in Table 13 [13].
Table 13 Changing the temperature of the passenger compartment, consumed energy

and the power of air conditioners in the period 11:48:29 - 11:49:57 am
Time Inside temperatre Air Cond.-Consumption energy Air Cond.- Power Energy – Cumulants
[hh:mm:ss] [˚C] [kWh] [kW] [kWh]
11:48:29 AM 31.9 0.0000 2.34 0.0000
11:48:41 AM 30.4 0.0072 2.161 0.0072
11:48:56 AM 30 0.0098 2.344 0.0170
11:49:11 AM 29.9 0.0088 2.103 0.0257
11:49:30 AM 29.8 0.0124 2.345 0.0381
11:49:41 AM 29.7 0.0066 2.161 0.0447
11:49:57 AM 29.7 0.0096 2.16 0.0543
From Table 13, it can be seen that for 88 seconds spent 0.0543 kWh of electricity and the temperature of the
passenger compartment is reduced from 31.9 °C to 29.7 °C.
The following measurements were made on the way that E-bus was left in an open parking space, directly exposed
to the sun. The aim was to determine the change in the interior temperature of the vehicle and the consumption of
electricity on the route from the depot "Dorcol" to the "Vukov monument" terminal. Except for the driver, there were
also two people who measured. While driving, the doors were closed. The E-bus was driven from the "Dorćol" depot
at 12:35:10.
After arriving at the Vuk's monument the E-bus was charged at 13:02:50, the air conditioning system was switch
’’on’’, all doors were closed. The current, voltage and power are shown in Figure 24 [13].
129
Figure 24 Current values of voltage, current and power of air conditioners on the route of
the depot Dorćol-Vukov monument in the period 12:35:10 - 13:02:50 hours
At this stage of measurement, the temperature change in the passenger compartment, consumption electrical energy
of the air conditioning system, the average power of the system and the time are presented in Table 14 and Figure
25 [13].
Тable 14 Changing the inside temperature, consumed energy and power of air conditioning
on the route of the depot "Dorcol" - Vuk's monument in the period 12:35:10 -13:02:50 hours
Inside
Period Temperature AC-consumpted energy AC-Average Power Energy – cumulants
[hh:mm:ss] [˚C] [kWh] [kW] [kWh]
12:35:10 37.6 0 2.1 0
12:43:40 30.6 0.291 2.06 0.288
12:55:21 25.1 0.416 2.13 0.707
13:02:50 18.8 0.274 2.19 0.981
Figure 25 Changing the temperature of the passenger compartment, consumed energy and power of air
conditioning on the route depot ''Dorćol'' – ''Vukov monument'' in the period 12:35:10 - 13:02:50 am
Based on these results it is possible to show the power consumption of air conditioners as a function of differences
in the temperature of the passenger compartment at the beginning and end of each measurement, which is shown
in Table 15 [13].
130
Тable 15 Consumption energy of air conditioners in function of difference temperature
of the passenger compartment T in the period 12:35:10-13:02:50 am
Tin – start T in - end T AC-Energy consumption Time
[˚C] [˚C] [˚C [kWh] [s]
31.9 30.4 1.5 0.0072 12
31.9 29.6 2.3 0.0543 88
30.6 25.1 5.5 0.416 701
37.6 25.1 12.5 0.707 1211
37.6 18.8 18.8 0.981 1660
The correlation between the consumed energy of the air conditioner and the operating time in function of the
temperature difference is shown in Figure 26.
Figure 26 Correlation between the consumed energy of the air conditioner and the operating time
in function of the temperature difference
These dependencies in a good way are described by a quadratic function with a high coefficient of regression, which
in both cases is greater than R2 = 0.98.
The effects of the air conditioning system - with passengers

Electrical energy consumption of air conditioning and the effect of thermal comfort in the passenger compartment E-
bus, in real conditions of exploitation, was carried out in a time of 13:08:10-13:32:56, in the direction of ''A''. The
number of passengers in the vehicle at every distance between stations is shown in Figure 27 [13].
Figure 27 Number of passengers in the E-bus, period 13:08:10-13:32:56, direction ''A''
The occupancy of the vehicle was relatively small (Figure 27) since the measurement was performed in the off-peak
period.
Changing the current temperature in the passenger compartment of the vehicle during the measurement period is
shown in Figure 28 [13].
131
Figure 28 Changing the current temperature in the passenger compartment during measurement, direction ''A''
The minimum registered the temperature in the passenger compartment was 21.6 °C and a maximum of 27.1 °C.
The average temperature was 24.2 °C. Compared with an outside temperature of 37 °C, the difference of 12.8 °C is
very significant. A health recommendation is that the difference between the outside temperature and the interior
temperature of the air-conditioned room should be within the limits of 5 °C to 8 °C. At the beginning of the
measurement, the vehicle was on the filling (without passengers), the doors were closed, and the effect of the air
conditioner operation was most effective since the temperature was 21.5 °C. When E-bus was moving, opening the
doors on the stations and the presence of passengers in the vehicle, the ambient temperature increased slightly in
the passenger compartment. At the time of switching ''on'' the air conditioning compressor, there was a drop in
temperature, while retention with open doors affected the temperature jumps in the interior of the vehicle.
In order to obtain the results of the consumption of the electric energy of the E-bus and the air conditioner, the results
of the change of the supercapacitor (SOC), which is recorded for the observed direction is shown in Table 16. [11]
Table 16 Total energy consumption of E-bus, direction "A" , period 13:08:10-13:33:09 for 11/8/2017
Outside
Time temperature SOC
Direction ''А'' [hhmmss] [°C ] [ %] [kWh] [kWh·km-1]
Vuk monument 13:08:10 38 96
Belvil 13:33:09 38 48
Total consumpted energy 9.84
Average consumption of E-bus 1.31
Average consumption of E-bus * 1.37
* with calculated energy loss when charging
Results of the total consumption of E-bus, which was recorded 7.8.2017. year, in the same period of the day, but at
an outside temperature of 20 ° C, when the air conditioner was not used, presented in Table 17.
Table 17 Total energy consumption of E-bus, direction "A" , period 12:55:02-13:25:11 for 7/8/2017
Outside
Time temperature SOC
.Direction ''А'' ' [hhmmss] [°C ] [ %] [kWh] [kWh·km-1]
Vuk monument 12:55:02 20 99
Belvil 13:25:11 20 65
Total consumpted energy 6.945
Average consumption of E-bus 0.93
Average consumption of E-bus * 0.976
* with calculated energy loss when charging
Results of the measurements of the average consumption of E-bus, expressed in kWh·km-1, as shown in Tables 16
and 17, during the operation of the air conditioning system was proved that total consumption of E-bus was higher
by 40.8%.
132
During July and August 2017, the change of SOC was recorded for the different dates, which the outside temperature
was in the range of 22 °C to + 39 °C [11]. Analyzing a sample of 75 measurements of total E-bus consumption (with
calculated losses in the charging phase) expressed in kWh·km-1, were in the range of 0.975-1.61 kWh·km-1.
Distribution of total consumption is presented in Figure 29.
Figure 29 Distribution of total electricity consumption in the period July-August 2017 on the line EKO 1,
direction ''A''
The results of total consumption include real exploitation conditions during peak load periods, off-peak loads as well
as first/last departures, are characterized by different vehicle loads. Figure 30 shows the total consumption of the E-
bus in the function of the outside temperature for the observed period.
Figure 30 Electricity consumption E-bus, expressed kWh·km-1, as a function of the outside temperature
(summer period), direction ''A''
Between energy consumption of E-bus and the outside temperature, there is a strong correlation (R2 =0.971) which
is described how quadratic function y=0.0007x2+0.068x-0.021
Based on the representative sample (Xi), the statistical analysis of the results of electric consumption of the E-bus
was performed using the mean value (Xs) and the standard deviation (sd).
The average value of the treated sample is 1.289 kWh·km-1, a standard deviation is 0.165 kWh·km-1.
Since the sample distribution corresponds to the Normal distribution, the expected interval of the E-bus consumption
based on the interval estimation of the mean value (μ) of the base set, with a probability of 95% for the sample n≥
30, P(1.251 <μ < 1.326) = 0.95 .
Based on the results of the total consumption of the E-bus and the resulting trend shown in Figure 30, the effect of
the air conditioning system on the total consumption of E-bus in the observed period in the function of the outside
temperature is shown in Table 18.
133
Table 18 Distribution of total consumption of E-bus in the summer period, direction ''A''
(motion, heating, losses in the filling phase)
Consumed
Consumed energy
Outside energy on the Energy losses TOTAL TOTAL Consumption
Temperature on the motion Air-condition during charging CONSUMPTION CONSUMPTION Air-condition
[˚C] [kWh [kWh [kWh [kWh [kWh km-1 [%]
22 6.947 0.977 0.417 8.341 1.117 11.7
23 6.947 1.222 0.430 8.599 1.151 14.2
24 6.947 1.457 0.442 8.847 1.184 16.5
25 6.947 1.681 0.454 9.083 1.216 18.5
26 6.947 1.895 0.465 9.307 1.246 20.4
27 6.947 2.098 0.476 9.521 1.275 22.0
28 6.947 2.291 0.486 9.724 1.302 23.6
29 6.947 2.472 0.496 9.915 1.327 24.9
30 6.947 2.644 0.505 10.096 1.352 26.2
31 6.947 2.805 0.513 10.265 1.374 27.3
32 6.947 2.955 0.521 10.423 1.395 28.3
33 6.947 3.094 0.529 10.570 1.415 29.3
34 6.947 3.223 0.535 10.706 1.433 30.1
35 6.947 3.342 0.542 10.831 1.450 30.9
36 6.947 3.450 0.547 10.944 1.465 31.5
37 6.947 3.547 0.552 11.047 1.479 32.1
38 6.947 3.634 0.557 11.138 1.491 32.6
39 6.947 3.710 0.561 11.218 1.502 33.1
In the summer period of exploitation at extremely high outside temperatures exceeding 35 °C, the air conditioning
system consumes more than 30.9% of the total energy consumption of the E-bus. For the temperature range of
outside temperature, which is usually 28-34 °C, the participation of the air conditioning system is from 23.6-30.1%.
The smallest effect of air-conditioning on the total consumption of the E-bus, is on the temperature range of outside
temperatures between 22-25 °C, when it is below 20%.
TRANSITION PERIOD OF EXPLOITATION

During the spring and autumn, the heating system is used with lower power, compared to the winter period for
additional heating of the passenger compartment, usually in the morning and evening when the daily temperature
minimum. In the periods of the day when the daily temperature reaches a maximum which is usually up to 20 °C,
depending on the number of passengers in the vehicle in certain cases is used of natural ventilation and forced
ventilation by the fan, which affects the increase of the total energy consumption of E-bus.
During September, October and April 2016 and 2017, the change of SOC for different dates was recorded, the
outside temperature was in the range of 10 °C to 22 °C [11]. Analysis of a sample of 84 measurements total
consumption of E-bus (with calculated losses in the filling phase) expressed in kWh·km-1, were in the range of 0.890-
1,230 kWh·km-1. Distribution of total consumption is presented in Figure 31.
Figure 31 Distribution of total electricity consumption in the transition period on the line EKO 1,
direction ''A''
134
The results of total consumption include real exploitation conditions during peak load periods, off-peak loads as well
as first/last departures, are characterized by different vehicle loads. Figure 32 shows the total consumption of the E-
bus in the function of the outside temperature for the observed period. Between consumption of E-bus and the
outside temperature in the transition period there is a high correlation (R2 =0.952) which is described by the function
of the 3rd degree polynomial.
Figure 32 Electricity consumption E-bus, expressed kWh·km-1, as a function of the outside temperature
(transition period), direction ''A''
Based on the representative sample (Xi), the statistical analysis of the results of electric consumption of the E-bus
was performed using the mean value (Xs) and the standard deviation (sd).
The average value of the treated sample is 1.045 kWh·km-1, a standard deviation is 0.082 kWh·km-1.
Since the sample distribution corresponds to the Normal distribution, the expected interval of the E-bus consumption
based on the interval estimation of the mean value (μ) of the base set, with a probability of 95% for the sample n≥
30, P(1.027 <μ < 1.062) = 0.95 . Based on the results of the total consumption of the E-bus and the resulting trend
shown in Figure 32, the effect of the heating and ventilation system on the total consumption of E-bus in the observed
period in the function of the outside temperature is shown in Table 19.
Table 19 Distribution of total consumption of E-bus in the transition period, direction ''A''
(motion, heating/ventilation, losses in the filling phase)
Consumed Consumed energy
Outside energy on the Energy losses TOTAL TOTAL Consumption
Temperature on the motion heating/ventilation during charging CONSUMPTION CONSUMPTION Heating/ventilation
[˚C] [kWh [kWh [kWh [kWh [kWh km-1 [%]
10 6.775 1.144 0.417 8.337 1.116 13.7
11 6.775 1.217 0.421 8.413 1.126 14.5
12 6.775 1.162 0.418 8.355 1.118 13.9
13 6.775 1.012 0.410 8.197 1.097 12.3
14 6.775 0.801 0.399 7.975 1.068 10.0
15 6.775 0.564 0.386 7.726 1.034 7.3
16 6.775 0.334 0.374 7.484 1.002 4.5
17 6.775 0.146 0.364 7.285 0.975 2.0
18 6.775 0.032 0.358 7.166 0.959 0.4
19 6.775 0.023 0.358 7.156 0.958 0.3
20 6.775 0.166 0.365 7.306 0.978 2.3
21 6.775 0.481 0.382 7.638 1.022 6.3
22 6.775 1.006 0.410 8.191 1.097 12.3
At outdoor temperatures between 17 °C and 19 °C there is the highest energy efficiency of the E-bus, since that
heating and air conditioning systems are not used. The average consumption of the E-bus is between 0.95 and 0.96
kWh·km-1, about 94.5% of the consumed energy refers to the movement of vehicles, auxiliary devices on the E-bus,
and losses in electric motors and transmissions.
135
The temperature range between 10 °C and 16 °C, requires the use of a heating system, so that the heating system
is impacted for consumption between 13.7% and 4.5%. Over 20 °C, on the E-bus ventilation fan, is used in the
passenger compartment, in the rare cases an air conditioner, so their influence on the electricity consumption can
be from 2.3% and 12.3%.
If all the obtained results of measuring the total consumption of the E-bus during the summer, winter and transition
periods were displayed in the function of outside temperature, the dependence which can be described by the
function of the polynomial of the 3rd degree with the coefficient of correlation R2 = 0.953, as shown in Figure 33.
Figure 33 Electricity consumption E-bus, expressed kWh·km-1, as a function of the outside temperature,
direction ''A''(winter, transition and summer period)
Average consumption of E-Bus on an annual level, taking into account the different consumption during different
periods of operation, can be shown in Table 20.
Table 20 Energy consumption of E-bus in different periods of exploitation, direction ''A''

Period of Duration Average energy
exploitation [months] consumption of E-bus
[kWh·km-1]
Summer 4 1.289
Winter 5.5 1.520
Transition 2.5 1.045
Total 12 1.34
The average electricity consumption of the E-bus on the EKO 1 line, direction ''A'', for all periods of exploitation is
1.34 kWh·km-1. In the winter period of exploitation, the consumption of E-bus is about 45.4% higher compared to the
transition period. In the summer period consumption of E-bus is higher at about 23.3% compared to the transition
period (spring, autumn).
CONCLUSIONS
The operation of the heating and air conditioning systems for electric buses are of high importance to the energy
efficiency of the vehicle. Unlike buses with diesel drive where the impact of these systems is not as pronounced,
buses on electric power are had a large variations in electricity consumption at different period of the exploitation.
In the case of exploitation of electric buses in Belgrade, the consumption of electric energy in the winter increase by
45.4% in the summer period by 23.3% compared to the transition period when these systems are not used or used
at a lower intensity.
136
The practice has shown, when are outside temperatures between 17 °C and 19 °C, the energy efficiency of buses
on an electric drive are highest since at these temperatures heating and air conditioning systems are not used. The
heating and air conditioning systems installed in the E-Buses which used in Belgrade, fully meet all the required
performance in terms of thermal comfort in accordance with the defined recommendations.
Knowing the performance of the heating and air conditioning system from the aspect of thermal comfort and electricity
consumption is important from the aspect of system selection and dimensioning of the required capacity of the
electrical energy storage system for electric buses in order to realize the planned autonomy of vehicle movement at
the working day level.
In order to maximize the energy efficiency of electric buses, the proper use of these systems is especially important
in some extreme conditions of low and high outside temperatures.
In summer period of operation when the air conditioner is intensively used, often open side windows in the passenger
compartment allow hot air flow, which reduces the efficiency of the air conditioning system as well as the energy
efficiency of the E-bus.
In winter conditions, the biggest problem is unplanned congestion on the streets, with the heating system, is operated
at full capacity. In such situations, it is necessary for the driver to switch off the heating system in order to prevent
the excessive discharge of the battery or supercapacitor and to prevent the risk of stopping the E-bus on the route.
ACKNOWLEDGMENTS
The authors are grateful to City Public Transport Company ’’Belgrade’’, especially fellow engineers: Dušan Savković,
Pavle Krstić and Slobodan Ivanov, for complete support, measuring equipment and additional time for measuring in
sometimes very extreme winter and summer conditions.
REFERENCES
[1] Klimatisierung von linienbussen der Zulassung klassen I und II, fur konventionell angetriebene Diesel-und
dasbusse als auch fur Hybrid und Elektrobuse, VDV-Schrift 236, Stand 03/2018,
[2] Christof, M.: "Thermal comfort for buses: standardization of requirements“, UITP- working group E-bus thermal
comfort, meeting 20.3.2018, Brussels,
[3] Higer Electric bus – Tehnička dokumentacija, 2016. godina,
[4] Статистички годишњак Београда 2016, Секретеријат за управу-Сектор статистике, Београд 2017,
[5] Radičević, B., Mikičić, D., Vukić D.: "Energetski potencijal sunca u Srbiji i primena energije sunca u poljoprivredi“,
Poljoprivredna tehnika, No. 4, 2009, pp 53-62,
[6] Ružić, D.: "Mikroklima u motornim vozilima“, Fakultet tehničkih nauka, Novi Sad, 2016,
[7] Fanger, P.O.: "Thermal confort“, McGraw-Hill, New York, SAD, 1970,
[8] Parsons, K.: "Human thermal environments: The effect of hot, moderate and cold environments on human
health, comfort and performance, 2nd ed.“, Taylor&Francis, London, 2003,
[9] Misanovic, S., Krstic, P.: "Measurement of the electrical energy consumption the heating system at E-bus in
Belgrade”, UITP- working group E-bus thermal comfort, meeting 20.3.2018, Brussels,
[10] JKP GSP Beograd, Меrenje potrošnje električne energije Е-busa u ekstremno zimskim uslovima eksploatacije,
Intreni dokument, 2016,
[11] JKP GSP Beograd, Baza podataka potrošnje električne enegije E-busa, interni dokument, 2017- 2018.godina,
[12] Vukadinović, S.: "Elementi teorije verovatnoće i matematičke statistike“, Privredni pregled Beograd, 1990,
[13] JKП ГСП Београд, Мерење потрошње електричне енергије Е-буса у екстремно летњим условима
експлоатације, Интерни документ, 2017.
137
138
Kragujevac, Serbia
MVM2018-011
Mikhail G. Shatrov1
Leonid N. Golubkov2
Andrey U. Dunin3
Pavel V. Dushkin4
Andrey L. Yakovenko5
ANALYSIS OF REALIZATION OF STEPPED FUEL

INJECTION WITHOUT CHANGING DESIGN OF COMMON
RAIL INJECTOR
ABSTRACT: A calculated-experimental analysis of realization of stepped front edge of injection rate (stepped
injection) was carried out. Common rail injectors (CRI) of three most widely used designs took part in the research:
CRI No 1 distinguished by its control valve with a locking cone and piston; CRI No 2 with a control valve unloaded
from fuel pressure, channel on the piston and enlarged internal volume; CRI No 3 with a control valve having a flat
lock and a needle which is not closing the drain channel when staying in the upper position. It is demonstrated that
friction in the couple – control valve piston/guiding surface of the CRI No 1 hampers realization of stepped front edge
of injection rate due to its smoothening. When using the CRI No 2, it is possible to obtain stepped injection only at
low fuel pressure in common rail (pcr ≤ 50 MPa). The CRI No 3 ensure the stepped injection at different pressures
up to pcr=200 MPa.
KEYWORDS: common rail injector, common rail fuel system, stepped fuel injection, electric impulse, control valve
INTRODUCTION
Advantages of Common rail (CR) type fuel systems with Common rail injectors (CRI) and electronic control are
explained by realization of the principle of sharing two functions: obtaining high fuel pressure and fuel injection
process organization. CR fuel systems also enables to carry out multiple injection, flexible control of injection advance
angle, injection pressure and amount of injections per cycle. Using CR fuel systems simplifies the tasks of cylinders
and cycles shutdown, exhaust gases neutralization system control, as well as sensors and actuators diagnostics.
The further strengthening of ecological rules and standards specifying content of toxic substances in exhaust gases
of diesel engines [13-15] accompanied by the requirement for improvement of their fuel efficiency predetermine
perfection of CR fuel systems.
1
Mikhail G. Shatrov, professor, MADI, 64, Leningradsky Prosp., Moscow, 125319, Russia, dvs@madi.ru
2
Leonid N. Golubkov, professor, MADI, 64, Leningradsky Prosp., Moscow, 125319, Russia,
dvsgolubkov@yandex.ru
3
Andrey U. Dunin, assistant professor, MADI, 64, Leningradsky Prosp., Moscow, 125319, Russia,
a.u.dunin@yandex.ru
4
Pavel V. Dushkin, assistant, MADI, 64, Leningradsky Prosp., Moscow, 125319, Russia, levvap@gmail.com
5
Andrey L. Yakovenko, assistant professor, MADI, 64, Leningradsky Prosp., Moscow, 125319, Russia,
iakovenko_home@mail.ru
139
Fields of these improvements: raising injection pressure up to 300 MPa [7, 8]; Ensuring the required shape of the
injection rate front edge [10, 11] taking into account wave propagation effects originating during fuel injection [12];
control of fuel distribution by the combustion chamber [9].
A boot-shaped front edge of the injection rate enables to decrease the pressure rise rate and maximal pressure in
the cylinder at many operating modes which makes it possible to decrease tailpipe toxic emissions and noise level.
RESEARCH OBJECTS
The research objects were three modern Common Rail injectors (CRIs) having different designs which are
distinguished by the following specific features.
Specific design feature of the CRI No 1 shown in Figure 1 – design of its control valve 2 including its locking cone 13
having internal diameter 2 mm and valve piston 12 having diameter 12 mm [3]. This CRI design is used by the Delphi
company and the Altai Factory of Precision Articles (AZPI)
Figure 1 A construction diagram of the CRI No 1 distinguished

by its control valve with a locking cone and valve piston:
1 – electromagnet; 2 – control valve; 3 – body; 4 – control chamber; 5 – nozzle needle valve; 6 – sack volume;
7 – valve spring; 8 – outlet fuel jet; 9 – inlet fuel jet; 10 – needle valve spring; 11 – internal volume; 12 – control
valve piston; 13 – control valve locking cone
The basic specific feature of the design of the CRI No 2 (Figure 2) – fuel pressure balanced control valve 2 (in its
closed position). In addition, the CRI No 2 is distinguished by the presence of the channel 14 in the piston 5 and
increased internal volume 6 [2, 6]. This construction diagram corresponds to the model CRI 2.6 of the Bosch
company.
Figure 2 A construction diagram of the CRI No 2 with fuel pressure balanced control valve,
valve channel and increased internal volume:
1 – electromagnet; 2 – control valve; 3 – body; 4 – inlet fuel jet; 5 – multiplier piston; 6 – internal volume;
7 – nozzle fuel jet; 8 – suck volume; 9 – valve spring; 10 – outlet fuel jet; 11 – control chamber; 12 – needle
valve spring; 13 – needle valve; 14 – multiplier piston channel
140
Figure 3 shows the construction diagram of the CRI No 3 which is distinguished by the design of its control valve 3
with a flat latch and also by the fact that the end surface of the needle valve 8 does not cover the fuel drain hole when
the needle valve is in the upper point (in contrast, for example, to the diagram of the CRI No 4 that will be mentioned
below) [1].
Figure 3 A construction diagram of the CRI No 3 which is distinguished by the control valve with a flat latch
and a needle valve which does not cover the fuel drain hole when it is in the upper position:
1 – control valve spring; 2 – electromagnet; 3 – control valve; 4 – outlet fuel jet; 5 – control chamber;
6 – CRI body; 7 – needle valve spring; 8 – nozzle needle valve; 9 – inlet fuel jet; 10 – CRI internal volume;
11 – sack volume
The construction diagram of the CRI No 3 was realized in the injector PLTD.387442.20.00 of the MADI Problem
Laboratory of Transport Engines (PLTD) jointly with the Noginsk Fuel Equipment Factory (NZTA).
SIMULATION RESULTS
Mathematical models of the Common Rail fuel systems with CRIs of the designs presented above were published in
a number of papers [1, 3-5].
Calculated analysis of the CRI No 1 (Figure 1) was carried out using the dependence of the electromagnet force Fem
versus time , as shown in Table 1. Parameters of the CRIs manufactured by the NZTA were used as input data.
Table 1 Durations of control impulses and intervals between them depending on time for the CRI No 1
Parameter Value
 (ms) 0 0.10 0.20 0.30 0.38 0.48 0.90 0.95
Fem (N) 0 200 200 0 0 200 200 0
Figure 5 and Figure 6 show the results of calculation at pressure in the common rail рcr = 100 MPa and fuel injected
mass Qc = 258,2 mg, where qcv – control fuel flow through the valve; qir – injection rate; hcv – control valve
displacement; hnv – needle valve displacement; kfr – friction coefficient.
Calculation results presented in Figure 5 were obtained in case when there was no friction (kfr=0 Ns/m) in the joints
of the CRI. A similar result was obtained at рcr=160 MPa, at that, the same dependence Fem=f() was used (see Table
1).
For calculation of the influence of friction forces Ffr during the control valve piston movement, data indicated in paper
[16] obtained for the nozzle needle valve was used. The calculation results demonstrated that the boot-shaped front
edge of the injection rate was smoothed completely at kfr=60 Ns/m (рcr=100 MPa) and at kfr=50 Ns/m (рcr=160 MPa).
Figure 6 shows the results of calculation at the same mode that in Figure 5, but with kfr=60 Ns/m (Ffr max=84 N).
Calculated analysis of the CRI No 2 (Figure 2) was carried out using parameters corresponding to the CRI 2.6 model
of the Bosch company for estimation of the opportunity of realization of the boot-shaped fuel injection rate. It was
shown that at рcr = 100 MPa, this may be realized only using the impulses differing by 1 ms (Table 2) which was out
141
of the limit of accuracy of modern Common Rail systems control. At the present operation mode of the CRI No 2, the
fuel injected mass Qc was 43.86 mg, and control fuel consumption Qcon = 37.28 mg.
Figure 5 Calculation results of fuel injection process parameters of the CRI No 1 using input data of the injector
produced by the NZTA at рcr = 100 MPa, Qc = 258,2 mg, kfr=0 Ns/m
produced by the NZTA at рcr = 100 MPa, Qc = 273,8 mg, kfr=60 Ns/m
142
at рcr = 100 MPa
Parameter Value
 (ms) 0 0.006 0.007 0.008 0.19 0.20 2.0 2.1
Fem (N) 0 200 200 0 0 200 200 0
To get the boot-shape injection rate at pcr = 200 MPa, one had to increase the tightening force of the valve spring
from 47 to 94 N. At these conditions, for organization of the boot-shaped injection, impulses differing by 0.01 ms were
required.
Table 3 presents the dependence of the electromagnet force Fem on time  at pcr = 50 MPa. At this operating mode
of the CRI No 2, the calculated fuel injection mass Qc was 31.1 mg, and control fuel consumption Qcon = 26.54 mg.
at рcr = 50 MPa
Parameter Value
 (ms) 0 0.005 0.015 0.020 0.19 0.20 2.0 2.1
Fem (N) 0 200 200 0 0 200 200 0
The results presented in Tables 2 и 3 show that it is possible to obtain the boot-shaped injection rate at a modern
accuracy of the control system signals (± 1 ms) only at рcr ≤ 50 MPa due to particular design of the control valve of
the CRI No 2 (Figure 2). One should also mention that the boot-shaped injection is used as a rule at close to maximal
loads [8] when a high fuel injection pressure is required.
For design of the CRI No 3 (Figure 3), a calculated analysis of the opportunity of getting a boot-shaped injection rate
in case of supply of one (Table 4) and two (Table 5) primary impulses was carried out. Parameters of the CRI
PLTD.387442.20.00. were used as input data.
at pcr = 200 MPa (one primary impulse)
Parameter Value
 (ms) 0 0.1 0.12 0.14 0.35 0.37 2.0 2.1
Fem (N) 0 200 200 0 0 200 200 0
at pcr = 200 MPa (two primary impulses)
Parameter Value
 (ms) 0 0.1 0.2 0.25 0.38 0.43 0.435 0.445 0.63 0.68 2 2.1
Fem (N) 0 200 200 0 0 200 200 0 0 200 200 0
Figure 7 and Figure 8 show the results of the CRI No 3 injection process calculation when one and two primary
impulses were supplied correspondingly.
In this way, the calculations proved the opportunity of realization of the boot-shaped fuel injection rate using the CRI
No 3 both by forming one primary impulse (Figure 7), as well as two primary impulses (Figure 8).
CONCLUSIONS
1. The authors carried out a calculated-experimental analysis of the opportunity of organization of boot-shaped
injection using Common Rail injectors (CRI) of three main designs that are used: CRI No 1 (Delphy and AZPI
design) distinguished by their control valve with a locking cone and valve piston; CRI No 2 (Bosch design, CRI
2.6 model) with fuel pressure balanced control valve, channel in the piston and increased internal volume; CRI
No 3 (MADI-NZTA design, PLTD.387442.20.00 model) distinguished by its control valve with a flat latch and
needle valve not closing the drain hole when staying in the upper position.
2. The boot-shape front edge of the CRI No 1 may be partially smoothed when attaining the friction coefficient
more than 60 Ns/m in the joint: control valve piston – CRI body.
3. Authors demonstrated that it was possible to obtain the boot-shaped injection rate of the CRI No 2 only at low
pressures in the common rail (рcr ≤ 50 MPa), which may be explained by specific features of the design of its
control valve.
4. The CRI No 3 ensure the opportunity of getting boot-shaped injection of fuel at various pressures in the common
rail.
143
PLTD.387442.20.00 at рcr = 200 MPa, Qc = 662,5 mg
Figure 8 Calculation results of fuel injection process parameters of the CRI No 3 with two primary impulses using
input data of the injector PLTD.387442.20.00 at рcr = 200 MPa, Qc = 675,24 mg
144
ACKNOWLEDGMENTS
Applied research and experimental developments are carried out with financial support of the state represented by
the Ministry of Education and Science of the Russian Federation under the Agreement No 14.580.21.0002 of
27.07.2015, the Unique Identifier PNIER: RFMEFI58015X0002.
REFERENCES
[1] Dushkin, P.V.: “Increasing Efficiency of the Working Process of Common Rail Fuel System with Injection
Pressure up to 300 MPa”, Master’s Thesis, 05.04.02, Moscow, 2017, 16 p.,
[2] Dushkin, P.V.: “Intercycle Instability of Small Injected Fuel Masses of Diesel Engine Fuel Systems”, Vestnik
MADI, Moscow, MADI, Issue 3 (46), 2015, pp 42-49,
[3] Golubkov, L.N., Dushkin, P.V.: “Mathematical model of Common Rail injector with incorporated common rail and
modified control valve”, Vestnik MADI, Мoscow, MADI, Issue 3(42), 2015, pp 11-18,
[4] Golubkov, L.M., Grishin, A.V., Yemenlianov, L.A.: “Results of Calculated Research and Optimization of the
Common Rail Fuel System with Electric Control of Injectors”, Porshneviye Dvigateli i Topliva v XXI veke:
Collection of research papers, MADI(GTU), Moscow, MADI, 2003, pp 37-52,
[5] Golubkov, L.N., Yemelianov, L.A., Mikhalchenko, D.A.: “Calculation-Theoretical Research of Common Rail Fuel
System for Improvement of Diesel Engine Ecological Parameters”, Dvigateli i Elologiya: Collection of research
papers, Moscow: NAMI, Issue 238, 2007, pp 103-109,
[6] Shatrov, M.G., Golubkov, L.N., Dunin, A.U., Dushkin, P.V.: “Experimental Research of Hydrodynamic
Parameters in Common Rail Fuel Systems at Multiple Injection”, Zhurnal avtomobilnyh inzhenerov, No. 2, (97),
2016, pp 15-17,
[7] Shatrov, M.G., Golubkov, L.N., Dunin, A.U., Yakovenko, A.L., Dushkin, P.V.: “Influence of high injection pressure
on fuel injection performances and diesel engine working process”, Thermal Science, Vol. 19, Issue 6, 2015, pp
2245-2253,
[8] Shatrov, M.G., Golubkov, L.N., Dunin, A.U., Yakovenko, A.L., Dushkin, P.V.: “Research of the impact of injection
pressure 2000 bar and more on diesel engine parameters”, International Journal of Applied Engineering
Research, Vol. 10, No. 20, 2015, pp 41098-41102,
[9] Shatrov, M.G., Malchuk, V.I., Dunin, A.U., Yakovenko, A.L.: “The influence of location of input edges of injection
holes on hydraulic characteristics of injector the diesel fuel system”, International Journal of Applied Engineering
Research, Vol. 11, No. 20, 2016, pp 10267-10273,
[10] Shatrov, M.G., Golubkov, L.N., Dunin, A.U., Dushkin, P.V., Yakovenko, A.L.: “The new generation of common
rail fuel injection system for Russian locomotive diesel engines”, Pollution Research, Vol. 36, No. 3, 2017, pp
678-684,
[11] Shatrov, M.G., Golubkov, L.N., Dunin, A.Yu., Dushkin, P.V., Yakovenko, A.L.: “A method of control of injection
rate shape by acting upon electromagnetic control valve of common rail injector”, International Journal of
Mechanical Engineering and Technology, Vol. 8, Issue 11, 2017, pp 676-690,
[12] Shatrov, M.G., Golubkov, L.N., Dunin, A.U., Yakovenko, A.L., Dushkin, P.V.: “Experimental research of
hydrodynamic effects in common rail fuel system in case of multiple injection”, International Journal of Applied
Engineering Research, Vol. 11, No. 10, 2016, pp 6949-6953,
[13] Shatrov, M.G., Sinyavski, V.V., Dunin, A.Y., Shishlov, I.G., Vakulenko, A.V.: “Method of conversion of high- and
middle-speed diesel engines into gas diesel engines”, Facta Universitatis. Series: Mechanical Engineering, Vol.
15, No. 3, 2017, pp 383-395,
[14] Shatrov, M.G., Sinyavski, V.V., Dunin, A.Y., Shishlov, I.G., Vakulenko, A.V., Yakovenko, A.L.: “Using simulation
for development of the systems of automobile gas diesel engine and its operation control”, International Journal
of Engineering and Technology, Vol. 7, No. 2.28, 2018, pp 288-295,
[15] Sinyavski, V.V., Alekseev, I.V., Ivanov, I.Y., Bogdanov, S.N., Trofimenko, Y.V.: “Physical simulation of high- and
medium-speed engines powered by natural gas”, Pollution Research, Vol. 36, Issue 3, 2017, pp 684-690,
[16] Trusov, V.I., Dmitrienko, V.P., Maslyani, G.D.: “Injectors for Automobile and Tractor Diesel Engines”, Moscow,
«Mashinostroyeniye», 1977, 167 p.
145
146
Kragujevac, Serbia
MVM2018-012
Mikhail G. Shatrov1
Vladimir V. Sinyavski2
Ivan G. Shishlov3
Andrey Y. Dunin4
Andrey V. Vakulenko5
EXPERIENCE OF DIESEL ENGINES CONVERSION FOR

OPERATION ON NATURAL GAS OBTAINED IN MADI
ABSTRACT: Analysis of different methods of diesel engines conversion for operation on natural gas was carried
out showing feasibility of these methods for engines of various applications. The KAMAZ V8 120/120 mm diesel
engine was converted to operate on natural gas by spark ignition cycle and Cummins Kama 6L 104/127 mm diesel
engine – by gas diesel (dual fuel) cycle. Gas feed and electronic engine control systems were developed for both the
methods of engine conversion. For gas diesel version, modular gas feed and engine control systems were developed
which could be mounted both on high-speed and medium-speed gas diesel engines. These systems were perfected
during engine tests. CR fuel system for the perspective medium-speed D200 6L 200/280 mm gas diesel engine was
developed jointly with the industrial partner. A computer model for simulation of engine operation was developed and
calibrated on the basis of engine tests and used for the development of gas feed and engine control systems,
experimental results analysis and forecasting of the parameters of the medium-speed D200 gas diesel engine.
Engine tests of the gas and gas diesel engines demonstrated considerable decrease of CO2 and NOx emissions,
practically no soot and lower fuel consumption compared to the base diesel engine.
KEYWORDS: gas engine, gas diesel engine, dual fuel engine, engine conversion, engine simulation
INTRODUCTION
Conversion of diesel engines to operate on natural gas is relevant because gas is almost twice cheaper than diesel
fuel in Russia. In our days, the main reason of replacement of diesel engines by gas and gas diesel (dual fuel)
engines on city transport is ecology. Most of the municipal vehicles operated in the towns: buses, delivery trucks,
garbage collectors, etc., as well as civil engineering machines which are permanently working on construction sites
are equipped with diesel engines which, in addition to nitrogen oxides, emit soot. Soot contains carcinogens such as
benzapyrene which threaten the health and life of millions of people living in megapolices. The influence of the
exhaust gases of diesel engines on human health is so evident that authorities of many European towns are already
putting bans on entrance of vehicles with diesel engines to the city centers or to the whole cities. Restriction for diesel
vehicles to enter Copenhagen is planned for 2019, and for entering Mexico, Paris, Madrid and Athens – for 2025.
There are many ways of conversion of diesel engines to operate on natural gas and they decrease significantly or
eliminate fully content of soot in the exhaust gases.
1
Mikhail G. Shatrov, MADI, 64, Leningradsky Prosp., Moscow, 125319, Russia, dvs@madi.ru
2
Vladimir V. Sinyavski, MADI, 64, Leningradsky Prosp., Moscow, 125319, Russia, sinvlad@mail.ru
3
Ivan G. Shishlov, MADI, 64, Leningradsky Prosp., Moscow, 125319, Russia, astra510@yandex.ru
4
Andrey Y. Dunin, MADI, 64, Leningradsky Prosp., Moscow, 125319, Russia, a.u.dunin@yandex.ru
5
Andrey V. Vakulenko, MADI, 64, Leningradsky Prosp., Moscow, 125319, Russia, ingener-avto@yandex.ru
147
Conversion of diesel engines to spark ignition gas engines operating on a stoichiometric gas-air mixture is described
in details in [8]. When the gas is supplied to the intake system, a portion of air entering the cylinders is substituted by
gas which results in decrease of filling efficiency and correspondingly engine power by about 10%. The engine power
is also lost due to reduction of the “diesel” compression ratio by 5-6 points to avoid knock. The chance for knock
origination in stoichiometric gas engines is much lower than in gas engines working on a lean gas-air mixture. Using
stoichiometric gas-air mixture instead of a lean mixture that was in the base diesel engine helps to compensate
considerably for power loss. Additional advantages are stable combustion of gas and possibility to use a three-way
catalyst similar to that mounted on petrol engines which decreases emissions of carbon oxide, nitrogen oxides and
hydrocarbons. Soot emissions are close to zero. Stoichiometric gas engines have a high temperature of exhaust
gases that is dangerous for turbine wheel and impedes considerable engine power augmentation. This is a good
method for naturally aspirated engines because of minimal loss of power. As modern diesel engines are
turbocharged, this method of conversion is reasonable in case of overhauling old diesel engines having no
turbochargers to give them the second “clean” life or for turbocharged diesel engines having low power augmentation.
Using a stoichiometric gas-air mixture does not improve fuel efficiency compared with the base diesel engine, but
taking into account the lower price of natural gas compared to diesel fuel, expenses for fuel decrease. Research of
fuel consumption by buses having gas engines with stoichiometric gas-air mixture in Serbia showed that despite the
growth of fuel consumption by 10-25% compared with buses having diesel engines, expenses for fuel decrease
considerable because natural gas in Serbia is by 52% cheaper than diesel fuel [3].
Lean mixture gas engines have low exhaust gases temperature and this enables to raise boost pressure and hence
to get a high power augmentation, high fuel efficiency and low NOx emissions which allows to meet the ecological
standards without a reduction catalyzer. For medium-speed gas engines such as Jenbaсher J624 (type 6) having
D/S=190/210 mm operating on a lean gas-air mixture and used for electric energy generation, prechamber with
enriched mixture is used to inflame the lean mixture in the main combustion chamber. The advanced Miller cycle is
implemented to avoid knock. It ensures high thermal efficiency and reduces NOx emissions. Two-stage turbocharging
system is used to compensate for filling efficiency drop caused by the Miller cycle [2, 7]. This is a good application
for power generation because the engine operates constantly at a high speed which enables it to avoid knock. The
use of this working process on medium-speed high boosted transport engines with large cylinder bore is impeded by
knock which originates at low speeds and transfer modes. But for high-speed engines having relatively small cylinder
bore, the origination of knock is less probable and it can be eliminated by engine electronic control system. Therefore
gas cycle with lean mixture is widely used on engines of trucks and buses.
The advantages of gas diesel engines compared to gas engines are high power, fuel efficiency and hence less CO2
emissions, no knock, no need to change the combustion chamber shape in order to decrease the compression ratio
[1, 11]. Modern high pressure Common Rail diesel fuel injection systems ensure small portions of ignition diesel fuel.
Disadvantages are higher price because the engine has gas and diesel fuel supply systems and the need to refuel
the vehicle with two types of fuel.
RESEARCH OBJECTS
All three methods of diesel engine conversion were realized in the Engine Laboratory of MADI for high-speed diesel
engines mounted in different times on KAMAZ trucks and buses. Also systems for a gas diesel version of the medium-
speed diesel engine D200 whose mass production has not yet started were developed and gas diesel engine
parameters were forecasted. The main data for the four engines are given in Table 1.
Table 1 Main data of the engines investigated

Number and Cylinder Cylinder Rated Rated break Comp-
Engine arrangement diameter stroke speed mean effective ression
of cylinders (mm) (mm) (rpm) pressure (MPa) ratio
Naturally aspirated KAMAZ V8
120 120 2200 0.704 13:0
gas engine
Turbocharged KAMAZ gas V8
120 120 2200 1.01 11.0:1
engine
Cummins KAMA gas diesel L6
107 124 2300 1.73 17.0:1
engine
D200 gas diesel engine L6
200 280 1000 2.00 15.0:1
(parameters were forecasted)
148
CONVERSION OF KAMAZ DIESEL ENGINES TO GAS ENGINES
In gas engines, the combustion chamber shape had to be changed to reduce the compression ratio from 17:1 of the
base diesel engine to 13.0:1 for naturally aspirated and 11.0:1 for turbocharged versions to avoid knock and a spark
plug was mounted in place of a fuel injector. Cross section of the KAMAZ gas engine cylinder is shown in Figure 1.
Compression ratio for the gas engines shown in Table 1 was selected on the basis of the world experience with the
aim to find the best compromise between high power and fuel efficiency, as well as absence of knock.
Figure 1 Cross section of the KAMAZ gas engine cylinder

1 – combustion chamber; 2 – piston; 3 – diesel engine injector mounting channel; 4 – spark-plug;
5 – spark-plug electrodes; 6 – exhaust port; 7 – steel sealing ring
The first engine converted for operation on gas fuel was naturally aspirated KAMAZ-7403.10 using a stoichiometric
gas-air mixture. A gas-feed system with electronic control was developed. Its basic diagram is shown in Figure 2.
Two injectors ensure a central gas supply to each row of cylinders which ensures the minimal difference of air-excess
ratio values in the cylinders and this is important for operation of a three-way catalyst. The stoichiometric mixture is
supported by electronic metering devices at loads higher than 70% and at idle mode. Rhodium was added to the
three-way catalyst to oxidize the emissions of unburned methane. Catalysts developed by specialists of the NAMI
Research Institute reduced the content of carbon oxide and also hydrocarbons including unburned methane. The
systems of the gas engine were calibrated to get the same rated power as the base diesel engine.
Figure 2 Basic diagram of the gas feed system for naturally aspirated KAMAZ engine
1 – reservoir for storage of gas; 2 – high pressure line; 3 – gas heater; 4 – high pressure reducer;
5 – electromagnet valve-filter; 6 – low pressure reducer; 7 – throttle valve; 8 – mixer; 9 – valve; 10 – idle
mode channel; 11 – control module; 12 – vacuum tube; 13 – main channel; 14 – metering unit; 15 – ventury;
16 – throttle valve; 17 – bypass channel; 18 – control valve; 19 – intake manifold; 20 – air filter;
21 – thermistor electric heater; 22 – electromagnetic valve; 23 – idle mode channel; 24 – idle mode needle;
25 – control valve; 26 – injector; 27 – position sensor; 28 – oxygen sensor; 29 – control unit; 30 – exhaust manifold
149
The second converted engine was a turbocharged KAMAZ-7409Т with a completely new electronic ejection gas
supply system shown in Figure 3. Gas was supplied to the inlet of the compressors. The electronic throttle valve was
mounted at the outlet of the compressors. The use of ejector gas supply makes it possible to decrease considerably
the time of engine perfection and simplifies the engine control system. Two-stage system of after-cylinder exhaust
gases treatment oxidation catalysts developed in the NAMI Research Institute was used.
Figure 3 Elements of gas feed and electronic control systems on the turbocharged KAMAZ gas engine
1 – ignition coil; 2 – special disc (60-2); 3 – cylinder position sensor; 4 - crankshaft position sensor;
5 - camshaft position sensor; 6 – throttle valve position sensor; 7 – engine start electromagnet gas valve;
8 –bypass channel electromagnet gas valve; 9 – engine coolant temperature sensor; 10 – acceleration
electromagnet gas valve; 11, 12, 13 – exhaust gas temeprturte sensors; 14 - throttle valve; 15 – absolute pressure
sensor; 16 – first step exhaust gas catalyzer; 17 – supply of natural gas from the gas reducer; 18 – air filter,
19 - second step exhaust gas catalyzer
As a part of intake air is replaced by gas, if the same turbocharger as on the base diesel engine is used on the gas
engine, this results in considerable decrease of engine power and efficiency, especially at low engine speed when
the boost pressure is low. Therefore on gas engines, turbochargers producing a higher boost pressure than on the
base diesel engine should be used. Few turbochargers having smaller turbine minimal flow section Ft0 were tested
and finally the Czech В 65-1 turbochargers having the Ft0 value 5.6 cm2 with bypass valve were selected (the base
diesel engine has turbochargers with the Ft0 value 12.0 cm2 and no by-pass valves). The bypass valves open
approximately at the middle of the speed range to avoid too high maximal combustion pressure and overrun of the
turbocharger. Turbochargers with smaller turbines enabled the gas engine to get even higher power and torque than
the base diesel engine (Figure 4). Though the gas engine efficiency was lower due to lower compression ratio [5].
All three gas engines were tested for emissions by the ESC cycle. The results are shown in Table 2 jointly with the
maximal permitted values of emissions for heavy-duty diesel engines, which generally include trucks and buses.
Table 2 EU Emission Standards for heavy-duty Diesel Engines (g/kWh)

Date of introduction CO HC NOx PT
(g/kWh) (g/kWh) (g/kWh) (g/kWh)
Euro-3 October 2000 2.1 0.66 5.0 0.1
Euro-4 October 2005 1.5 0.46 3.5 0.02
Euro-5 October 2008 1.5 0.46 2.0 0.02
Euro-6 September 2014 1.5 0.13 0.4 0.01
KAMAZ-7403.10 naturally aspirated 3.75 5.45 6.85 -
engine without catalyst
KAMAZ-7403.10 naturally aspirated 0.2 0.56 2.62 -
engine with catalyst
KAMAZ-7409 turbocharged engine 0.485 0.545 1.741 -
with catalyst
150
Figure 4 Speed performance of the base diesel engine and two gas engines having different turbochargers
Operation of the gas engine on a lean mixture with gas-air ratio 1.55-1.7 ensured the minimal emissions of NOx.
Emissions of NOx and CO comply with Euro-5 emission standard and emissions of HC are only by 18% higher than
the value set by Euro-5.
The naturally aspirated KAMAZ gas engine was mounted on the А-4216/17 bus which was successfully operated for
three years in the town of Nevinnomyssk, Stavropolski region, without any damages. The average gas consumption
was 48 m3/100 km. The expenses for the diesel engine conversion were paid back within 8 months [6].
CONVERSION OF CUMMINS KAMA DIESEL ENGINE TO GAS DIESEL

ENGINE
Cummins KAMA diesel engine was converted to operate on natural gas by the gas-diesel cycle [9]. The gas-feed
and electronic engine control systems were developed within the framework of the State program of creation of high-
and medium-speed engines fed with natural gas. Therefore both the systems were designed in such a manner that
they could be mounted on high- and medium-speed engines with minor modifications.
The gas feed system has a modular architecture. Each module (Figure 5) ensures pressure reduction and supply of
natural gas. This enables to use a different number of modules depending on the engine size. One module is intended
for the high-speed Cummins KAMA gas diesel engine and three modules – for the medium-speed D200 gas diesel
engine. The gas feed system supplies natural gas with a working pressure of 1 MPa to the compressor inlet. It has
metering valves with electronic control. When three modules are mounted on the D200 engine, three gas supply
valves for each cylinder are used: two small valves of the Cummins KAMA engine for injection of small portions of
gas at idle and one large valve – at high loads.
An electronic engine control system for 6-cylinder gas diesel engines was developed which controls supply of natural
gas and diesel fuel [9, 12]. The system generates electric control impulses to control actuators and carries out
synchronization and distribution of impulses by the cylinders depending on the engine operation mode on the basis
of information received from many sensors.
151
Figure 5 One module of gas supply system for the gas diesel engine:
1 – main high pressure solenoid valve; 2 – two-stage gas pressure reducer; 3 – pressure and temperature
sensors in the reducer; 4 – cooling agent controller; 5 – gas pressure sensor; 6 – gas supply valves; 7 – gas
receiver; 8 – gas temperature sensor
A high pressure Common Rail diesel fuel supply system of the base diesel engine was used. The new electronic
engine control system after its thorough calibration ensured injection of small ignition portions of diesel fuel capable
to ignite reliably the gas.
After experimental perfection of the modular gas feed and electronic engine control systems on the Cummins KAMA
gas diesel engine, pretty high engine parameters were obtained (Table 3). The maximal effective efficiency ηe=0.46
was the same as on the base diesel engine.
Table 3 Comparison of ecological parameters of diesel and gas diesel versions of Cummins KAMA engine
Engine Engine Diesel fuel CO2 emissions (%) NOx emissions (ppm)
speed torque portion Diesel Gas diesel Diesel Gas diesel
(rpm) (Nm) (%) engine engine engine engine
1220 660 4.5 9.0 7.5 2100 1570
280 8.8 4.9 3.2 1260 162
1420 840 6.2 8.8 7.7 1123 949
285 8.9 4.8 2.8 766 207
1625 940 6.0 8.5 7.6 1030 504
260 8.7 4.6 3.9 540 51
Stable engine operation at a pretty small portion of igniting diesel fuel was attained: 4.5-6.2% at full load, 8.7-8.9%
at 28-42% load and 33% at idle. On the average, emissions of NOx decreased 1.52 times and of CO2 – 1.18 times.
DEVELOPMENT OF FUEL SUPPLY SYSTEM FOR THE MEDIUM-SPEED D200

GAS DIESEL ENGINE AND FORECASTING OF ITS PARAMETERS
Common Rail fuel supply system for the perspective medium-speed D200 diesel/gas diesel engine was developed
jointly with the industrial partner Noginsk Factory of Fuel Systems OAO (NZTA) [9, 10]. The CR system includes a
high pressure fuel pump with six plunger sections located in radial direction by pairs along the circle over 120o. The
plungers located in one row operate in reversed phase – delivery cycles take place every 180o of the crankshaft
rotation. In this case, the HP fuel pump cycles of fuel delivery into the high pressure line occur every 60°. Such a
sequence of working strokes of the plungers assures a low (compared with the in-line HP fuel pump) and uniform
load on a drive camshaft and reduction of power consumption. The high pressure common rail is integrated into the
injector body to smooth pressure oscillations which originate due to a fluctuating fuel supply from the HP pump, as
well as due to operation of the injectors during injection process. As mass production of the D200 diesel engine has
not yet started, parameters of its gas diesel version with the systems developed were forecasted using the MADI
simulation model.
Simulation model
A one-zone simulation model of diesel/gas diesel/gas engine was developed which calculates the 4-stroke engine
cycle and joint engine/turbocharger operation [4, 12]. Combustion is modeled by the I.Vieve formula and heat losses
152
– by the empirical formula of G.Woschni. Gas exchange processes are calculated using quasi-stationary method.
Simulation model was used for the following:
 Simulation of gas engines and gas diesel engines parameters required for the developement of gas-feed and
engine control systems.
 Analysis of parameters obtained during engine tests
 Forecasting of parameters of the medium-speed D200 gas diesel engine with the modular gas-feed,
electronic engine control and fuel supply systems developed.
It was demonstrated in [14] that if engines having different size (high- and medium-speed gas diesel engines) use
the same method of their working process organization (including the same modular gas feed system and electronic
engine control system), in certain comparison conditions: equal mean piston speed, air-access coefficient and brake
mean effective pressure, their indicated parameters are pretty close. This enables to carry out the primary
experimental perfection of the modular gas feed and electronic engine control systems for medium-speed engines
on high-speed engines. This saves considerable time and money.
The locomotive performance of the medium-speed gas diesel engine D200 was calculated using 5% of the ignition
portion of diesel fuel at all the points [13]. Turbocharger and intercooler of the base diesel engine were used which
ensured a pretty high boost pressure 0.31 MPa and low boost air temperature 335 К at the rated mode. This resulted
in a high air excess coefficient 2.22, ensuring good fuel efficiency, low emissions and thermal strength. At the rated
mode, the required brake mean effective pressure pe=2.0 MPa was attained with a safe value of maximal combustion
pressure 18.5 MPa (maximal is 22 MPa for this engine). At engine speeds 1000, 750 and 500 rpm, high values of рe
were obtained: 2.0, 1.7 and 1.0 MPa ensuring good traction of the locomotive. The highest effective efficiency 0.450
was obtained at the rated mode corresponding to low brake specific fuel consumption 164 g/kWh.
CONCLUSIONS
1. A naturally aspirated KAMAZ diesel engine was converted for operation on natural gas by gas cycle with
stochiometric gas-air mixture and was successfully operated on a city bus during 3 years.
2. A turbocharged KAMAZ diesel engine was converted to operate by gas cycle with a lean gas-air mixture. Using
turbochargers with smaller tubine flow area enabled to get engine power and torque not less than of the base
diesel engine. Special catalysts decreased emissions considerably. Emissions of CO and NOx met the
requirements of Euro-5 ecological standard.
3. Cummins KAMA diesel engine was converted to operate by gas diesel cycle. Modular gas feed and engine
electronic control systems were developed which can be used both on high- and medium speed-engines. After
experimental perfection of the systems, Cummins KAMA gas diesel engine showed the same effective efficiency
as the base diesel engine, high degree of substitution of diesel fuel by gas (from 5.6% to 33.0% of diesel fuel
as the load decreased from full value to idle), considerable decrease of toxic emissions: NOx – 1.52 times and
CO2 – 1.18 times on the average.
4. Diesel fuel supply system for the perspective medium-speed D200 diesel/gas diesel engine was developed
jointly with the industrial partner. Parameters of the D200 gas diesel with the systems developed in MADI were
forecasted using the MADI one-zone model of diesel/gas diesel/gas engine.
ACKNOWLEDGMENTS
Applied research and experimental developments of the diesel fuel equipment were carried out with financial support
of the state represented by the Ministry of Education and Science of the Russian Federation under the Agreement
No 14.580.21.0002 of 27.07.2015, the Unique Identifier PNIER: RFMEFI58015X0002.
REFERENCES
[1] Grehov, L.V., Ivaschenko, N.A., Markov, V.A.: “On Ways to Improve the Gas Diesel Cycle”,
AvtoGasoZapravochniy komplex + Alternatvnoye toplivo, Vol. 100, No. 7, 2010, pp 10-14,
[2] Grotz, M., Böwing, R., Lang, J., Thalhauser, J., Christiner, P., Wimmer, A.: “Efficiency Increase of a High
Performance Gas Engine for Distributed Power Generation”, 6th CIMAC Cascades. Dual fuel and gas engines
– Their Impact on Application, Design and Components, 2015,
[3] Ivković, I.S., Kaplanović, S.M., Milovanović, B.M.: “Influence of Road and Traffic Conditions on Fuel
Consumption and Fuel Cost for Different Bus Technologies”, Thermal Science, Vol. 21, No 1B, 2017, pp 693-
706,
[4] Khatchijan, A.S., Sinyavskiy, V.V., Shishlov, I.G., Karpov, D.M., “Modeling of Parameters and Characteristics of
Natural Gas Powered Engines”, Transport na Alternativnom Toplive, Vol. 15, No. 3, 2010, pp 14-19,
153
[5] Khatchiyan, A.S., Kuznetzov, V.Ye, Shishlov, I.G: “Conversion of domestic diesel engins being in operation to
be fed by naturtal gas - a rational way of improving their ecological characteristics”, Avtozapravochni complex +
Alternativnoye toplivo, Vol. 44, No 2, 2009, pp 34-40,
[6] Khatchiyan, A.S., Kuznetzov, V.Ye, Vodejko, V.F., Shishlov, I.G.: “Results of development of gas engines in
MADI (GTU)”, Avtozapravochni complex + Alternativnoye toplivo, Vol. 21, No 3, 2005, pp 37-41,
[7] Klausner, J., Lang J., Trapp, C.: “J624 – Der weltweit erste Gasmotor mit zweistufiger Aufladung, MTZ –
Motortechnische Zeitschrift Ausgabe, No. 04, 2011,
[8] Luksho, V.A.: “A Complex Method of Increasing Energy Efficiency of Gas Engines with High Compression Ratio
and Shortened Intake and Exhaust Strokes”, Ph.D. thesis, NAMI, Moscow, 2015, p 365,
[9] Shatrov M.G., Sinyavski V.V., Dunin A.Yu., Shishlov I.G., Vakulenko A.V.: ”Method of conversion of high- and
middle-speed diesel engines into gas diesel engines”, Facta Universitatis, Series: Mechanical Engineering, Vol.
15, No 3, 2017, pp 383-395,
[10] Shatrov, M.G., Golubkov, L.N., Dunin, A.U., Dushkin, P.V., Yakovenko, A.L. “The new generation of common
rail fuel injection system for Russian locomotive diesel engines”, Pollution Research, Vol. 36, No. 3, 2017, pp
678-684,
[11] Shatrov, M.G., Khatchijan, A.S., Shishlov, I.G., Vakulenko, A.V.: “Analysis of Conversion Methods of Automotive
Diesel Engines to be Powered with Natural Gas”, Transport na Alternativnom Toplive, Vol. 34, No. 4, 2008, pp
29-32,
[12] Shatrov, M.G., Sinyavski, V.V., Dunin, A.Yu., Shishlov, I.G., Vakulenko, A.V., Yakovenko, A.L.: “Using simulation
for development of the systems of automobile gas diesel engine and its operation control”, International Journal
of Engineering & Technology, Vol. 7, No 2.28, 2018, pp 288-295,
[13] Shatrov, M.G., Sinyavski, V.V., Shishlov, I.G., Vakulenko A.V.: “Forecasting of parameters of boosted locomotive
diesel engine fed by natural gas”, Naukograd Nauka Proizvodstvo Obschestvo, Vol. 4, No 2, 2015, pp 26-31,
[14] Sinyavski, V.V., Alekseev, I.V., Ivanov, I.Ye., Bogdanov, S.N., Trofimenko Yu.V.: “Physical simulation of high-
and medium-speed engines powered by natural gas”, Pollution Research, No. 3, Vol. 36, 2017, pp 684-690.
154
Kragujevac, Serbia
MVM2018-014
Hristo Stanchev1
Krasimir Markov2
INVESTIGATION OF DIESEL ENGINE OPERATED WITH

DIESEL FUEL-BUTANOL BLENDS AND CETANE
IMPROVER
ABSTRACT: Experimental investigation of a common rail direct injection diesel engine is carried out, when
working with diesel fuel in blends with 10% (B10), 15% (B15) and 20% (B20) n- butanol at different loads and
speeds. The cetane number of the blends is lower than basic fuel and is corrected to these of the neat diesel fuel.
The differences in brake specific fuel consumption (BSFC) and exhaust emissions are compared with the baseline
operations ( when working with the neat diesel fuel). The cetane improver 2-EthylHexil Nitrate is used. Six types of
fuels are investigated: B10 and B10 with cetane improver, B15 and B15 with cetane improver, B20 and B20 with
cetane improver. This approach shows that the BSFC for fuel blends with corrected cetane number is closer to this
with neat diesel fuel, especially at higher engine speeds. The exhaust smoke (D) for blends with corrected cetane
number is lower than the neat diesel fuel and blends without cetane improver. There is no clear trend for NOx
emission.
KEYWORDS: diesel fuel, n-butanol, fuel blends, cetane number, cetane improver, emission
INTRODUCTION
The high emission standards for greenhouse gases exist using of different type of high alcohols as an alternative
fuels (blends with standard fuel). The new low cost technology for butanol production and its very close characteristics
to fossil fuels lead to active scientific work all around the world for replacement of part of these fuels with butanol.
There are lot hurdles to be overtaken to use high percentage of butanol in blends with diesel fuel. Butanol has lower
cetane number which resulted in different type of combustion process.
The main advantage of butanol as a fuel for internal combustion engines are higher flash point (safer for distribution
and storage), less corrosive influence on fuel system, less evaporative emission, lower emission of smoke, CO and
HC due to high oxygen content and higher laminar flame speed resulted in enhance reactions. There is no clear
tendency in NOx emission changes because of concurrent between lower temperature and higher oxygen content,
but NOx emission can be effectively reduced due to high EGR rate without increasing in smoke emission [1, 3, 5, 6].
Using butanol as a fuel for diesel engines leads to extended ignition delay period and this increases heat release
quantity. As a result the specific effective fuel consumption increases with higher percentage of butanol. The
temperature of exhaust gases is lower and effective efficiency is higher [2, 4].
1
Hristo Stanchev, Prof. PhD, University of Ruse, 8 Studentska St, 7017 Ruse, Bulgaria, hstanchev@uni-ruse.bg
2
Krasimir Markov, Ass. Prof. PhD, University of Ruse, 8 Studentska St, 7017 Ruse, Bulgaria,
krmarkov@uni-ruse.bg
155
EXPERIMENTAL RESULTS
Test rig description
An engine Fiat 1.9 JTD has been tested in laboratory condition on DC dynamometer. The fuel consumption is
determined by mas flow method. For evaluation of emission components a 5-components gas analyzer AutoLogic is
used. For more precise NOx concentration measurement special analyzer ECM NOx 5210 (g) is used. The exhaust
smoke is measured by OPA-101 Nextech diesel smoke opacity meter. The technical specifications of equipment is
given in next tables:
Table 1 AutoLogic 5-components gas analyzer

Component Range Accuracy (rezolution)
HC 0-2000 ppm 1 ppm
CO 0-15% 0.001 vol.%
CO2 0-20% 0.01 vol.%
O2 0-25% 0.01 vol.%
NOx 0-5000 ppm 1 ppm
A/F ratio 0-72 0.01
Table 2 ECM NOx 5210(g) analyzer

Component Range
NOx 0 - 8000 ppm
λ 0.4 - 25
AFR 6.0 - 364
Ф 0.04 – 2.5
O2 0.0 – 25 Vol.%
FAR x 10 000 27 - 1667
The data from ECM gas analyzer are compensated depend on the pressure in place of NOx /O2 sensor.
Table 3 OPA – 101 smoke opacity meter

Parameter Range (accuracy)
Opacity 0,0 – 100% (0,01%)
Absorbtion coefficient K 0,00– 21,42 m-1 (0,01 m-1)
Engine oil temperature 0 – 120 oС
Engine speed 0 – 5000 min-1
Experimental study has been carried out to investigate the impact of the different butanol-diesel blends on the
engine performance and emission levels under a variety of operating conditions (load and speed). The cetane
number of fuel mixture is corrected to these of neat diesel fuel – 51 points.
For evaluation of influence of the lower cetane number of butanol-diesel blends on engine performance were
conducted experiments with different fuel blends without cetane number corrections first. In references there is not
exact cetane number of n-butanol. The data are as between 17 and 26 points. The cetane number of the butanol has
been assumed 25 points in our case. This cetane number is used for calculation of quantity of cetane improver to
be added to different fuel blends.
Cetane improver 2-EthylHexil Nitrate, produced by “Very One” company is used. According to the technical
specification of this product 100 ppm volume concentration increases cetane number with 0,28 points or 400 ppm
approximately for 1 point increasing. Taking this into account, for the three fuel blend with 10, 15 and 20% butanol,
the quantity of cetane improver for a liter fuel is given in Table 4.
Table 4 Quantity of cetane improver for reaching of fuel blends cetane number 51 points
% butanol Cetane number (CN) Difference to 51 Quantity
of the blend points 2-Ethylhexil Nitrate, ml
10 48.4 2.6 1.04
15 47.1 3.9 1.56
20 45.8 5.2 2.08
156
The results from investigation are shown in graphical form. For comparison at every figure for certain concentration
of butanol are shown the results, obtained for neat diesel fuel, the diesel-butanol blend and the diesel-butanol-cetane
improver blend.
Analysis of the results

At engine speeds 1500 min-1 and 2000 min-1 cetane number of fuel does not have significant influence on the
combustion process and consequently on the power and brake specific fuel consumption (BSFC) for all types of
blends. It is observed increase in NOx emissions and decrease in smoke emissions. At engine speed 2500 min-1
(Figure 1 – 3) the correction of fuel blends cetane number has visible effect on the BSFC in comparison with the
operation of the engine with fuel blends without cetane number correction. It is visible from the results that the BSFC
does not differ from the engine operation with neat diesel fuel.
Figure 1 Change of BSFC (ge) at neat diesel fuel (B0), addition of 10% butanol (B10) and B10 with cetane
improver (B10K)
Figure 2 Change of BSFC at neat diesel fuel (B0), addition of 15% butanol (B15) and B15 with cetane improver
(B15K)
From the obtained result it can be concluded that with addition up to 20% butanol to diesel fuel is possible to keep
engine performance the same as with neat diesel fuel, without any modifications in exist injection program. Any
modification requires reprogramming of electronic control unit on the base of experimental defined optimal injection
program (pilot injection quantity and timing, main injection timing and late injection timing) at operation with diesel-
butanol blends.
157
As is mentioned above, addition of butanol to diesel fuel has contradiction influence over the NOx emissions. Lower
cetane number of diesel-butanol blends leads to extended ignition delay and increased in-cylinder temperature and
accordingly higher NOx emissions. Higher content of oxygen in butanol has the same influence. Conversely, higher
boiling point of butanol reduces the flame temperature in the combustion chamber, which leads to decrease in NOx
emission. This effect is more crearly visible at low engine loads when the heat for evaporation exceeds the
combustion temperature. Also, the presence of 2-ethylhexyl nitrate cetane improver introduces a new fuel-borne NOx
formation mechanism to the combustion process, which increase NOx emissions. The increase in NOx emission is
formed by the decomposition of the cetan improver.
Reduction of smoke emissions, when using diesel-butanol blends, is due to increased content of oxygen in fuel
blends.
Figure 3 Change of BSFC at neat diesel fuel (B0), addition of 20% butanol (B20) and B20 with cetane improver
(B20K)
Results for smoke (D) emissions at all investigated working regimes and fuel blends are shown in Figure 4 to Figure
12. There is not clear trend at different operation conditions. The main reason for this is different influence of ignition
delay period at different engine speeds with correction of blend’s cetane number. At higher loads, when cetane
improver is added, the smoke emission is close to that without cetane improver. This effect reduces with increasing
of butanol percentage in blends. For full estimation of the effect of fuel cetane number on the combustion process
must be proceed in-cylinder pressure measurement (indicator diagram).
Figure 4 Change of smoke (D) at neat diesel fuel (B0), addition of 10% butanol (B10) and B10 with cetane
improver (B10K)
158
improver (B15K)
D, % n = 1500 rpm
30 25.2
23.3 23
25
20
15
9.4
10 5 4 6.2
3.8 2.2 3 2.4 4.2 4 3.4
5 1.4
0
3.4 6.5 9.7 12.5 16.2
Ne, kW
B0 B20 B20К
improver (B20K)
D, % n=2000 rpm
20 16.7 17.2
12.4 13.7
15
9 8.7 8.5
10 6.5 6.7 6.7 7.3 7
5.6
3.6
5 1.8
0
4.7 8.4 12.5 17.3 21
Ne, kW
B0 B10 B10К
improver (B10K)
159
D, % n = 2000 rpm
20 17.2
16.7
15 13.7
12.4 11.7 12.5
9 10
10 8.4
6.5 7.5 5.7
7.2 7.4
5 2.5
0
4.7 8.4 12.5 17.3 21
Ne, kW
B0 B15 B15K
improver (B15K)
D, % n=2000 rpm
20 16.7 17.2
15 13.7
12.4
10
6.5 5.5 5.6 6.1
4.2 3.6 4
5 2.2 2.8
1.2 2
0
4.7 8.4 12.5 17.3 21
Ne, kW
B0 B20 B20K
improver (B20K)
D, % n=2500 rpm
20
15
15 12.1 12 12.5
10.3 10.7
8.1 7.4 9.2 9.1
10 6.6 7.3
5.2 5.9
3.8
5
0
5 11 15.4 22 27
Ne, kW
B0 B10 B10K
improver (B10K)
160
D, % n=2500 rpm
20
15
15 12.1 12 12.5
9.4
10 8.1 7.6
5
5 3 2.7 3.3 4.2 4
1.2 1.5
0
5 11 15.4 22 27
Ne, kW
B0 B15 B15K
improver (B15K)
improver (B20K)
As an example, results for NOx emissions at engine speed n=2500 min-1 and for all fuel blends are shown in Figure
13 to Figure 15. It is evident, that there is not clear trend in NOx concentration at different load conditions and fuel
blends.
NOx, ppm n=2500 rpm

600
500
NOx, ppm
400
300
200
100
0
5 11 15.4 22 27
Ne, kW
B0 B10 B10K
Figure 13 Change of NOx concentration at neat diesel fuel (B0), addition of 10% butanol (B10) and B10 with
cetane improver (B10K)
161
NOx, ppm n=2500 rpm
600
500
NOx, ppm
400
300
200
100
0
5 11 15.4 22 27
Ne, kW
B0 B15 B15K
NOx, ppm n=2500 rpm

600
500
400
NOx, ppm
300
200
100
0
5 11 15.4 22 27
Ne, kW
B0 B20 B20K
CONCLUSIONS
1. Brake specific fuel consumption, when engine work with addition of cetane improver is almost equal to this with
neat diesel fuel. The effect is better at higher engine speed due to stronger influence of fuel cetane number on
the combustion process at these speeds.
2. It is evident from experiment that fuel (fuel blends) cetane number has different influence on smoke emission
at different engine speed. It is observed saving or slightly more or less smoke emissions, compare with engine
operation without cetane improver.
3. There is no need of engine control unit reprogramming when up to 20% n-butanol is added to diesel fuel and
fuel blends cetane number is corrected.
REFERENCES
[1] Algayyim, S., Wandel, A., Yusaf, T., Hamawand, I.: “The impact of n-butanol and iso-butanol as components of
butanol-acetone (BA) mixture-diesel blend on spray, combustion characteristics, engine performance and
emission in direct injection diesel engine”, Energy, Vol. 140, Part 1, 2017, pp 1074-1086,
162
[2] Chen, Z., Liu, J., Han, Z., Du, B., Liu, Y., Lee, C.: “Study on performance and emissions of a passenger-car
diesel engine fueled with butanol-diesel blends”, Energy, Vol. 55, 2013, pp 638-646,
[3] Dobre, A., Pană, C., Nuţu, N., Cernat, Al.; Bondoc, I. D.: “Bio-butanol-alternative fuel for Diesel engine”,
INMATEH-Agricultural Engineering, Vol. 42, Issue 1, pp 145-152, 2014,
[4] Giakoumis, G., Rakopoulos, C., Dimaratos, A., Rakopoulos, D.: “Exhaust emissions with ethanol or n-butanol
diesel fuel blends during transient operation”, Renewable and Sustainable Energy Reviews, Vol. 17, 2013, pp
170-190,
[5] Karabektas, M., Hosoz, M.: “Performance and emission characteristics of a diesel engine using isobutanol–
diesel fuel blends”, Renewable Energy, Vol. 34, Issue 6, 2009, pp 1554-1559,
[6] Rajesh Kumar, B., Saravanan S., Rana D., Nagendran A.: “A comparative analysis on combustion and emissions
of some next generation higher-alcohol/diesel blends in a direct-injection diesel engine”, Energy Conversion and
Management, Vol. 119, 2016, pp 246-256.
163
164
Kragujevac, Serbia
MVM2018-016
Igor V. Alekseev1
Mikhail G. Shatrov2
Andrey L. Yakovenko3
METHOD AND SOME RESULTS OF INVESTIGATION OF

STRUCTURE-BORNE NOISE OF DIESEL ENGINE AT
ACCELERATION MODE
ABSTRACT: A vehicle engine operates most time at transient modes. Therefore the problem of engine noise
investigation in real-life operation conditions is important. The paper presents a method and results of calculation of
the engine structure-borne noise at transient operation mode using the components of the IC Engines Single
Informational Environment. Investigation of the noise of the V8 120/120 mm diesel engine at the acceleration mode
was carried out. The transient mode was presented as an ensemble of steady modes for which the working cycle
was calculated and noise calculation was carried out successively. Finally the engine acceleration performance
was formed. Analysis of the investigation results demonstrated that at the acceleration mode, the diesel engine
noise was by 1.5-2.0 dBA higher than at the similar engine speeds at stationary conditions. It was determined that
the basic factors which cause the difference in the diesel engine acoustic performances at the acceleration mode
were the injection start angle, ignition delay period, thermal inertia of engine parts and some others. Consequently
the pressure rise rate in the cylinder increases which causes the engine noise growth. The approach presented in
the paper enables to forecast the engine acoustic performance during its development and perfection.
KEYWORDS: structure-borne noise, diesel engine, transient mode, acceleration mode, sound power
INTRODUCTION
The relevance of reducing the noise of road transport increases with the growth of the engine power augmentation.
Operation mode exerts a great influence on the noise level of the engine. In urban traffic, the vehicle is moving at
variable speed a significant part of the time. Therefore, the search for effective solutions to reduce engine noise at
a transient mode, especially during acceleration, requires the development of appropriate modelling techniques.
It should be noted that the measurement of the external noise level of the vehicle according to GOST 41.51-2004 is
carried out on a section of the road by a special method at the acceleration mode.
THE MODEL OF FORMATION OF STRUCTURE-BORNE NOISE OF THE

INTERNAL COMBUSTION ENGINE
With appropriate design solutions that provide a specified level of aerodynamic noise, the main noise in the internal
combustion engine (ICE) is its structure-borne noise.
1
Igor V. Alekseev, professor, MADI, 64, Leningradsky Prosp., Moscow, 125319, Russia, igor_alexeev@umail.ru
2
Mikhail G. Shatrov, professor, MADI, 64, Leningradsky Prosp., Moscow, 125319, Russia, dvs@madi.ru
3
Andrey L. Yakovenko, assistant professor, MADI, 64, Leningradsky Prosp., Moscow, 125319, Russia,
iakovenko_home@mail.ru
165
Its level is influenced by excitatory factors: the working process, piston tilting, etc., specific freatures of the design,
as well as its radiating properties.
In accordance with this, the mathematical model of structure-borne noise of ICE, formed at the Department "Heat
Engineering and Automotive Engines" of MADI, proposes the following steps for its implementation.
Engine working process modelling

The model used to calculate the working cycle of an internal combustion engine takes into account the actual
chemical composition of the working fluid, heat exchange processes in the cylinder, dependence of the heat
capacity of the working medium on temperature, heat release characteristics, parameters of the timing mechanism
(diameter and number of valves, their flow sections, valve timing), calculation of gas exchange processes. Heat
release characteristic is calculated according to the method of I. Viebe.
It does not work out in detail oscillatory processes in the manifold and the mixing process. Surface temperatures in
the cylinder depend only on the operation mode of the engine, the size and material of the parts. For calculation of
heat-transfer processes, the G. Voschni model [1] which has proved many times its efficiency is used.
Engine structure modelling

When calculating structure-borne noise, a number of mass-geometric parameters of the structure is used: weight,
length and area of the external surfaces of the engine. Simplified analytical models, as well as three-dimensional
models of the engine structure, can be used to calculate these parameters depending on the design stage [2].
The use of three-dimensional modeling can improve the quality of information about the engine, visualize its design
and individual parts. The use of parameterization in the development of models makes it possible to change rapidly
the design and evaluate the results of changes.
In the model of structure-borne noise, the engine emission, the properties of its design are integrated into an
equivalent cylindrical shell, for which the oscillatory characteristics are known. Equivalence conditions are: equality
of mass, length and area of the outer surface of the engine and of such a shell [3].
Simulation of engine structure-borne noise from individual sources

To simulate the structure-borne noise of the engine, a physical model is proposed, according to which the sound
power is emitted by its oscillating outer surfaces under the influence of exciting force factors [3].
Calculation of sound power is carried out by the formula
PW (kf0 )  zS (kf0 )  c  SICE v e2(S ) (kf0 ) (1)
zS(kf0) - normalized by the area of the outer surfaces of the engine SICE relative coefficient of resistance to radiation;
 - air density; c - speed of sound in the air; с - wave resistance of the air; SICE - area of the outer surfaces of the
engine; v e2(S ) (kf0 ) - the average squared effective vibration velocity on the outer surface; k - counting number of
harmonics; f0 - crankshaft speed.
The vibration velocity of the outer surface is calculated by the formula
N
1 1
v e2(S ) (kf0 )   G 2 (kf0 ) z (2)
2  MICE  Т V (kf0 )   (kf0 )  (kf0 )
2
k A
G(kf0) - spectral density of the force factor at the frequency of the kth harmonic; T - working cycle follow-up period;
A - number of the lowest harmonic force factor (provided that the first is f0); N - number of the highest harmonic
force factor (provided that the first is f0); MICE – mass of the engine; η(kf0) - inelastic loss ratio; zV(kf0) - input
resistance of engine design.
Equivalent cylindrical shell model is used to calculate the resistance of the engine structure.
It was found that the most active sources of structure-borne noise are the working process and piston tilting. For
each of these sources, mathematical model was developed to calculate their force factor and its spectral density.
166
The presented technique was tested experimentally when assessing the noise level of the V8 120/120 mm diesel
by the full-load performance (Figure 1). The discrepancy between the calculated and measured sound power levels
was 0.5 ... 2.0 dB, which is sufficient accuracy for vibroacoustic studies.
Figure 1 Calculated and experimental structure-borne noise levels of the V8 120/120 mm diesel by the full-load
performance
ASPECTS OF MODELLING OF WORKING PROCESS AT THE TRANSIENT

MODE
In case of unsteady operation of the engine and continuous variation of its speed compared to the stationary mode,
the main difference is observed in the organization of the working process. This stipulates the need for changing
the method of modelling of structure-borne noise of the internal combustion engine in these conditions. Taking into
account a high dynamics of the working process of the diesel engine in comparison with the spark ignition engine,
the study of the structure-borne noise of the diesel engine in transient conditions is more relevant.
Analysis of the working process of the diesel engine in transient conditions stipulates the need to take into account
the features of the processes occurring in the cylinder: fuel and air supply, mixing, combustion and heat transfer
with a continuous variation of speed.
The main factors affecting the working process of the diesel engine at the transient mode are:
 the amount of fuel supplied to the cylinder during the ignition delay period;
 fuel injection characteristics;
 thermal state of the engine parts forming the combustion chamber.
When the engine is running at transient mode (TM), characteristics of fuel injection in comparison with the same
speed at the steady-state mode (SSM) change. For different types of fuel systems, this can be caused by different
reasons. In traditional fuel equipment, there are oscillations of the fuel pump rack, and in Common Rail fuel
systems – wave propagation effects in the rail and fuel lines.
With an increase of the engine speed, the delay period of the ignition increases up to the state when the entire
cycle fuel portion is injected during this period, which results in the increase of the rate of pressure growth in the
cylinder and the noise level of the ICE, respectively.
Also at TM, there are problems with supply of air to the cylinder due to its excessive turbulence in the manifold.
This is especially evident in turbocharged diesel engines as the turbocharger has a certain inertia. As a result,
composition of the fuel-air mixture at the transient mode differs from the steady state.
The process of heat transfer to the walls of the combustion chamber also affects the character of the change of
pressure in the cylinder. It depends on the temperatures of surfaces forming the combustion chamber. It was
established in [4] that at the same speeds, the combustion chamber walls temperatures for TM and SSM are
different due to thermal inertia of the engine body parts.
167
FEATURES OF THE METHOD OF MODELLING THE STRUCTURE-BORNE
NOISE OF INTERNAL COMBUSTION ENGINES AT THE TRANSIENT MODE
At the Department "Heat Engineering and Automotive Engines" of MADI, a method of modelling and calculation of
structure-borne noise of the internal combustion engine at the transient mode was developed.
The method includes the following steps (Figure 2):

 Formation of conceptual parameters of the engine. The following input data are used: engine type (diesel
or spark ignition engine), cylinder diameter D, piston stroke S, compression ratio ε, etc.
 Modeling of internal combustion engines. It is performed depending on the degree of study of the engine
design or using analytical dependencies or three-dimensional models. As a result, the required mass-
geometric parameters (mass MICE, the area of the outer surface SICE and the length of the engine LICE, the
material parameters of the structure) are determined which are also transmitted to the subsystem for noise
calculation.
 Calculation of the working cycle taking into account the transient process. In this case, the transient
process itself is quasi-stationary. In this case, the difference between calculation of the cycle from the
traditional stationary mode lies in taking into account the changes of individual factors: fuel supply, thermal
state of the parts forming the combustion chamber. The following input data are assigned depending on
the time of the HP fuel pump rack movement h, engine speed n, temperature t.
For naturally aspirated engines with conventional fuel equipment, appropriate algorithms for calculating the
changes in the specified factors were proposed [4].
Figure 2 The sequence of calculation of structure-borne noise at the transient mode
However, for modern turbocharged diesel engines with Common Rail fuel systems, the methods of calculating the
injection characteristics and the thermal state at the transient mode only have to be formed, they require
168
experimental data on the transient process of the diesel engine. Research in this field is currently performed at the
Department.
Therefore, at present, for a given time interval from nS (tS) to nE (tE) for each selected engine speed (timepoint) of a
quasi-stationary transient mode, calculation of the working cycle is performed and the indicator diagram is
determined. Then the totality of the obtained data is transferred to the ICE CAD subsystem for noise calculation.
 Calculation of spectrum and total sound power levels of the main engine noise sources. At this stage, the
existing model of formation of structure-borne noise from the working process and piston tilting for the
stationary mode is applied, which is implemented in the corresponding ICE CAD subsystem.
First, calculation of the spectrum and overall sound power level for each engine speed comprising the transient
process is carried out. Then, using the dependence of the speed variation on time and the set of spectrum and
noise levels obtained for each mode, a graph of the change in the acoustic characteristics of the engine depending
on time or engine speed at the transient mode is formed.
CALCULATION OF THE STRUCTURAL NOISE OF THE V8 120/120 MM

DIESEL ENGINE AT THE ACCELERATION MODE
Using the developed technique, calculation of the noise level of the V8 120/120 mm diesel engine for acceleration
mode by the full-load performance was made.
Experimentally determined parameters of the engine [4] describing the transient process were used as input data.
With the use of a set of input data, calculation of the working cycle of the diesel engine for a number of modes of its
operation by the full-load preformence was carried out. After that, the noise level of the V8 120/120 mm diesel
engine was estimated at the acceleration mode, the spectrum and the overall sound power levels were determined
(Figure 3).
Figure 3 The results of calculation of the noise level of the V8 120/120 mm diesel engine for acceleration mode by
the full-load performance
169
Analysis of the calculation results showed that the noise level during acceleration by the full-load performance is by
1.5 ... 2.5 dBA higher compared to similar speeds at the steady state.
CONCLUSIONS
The method of modelling of the structure-borne noise of the engine in the transient mode is described.
Specific features of calculation of structure-borne noise emissions of a diesel engine in the transient mode were
worked out in detailes.
Parameters of the transient process of the V8 120/120 mm diesel engine were obtained experimentally and its
noise levels were calculated which showed the divergence of the noise levels for steady-state and transient modes
by 1.5...2.5 dBA.
REFERENCES
[1] Lukanin, V., Shatrov, M., Krichevskaya T., etc.: “Internal combustion engine. In 3 Books. Book 3. Computer
workshop, Modeling of processes in internal combustion engines: Textbook for universities”, 3rd ed., 2007,
Moscow, 414 p.,
[2] Shatrov, M., Yakovenko, A.: “Methods and some results of calculation of the internal combustion engine
structural noise for the formation of components of a single information space of the" ice", Vestnik MADI
(GTU), Vol. 1, No. 16, 2009, Moscow, pp 10-18,
[3] Shatrov M.: “Formation of components of a single information space to ensure the life cycle of internal
combustion engines”, dissertation of doctor of technical Sciences, 2007, MADI (GTU), 403 p,
[4] Gadir, N.: “Analysis of the processes of formation of diesel noise on unsteady modes of operation and
development of measures to reduce it”, dissertation of candidate of technical Sciences, 1991, MADI (GTU),
135 p.
170
Kragujevac, Serbia
MVM2018-017
Branimir Milosavljević1
Radivoje Pešić2
Žarko Đorđević3
Nikola Kostić4
Jelena Erić Obućina5
Saša Babić6
STATISTICAL ANALYSIS OF INFLUENTIAL PARAMETERS

ON THE PROBABILITY THAT CARS MEET THE CO
EMISSION TEST
ABSTRACT: Despite modern technological improvements in the control of traffic emissions from individual
sources and increasingly strict regulation of the regulation, traffic pollution is one of the most important problems of
urban areas. The complexity of its accurate assessment is the product of traffic emissions’ fluctuations due to the
unknown contribution of large emitters and the percentage of cold-started engines that even at the level of one vehicle
emission characteristics would vary over time due to changes in the engine operation and devices for exhaust gases
after treatment. Mentioned changes have become especially important in short-term testing studies because the
empirical data are collected mainly in a limited period of time. In this paper the level of CO emissions at idle speed
was experimentally determined for 1,785 vehicles with ignition engine in order to verify the emission correctness of
a representative sample of the Republic of Serbia vehicle fleet. Permissible emission limit has been considered for 3
fitted binomial logistic regression (BLR) models, and the main reason for such an analysis is to identify those
predictors that may essentially affect the accuracy of assessing whether the vehicle emission is correct or not. Current
emission standards are met by 85% of the vehicles; while with the adoption of more stringent standards only 25% of
the vehicles would be correct. By drawing conclusions on final results of the survey, domestic cars are much bigger
polluters in relation to the foreign ones, although on average, are younger and cross less kilometers. Also, there has
been a noticeable trend that a low power motor vehicles are potentially bigger polluters.
KEYWORDS: vehicle, idle CO emission, binominal logistic regression (BLR), ROC curve
INTRODUCTION
Despite considerable improvements in engine technology, as well as the quality of fuel [1], several studies have
confirmed that traffic emission continues to be a major source of directly emitted pollutants in urban areas [2], [3].
Thus, it has also become the most responsible for acute and chronic diseases of the human body [4]. However, in
addition to health consequences costs of air pollution can be very significant as well. Thus, for example, the Ontario
Medical Association [5] estimates that the cost due to smog cost Ontario more than one milliard dollars a year in
1
Branimir Milosavljević, Ph. D., Prof., College of Tehnical and Mehanical Engineering Trstenik,
lenka1809@gmail.com
2
Radivoje Pešić, Ph. D., Prof., University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić STR., 34000
Kragujevac, Serbia, pesicr@kg.ac.rs
3
Žarko Đorđević, Faculty of Mechanical Engineering in Kosovska Mitrovica, Serbia, zaredj@gmail.com
4
Nikola Kostić, College of Tehnical and Mehanical Engineering Trstenik, kosticn83@gmail.com
5
Jelena Erić Obućina, College of Tehnical and Mehanical Engineering Trstenik, jericobucina@gmail.com
6
Saša Babić, Ph. D., Prof., College of Tehnical and Mehanical Engineering Trstenik, babicsf@nadlanu.com
171
terms of lower productivity, health care costs and premature deaths. In addition, each year 13,000 people visit the
shock room due to the consequences of air pollution, 9800 are hospitalized, while 47 mil. of patient days are
registered. In Ottawa, more than 100 people die annually due to poor air quality, the cost of medical services reach
23 mil. $ per year due to air pollution, and lower productivity costs the city about 29 mil. $.
An emission from motor vehicles varies in different driving conditions [6, 7]. One of the specific regime is idling the
engine and the vehicle in it may spend more than 25% of travel time [8]. The emission at idle speed is considered to
be harmful to human health, because in this case there is no aerodynamic excitation generated by a moving vehicle,
so the concentration of pollutants around the vehicle will be higher.
Testing of emission in the idle speed was originally developed for vehicles that did not have or had very little emission
control, whereby it was possible to detect the engines that are misconfigured or inoperative. Such vehicles typically
have a mechanical carburetor or fuel injection system, where the ratio of air-fuel mixture at idle speed corresponds
to ratio of air-fuel mixture under load. Therefore, measurements of CO and HC emission during idle speed (2,500
rpm) provide a reasonable indication of emission under normal operating conditions for vehicles with mechanical
control system of fuel supply. Such vehicles include cars of older technologies that are used mainly in developing
countries, such as the Republic of Serbia.
In modern vehicles fitted with emission control, a large part of the total emission is caused by a small number of
vehicles with improper emission control system. The research by Guensler [9] indicated that 5% of the vehicle fleet
makes up to 25% of total emissions, 15% causes 43% and 20% is responsible for 60% of emissions. The problem is
further exacerbated by the super emitters (vehicles with a five-fold or higher emission than the prescribed). If we
want to consider emissions by type of harmful substance Faiz et al. [10] have presented that only 20% of vehicles
have HC emissions above 1.2 g km-1, but are therefore responsible for 40% of the total HC emission. Other pollutants
also showed a similar behavioral pattern: 10% of vehicles are responsible for 37% of total emission PM, 20% for 43%
of the total CO emission and 20% of vehicles are responsible for 35% of total NOx emissions.
Numerous inspections in Germany and Sweden, supplement emissions measurement at idle speed with another
measurement that is carried out with the engine running at 2,500 rpm with no load. Austria has adopted a procedure
idle/3000 rpm with a mandatory three-minute conditioning before the measurement [11], while studies in Finland
showed that 95% of the vehicles equipped with three-component catalyst reach a stable reading within 3 minutes.
More than 1/3 of vehicles that do not pass the initial test pass it after conditioning. Since the effective I/M programs
can reduce average emission in the range of 30-50% by identifying high-polluting vehicles, Table 1 shows the
allowable CO emissions limits at idle speed.
Table 1 Global standards for idle CO emission

State Idle CO emission (%vol)
 < 1975: no check
USA  1975-1980 (no catalytic converter): ≤ 2.5
(state Oregon)  1975-1980 (with catalytic converter): ≤ 1.0
 ≥1981: ≤ 1.0
 ≤ 01.07.1995: < 4.5
PR China
 ≥ 01.07.1995: < 4.5
Directive 70/220/EC <3.5
 No catalytic converter: <3.5
 With catalytic converter:
Directive 92/55/EEC
o Low idle : < 0.5
EU o High idle (min. 2000 rpm): < 0.3
 No catalytic converter: <4.5
 With catalytic converter:
Directive 96/96/EC
o Low idle : < 0.5
o High idle (min. 2000 rpm): < 0.3
 < 01.03.2014: < 4.5
 ≥ 01.03.2014:
Serbia
o With mechanical carburetor: ≤ 3.5
(first time
o Injection:
vehicle registrate)
 Low idle : ≤ 0.3
 High idle (min. 2000 rpm): ≤ 0.2
In fact the main reason for this study is the fact that the average age of the vehicle fleet of the Republic of Serbia
(2011) 17.5 years [12]. By identifying the dominant predictor of the impact on the resident CO emissions, as well as
acting on them, a possibility is created of early establishment of stricter emission standards and thus also the better
protection of the environment and human health. The following subsections describe the experiment and the
methodology of results processing.
172
METHODS AND EQUIPMENT OF EMISSION MEASUREMENT
Experimental measurements of CO emissions in passenger cars with ignition engine were implemented in 13
locations of the city of Kragujevac and technical inspection at Auto Moto Club Trstenik during 2008. The methodology
of procedure is harmonized with the following EU directives: Council Directive 92/55/EEC and Council Directive
72/306/EEC of 2 August 1972 Last amended by Directive 91/441/EEC (OJ No L 242, 30 8 1991, p. 1). To measure
the emission AVL DiCom 4000 device is used, that is the combined device for analyzing emissions and diagnostics
of petrol and diesel engines (5 gas analyzer for petrol engines and opacimeter for diesel engines).
METHODOLOGY OF RESULTS PROCESSING
Binominal Logistic Regression

The goal of regression analysis is to find the model that is best adapted to the data, but still acceptable model that
describes the relationship between the dependent and independent variables describing it. The dependent variable
in the logistic regression model is discrete, usually binary (BLR), but it can have more than two categories (MLR).
Independent variables can be categorical or a combination of categorical and continuous, with no assumptions about
the laws of distribution of variables in the BLR. The dependent variable will usually be marked Yi and independent xi.
We can present the probability π of one of the binary events to happen for the case where we have multiple
independent variables, in the following form [13]:
𝑒 𝛽0 +𝛽1 𝑥 1 +⋯+𝛽𝑝 𝑥 𝑝
𝜋 = 𝑃 𝑌𝑖 = 1 𝑋1 = 𝑥1 , … , 𝑋𝑝 = 𝑥𝑝 = (1)
1 + 𝑒 𝛽0 +𝛽1 𝑥 1 +⋯+𝛽𝑝 𝑥 𝑝
Compliance of the model with output data
Assuming that we are satisfied with efforts of developing the model, there are several measures that can give us an
answer as to what extent our model effectively describes the outcome variable (GOF - goodness-of-fit). Is necessary
to emphasize that such measures do not speak of whether the model is better than some other model. Pearson chi-
square statistics of residuals and their deviation may constitute a quality measure of rating compliance of the model
with the output results. However, a problem arises when the J ≈ n (the number of single experiments is equal or
approximately equal to the number of experiments (J - the number of different values for registered x, i.e. the
variable)), because the distribution is n-asymptotic, and thus the number of parameters increases in the same
proportion as the sample size. Thus, the p - values calculated for the above-mentioned measures when J ≈ n and
2
using 𝜒𝑗 −𝑝−1 distribution are incorrect [14].
Since in this experiment J ≈ n that is 1771 ≈ 1785 in the following paragraphs the statistics that avoids the problematic
use of Pearson's chi-square statistics of residuals and their deviation as a measure of compliance of model with
outputs, that is Hosmer and Lemeshow test (HL test) will be briefly explained.
Hosmer-Lemeshow test (HL test)
Hosmer and Lemeshow [15] have proposed two procedures of grouping based on the values of estimated
probabilities by forming a table based on percentiles of estimated probabilities or fixed values of estimated
probabilities. In this paper we will use the first method, using Qs = 10 groups, so that the first group contains
𝑛1, = 𝑛/10 subjects that have the lowest-rated probabilities, while the last group includes 𝑛10
,
= 𝑛/10 subjects with
the highest-rated probabilities. Regardless of the grouping procedure concerned, HL GOF statistics 𝐶 is obtained by
2
calculating the Pearson 𝜒 statistics from the table Qs×2 with observed and evaluated expected frequencies.
Statistics 𝐶 is defined by [16]:
𝑄𝑠 2
𝑜𝑘 − 𝑛𝑘, 𝜋𝑘
𝐶= (2)
𝑛𝑘, 𝜋𝑘 1 − 𝜋𝑘
𝑘=1
where:
𝑛𝑘, - the total number of subjects in the kth group;
𝑜𝑘 - the number of units among 𝑐𝑘 covariant forms:
173
𝑐𝑘 - the number of covariant forms in kth decile;
𝜋𝑘 - average estimated probability in the kth group.
Let 𝑐𝑘 denote the number of covariant forms in kth decile, and then it applies that the number of units among 𝑐𝑘
covariant forms [16]:
𝑐𝑘
𝑜𝑘 = 𝑦𝑗 (3)
𝑗 =1
Hosmer and Lemeshow showed that for J = n (as well as J ≈ n) and when fitted logistic model is the correct model,
2
the distribution of 𝐶 statistics approximates well with the 𝜒𝑄𝑠−2 distribution. Additional research was carried out by
Hosmer et. al. [23] have shown that the method of grouping based on percentiles of estimated probabilities has the
2
advantage over those based on fixed levels of decision-making in terms of better compliance with the 𝜒𝑄𝑠−2
distribution. As the distribution of the statistics 𝐶 depends on m - asymptotic distribution, suitability of p - value
depends on the validity of the assumption that the rated expected frequencies are high. It is believed that the model
2
agrees well with the data if the p - value of the corresponding 𝜒𝑄𝑠−2 statistics is greater than 0.05 [14].
THE RESULTS OF CO EMISSION MEASUREMENT

When setting the binomial logistic model first it is necessary to select those parameters (predictors) that by intuition
or experience of modelers can affect the test results, but also that there are no multicollinearity among them (do not
have a high correlation). The reason is that if we adopt the predictors that are strongly related, they will constantly
interfere in the model that is they will divide the dispersion and produce errors in the output results. Table 2 presents
the contingency table that is the results of the Pearson correlation coefficient (R) of the assumed variables affecting
the results of the aforementioned emission test.
Table 2 Contingency table of presumed predictors that can affect the CO emissions test results
Predictors D L Y ccm kW FI K
D 1 -0.413** 0.074** -0.767** -0.703** 0.590** 0.557**
L 1 0.512** 0.511** 0.451** -0.129** 0.022**
Y 1 -0.049* -0.140* 0.427** 0.519**
ccm 1 0.899** -0.594** -0.461**
kW 1 -0.615** -0.475**
FI 1 0.792**
K 1
Notes: ** p-value ≤0.01; * p-value ≤0.05; D - Territorial affiliation; L - Odometer readings;Y - Vehicle old;
ccm - Engine Displacement; kW - Engine power; FI - System of fuel supply; K - Catalytic converter (CAT)
Contingency table clearly shows highly significant correlation between the engine power and volume. Due to the
smaller standard deviation (SDkw=20; SDccm=305) and standard error (SEkw=0.4; SEccm=8.1) of engine power
compared to its volume, the other enumerated variable is eliminated as a possible predictor of model. The second
reason for the rejection are the conditions imposed by the binomial logistic regression, that is, the number of samples
should not be less than 400 [15]. For the purpose of model accuracy, thus each subcategory of potential independent
variable needs to fulfill the above mentioned, which is not allowed in the division of predictors, “ccm". During the
modeling process such a division could be made which would meet the requirements, but then it would fail to meet
professionally preferred classification. Table 3 presents the adopted predictors of BLR starting model:
Table 3 Adopted predictors for BLR modeling

Symbol of Definition of Type of Dummy
predictor- (xi, i = 1,...,6) predictor predictor coding
0 - foreign a
D – (x1 ) Territorial Affiliation (bi)nominal
1 - domestic
L – (x2 ) Odometer Readings Discrete Scale
Y – (x3 ) Vehicle Old Discrete Scale
0 - (≤ 40 kW) a
Categorical
kW – (x4) Engine Power 1 - (41–70 kW)
interval
2 - (≥ 71 kW)
FI – (x5 ) System of Fuel Supply (bi)nominal 0 - injection a
1 - carburetor
174
Table 3 (continued)
Catalytic converter 0 - with CAT a
K – (x6 ) (bi)nominal
(CAT) 1 - no CAT
Notes: a Reference subcategory in related category
The allowed limit for vehicle passing the emission test has been considered for 3 models, with the engine running at
idle and high number of revolutions (2000 rpm). The first model adopts strictest standard or limit for CO in the state
of Oregon (USA model (≤ 1.0% volCO)). The second model adopts the limit specified in Article 82 of the Rules of the
division and the technical conditions of motor vehicles in traffic of 26.04.2012, specifying that the vehicle first
registered in the Republic of Serbia until March 1st, 2014, must have less than 4.5 vol% CO, at engine speeds at
operating temperature and idling (SRB model). It should be noted that the mentioned provision is less favorable when
compared with international standards (Table 1). However, since the RS is a candidate for membership in the
European Union, the mentioned Rules comply with EU directives for vehicles registered for the first time after March
1st, 2014. That is why the third model is to adopt its regulations (EU model (92/55/ECC: vehicles with carburetor: ≤
3.5 and with injection: ≤ 0.3 vol%CO)).
The main reason for considering the three models is the recognition of predictors that essentially affect the accuracy
of the assessment of whether the vehicle emission is proper. The models are not comparable in results of event’s
probability, because limits of allowed CO emission are differently set. In addition, it is possible that in some of the
models certain predictors are influential, while in other models the same predictors are not influential, and yet the
models still have similar probability values of the vehicle not passing the emission tests. Therefore, it is possible to
compare only the influence of the same predictor in the second model.
The base consists of 1785 vehicles with registered CO emission in the territory of municipalities of Kragujevac and
Trstenik, where vehicles are classified to domestic (754) and foreign (1031). Such divided analysis was not necessary
from the point of logistic regression purpose, because three models do not have independent results by territorial
belonging. However, the results of divided analysis may represent an important complement to the basic conclusions,
or substantial improvement of the analysis, and thus the quality of the recommendations of the domicile measures
to reduce harmful emissions from road transport. Also, this analysis should prevent the generalization of specific
findings, as the domestic vehicles have been until recently uncompetitive on the market compared to most foreign
ones by technologies and devices for controlling harmful emissions of the engine.
Sub category of engine power up to 40 kW is of similar average age in domestic and foreign vehicles, and since the
mentioned category moves from lower to higher, the average age, and reasonably the average mileage is indeed
higher in foreign cars. Foreign vehicles with carburetors are on average older by 3 years, and with the injection
system 9 years compared to domestic vehicles. Also, in foreign vehicles the average mileage is greater by 84%.
Further, 45% of sampled foreign cars do not have a catalyst, and 41% does not have a modern type of fuel supply,
vehicles of foreign manufacturers in the previous analysis may be higher pollutants in relation to domestic.
Domestic vehicles of I subcategory kW (<40 kW) that failed the emission test, on average pass the maximum
kilometers (about 110,000 km) and are the oldest with an average of 20 years. Their average CO emission level is
extremely high (6.03 vol%), however, the result of 7.27 vol% CO of II subcategory (40-70 kW), appears as essential
from the point of the amount of harmful emission, although they are on average 8 years younger and on average
pass 25,000 km less relative to I subcategory of engine power. Analyzing the results of the foreign vehicles group
that did not pass the test, the oldest are from III subcategory of engine power (> 70 kW), with an average of 23 years
and mileage of 270,000 km. However, when reviewing the results of the average CO emission, none of the
subcategories stands out, that is the emissions amount to 6.15, 6.52 and 6.20% volCO, respectively. As previously
analyzed, predictors do not provide sufficiently precise answer to the question which of the vehicles by territorial
affiliation are greater polluters, the dependency of the remaining two variables (x2 and x3) (Figure 1) is represented
by 3D plot diagram, and the results of CO emission are set on the z axis.
From Figure 1 it is clearly seen that foreign vehicles have lower average emission and pass more kilometers, so we
can conclude that the domestic vehicles are greater pollutants in relation to the foreign. This is further confirmed by
the mean value of CO (average Xdomestic=3.6 and average Xforeign=1.99). One of the reasons is that the 97% of sampled
domestic vehicles have a carburetor and no CAT.
Figure 2 shows the number of vehicles that meet emission standards. If we used the SRB model, 85% of vehicles
would meet the required norms. If we adopt the sharpest emission limit (the USA model), the number of emission
correct vehicles drastically decreases to 25% of the sampled vehicles. Thus, by stricktening emission standards by
4.5 times, the number of emission improper vehicles further increases by 60%. The value of the EU model of 70%
should be interpreted with caution because included into the limits to 3.5 vol% CO are also the cars with direct
injection which does not meet the set criteria. Realistically, the EU model meets 44% of sampled vehicles.
175
300 100%
90%
Cimulative frequency [%]

250
85,15% 80%
Frequency [vehicle]
(model SRB)
70%
200
60%
Frequency
150 50%
Cumulative 40%
100 frequency 30%
24,82% 20%
50
(model USA)
10%
0 0%
0,5
3,5
6,5
9,5
1,5
2,5
4,5
5,5
7,5
8,5
More
2
8
0
10,5
11,5
11
10
12
Idle CO emission (vol%)
Figure 1 3D plot diagram Figure 2 Cumulative and absolute frequency of tested

vehicles
The following paragraphs present the methodology of calculation of SRB, USA and the EU BLR models. The final
models having to meet the following requirements:
1. Independent variables have a significant impact on the dependent variable, if their Rao`s score statistic is
2
significant, i.e. when it is tested by 𝜒 distribution with k-1 degree of freedom and at the same time p≤0.05 is
met [17];
2. The model is statistically significant if the difference of the credibility logarithms of the new (L(β)) and the
2
base model (L(0)) is such that p≤0.05 is met when it is tested by 𝜒 distribution with k+1- n degree of freedom
[18], that is it has to be checked whether the model with the sample is good enough to apply to the whole
population (Omnibus test);
3. In assessing the significance of the coefficients of independent predictors using Wald’s test, all have to be
2
significantly different from 0, which is tested on 𝜒 distribution. The condition is fulfilled for p≤0.05 [19], and
2
4. If HL GOF statistics shows good agreement of the model with the data, p-value of proper 𝜒𝑄𝑠−2 statistics
must be > 0.05 [14].
After the selection of a satisfactory combination of the predictors, it is necessary to rank them so that the predictors’
group for which the highest % of correct classification (Hit ratio - HR or Percentage Analyses - PA) [20] and the
𝟐
highest value of Nagelkerke determination coefficient 𝑹𝑵 [21], are the group of greatest impact on the probability
results of Yi that the vehicle failed CO emission test.
The power of the individual impacts of predictors in the models will be evaluated according to the following:
I criteria - The greater Wald’s statistics of predictors than its standard error, the greater the probability that it is
significant according to the impact [22];
II criteria - Farther odds ratio of the independent variable from 1, greater is the predictor’s impact [18], and
III criteria - The area under the curve (AUC) essentially assesses the quality of predictors in the established
model, whose calculation requires the construction of ROC curve [15].
In Table 4 Roa`s score statistics of univariate logistic regression for the three models is presented. Apart from the
2
variable L in the model SRB( 𝜒 2.55; 1 = 0.1111 > 𝑝 = 0,05 ), all other predictors have an impact on model results.
Therefore, the variable L will not be considered as influentially possible in the SRB model.
Table 4 Score statistics for the 3 emissions BLR model

Score statistics (df)
Predictors
SRB MODEL EU model USA model
D (1) 84.14 (1)** 53.79 (1)** 328.79 (1)**
* **
L 2.55 (1) 198.26 (1) 75.54 (1) **
** **
Y 109.38 (1) 55.37 (1) 669.92 (1) **
kW 127.68 (2) ** 98.99 (2) ** 354.11 (2) **
** **
kW (1) 34.79 (1) 47.18 (1) 42.27 (1) **
** **
KW (2) 40.04 (1) 88.31 (1) 178.95 (1) **
** **
FI (1) 154.73 (1) 172.46 (1) 1011.86 (1) **
K (1) 148.62 (1) ** 46.33 (1) ** 1133.36 (1) **
∗∗ p ≤ 0.01 ; ∗ p > 0.05 ; df - degree of freedom
176
After selecting a combination of a set of predictors, classification and ranking according to defined criteria the results
in Table 5 were obtained.
Table 5 Fitted BLR models according to the defined requirements

Model Predictors - 2LL(ϴ) 𝝌𝟐𝒅𝒇=𝒌+𝟏−𝐧 PA [%] 𝑹𝟐𝑵 𝝌𝟐𝑸𝒔−𝟐 (𝒅𝒇)
SRB x3x4x5x6 1194.8 304.6 (5)a 85.2 0.28 11.947(8)b
a
EU x2x4x5x6 1836.8 612.1 (5) 74.6 0.39 11.534(8)b
a
USA x2x3x4x5x6 174.2 1805.8 (6) 97.3 0.95 1.364 (8)b
Notes: p-value ≤0.01; p-value >0.05; df - degree of freedom; 𝑹𝑵 - Nagelkerke
a b 𝟐
determination coefficient; –2LL(ϴ) - log likelihood of the new model
Key predictors of the impact on the event probability of the SRB model that the vehicle fails the emission test are: Y,
kW, FI and K. The probability calculation can be expressed in the following mathematical form:
𝑒 (−7.28+0.077𝒀−0.77𝒌𝑾(𝟏)−1.14𝒌𝑾(𝟐)+1.94𝑭𝑰(𝟏)+2.94𝑲(𝟏))
𝑃𝑆𝑅𝐵 (𝑌 = 1|𝑋𝑖 = 𝑥𝑖 ) = (4)
1 + 𝑒 (−7.28+0.077𝒀−0.77𝒌𝑾(𝟏)−1.14𝒌𝑾(𝟐)+1.94𝑭𝑰(𝟏)+2.94𝑲(𝟏))
Key predictors of the impact on the event probability of the EU model that the vehicle fails the emission test are: L,
kW, FI and K, respectively represented in the following mathematical form:
𝑒 (1.08+0.000017𝑳−1.64𝒌𝑾(𝟏)−2.32𝒌𝑾(𝟐)−7.11𝑭𝑰(𝟏)+4.50𝑲(𝟏))
𝑃𝐸𝑈 (𝑌 = 1|𝑋𝑖 = 𝑥𝑖 ) = (5)
1 + 𝑒 (1.08+0.000017𝑳−1.64𝒌𝑾(𝟏)−2.32𝒌𝑾(𝟐)−7.11𝑭𝑰(𝟏)+4.50𝑲(𝟏))
Key predictors of the impact on the event probability of the USA model that the vehicle fails the emission test are: Y,
L, kW, FI and K, respectively represented in the following mathematical form:
𝑒 (−10.61+0.000045𝑳+0.59𝒀−6.44𝒌𝑾(𝟏)−8.53𝒌𝑾(𝟐)+9.5𝑭𝑰(𝟏)+6.41𝑲(𝟏))
𝑃𝑈𝑆𝐴 (𝑌 = 1|𝑋𝑖 = 𝑥𝑖 ) = (6)
1 + 𝑒 (−10.61+0.000045𝑳+0.59𝒀−6.44𝒌𝑾(𝟏)−8.53𝒌𝑾(𝟐)+9.5𝑭𝑰(𝟏)+6.41𝑲(𝟏))
In Table 6 certain logistic regression parameters are presented that cannot be detected from the mathematical
formulation of the model; however can find its certain features can be essentially found by their interpretation. In the
third column the logistic regression coefficients are given, in the fourth their standard error, while in the last Odds
Ratio (OR) that is coefficient exponent of considered variable.
Table 6 Logistic estimates for CO emission test failure

Wald
Model Predictors 𝜷𝒊 𝐒𝐄𝜷𝒊 𝐞𝜷𝒊
statistics (df)
Y 0.077 0.014 31.08 (1)a 1.080
kW(1) -0.766 0.173 19.580 (1)a 0.465
SRB kW (2) -1.14 0.29 15.453 (1)b 0.319
FI (1) 1.938 0.585 10.961 (1)a 6.944
K(1) 2.943 1.042 7.969 (1)b 18.969
L 0.000017 0.000001 217.803 (1)a 1.000017
kW (1) -1.636 0.157 108.538 (1)a 0.195
EU kW (2) -2.319 0.24 93.342 (1)a 0.098
FI (1) -7.106 1.05 45.829 (1)a 0.00082
K(1) 4.502 1.031 19.067 (1)a 90.181
L 0.000045 0.000006 56.939 (1)a 1.000045
Y 0.597 0.078 59.413 (1)a 1.822
kW (1) -6.436 0.98 43.167 (1)a 0.001603
USA
kW (2) -8.525 1.166 53.485 (1)a 0.000199
FI (1) 9.5 1.366 48.339 (1)a 1,3360.72
K(1) 6.41 1.246 26.483 (1)a 608.141
Notes p-value ≤0.01; p-value ≤0.05; df - degree of freedom; kW(1) - first subcategory in
:a b
category of engine power (41-70 kW); kW(2) - second subcategory in the category of engine
power (>71 kW); FI (1) - first subcategory in category of the system of fuel supply (vehicles
with carburetor); K(1) - first subcategory in the category of the catalytic converter (CAT)
(vehicles without CAT)
Since all predictors of the fitted SRB, EU and the USA models are identified, the significance of individual impacts
remains to be assessed. If in the SRB model the independent variables impact is analyzed through the ratio of Wald’s
statistics and the standard deviation of the variables coefficient, the vehicle age is the greatest. Since it is continuous,
it is necessary to calculate the odds of test failure for 10 or 20 years old vehicle. If the vehicle is older by one year,
177
the odds of failing increase 1.08 times. For 10 years older vehicle, the odds increases by 2.15 times, while for 20
years older vehicle the odds of failing increases by 4.7 times. If we want to find out % of increase in the probability
of not passing the emissions test for 5, 10 or 20 years old car, it is necessary to obtain a new value of odds as a
product of the old odds, the ratio of odds and changes in the values of the input variable. Thus, 5 years old vehicle
has the increase of the probability of test failure for an additional 31%, for 10 years old vehicle by 41%, while for 20
years old vehicle, the probability of failing would increase by an additional 45%.
The exponent of the logistic regression coefficient represents the change in the odds resulting from a unit change in
the predictor variable. Therefore, increasing the age of vehicle for one year, the odds of the vehicle falling the test
increases by 8%, while for the USA model the increase would amount to 82.2%. Practically, by making the emissions
standards 4.5 times stricter, a one-year increase in age affects the odds of emissions test failure by increasing it by
~ 10 times with only one-year increase of age.
Based on the relation of Wald’s statistics and the standard deviation of the coefficient of variables, total vehicle
distance traveled is emerging as the most influential predictor in the models of the EU and the USA. The explanation
that for a 1 km further, the vehicle has a greater chance of emission test failure by 1.000017 times is impractical. It
is more realistic to consider how much would a change of odds amount to for the increased 5,000; 10,000; 50,000 or
100,000 additional distance travelled. Table 7 provides the required answer.
Table 7 Odds failure changings of predictor L

Additional distance EU USA
covered [km] model model
5,000 1.09 1.25
10,000 1.18 1.57
50,000 2.34 9.49
100,000 5.47 90
From the above described, it is possible to deduct that by applying the EU model, which has been in force since 1
March 2014, the sampled vehicles used in Serbia have 5.47 times greater odds of emissions test failure having
covered each 100,000 km, while the adoption of stricter emission norms (USA) would increase the odds of failing by
90 times, or 9.5 times for each 50,000 km covered.
Comparing the first two tests of influence of the predictors in Table 4, we can see that there is a significant divergence
in the results, so it is necessary to consider the third criterion as well. However, before we do that by using ROC
curve, we shall analyze the only predictor of the category with more than two subcategories, the engine power.
Reference subcategory comprises engines up to 40 kW (Table 4). The coefficient exponent of subcategory 1 (0.465)
and subcategory 2 (0.319) in model SRB (Table 6) is the odds ratio of a vehicle failing emissions test compared to
reference subcategory. Since these values are lower than 1, it is necessary to observe the inverse proportional
relationship, i.e. 1 is to be divided by the obtained/resulting coefficient exponent (odds ratio value lower than 1
reduces the probability of the vehicle having failed the test, i.e. it increases the probability of the vehicle having correct
emissions values).
Thus, in the SRB model, the engines in subcategory 1 has 2.15 times greater odds of fulfilling emission standards
than the engines in reference subcategory, while the engines in subcategory 2 have 3.13 times greater odds to fulfill
the emissions standards. Since this predictor is applied in all 3 models, Table 8 shows the analytically summarized
relations. It clearly indicates that the stricter the emission standards are, the greater are the odds of high-power
engines to have correct emission values.
Table 8 Odds relation analysis of emission test pass

Subcategory SRB EU USA
of predictor kW model model model
kW (1) 2.15 5.13 623
kW (2) 3.13 10.21 5,025
Also, another important indicator of the predictor’s characteristics can be a sign in front of its coefficient. In the SRB
model, the sign in front of the coefficients Y, FI and K is positive, which means that older vehicles, with a carburetion
system and without catalyst have higher odds of emission test failure. The same conclusion is valid for the USA
model, except that it is supplemented by the fact that the vehicles have travelled more kilometers. The coefficient FI
in the EU model has a negative sign.
If we analyze the exponents of the mentioned predictors’ coefficients in all models, the absence of a catalyst and
carburetor system of power supply in all models implies greater odds of failing the test, except in the aforementioned
178
EU model for FI, in which vehicles with carburetors have ~ 1,219 times higher odds of passing the test of allowed
emission.
As already noted, disagreement between the first two criteria of the evaluation of predictors’ impact is evident. The
answer can be obtained by using a ROC curve. The area under the curve (AUC) provides a value that can be
compared and placed in defined boundaries of separating quality.
The AUC values for influential predictors of SRB model are: Y (0.702), kW (0.306), FI (0.698), and K (0.691). It is
notable that only the Y predictor has acceptable separation, while FI and K belong to a group that separates poorly.
For this reason, the point of intersection is determined (see Figure 3 a), showing that the SRB model has an
acceptable separation on vehicles which have passed the test as compared to those that have not at the 19 year old
vehicle, and with the probability of such an event of 50 %. The predictor kW is far below the reference curve, but this
can be explained by the above-mentioned variable having a greater impact on the assessment of vehicles that can
pass the test, i.e. that the emission is proper. That's why we can ascertain that in the SRB model the highest quality
variable of accuracy assessment of the output results whether the vehicle failed the emission test is represented by
Y.
The values of area under the curve for the influential predictors of the EU model are: L (0.711), kW (0.568), FI (0.350)
and К (0.424). Predictor L has an acceptable separation, while kW belongs to a group of poor separation. For this
reason, the point of intersection (the so called, “cutpoint”) is determined (Figure 3 b), which clearly shows that the
EU model has acceptable and the most accurate separation of the vehicles that have passed the test compared to
those that have not when they have travelled 110,000 kilometers. The predictors K and FI are below the reference
curve, but this behavior can be explained by the aforementioned independent variables having a greater influence
on the assessment of vehicles that can pass the test, i.e. that they are emission proper. In the conclusion, we can
ascertain that a predictor of vehicle distance travelled is the highest quality variable of output results accuracy
assessment if the vehicle failed the emission test.
Figure 3 c shows the results of USA models and the values of area under the curve for influential predictors of the
U.S. model are: L (0.623), Y (0.890), kW (0.222), FI (0.919), and K (0.936). Considering the AUC value, the predictors
FI and K, have an excellent capability of separation of correct answers relative to the incorrect, while the independent
variable Y, now in the U.S. model and related to SRB model, in the estimate of the correct answers rises by one bar
of separation quality and now belongs to a group that performs the excellent separation. Total distance travelled in
the U.S. vehicle model, does not belong to quality indicators, whereas the predictor of power-supply type better
assesses which of the vehicles passed the CO emissions test. The USA model has the most accurate separation to
the vehicles that have passed the test compared to those that have not passed it at 11 years old vehicle. Thus, we
can conclude that by stricktening of allowed emissions standards, the predictor Y improves its quality in terms of
accuracy of properly graded vehicles that have failed the test. If we finally compare the 3 quality assessment tests,
i.e. the significance of predictors in the model, we can see that the conclusions I and III criteria fully match, which
further reinforces the conclusion that the age of the vehicle is a very influential variable in the extreme emission limits,
while the EU model, the most influential predictor is L.
L
kW L
Y Y
FI
kW kW
K
FI FI
reference line
K K
reference line reference line
Figure 3 ROC curve and “cut-point” of predictors OR and VO in the SRB, EU and USA model
179
Finally, consider the probability results of a vehicle failing the emission test for the reference model SRB should be
considered. Thereat, first shown is Figure 6 with the division on domestic and foreign vehicles for predictors of Y and
W. The main reason for selecting these predictors is that K and FI do not have a great AUC when it comes to vehicles
passing on the emission test, while the predictor L, as already set forth does not participate in the formation of the
SRB model. Presented on the ordinate is the mean probability of failing emission test, the dashed lines represent its
mean value, while the color lines represent belonging to domestic or foreign vehicles.
It is obvious that domestic vehicles have a greater probability of failing the emission test compared to vehicles of
foreign manufacturers, while vehicles of less power, and therefore a smaller volume have a greater probability of
failing the test.
Shown in Figure 4 is a probability diagram of SRB model, and depending on the most influential predictor Y. Also
represented on the ordinate is the mean value of CO emission and its standard deviation with the confidence interval
for the probability of event occurrence in 95% of the population.
Figure 4 Actual CO emission test failure rates by predictor Y and kW in model SRB
From the figure we can conclude that on the sample of the vehicle fleet of the Republic of Serbia there is obviously
a big difference in the probability of the vehicle not passing the emission test for vehicles aged between 16 and 20.
With the averaged probability of failing the test of 9%, the critical vehicle of age 16, the mentioned probability would
increase to 21% in the next four years. The difference of 12% in just four years clearly indicates the intense increase
in the number of emission defective vehicles from 1992.
CONCLUSION
In the conclusions of the experimental studies of emission we can state that under current legislation, i.e. the model
SRB (≤ 4.5 vol% CO), the vehicle age is the most influential factor on the probability of vehicle failing on the CO
emission test. Thereat, 19 age vehicle has the most accurate assessment of whether it will fail the emission test,
while by stricktening the emission criteria by 4.5 times the estimate accuracy of the probability of failing the test
lowers to the vehicle of the age of 11.
Domestic vehicles in relation to the foreign are bigger polluters and have a greater probability of failing the test,
although they are, on average, younger and have lower total distance travelled (97% of sampled domestic vehicles
without a catalytic converter and with a carburetor) With the current emission limit of 4.5 vol% CO, in domestic
vehicles to every 50,000 kilometers travelled the odds of failing the emission test increases 2.3, while in the foreign
it increases 9.5 times. With the strictening of current emission standards by 4.5 times, for 1 year of age difference
the odds of vehicle failing the emission tests increases by 10 times, while the number of vehicles that will meet new
emission standards reduces by 60%.
The vehicle aged 5 years has 31% greater probability of failing the CO emission test related to one-year old vehicle,
while the vehicle aged 20 years has 45% greater probability of failing the test related to the one-year old reference
vehicle. From the sampled 1785 vehicles in operation in the Republic of Serbia, 85% of the vehicles meet the current
emission standard of CO limits, while with the administration of one of the world's most restrictive standard (State of
Oregon, USA), only 25% of vehicles in operation passes the emission test. Meaning that in that case 60% of the
vehicles would be additionally emission incorrect.
180
Due to the great age of the vehicle fleet of the Republic of Serbia (17.5 years) and potentially serious problems with
air quality, it is necessary to adopt stringent emissions test procedures, or import inspection and maintenance
programs of vehicles in operation (I/M programs). It would be ideal if the test procedures would be defined on the
basis of the strong correlation with actual emissions, and that the boundaries of each procedure were determined so
as to take into account the technical capabilities of the vehicle and emission standards. With the level of traffic
emissions monitoring and analysis of technical failure rates standards can be stricktened, because due to advances
in engine and vehicles technology, the technical accuracy improves over time.
Oblige the owners of passenger cars with internal combustion engines that annually perform the procedure of I/M
program when the vehicle turns 3 years, LTV with a diesel engine when the vehicle turns 4 years, and to continue
every second year, while TTV would require testing once a year (virtually compliance with European I/M program).
As an example of comparison, the vehicles in Japan have to undergo testing three years after purchase, and
thereafter every two years. Until recently, in the U.S. most of the I/M programs required testing on an annual basis,
but now the US EPA forces testing every two years, because it is equally effective and the costs are significantly
lower as well.
REFERENCES
[1] Fenger, J.: “Urban air quality”, Atmospheric Environment, Vol. 33, No. 29, 1999, pp 4877-4900,
[2] Jo, W.K., Park, J.H.: “Characteristics of roadside air pollution in Korean metropolitan city (Daegu) over last 5 to
6 years: temporal variations, standard exceedances, and dependence on meteorological conditions”,
Chemosphere, Vol. 59, No. 11, 2005, pp 1557-1573,
[3] Fujita, E.M., Stockwell, W.R., Campbell, D.E., Keislar, R.E., Lawson, D.R.: “Evolution of the magnitude and
spatial extent of the weekend ozone effect in California’s south coast air basin, 1981-2000”, Journal of the Air &
Waste Management Association, Vol. 53, 2003, pp 802-815,
[4] WHO, Guidelines for Air Quality, World Health Organization, Geneva, 2000,
[5] Armstrong, J.: “Development Of A Methodology For Estimating Vehicle Emissions”, PhD Thesis, Department of
Civil and Environmental Engineering Carleton University Ottawa, Ontario, 2000,
[6] Hung, W.T., Cheung, C.S., Tong, H.Y.: “On-road motor vehicle emissions and fuel consumption in urban driving
conditions”, Journal of the Air and Waste Management Association, Vol. 50, 2000, pp 543-554,
[7] Frey, H.C., Unal, A., Rouphail, M., Colyar, J.D.: “On-road measurement of vehicle emissions using a portable
instrument”, Journal of the Air and Waste Management Association, Vol. 53, No. 8, 2003, pp 992-1002,
[8] Tong, H.Y., Hung, W.T., Cheung, C.S.: “Development of a driving cycle for Hong Kong”, Atmospheric
Environment, Vol. 33, 1999, pp 2323-2335,
[9] Guensler, R.: “Loop Holes for Air Pollution”, ITS Review, Vol. 18, No. 1, University of California, 1994,
[10] Faiz, A., Weaver, C.S., Walsh, M.P.: “Air Pollution from Motor Vehicles Standards and Technologies for
Controlling Emissions”, The World Bank Washington, D.C., 1996,
[11] Laurikko, J.: “In-use Vehicle Emissions Control in Finland: Introduction and Practical Experience”, The Science
of the Total Environment, Vol. 103, 1994, pp 557-564,
[12] Аnnual report of the interior ministry of the republic of Serbia, 2011,
[13] Pohar, M., Blas, M., Turk, S.: “Comparison Logistic Reression Models and Linear Discriminant Analysis: A
Simulation Study”, Metodološki zvezki, Vol. 1, No. 4, 2004, pp 143-161,
[14] Kleinbaum G., Klein M.: “Logistic regression: A Self Learning Text - third edition”, Springer, 2010,
[15] Hosmer W.D., Lemeshow S.: “Applied Logistic Regression-second edition”, Wiley Series, 2000,
[16] Stevenson, M.: “An Introduction to Logistic Regression”, EpiCentre, 2008,
[17] Rao, C.R.: “Linear Statistical Inference and its Application, Second Edition”, Wiley Inc., N.Y., 1973,
[18] Tabachnick, G.B., Fidell, S.L.: “Using Multivariate statistics, Pearson Education”, Inc., USA, 2013, ISBN-10: 0-
205-89081-4,
[19] Hosmer, W.D., Lemeshow, S., Sturdivant, X.R.: “Applied Logistic Regression (III edition)”, John Wiley & Sons,
Inc., Hoboken, New Jersey, 2003, ISBN 978-0-470-58247-3,
[20] Ott, L. R., Longnecker, M.: “An Introduction to Statistical Methods and Data Analysis, Sixth Edition”,
Brooks/Cole,Belmont, Canada, 2010, ISBN-13: 978-0-495-01758-5,
[21] Nagelkerke, N.J.D.: “A note on a general definition of the coefficient of determination”, Biometrika, Vol. 78, 1991,
pp 691–692.
[22] Tošković, O.: “Lectures from the subject: Statistics in psychology 2 (theme: Binary logistic regression)”, Faculty
of Philology, Belgrade, 2012,
[23] Hosmer, W.D., Lemeshow, S., Klar, J.: "Goodness-of-Fit Testing for the Logistic Regression Model when the
Estimated Probabilities are Small“, Biometrical Journal, Vol. 30, Issue 8, 1988, pp 911-924.
181
182
Kragujevac, Serbia
MVM2018-023
Almir Blažević1
Dževad Bibić2
Ivan Filipović3
INFLUENCE OF DETERMINING THE PERIOD OF IGNITION

DELAY ON THE HEAT RELEASE CHARACTERISTIC OF
DIESEL ENGINES FUELLED WITH DIESEL AND BIODIESEL
ABSTRACT: Period of ignition delay in diesel engines is a very important parameter of the combustion process
due to its direct impact on the intensity of heat release, as well as its indirect impact on the noise and exhaust
emissions of the engine. Due to the complexity of the self-ignition process of heterogeneous mixtures of fuel and air
in the diesel engine, and because the period of ignition delay dependence of many factors, both physical and
chemical, applied methods of determining the period of ignition delay are very varied and include empirical, semi-
empirical and experimental methods. The influence of determining the period of ignition delay on the heat release
characteristic, respectively the effective power and torque of a diesel engine is discussed in the paper. For the
purpose of the analysis in literature, the known semi-empirical and experimental methods for determining the period
of ignition delay are used. The experimental methods for determining the period of ignition delay, that are used in
this paper, are based on measured values of the IC engine in-cylinder pressures and finding the first or second
derivations of the indicated pressure curves, in other words, the extremes of indicated pressure curves. Also the
change of ignition delay period with change of fuel type (diesel and bio diesel fuel), and the ability of the semi-
empirical expressions for determining the period of ignition delay when using the biodiesel fuel is analysed.
KEYWORDS: bio fuel, IC diesel engine, heat release, period of ignition delay
INTRODUCTION
The internal combustion (IC) engine is a very complex working machine that bounded chemical energy in fuels
transforms through the combustion process into mechanical work. The transformation process is highly complex and
incorporates correlated physical and chemical processes. For the sake of better analysis, the combustion process
can be divided into phases, as well as with different approaches depending what the analysis goals are. The most
common division of a combustion process for a diesel IC engine is: ignition delay, rapid combustion, regulated
combustion and afterburning [1]. In this paper the focus is on the period of ignition delay respecting the way of its
estimation on IC engine performance calculation.
In general, all methods that are into solving this problem can be put into two groups: the experimental methods for
determining the period of ignition delay and the correlation functions for estimating the period of ignition delay.
The experimental methods are based on performing measurements of characteristic parameters on which the start
and ending of the ignition delay period can be determined. A common approach for the diesel engine is to observe
the injection needle lift (as a parameter that defines the start of fuel injection and so the start of ignition delay period),
and further the in-cylinder pressure (change in pressure gradient as a parameter for defining the end of ignition delay
period).
1
Mr. Almir Blažević, Faculty of Mechanical Engineering, Vilsonovo š. 9, 71000 Sarajevo, blazevic@mef.unsa.ba
2
Prof. Dr Dževad Bibić, Faculty of Mechanical Engineering, Vilsonovo š. 9, 71000 Sarajevo, bibic@mef.unsa.ba
3
Prof. Dr Ivan Filipović, Faculty of Mechanical Engineering, Vilsonovo š. 9, 71000 Sarajevo, filipovic@mef.unsa.ba
183
For defining the end of ignition delay period other technics can be used, in ex. detection of luminescent appearances,
what for released IC engines isn’t a quite appropriate method [2]. Use of such methods for determining the ignition
delay period exclude the need of knowing parameters that influence the duration of this period.
The second approach, the use of correlation functions for ignition delay period estimation, demands analyses in detail
of physical and chemical influencing parameters and acceptability verification. On basis of experiments performed
on realised engines the correlation functions are obtained. Their advantages arise in the fact that once a correlation
function for a specific engine is developed and verified it can with care also be used for similar engines.
EXPERIMENTAL SET UP AND BOUNDARY CONDITIONS

To determine the influence of ignition delay period definition on the engine performance calculation a six cylinder,
water cooled diesel engine with M type fuel injection system was observed. The general data on the diesel engine
and the fuel delivery system are given in the Table 1 and in the Table 2, respectively.
Table 1 IC engine data

Engine model MAN
Engine type D2566 MUM
Nominal engine power/engine speed 160 kW/2200 min-1
Max. engine torque/engine speed 775 Nm/1400 min-1
Number of cylinders 6 in line, horizontal set up
Piston diameter 125 mm
Piston stroke 155 mm
Length of the connecting rod 244 mm
Swept volume 11,413 dm3
Intake valve diameter 58 mm
Exhaust valve diameter 51 mm
Intake valve lift 14,3 mm
Exhaust valve lift 14,3 mm
Intake valve opening/closing angle 8° before TDC/40 DEG CA after BDC
Exhaust valve opening/closing angle 50° before BDC/12 DEG CA after TDC
Compression ratio 18:1
Firing order 1-5-3-6-2-4
Static fuel injection angle 23 DEG CA before TDC
Table 2 Fuel injection system data

Make and type of high pressure pump Bosch PES 6A 95 D410 LS2541
Max. revolution 1230 min-1
Nominal cyclic fuel delivery 131+2 mm3/cycle/cylinder
Diameter x stroke of one section piston Φ 9,5 x 8 mm
Pre lift of the pump piston 1,8 mm
Supply pressure 1,5 bar
Make and type of speed governer Bosch RQ 250/1100 AB 965 DL
Make and type of injector Bosch DLLA 25 S 834
Opening pressure 175 bar
Needle lift 0,30 mm
Make and type of relief valve Bosch Atlas 50 mm3
Opening pressure 15 bar
Relief volume 50 mm3
Diameter and length of high pressure tubes Φ6/Φ1,8 x 900 mm
The experiments were performed on an electrical engine test bench at laboratory conditions. Basically the
experiments were performed to get data for a very bright spectre of engine parameters, so that the measuring results
excide the purpose of this paper. The parameters that were measured, among others, are: engine speed, engine
torque, hourly fuel consumption, fuel temperature, ambient air temperature, water temperature at the intake of the
engine, water temperature at the engine block, engine oil temperature, exhaust gas temperature, in cylinder pressure
at the first cylinder, pressure at the engine manifold, pressure drop at the air filter, injector needle lift at the first
cylinder, air consumption, air humidity, CO, NOx and HC emissions, engine smoke, air temperature at the intake
manifold, pressure before the air flow sensor, pressure in high pressure fuel lines directly before the injector of the
first cylinder, TDC capturing of the first cylinder.
184
The data acquisition was performed throughout a software solution in LabView made by the laboratory stuff, and
using commercial programs for table calculations [3].The engine simulations were performed using AVL Boost [4].
The experiments were performed with two types of fuel, diesel and bio-diesel, with basic fuel properties given in the
Table 3.
Table 3 Diesel and bio-diesel fuel properties

Fuel Diesel Biodiesel
Kinematic viscosity at 30 °C [mm2/s] 3,34 5,51
Surface tension at 30 °C [N/m2] 0,0255 0,028
Heating value [kJ/kg] 43.800 38.177
Cetan number [-] 45-55 >51
By using biodiesel, the cyclic fuel delivery rate was increased to meet boundary conditions in form of equal engine
performance as close as possible to the case of using diesel.
Considered operating modes of the engine were at full load and alternating engine speed from idle to maximum. The
beginning of fuel injection of the high pressure fuel pump was set for a static value of 21 °CA before TDC.
DETERMINING THE IGNITION DELAY PERIOD

To determine the ignition delay period, it’s necessary to identify the real start of the in cylinder fuel injection and the
start of combustion that defines the end of the ignition delay period.
In practice, the dynamic angle of in cylinder fuel injection start is defined in two ways. The first is based on the known
geometry of the fuel injection system, needle opening pressure and the shifting of the injection timing as a result of
engine speed changing [5]. The second is based on observing the injector needle lift. The beginning of needle lift
defines the start of fuel injection.
Start of combustion, respectively the end of ignition delay period, is commonly identified observing the in-cylinder
pressure.
Experimental approach to determine the ignition delay period

Identifying the start of ignition delay is performed analysing the diagram of the needle lift. For the M-type fuel injection
system it is relatively easy to observe the point of the beginning of needle lift. The needle of the fuel injector begins
to rise from its seat when the fuel pressure at the needle seat is high enough to overcome the needle spring force.
After the needle starts to lift the fuel flows to the injector suck and further into the cylinder and causes a small drop
in fuel pressure at the needle seat and the needle descend slightly. The needle lift diagram with marked point when
the needle starts to lift from the seat and the digitalised value expressed in degree of DEG CA is shown in the Figure
1. Repeating this procedure for all considered engine operating modes including diesel and biodiesel the start of real
fuel injection, respectively the start of ignition delay is determined.
Figure 1 Determining the start of ignition delay period observing injector needle lift; for diesel fuel at 1500 min-1
185
Point of the end of ignition delay period is not so easy to determine unambiguously. To avoid mistakes it’s inevitably
to have knowledge of thermodynamic and chemical processes, about techniques and processes in engineering
measurements and numerical operations. The most common methods to determine the end of ignition delay period
are: locating the point of separation of the in-cylinder pressure curve and the curve of pure compression stroke,
locating the extreme value of the first derivation of the in-cylinder pressure, locating the extreme of the second
derivative of the in cylinder pressure.
To locate the end of ignition delay period using the point of separation of in-cylinder and pure compression stroke
pressure curves requires additional measurements. Additional measurements are made without fuel injection,
respectively the in cylinder pressure is measured without combustion. It is possible to undergo measurements only
for few engine speeds and approximate the pressure curves for the other engine speeds and put errors advisedly
into calculation.
The use of the other two methods is highly represented because there is no need for measuring of pure compression
stroke pressure and they use only the first and second derivative of the in cylinder pressure. In mathematical meaning
the problem of locating the end of ignition delay is reduced to determining the extreme values of the analysed in
cylinder pressure curves. This extreme values are result of a sudden disturbance in the in cylinder pressure rise, and
that can graphically be recognised as discontinuities in first and second derivative curves of the in cylinder pressure.
This disturbance is caused by heat release of the started combustion. In this way the located points can be considered
as end points of the ignition delay period. The end points of the ignition delay period determined with the three
previous mentioned methods for using diesel and an engine speed of 1500 min-1 are given in the Figure 2. For all
other engine modes, the same methodology for determining the end of ignition delay period was used.
Analysing trends and shapes of the first and second derivatives of the in cylinder pressure curve, the location of
previous mentioned disturbances are marked with symbols. This points can be classified as points where the real
combustion starts. Depending on the used method different values for the end point of ignition delay were obtained.
The difference is about 1 ˚CA. Relying on experience and recommendation of many researchers the results, that are
gained using the method that involves the second derivative, should be taken as the most credible one.
Figure 2 Determining the end point of ignition delay period using pure stroke pressure curve, in cylinder pressure
curve and its first and second derivatives for using diesel and at an engine speed of 1500 min-1
The methods to determine the end of ignition delay using curves of first and second derivatives are affected with
oscillations as a consequence of in cylinder pressure curve oscillations. It can be seen at the magnified part of
derivative curves at Figure 3. Therefore, in this methods certain errors are introduced. The oscillations of the in
cylinder pressure curves depend on the used measurement equipment, analog-to-digital converters, measuring
performance and mostly on electrical disturbances. To maintain the oscillations amplitudes as small as possible
numerical method known as cubic spline smoothening technic is used. The curve smoothening was performed in
commercial software Grapher.
In order to clarify the influence of numerical derivation methods on determining the end point of the ignition delay
period, forward, backward and central derivation were performed [6]. The results of using these numerical derivation
technics are shown in the Figure 3.
186
The influence of the used numerical derivation technics and curve smoothening on determining the location of the
end point of ignition delay can be easily seen in the Figure 3. The point of interest has different values depending on
the used derivation method and varies in the range of 0.5 °CA. In this particular case, for the given engine operation
mode, values coincided for using backward and central derivation of the in-cylinder pressure spline. At other engine
modes this was not the case.
The curve obtained by using central derivation shows the smallest oscillation amplitudes, and it’s marked as a black
line in Figure 3. Using this derivation technique is most acceptable because it defines the start of combustion in °CA
the best. In order to obtain the influence of line smoothening on the results it is appropriate to compare the first
derivation of the origin pressure curve, blue line at Figure 3, with the derivative of the pressure spline, black line in
the Figure 3. The same comparison is shown at the magnified part of the Figure 3 for the second derivation. The
same techniques were used for obtaining the end of ignition delay period for all observed engine operation modes
by using diesel and bio-diesel fuels.
Figure 3 Influence of derivation methods and smoothening of the in cylinder pressure curve on determining the
end point of ignition delay period; diesel and bio-diesel fuel, engine speed 1500 min-1
Correlation expressions for determining the ignition delay period
A significant number of expressions for determining the ignition delay period in dependence of engine parameters
can be found in the literature. The most of them are based on the Arrhenius equation for combustion velocity [1]:
E
 id  Ap n exp( ~A ) (1)
R
where A and n are fuel dependent coefficients, EA activation energy, 𝑅̃ universal gas constant.
Unfortunately, the expression (1) can only be used for determining the chemical part of ignition delay. To obtain the
whole period of ignition delay it’s necessary to define the physical delay as well. Expressions that incorporate, both
physical and chemical delay, have a much more complex form than expression (1). When using expression (1) some
dilemmas appear like what is the value of temperature T that is supposed to use. Considering different references, it
can be seen that some authors suggest to use the temperature of the in-cylinder gas at fuel injection, while others
suggest to use the average temperature of the in-cylinder gas at ignition delay period. Similar suggestions prevail in
the case when analysing pressure values.
The widely used expressions for determining the ignition delay period are based on selection of temperature and
pressure values related with IC engine states, e.g. pressure and temperature at the beginning of fuel injection, pfi and
Tfi. Moreover, the influence of engine load on ignition delay is taken into account through equivalent air fuel ratio,
187
using temperature and pressure values (Tic and pic) at the beginning of the compression stroke (indirectly the
coefficient of residual gas), while the load rise has a direct influence on Tfi and pfi [7]. It is obvious that an inadequate
selection of temperature and pressure values can lead to errors in determining ignition delay period when using
correlation expressions.
Having in mind that correlation expressions that can be found in the literature are developed on basis of experiments
on certain type of engines, at specific engine modes, using specific type of fuels, and their universality can be
questioned.
Usability of semi empirical expressions that can be found in literature [7] for a diesel engine with M-type fuel injection
for determining the ignition delay period is analysed in this paper. The analysed correlation expressions are given in
the Table 4.
Table 4 Correlation expressions for determining the ignition delay period
Results for determining the ignition delay period using experimental approach and correlation expressions from the
Table 4, by using diesel and bio-diesel and a start of fuel injection at a static angle of 21 DEG CA before TDC are
given in the Figure 4 and the Figure 5.
188
Figure 4 Ignition delay period using diesel and a start of fuel injection at a static angle of 21 DEG CA before TDC;
expressed in DEG CA a); expressed in ms (b)
Figure 5 Ignition delay period using bio-diesel and a start of fuel injection at a static angle of 21 DEG CA before
TDC; expressed in °CA a); expressed in ms (b)
The best matching with experimental results, in case of diesel and bio-diesel use, can be achieved with expressions
by Bibić and Sitkei variant III. The expression by Sitkei variant III for the chemical part of the ignition delay period
uses the values of pressure and temperature at the start of fuel injection (pfi, Tfi), while for the physical part a constant
value is used. Also, the influence of fuel type is not explicit given in the expression.
CALCULATION OF ENGINE PERFORMANCE USING DIFFERENT METHODS

FOR DETERMINING THE IGNITION DELAY PERIOD
The well-known software AVL Boost for engine cycle simulation were used for research. The heat release
characteristic is defined in form of sheet data. The data for dimensionless heat release characteristic are obtained
using an additional program the Burn in the AVL Boost. The cycle simulation for the observed engine were performed
for different input parameters. This is mainly associated with parameters that are different for different engine modes.
A very helpful feature of AVL Boost is the possibility to parametrically define input variables that characterise the
engine operating mode. Variables defined parametrically to carry out engine cycle simulation for a given engine
speed were: cyclic fuel delivery, heat release characteristic and ignition delay. Results of effective engine power
calculation by using diesel and bio-diesel, and ignition delay period obtained using experimental data and correlation
expressions are shown in the Figure 6.
189
Figure 6 Effective engine power calculated using different approaches to determine the ignition delay period; for
using diesel a) and bio-diesel b)
Based in the results presented in the Figure 6 it can be seen that the best matching with experimental obtained
engine power have the calculated engine power curves that were determined using correlation expressions by Bibić
and Sitkei variant III. Having in mind that these correlation expressions have shown a very good matching with
experimental results obtained ignition delay period, it can be concluded that either greater or smaller values of the
real ignition delay period cause a drop in calculated effective engine power.
Comparison of calculated and experimental obtained effective engine torque, for the same engine operation modes,
considering use of diesel and bio-diesel and a static fuel injection angle of 21 °CA before TDC is shown in the Figure
7.
Figure 7 Effective engine torque calculated using different approaches to determine the ignition delay period; for
using diesel a) and bio-diesel b)
Taking into account the facts that engine power depends on engine torque, and that there are connected through a
mathematical formulation in which the engine speed has a dominant role, a parallel between this two variables can
be withdrawn. Based on the results shown in the Figure 7 once again, it can be seen that the best matching with
experimental torque curves are achieved with using correlation expressions for ignition delay period by Bibić and
Sitkei when performing engine cycle simulation. Matching values of this curves are in the range of 3%. The best
matching can be observed at lower engine speeds. This confirms the conclusion that at lower engine speeds the
ignition delay period isn’t so influential on determining the start of combustion when performing engine cycle
simulation.
CONCLUSIONS
For the most realistic definition of processes at an actual IC engine it’s necessary to define the input parameters as
detailed as possible. From the aspect of combustion the greatest attention should be given to define the heat release
190
characteristic and heat transfer as accurate as possible. In this process the ignition delay period has an important
role because it defines the start of combustion.
Analysing experimental data the ignition delay period was determined. For this purpose experimental data like injector
needle lift, in-cylinder pressure, pressure of pure compression stroke, fuel pressure in front of the fuel injector,
pressure in the fuel line and other have been used.
The prediction of the ignition delay period can be faster achieved using correlation expressions. However, their
versatility and their accuracy in the application of different IC engine types is questionable. For the given IC engine
the best match with experimental results, for using diesel and bio-diesel, were observed for using correlation
expression developed on hand of its type. An important thing when using correlation expressions is their applicability
to the entire operating range of the engine. Some correlation expressions used in this paper showed a good matching
with experimental results at lower engine speeds, but not at higher engine speeds.
REFERENCES
[1] Dobovišek, Ž., Černej, A.: “Procesi sagorijevanja”, Mašinski fakultet Sarajevo, Sarajevo, 1978,
[2] Soon-Ik Kwon, Masataka Arai, Hiroyuki Hiroyasu, Effects of Cylinder Temperature and Pressure on Ignition
Delay in Direct Injection Diesel Engine, Bulletin of the M.E.S.J., Vo1. 18, No.1, 1990,
[3] Bibić, Dž.: “Karakteristike sagorijevanja bio dizela i njegovih mješavina sa fosilnim gorivima u dizel motorima”,
doktorska disertacija, Mašinski fakultet Sarajevo, 2007,
[4] AVL Boost, User's guide series, 2013,
[5] Bibić, Dž., Filipović I., Hribernik A., Pikula B.: “Investigation Into The Effect of Different Fuels on Ignition Delay
of M-Type Diesel Combustion Process”, Thermal Science, Vol. 12, No. 1, 2008, pp 103-114, Belgrad, Serbia
(ISSN 0354-9836),
[6] Demirdžić, I.: “Numerička matematika”, IP Svjetlost, Sarajevo, 1997,
[7] Alkhulaifi, K., Hamdalla M.: “Ignition Delay Correlation for a Direct Injection Diesel Engine Fuelled with
Automotive Diesel and Water Diesel Emulsion”, World Academy of Science, Engineering and Technology, Vol.
5, No. 10, 2011, pp 884-896.
191
192
Kragujevac, Serbia
MVM2018-029
Ivan Blagojević1
Saša Mitić2
HYDROGEN AS A VEHICLE FUEL
ABSTRACT: Limited oil reserves and the ecological consequences of its use in motor vehicles impose an
increasing need for alternative energy and drives. The use of hydrogen as the fuel of the future is one of the
solutions whose commercialization has begun in the past few years. Although its use has no impact on the
environment, the production process, which is still expensive, does. Transport and storage of hydrogen, as well as
advanced, but expensive technologies for its use, also present additional challenges, which are examined by the
authors of this review paper through few topics. The properties of hydrogen, the way it is obtained, stored and
transported are presented. Particular attention is paid to the method of obtaining the propulsion energy by
combustion or by using different types of fuel cells. The paper also covers some examples of vehicles powered by
hydrogen.
KEYWORDS: hydrogen fuel, fuel cells, ecology, logistics problems
INTRODUCTION
Oil, coal and natural gas are three major energy sources of the present. The share of sources such as nuclear
energy, geothermal springs and wind is significantly smaller, whereas so-called alternative fuels and innovative
technologies take only a small part. Unfortunately, oil is not a renewable source of energy. According to the
forecast [8], drastic changes will happen in year 2040 (Figure 1). By that time, the production of crude oil will
increase, but it will be followed by a sudden decrease; thus, it is expected that in 2100 major sources of energy will
be coal, renewable liquid and gaseous biofuels, solar energy, wind energy, as well as hydropower and nuclear
energy. Also, there is the problem of unequal distribution of oil as a natural resource - for many countries, it is an
import energy source whose price varies.
Figure 1 Forecast of energy demand
1
Prof. dr Ivan Blagojević, University of Belgrade - Faculty of Mechanical Engineering, iblagojevic@mas.bg.ac.rs
2
Prof. dr Saša Mitić, University of Belgrade - Faculty of Mechanical Engineering, smitic@mas.bg.ac.rs
193
Estimates from 2014 point to the fact that 98% of motor vehicles on the planet use IC engines running on
hydrocarbon-based fuels. According to the statistical data [32], the number of motor vehicles in the world in 2014
was over 1.23 billion (Figure 2). The number of vehicles in 1970 was approximately five times smaller and in only
16 years (by 1986) the number of vehicles has doubled. In 2009, the number of motor vehicles on the planet
already exceeded 1 billion. In only five years (from 2009 to 2014), the number of vehicles increased by more than
200 million – more than 20%. It is estimated that, by the year of 2035, the number of vehicles on the planet will
reach two billion [31]. Such a forecast is not encouraging at all, because of the fact that limited sources of energy
are being used, primarily oil, which still represents the world leading energy source for the motor vehicles, and also
because of the global pollution of our planet due to exhaust emission of IC engines.
Figure 2 Total number of vehicles in the world
It should be emphasized that the research has shown that the increase in fuel consumption while driving is followed
by the increase in harmful exhaust emission, but in different proportions for various emission components. For
example, if the difference in fuel consumption caused by different driving techniques is 21%, the difference in
emission can be 72% for hydrocarbons and 48% for nitrogen oxides [22]. The content of carbon dioxide in
exhausts is proportionate to the fuel consumption. According to the data provided by the U.S. Environmental
Protection Agency, transportation represented 27% of total U.S. GHG emissions in 2015, 83% of which is caused
by on-road vehicles [10]. Each liter of petrol combusted in the vehicle engine produces 2.35 kg of CO2 on average,
whereas each combusted liter of diesel releases in to the atmosphere 2.69 kg of CO2 on average [6]. Average
passenger vehicle emits 4.7 tons of carbon dioxide over one year [11]. This clearly shows that fuel-efficient driving
not only saves fuel, but also contributes to the environment protection. Carbon dioxide is present in the air as a
natural ingredient and it is not considered a pollutant in terms of exhaust emission of motor vehicles. However, it is
considered as one of the major causes of climate changes due to “greenhouse effect”. Atmospheric CO2 levels
rose to 395 ppm in 2012, making the second highest jump since 1959, when the measurements of atmospheric
carbon dioxide levels began [5]. The direct relationship between exhaust emission and human health points to the
significance of the reduction of the emission produced by fuel combustion.
Limited sources of oil and increased exhaust emission of motor vehicles represent the global challenge for the
researchers. The problems are solved by applying new solutions on existing IC engines, as well as by using
alternative fuels such as liquefied petroleum gas, compressed natural gas or certain kinds of biofuels. However, a
special way to solve the above-mentioned problems is to develop and use modern solutions, such as hybrid
vehicles, electric vehicles and vehicles running on hydrogen.
HYDROGEN
If we want to decarbonise our energy footprint, renewable energy source are the way forward. They can either be
consumed directly, like biogas, or can be transformed into electricity like hydro, solar and wind power. Renewable
energy sources bring their challenges though: when the electricity is produced, it does not necessarily match the
demand from the grid. They can be “controllable”, like hydro power for instance, meaning you can adjust the
electric power by regulating the water flow, or they can be weather-dependent like solar and wind power with a little
control on the electricity generation. If we want to fully exploit their power potential, we need to store the electricity
they produce and hydrogen has the biggest potential.
194
So what are the hydrogen properties that make it so interesting? First of all, it is the most abundant element on
earth. It has potential for zero carbon footprints if produced from renewable energy, and it can be transported on
long distance. It has a high energy density, actually higher than batteries. Finally, it can serve as a feedstock to
capture CO2, to produce methane and other gases.
Hydrogen has some very specific characteristics that determine its behavior. It consists of very stable molecules.
Whilst hydrogen can burn by itself, it cannot easily detonate, and in the case it does, it has less energy content
than other gases, so it is much safer. Since hydrogen molecules are the lightest and smallest they disperse very
quickly and doesn’t gather in a cloud at floor level like natural gas does. It requires between 4% and 74% oxygen
concentration to burn. This means that a hydrogen tank cannot explode, since there is no oxygen present. And
when it leaks, hydrogen disappears easily in the surrounding air, as long as it has a way to escape.
Hydrogen can be produced in many ways. Most of the hydrogen produced (95%) is obtained from fossil fuels by
reforming steam, partial oxidation of methane or gasification of coal. The remaining percentage is reserved for
other, less widely used methods such as biomass gasification or water electrolysis.
Hydrogen has long history of mass production and usage. It has been used for over two centuries – the first street
lighting, for example, used hydrogen. Nowadays, hydrogen is mainly produced for usage in refineries to desulfurize
fuels and for the production of ammonia.
HYDROGEN-PROPELLED VEHICLES
Hydrogen is one of the most interesting solutions for propelling the motor vehicles even though, globally speaking,
we are still far from using it massively. Many car manufacturers have presented their prototypes and models driven
by hydrogen with the aim to emphasize that there is a potential in it. The greatest advantage of hydrogen is the fact
that it does not pollute the environment. It does not emit carbon in any form, which also applies to carbon dioxide,
nitrogen oxides or particulates.
There are two types of hydrogen vehicles: vehicles using hydrogen as direct energy source and vehicles using fuel
cells. The former use IC engines modified from standard petrol engines to burn hydrogen. However, the most
popular hydrogen vehicles nowadays are the ones using fuels cells.
Propulsion using hydrogen combustion

Hydrogen can be used in conventional petrol engines as its flame spreads quickly from the ignition core throughout
the chamber. However, due to smaller energy density of hydrogen compared to petrol, at pressures suitable for
cylinder pistons, engine displacement must be two or even three times bigger than that of petrol engines (around 4
liters and 8 to 12 cylinders are needed), which is a problem in terms of space needed. The comparative overview of
some characteristics of hydrogen and petrol is presented in Table 1 [17, 13].
Since large amounts of hydrogen are required to fill larger engine displacement, hydrogen needs to be much
denser than it is possible in its gaseous form. Therefore, it is necessary to use liquid hydrogen cooled down to the
temperature of around 20 K, as well as special filling stations. Such technologies are already being developed and
special tanks for liquid hydrogen have already been designed (with multiple metal cylinders with appropriate
insulation). However, it is also necessary to solve the problems of heat dissipation and hydrogen leakage.
Table 1 Comparative overview of certain characteristics of hydrogen and petrol

Characteristic Hydrogen Petrol Unit
Minimum ignition energy in air 0.018 0.2 – 0.3 mЈ
Flame temperature 2,207 2,307 °С
Auto ignition temperature 575 – 580 480 – 550 °С
Flame velocity in air (λ = 1) at 20 °С and
2.37 0.12 m/s
atmospheric pressure
Octane number > 130 90-98 -
Flammability limits in air 4.1 – 75.6 1.48 – 2.3 % of volume
Hydrogen as a vehicle motor fuel is easier to use in public transportation vehicles. The engine and tank size
problems are relatively easy to solve in city buses, because of more space is available, including the roof.
Hydrogen propulsion using IC engine can be achieved by modifying existing engine in a certain way: valve thermal
treatment, installation of non-platinum tip spark plugs, higher coil voltage, injectors designed for gas (not liquid)
195
usage, more durable gasket materials, higher temperature engine oils etc. Many manufacturers have experimented
with hydrogen engines. Mazda has developed Wankel engine for hydrogen combustion, an Austrian company
named Alset has developed hybrid hydrogen/petrol system for Aston Martin Rapide S which participated in 24
Hours Nürburgring race [19], and BMW has developed their own supercar reaching the speed of 301 km/h burning
hydrogen [16].
Propulsion using fuel cells

The increasingly popular technology widespread with hydrogen cars is the fuel cell technology. The fuel cell
converts chemical energy into electric energy with the help of chemical reaction of positively-charged ions of
hydrogen with oxygen or another oxidation agent. It is important to distinguish fuel cells from batteries because fuel
cells need the flow of fuel and oxygen (air) in order to maintain the chemical reaction and produce electricity. This
continuity in the production of electricity, as long as the fuel cell is supplied with fuel and oxygen (air), is its
advantage.
Figure 3 Fuel cell with polymer electrolyte membrane (PEM) - principle of operation
There are many types of fuel cells [7], but the principle of their operation is essentially the same (Fig. 3). Each of
them has a cathode, an anode and an electrolyte which enable positively-charged hydrogen ions to flow inside the
fuel cell. The cathode and anode have the catalyst which causes the fuel reaction so that it can generate positively-
charged ions and electrons. While hydrogen ions flow through the electrolyte after the reaction, electrons flow from
the anode to the cathode through the external circuit, thus producing direct current. Fuel cells are mostly classified
by the type of used electrolyte and by the time of the reaction initiation, which can vary from one second to several
minutes. Besides electricity, fuel cells produce water and heat, as well as the small amount of nitrogen oxides and
other emissions depending on the used fuel.
In theory, fuel cells would function without losses, but that is not the case in real conditions. During their normal
function, the major losses of the fuel cell are: activation losses which directly depend on the degree of the chemical
reaction; voltage drop due to the resistance of the medium to the flow of ions and electrons; and concentration
losses depending on the reactant concentration and their changes. Theoretical voltage which the fuel cell can
achieve is somewhere around 1.2 V. However, due to the energy efficiency of the cell, which ranges from 40% to
60% [29], real operating voltage of the cell is between 0.45 and 0.72 V. Depending on the purpose of the fuel cells,
they can be connected in two ways: in series (which produces higher voltage) or in parallel (which produces higher
electric current). For the needs of motor vehicles, fuel cells are arranged in a way in which reacting gases must be
uniformly distributed in all cells in order to gain maximum power output.
The type of electrolyte and its chemical composition determine the type of the fuel cell itself. The fuel it uses is also
very important, but pure hydrogen is used most commonly. The anode catalyst, which is usually fine platinum
powder, initiates the fuel and dissolves it into electrons and ions. Cathode catalyst, which is most often made of
196
nickel, transforms ions into chemical compounds such as water or, rarely, dioxides of other elements. For the
needs of hydrogen vehicle propulsion systems, fuel cells with polymer electrolyte membrane (PEM) represent an
optimum choice.
Figure 4 2015 Toyota Mirai
At this point of hydrogen vehicles development, Toyota has made the greatest progress by presenting the first
commercial vehicle propelled by fuel cells at the end of 2014. Toyota Mirai (Figure 4) uses PEM fuel cells – there
are 370 of them in the package, arranged in series. One fuel cell is 1.34 mm wide and weights 102 g with the total
weight of all cells in the package being 56 kg. The curb weight of the vehicle is 1,850 kg. Increased fuel-cell output
voltage enables reduction in the size of the electric motor, as well as in the number of fuel cells. Hydrogen tank
refilling takes 3 to 5 minutes, and the vehicle with full tank can cover almost 500 km. This type of vehicle has two
tanks made of plastic reinforced with carbon fibers weighing together 88 kg. Fuel cells, arranged in a three-
dimensional grid for better dispersion of oxygen, can produce a total maximum power output of 114 kW [1]. Current
researches are pointed at making the fuel cells more efficient using advanced control strategies [14, 3].
PROBLEMS IN HYDROGEN EXPLOITATION

The main problem is the production of hydrogen, which is estimated to be at least five times more expensive than
the production of petrol. The fact that car manufacturers are interested in this technology and its improvement is
encouraging, so it is rightfully expected that, in the future, the process of hydrogen production will be significantly
cheaper. Another disadvantage is that the hydrogen is separated through the process of electrolysis that needs
electricity, which is usually largely produced by environment-unfriendly processes. If the hydrogen is to be the true
environment-friendly source of energy, the electricity needed for its production must be produced by environment-
friendly processes.
Hydrogen storage is also a big problem. There are several ways to store hydrogen and the most popular are
compression, liquefaction and storage in underground pits. Hydrogen storage in a liquid state is particularly difficult
because it must be carried out in a cryogenic state, or at very low temperatures, since the boiling point is about 20
197
K (-253 °C). Therefore, the process of liquefaction is a negative energy process, because a large quantity of energy
is necessary for the hydrogen to cool down at such a low temperature. Reservoirs must be extremely well protected
in order to avoid leakage, and their insulation leads to additional costs.
Compressed hydrogen is stored quite differently. It is known that hydrogen has a good mass energy density, but
very low volumetric energy density, especially when compared to hydrocarbons, and therefore a larger storage
tank is needed, heavier than the reservoir for conventional fuel for the same amount of energy. Increasing the
pressure will increase the volumetric energy density. A high degree of compression without an energy recovery
plan will lead to losses in the compression process, so attention should be paid to this factor. Compressed
hydrogen may also be exposed to a low permeation rate. For the vehicle's needs, hydrogen is compressed to a
pressure between 350 and 700 bar.
The great problem is the infrastructure itself (factories, transport, filling stations). Today, hydrogen is most often
transported in compressed or liquid form in containers with gas cylinders. Transport by pipes is also possible, only
by land using pipelines used for the transport of natural gas. This would be possible with minimal modifications of
existing infrastructure. On the other hand, the construction of a pipeline solely for hydrogen is currently too
expensive.
FUEL CELL AND BATTERY ELECTRIC VEHICLES COMPARED

The transition to a more sustainable personal transportation sector requires the widespread adoption of electric
vehicles powered by batteries (Battery Electric Vehicle - BEV) or fuel cells (Fuel Cell Vehicle - FCV). Automotive
manufacturers are now confronted with decisions to invest in technologies that will become adopted in the future.
While some manufacturers have chosen to invest in either BEVs or in FCVs, most companies have invested in both
and/or have formed partnerships to develop both batteries and fuel cells. Academic literature has not yet
sufficiently addressed the battle between BEVs and FCVs [27].
All-electric vehicles, either powered by batteries or by hydrogen fuel cells based on hydrogen produced from
renewable energies seem to be the only viable option to meet the future CO2 emission targets of less than 95 of
gCO2/km. An analysis of the system-level energy density of lithium-ion batteries (LiBs) suggests that the
gravimetric energy density of advanced LiBs is unlikely to exceed 0.25 kWh/kg, which would limit the range of
BEVs for the compact car to ca. 200 miles (320 km), with recharging times substantially larger than that of
conventional vehicles. Higher energy densities would only be possible, if one were able to develop durable and
safe metallic lithium anodes. While the so-called post-LiBs, lithium-air and lithium-sulfur batteries have been
assumed to revolutionize battery energy storage, cell- and system-level gravimetric energy densities are not
expected to substantially exceed that of advanced LiBs; volumetric energy densities will most definitely be lower. In
contrast to BEVs, hydrogen powered FCVs are capable of large driving ranges (more than 480 km) and can be
refilled within several minutes. Besides the need for a hydrogen infrastructure based on hydrogen produced from
renewable energy, a reduction of the platinum requirement per vehicle (currently ≈20–40 g Pt/FCEV) still requires
further development [12].
The importance of technological superiority is confirmed in the literature on the technology battle between BEVs
and FCVs. FCVs are considered to suffer from hydrogen storage and safety issues [6,9,15]. BEVs face challenges
in range, i.e. battery capacity, and long charging times [2,21,23]. These limitations to technological performance
make BEVs and FCVs less attractive in the eye of potential buyers, posing a barrier to market acceptance.
Batteries are widely used in a wide variety of appliances, whereas fuel cells remain relatively unknown to the
broader audience. Batteries are simply a proven technology. Concerning the latter, academic literature [28,25,20]
indicates many technical specifics that determine superiority of BEVs, ranging from fuel costs and battery/fuel cell
life cycle to the performance indicators mentioned above and more advanced factors such as possibilities to use
the car for energy storage. An additional consideration worth mentioning at this point is whether technological
superiority will remain important. According to Suarez [24], when the first commercial product has been introduced
within a product category, technology related factors for standard dominance become less important and marketing
and business strategies become increasingly relevant. In other words, once both options have proven themselves
in the market, other factors may become more important.
International compatibility is considered critical for the success of BEVs; the difference in charging systems reduces
the attractiveness, and therefore, the adoption of BEVs [23,4,18]. The scattered process of development (in time
and space) and the current lack of compatibility standards have led car manufacturers to produce vehicles with
their own electrical connector types for DC fast charging [30]. In contrast, FCVs hardly have any compatibility
issues between FCVs and fuel dispensers as hydrogen dispensing nozzles adapt to car receptacles [26].
Renewable electricity was considered for charging BEVs and the production of hydrogen, but still it’s hard to call
FCVs environmentally optimal if they ultimately still waste 78% of the net energy (Fig. 5). Traditional electrolysis
has an efficiency of around 70%, whereas a newer technology called proton exchange membrane electrolysis can
198
reach 80%. The transport, storage and distribution of hydrogen cost about 26% of the energy. By contrast, BEVs
only have to contend with grid losses, which average around 5%. Once it’s in the vehicle, hydrogen has an
efficiency which ranges from 40% to 60% - much better than efficiency of a gas or diesel engine, but lower than the
75% for a BEV. So FCVs are less efficient than BEVs at every stage of the process: generating hydrogen;
transportation and storage; and converting it back to energy in the vehicle.
In the future, a hydrogen infrastructure is likely to be set up for heavy duty vehicles, and in the shipping and
aviation sector, which might enable an easy addition of fuel stations for personal transportation. However, as this
infrastructure still needs to be built, and the electricity grid and an increasing number of charging stations are
already in place, it is not surprising that experts believe that BEVs still have a substantial advantage over FCVs. In
addition, the relative presence of BEVs on the road compared to relatively few FCVs could be responsible for a
bias among experts that BEVs have substantially fewer problems related to compatibility.
Figure 5 Overall energy efficiency BEV vs. FCV [33]
CONCLUSIONS
It should be emphasized that almost all forecasts so far, especially the ones from the end of the previous century,
have anticipated a significantly greater usage of electric vehicles in first two decades of 21st century, but that did
not happen. The reasons for this may include the existence of the “fossil fuels lobby”, high prices of today’s electric
vehicles and undeveloped infrastructure of charging stations.
It is clear that future solutions must overcome many obstacles standing in the way of commercialization of
hydrogen as a motor fuel. First of all, it implies the reduction of hydrogen production and transport costs, as well as
building of a well-spread network of filling stations. Experts anticipate that, despite the presence of certain models,
the mass production of hydrogen vehicles will not start in the following 15 years. However, hydrogen, being the
most widespread chemical element, is considered a very promising energy source.
One interesting advantage is that hydrogen fuel cell technology may have many more uses outside of
transportation – the promise of a full hydrogen economy. Fuel cells could power buildings, for example. So
hydrogen power may prove more versatile, and its use in transportation may benefit from this, even if by itself it is
not inherently better than battery technology.
It is also possible that we will end up with both. Some manufacturers are focusing on long-haul vehicles, like trucks,
buses, and even trains, propelled by hydrogen. Further, they can build a much more limited infrastructure for
predictable travel routes of these vehicles. It seems like car manufacturers are putting most of their resources in
battery technology for now, but are hedging their bets by continuing to develop hydrogen fuel cells, which they think
will be the long term winner.
199
ACKNOWLEDGMENTS
This paper presents some of the results obtained through the project supported by Serbian Ministry of Education,
Science and Technological Development (TR 35045 - "Scientific-Technological Support to Enhancing the Safety of
Special Road and Rail Vehicles").
REFERENCES
[1] 2016 Mirai Product Information, Toyota Motor Corporation,
[2] Adepetu, A., Keshav, S.: “The relative importance of price and driving range on electric vehicle adoption”, Los
Angeles case study, Transportation, Vol. 44, 2017, pp 353–373,
[3] Ahmadi, S., Bathaee, S.M.T., Hosseinpour, A.: “Improving Fuel Economy and Performance of a Fuel-Cell
Hybrid Electric Vehicle (Fuel-Cell, Battery, and Ultra-Capacitor) Using Optimized Energy Management
Strategy”, Energy Conversion and Management, Vol. 160, 2018, pp 74-84,
[4] Boulanger, A.G., Chu, A.C., Maxx, S., Waltz, D.L.: “Vehicle electrification: Status and issues”, Proceedings of
the IEEE, Vol. 99, 2011, pp 1116–1138,
[5] Carbon Dioxide Levels Rose at Record Pace for 2nd Straight Year, National Oceanic and Atmospheric
Administration, 2017,
[6] Chalk, S.G.; Miller, J.F.: “Key challenges and recent progress in batteries, fuel cells, and hydrogen storage for
clean energy systems”, Journal Power Source, Vol. 159, 2006, pp 73–80,
[7] Chan, C.C., Chau, K.T.: “Modern Electric Vehicle Technology”, Oxford University Press, Oxford, 2001,
[8] Cobb, L.: “The History and Future of World Energy”, The Quaker Economist, Vol. 7, 2007, pp 155,
[9] Edwards, P.P., Kuznetsov, V.L., David,W.I., Brandon, N.P.: “Hydrogen and fuel cells: Towards a sustainable
energy future”, Energy Policy, Vol. 36, 2008, pp 4356–4362,
[10] Fast Facts - US Transportation Sector Greenhouse Gas Emissions 1990-2015, US Environmental Protection
Agency, 2017,
[11] Greenhouse Gas Emissions from a Typical Passenger Vehicle, US Environmental Protection Agency, 2014,
[12] Gröger, O., Gasteiger, H., Suchslandc, J-P.: “Review—Electromobility: Batteries or Fuel Cells?”, Journal of
The Electrochemical Society, Vol. 162, Issue 14, 2015, A2605-A2622,
[13] Gupta, B.R.: “Hydrogen Fuel Production, Transport and Storage”, CRC Press, 2008, Boca Raton, USA,
[14] Hames, Y., Kaya, K., Baltaciouglu, E., Turksoy, A.: “Analysis of the Control Strategies for Fuel Saving in the
Hydrogen Fuel Cell Vehicles”, International Journal Hydrogen Energy, Vol. 43, Issue 23, 2018, pp 10810-
10821,
[15] Mori, D., Hirose, K.: “Recent challenges of hydrogen storage technologies for fuel cell vehicles. Int. J.
Hydrogen”, Energy, Vol. 34, 2009, pp 4569–4574
[16] Müller, C., Fürst, S., von Klitzing, W.: “Hydrogen Safety: New Challenges Based on BMW Hydrogen 7”,
Proceedings, Second International Conference on Hydrogen Safety, 2007, San Sebastian, Spain,
[17] Negurescu, N., Pana, C., Popa, M.G., Cernat, A.: “Performance Comparison between Hydrogen and Gasoline
Fuelled Spark Ignition Engine”, Thermal Science, Vol. 15, No. 4, 2011, pp 1155-1164,
[18] Nemry, F., Brons, M.: “Plug-In Hybrid and Battery Electric Vehicles: Market Penetration Scenarios for Electric
Drive Vehicles, JRC European Commission, 2010, Seville, Spain,
[19] Propel Technology; Available from: news.cision.com/propel-technology-ltd/r/history-is-made-as-reliability-and-
performance-of-unique-hybrid-hydrogen-technology-is-proved-at-the, Accessed 25.08.2018.,
[20] Offer, G.J., Howey, D., Contestabile, M., Clague, R., Brandon, N.P.: “Comparative analysis of battery electric,
hydrogen fuel cell and hybrid vehicles in a future sustainable road transport system”, Energy Policy, Vol. 38,
2010, pp 24–29,
[21] Perujo, A., Van Grootveld, G., Scholz, H.: “Present and Future Role of Battery Electrical Vehicles in Private
and Public Urban Transport In New Generation of Electric Vehicles”, InTech, 2012, London, UK, pp 3–28,
[22] Silva, C.M., Farias, T.L., Mendes-Lopes, J.M.C.: “EcoGest – Numerical Modelling of the Dynamic Fuel
Consumption and Tailpipe Emission of Vehicles Equipped with Spark Ignitions engines”, Urban Transport VIII,
WIT Press, 2002, Southampton,
[23] Steinhilber, S., Wells, P., Thankappan, S.: “Socio-technical inertia: Understanding the barriers to electric
vehicles”, Energy Policy, Vol. 60, 2013, pp 531–539,
[24] Suarez, F.F.: “Battles for technological dominance: An integrative framework”, Research Policy, Vol. 33, 2004,
pp 271–286,
[25] Thomas, C.E.: “Fuel cell and battery electric vehicles compared”, International Journal of Hydrogen Energy,
Vol. 34, 2009, pp 6005–6020,
[26] Toyota Fueling the Toyota Mirai, Available online: https://ssl.toyota.com/mirai/Mirai_Fueling.pdf, Accessed
26.08.2018.,
[27] van de Kaa, G., Scholten, D., Rezaei, J., Milchram, C.: “The Battle between Battery and Fuel Cell Powered
Electric Vehicles: A BWM Approach”, Energies, Vol. 10, Issue 11, 2017,
[28] Van Mierlo, J., Maggetto, G., Lataire, P.: “Which energy source for road transport in the future? A comparison
of battery, hybrid and fuel cell vehicles”, Energy Conversion and Management, Vol. 47, 2006, pp 2748–2760,
200
[29] Welaya, Y.M.A. et al.: “A Comparison Between Fuel Cells and Other Alternatives for Marine Electric Power
Generation”, International Journal of Naval Architecture and Ocean Engineering, Vol. 3, Issue 2, 2011, pp 141-
149,
[30] Wittenberg, A.: “Fast-Charge Plugs Do Not Fit All Electric Cars”, Available online:
https://www.scientificamerican.com/article/fast-charge-plugs-do-not-fit-all-electric-cars/, Accessed 25.08.2018.,
[31] Green Car Reports; Available from: www.greencarreports.com/news/1093560_1-2-billion-vehicles-onworlds-
roads-now-2-billion-by-2035-report, Accessed 26.08.2018.,
[32] The Statistics Portal, Available from: www.statista.com/statistics/281134/number-of-vehicles-in-useworldwide,
Accessed 25.08.2018.,
[33] Transport and Environment, Available from: www.transportenvironment.org/newsroom/blog/renewable-
electricity-must-decarbonise-land-freight-transport, Accessed 25.08.2018.
201
202
Kragujevac, Serbia
MVM2018-044
Ivana Terzić1
Vanja Šušteršič2
Aleksandar Nešović3
Mladen Josijević4
POSSIBILITIES FOR THE USE OF BIOFUELS IN EUROPE

AND IN SERBIA
ABSTRACT: Biomass is one of the major potentials of renewable energy sources in Serbia. Due to the lack of
legal acts, regulations, and incentive systems by the state, this potential has not been sufficiently utilized. According
to some estimates, the theoretical potential of biomass in Serbia amounts to about 3.2 million equivalent tonnes of
oil. This paper gives an overview of the existing situation in Serbia, including a set of measures necessary to stimulate
the production of biofuels, primarily for the purpose of more intensive development of agriculture, reduction of
pollution of the environment, reduction of the import of crude oil and for the creation of new jobs. The aim of this
paper is to present the possibilities, conditions, and obstacles in the production and use of biodiesel in Serbia, as
well as the comparison of the use of biodiesel in the countries of the European Union. Special attention is paid to the
economics of production and the ecological motive for the production of biodiesel, as well as the ability to reduce
CO2 emissions and the environmental impact of transportation vehicles.
KEYWORDS: biofuel, renewable energy sources, transportation vehicles, environmental impact
INTRODUCTION
The total global oil consumption is almost 4 billion tonnes a year, whereas its reserves are estimated at around 120-
160 billion tonnes, which represents one of the major problems in the field of energy. On the other hand, the natural
and technical potential of renewable energy sources on a daily basis is about 20,000 times greater than the daily
consumption of nuclear fuels and fossil fuels, together. Fossil fuels are largely the cause of global warming,
generating the greenhouse effect through combustion. Spending on crude oil reserves and environmental problems
related to its processing and utilization are the two main reasons why we should turn to bio-sources that convert into
biodiesel and bioethanol.
Biofuels in the narrower sense include fuels derived from biomass and used for transportation purposes. This
category of fuel includes methane, gas produced by the gasification process, alcohols, esters and other chemicals
derived from cellulosic biomass. The basic raw materials for the production of biofuels are plant and tree plantations,
sugar and oil crops, agricultural and forestry residues, as well as individual fractions of municipal and industrial waste.
1
Ivana Terzić, PhD student, University of Kragujevac, Faculty of Engineering, Sestre Janjić 6, 34000 Kragujevac,
ivana.aa.terzic@gmail.com
2
Vanja Šušteršič, full professor, University of Kragujevac, Faculty of Engineering, Sestre Janjić 6, 34000 Kragujevac,
vanjas@kg.ac.rs
3
Aleksandar Nešović, PhD student, University of Kragujevac, Faculty of Engineering, Sestre Janjić 6, 34000
Kragujevac, aca.nesovic@gmail.com
4
Mladen Josijević, PhD student, University of Kragujevac, Faculty of Engineering, Sestre Janjić 6, 34000 Kragujevac,
mladenjosijevic@gmail.com
203
Oil reserves are steadily decreasing, so developed countries have been intensively engaged in biomass production
processes over the past decade. In Europe, rape oil (82.8%) and sunflower oil (12.5%) are used as the dominant
raw material for the production of biodiesel, whereas soybeans are used in the USA for the production of biodiesel
[14].
Biodiesel is a non-toxic and biodegradable motor fuel derived from rapeseed oil or other vegetable oils (sunflower,
palm, soybean oil) through esterification with methanol as well as from animal fats and recycled oils (Fig 1) [2, 14].
1%
1%
1%
13% 1% Sunflower
Soybean
Palm
Other
Unclassified
83%
Rape
Figure 1 Graph of the most important raw materials for the production of biodiesel
In developed countries, already used oils and animal fats can be futher utilised in the production of biodiesel. Such
an environmentally friendly and renewable energy source is used as a substitute for mineral diesel. The use of pure
biodiesel reduces the emission of harmful gases (NOx, CO2, SO2) and solid particles, does not contain sulfur, lead,
nitrogen, toxic aromatic compounds, and is biodegradable so that its use reduces environmental pollution. Biodiesel
is commonly used as a mineral oil additive to reduce emissions of solid particles, carbon dioxide, aromatic
hydrocarbons, and other air pollutants resulting from the combustion in conventional diesel engines. Mixtures of a
maximum of 20% biodiesel with mineral diesel can be used without the requirement to modify the engine in almost
all conventional diesel engines. By-products in biodiesel production (such as glycerine, lecithin, and fatty acids) can
be used in the pharmaceutical and cosmetic industry, whereas by-products from bioethanol production can be used
as protein-rich foods [17].
One of the goals of the European Commission Directive 2009/28/EC [7], is to have 10% of energy demand in the
transportation sector coverd by the renewable energy sources (RES) by 2020. In other words, 50 billion litres of fossil
fuels in the European Union (EU) will be replaced by biofuels [3]. In the United States, the Energy Act [4] (passed in
2005) and the 2007 Energy Security and Safety Act [5] both promote RES, including biomass primarily through liquid
biofuels, and set the target of producing 136 billion litres of biofuel for the transportation sector in 2022. According to
some forecasts, it is expected that more than 5% of the fuel in the road sector and 1% in air transportation in the
world will be replaced by biofuels by 2030 [6]. Considering the great importance the biofuels have in the transportation
sector, as well as the forecasts for their higher demand, the aim of this paper is to review currently available
conventional liquid biofuels including advanced liquid biofuels and compare their properties with fossil gasoline and
diesel, indicating the appropriate raw materials and the paths of their conversion. The work additionally contributes
to the definition and differentiation in terminology of the basic concepts related to biomass and the classification of
biofuels.
BIOFUELS AND THEIR DIVISION

According to the European Directive 2009/28/EC [7] and the National Action Plan for the Use of Renewable Energy
Sources of the Republic of Serbia [13], based on the Directive, biofuels represent liquid or gaseous fuels produced
from biomass and used in transportation. There are different approaches to the classification of biofuels considering
the large variety of raw materials and the diversity of the involved conversion processes, which are evolving towards
sustainability and the required fuel quality standards. The most commonly used biofuel classification in the literature
[8, 14, 15], includes the following generations:
 1st generation biofuels: bioethanol, biodiesel, vegetable oils (SVOs or "Pure Plant Oil" - PPO) and biogas /
landfill gas;
 2nd generation biofuels (2G): bio alcohols (cellulose ethanol, bio methanol), BtL fuels (synthetic liquid fuels
produced by thermo-chemical conversion of biomass, for example FT diesel and FT kerosene), biodiesel
from waste materials, HTU diesel (bio DMF), hydro treated vegetable oils (renewable diesel), bio-dimethyl
ether (bio DME), bio-synthetic gas (bio SNG) and bio-hydrogen;
 3rd generation biofuels (3G): algae biofuels (for example, algae bio alcohols, biodiesel and BtL fuels);
204
 4th generation biofuels (4G): biofuels produced using genetically modified organisms, such as
photosynthetic microalgae with a higher proportion of lipids and the ability to use larger amounts of carbon
dioxide than those commonly used for photosynthesis processes, this way participating in carbon
sequestration processes (removing carbon dioxide from the atmosphere) and consequent reduction of the
greenhouse effect [8, 12, 18].
Figure 2 shows the classification of energy sources and their share in total global energy consumption [9].
Energy
Renewable energy (16 %) Non-renewable energy (84 %)
Clean energy (6 %) Bioenergy (10 %) Nuclear (4 %) Fossil fuels (80 %)
Wind energy,
hydro, solar,
geothermal
Torefaction Combustion Animal fat, Microalgae Biological

Bio coal Heat energy Crops rich in lignocellulosic organisms +
sugars, starch material processes
and oils separating
carbon
Biofuels Biofuels Biofuels Biofuels

1. generation 2. generation 3. generation 4. generation
Figure 2 Classification of energy sources with the contribution to the global energy consumption [9]
Depending on the soil quality and climatic and economic conditions, different raw materials can be used to obtain
biodiesel (see Table 1). In order to reduce the costs of biodiesel production, alternative raw materials are considered,
such as inert vegetable oils and used oils and fat [9, 14].
Table 1 Row materials for biodiesel production [9]

Raw material for biodiesel production Country
Soybean oil USA, Brazil
Rape oil (80%) and sunflower oil (20%) Europe
Palm oil Indonesia, Malaysia
Flax and olive oil Spain
Vegetable oils / animal fats Canada
Animal fats, beef cattle and rapeseed Australia
World total biofuel production and consumption

World total biodiesel production for the year 2017 in different countries is shown in Figure 3. With a production volume
of around 3.3 billion litres, Argentina was ranked fourth that year. The United States and Brazil were among the
largest biodiesel producers in the world, totaling some 6 and 4.3 billion litres, respectively. The United States are
projected to reach production levels of over 1 billion gallons5 of biodiesel by 2025. After the implementation of the
Energy Policy Act of 2005 [4], which provided tax incentives for certain types of energy, biodiesel production in the
USA began to increase. The Volumetric Ethanol Excise Tax Credit is currently one of the main sources of financial
support for biofuels in the USA. In 2010, the USA exported about 85 million gallons of their biodiesel products.
Comparatively, Argentina accounted for over half of the world’s total exports. The United States have one of the
highest bioenergy capacities in the world, totaling 13,151 megawatts in 2017. Today, biofuels provide around 3% of
total road transportation fuel consumption globally on an energy usage basis. Consumption of these fuels in some
countries is considerably higher. According to International Energy Agency (IEA), Brazil met about 23% of the road
transportation fuel demand with biofuels in 2009. IEA analysis shows that biofuels may have to play an important role
5
1 gallon=3.79 l
205
if the world is to make meaningful reductions in carbon dioxide emissions and reduce reliance on crude oil at costs
similar to those of gasoline and diesel in the medium-term. Global biofuels production and consumption are on the
rise. According to the US Energy Information Administration (USEIA), there is no record found on biodiesel production
and consumption before 2000. Most of the countries started using biofuel in the last decade. Many more countries
are in the early stages of using biofuel and they are trying to develop the infrastructure and production technologies
to follow the suite. Leading countries based on the biofuel production in 2017 (in 1,000 metric tonnes of oil equivalent)
are presented in Figure 4, which indicates that the USA produced around 80% of the total biofuel in the world. Europe,
Asia and Oceania are increasing their biofuel production. The remaining countries are in their initial stages of
production.
Figure 3 World total biofuel production in 2017 [24]
Figure 4 The leading countries based on biofuel production in 2017 (in 1,000 metric tonnes of oil equivalent) [24]
This statistic represents the leading countries in biofuel production in 2017. In Germany, production reached around
3.3 million metric tonnes of oil equivalent that year. That means, Germany is among the global top three countries in
biofuel production.
Biofuel production in the Americas

Biofuels are characterized by fuels that obtain their energy through the process of biological carbon fixation. These
hydrocarbons are made by or from living organisms in a relatively short period of time - in comparison to the formation
of fossil fuels which requires millions of years. The United States were by far one of the largest producers of biofuel
in the world in 2016, accounting for 43.5 percent of global biofuel production (Table 2, Figure 5) The country produced
206
35 million metric tonnes of oil equivalents that year, while Brazil produced 18.2 million metric tonnes of oil equivalents.
Global biofuel production has gradually increased from 9.2 million metric tonnes of oil equivalents in 2000 to 84.12
million metric tonnes of oil equivalents in 2017.
Biofuels are commonly used as part of mixtures with fossil fuel sources or as additives. One of the largest consumers
of biofuels in the USA is the national army. Many vehicles can be fueled using blends containing up to 10 percent of
ethanol. In the beginning of the 20th century, many Ford T models were fueled with ethanol. Biofuels can also be
generated through the consumption or conversion of biomass material. This conversion can occur thermally,
chemically, or biochemically. Biomass consumption in the United States totaled 4,696 trillion British thermal6 units in
2015.
Table 2 Biofuels production [22]

Thousand Growth rate per annum
tonnes oil Share
2016 2016 2005-15
equivalent 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
US 10,670 14,709 20,934 23,761 28,044 31,184 29,808 31,057 32,890 33,849 35,779 5.4% 15.2% 43.5%
Canada 174 503 546 786 809 950 1,017 1,056 1,188 1,142 1,160 1.2% 22.8% 1.4%
Mexico - 5 5 5 14 13 15 58 58 58 58 - - 0.1%
Total North
10,844 15,216 21,485 24,552 28,866 32,147 30,840 32,171 34,137 35,049 36,997 5.3% 15.4% 45%
America
Argentina 30 173 635 1,055 1,670 2,234 2,295 2,014 2,644 2,038 2,828 38.4% 71.7% 3.4%
Brazil 9,590 12,427 15,486 15,277 16,866 14,403 14,739 17,114 18,005 19,332 18,552 -4.3% 8.4% 22.5%
Colombia 144 155 158 320 455 572 627 650 676 693 626 -10% 46.2% 0.8%
Other S.& Cent.
513 596 806 634 229 310 300 354 378 379 373 -1.9% 6.5% 0.5%
America
Total S. Cent.
10,278 13,351 17,085 17,285 19,220 17,519 17,961 20,131 21,703 22,442 22,378 -0.6% 9.8% 27.2%
America
Total Europe &
5,269 7,021 8,482 10,646 11,604 10,876 11,734 12,503 14,445 14,012 13,777 -1.9% 15.5% 16.7%
Eurasia
Total Africa 9 6 11 18 8 8 23 32 40 40 40 - 20.5% 0.05%
Total Asia
1,446 1,876 3,074 3,435 4,306 5,280 6,300 7,450 9,374 8,476 9,110 7.2% 25% 11.1%
Pacific
Total World 27,848 37,471 50,138 55,936 64,008 65,834 66,863 72,293 79,703 80,024 82,306 2.6% 14.1% 100%
Figure 5 World biofuels production (million tonnes oil equivalent) [22]
Global biofuels production rose by 2.6 % in 2016, well below the 10-year average of 14.1 %, but faster than in 2015
(0.4%). The USA provided the largest increment (1,930 thousand tonnes of oil equivalent, or ktoe). Global ethanol
production increased by only 0.7 % partly due to falling production in Brazil. Biodiesel production rose by 6.5 % with
Indonesia providing more than half of the increment (1,149 ktoe).
World biofuels production increased by 3.5 % in 2017, well below the 10-year average of 11.4 %, but the fastest for
three years. The US provided the largest increment (950 thousand tonnes of oil equivalent, or ktoe). By fuel type,
global ethanol production grew by 3.3 %, contributing over 60 % to total biofuels growth. Biodiesel production rose
by 4 % driven mainly by growth in Argentina, Brazil and Spain [22].
Global biofuel market is dominated by the USA in North America and Brazil in Central and South America. They have
been producing huge amount of biofuels annually and trying to meet their major part of transportation fuel demand
with biofuel. Currently, Brazil produces 27 billion litres bioethanol annually which is supported by the development of
new sugarcane varieties and agricultural technologies. The USA widely used bioethanol from corn and grain, and
6
1Btu=1.054 kJ
207
biodiesel from soybean as an alternative fuel. They also aimed to replace 30% of fossil fuel with biofuel while Europe
is aiming for 5.75% [1].
Production of biodiesel in the EU

The USA is the world’s largest biodiesel producer. Biodiesel is also the most important biofuel in the EU and, on an
energy basis, represents about eighty percent of the total transport biofuels market. Biodiesel was the first biofuel
developed and used in the EU in the transportation sector in the 1990s. At the time, rapid expansion was driven by
increasing crude oil prices, the Blair House Agreement and resulting provisions on the production of oilseeds under
Common Agricultural Policy set-aside programs, and generous tax incentives, mainly in Germany and France. EU
biofuels goals set out in Directive 2003/30/EC (indicative goals) and in the RED 2009/28/EC (mandatory goals) further
pushed the use of biodiesel.
In 2017, EU FAME7 and HVO8 production did benefit from higher domestic consumption, as elevated imports only
commenced in September that year. FAME production increased by 5 percent, mainly due to expansion in Spain,
Italy, Portugal, Belgium, and Poland. HVO production increased by 5 percent, driven by elevated production in the
Netherlands and Spain, and a new co-processing unit coming into production in Portugal (Tables 3 and 4).
Table 3 EU FAME Main Producers (Million Litres) [23]

2011r 2012r 2013r 2014r 2015r 2016r 2017e 2018f
Germany 3,408 3,106 3,307 3,911 3,555 3,592 3,522 2,610
France 2,090 2,175 2,170 2,386 2,442 2,215 2,181 1,700
Spain 787 538 659 1,017 1,103 1,319 1,680 1,200
Netherlands 558 974 790 1,056 795 638 568 570
Poland 414 673 736 786 861 985 1,029 1,030
United Kingdom 261 352 640 554 572 496 503 510
Italy 704 326 521 452 625 398 599 560
Belgium/Luxemburg 536 568 568 568 535 521 568 570
Austria 352 301 247 332 386 349 352 365
Portugal 419 356 329 349 386 333 388 400
Other 1,667 1,214 604 203 811 977 1,007 1,375
Total 11,197 10,582 10,570 11,614 12,072 11,823 10,397 10,890
Ranked by production in 2018 r = revised / e = estimate / f = forecast.
Source: FAS EU Posts based on information in MT and converted to liters using a conversion rate of 1 MT = 1,136 liters.
Table 4 EU HVO Production (Million Litres) [23]

2009r 2010r 2011r 2012r 2013r 2014r 2015r 2016r 2017e 2018f
Netherlands - - - 410 872 1,013 1,192 1,154 1,218 1,220
Finland 281 365 250 317 392 438 536 545 545 545
Spain - - 28 73 179 377 262 418 465 470
Italy - - - - - 323 323 323 323 445
Portugal - - - - - - - - 32 32
France - - - - - - - - - 128
Total 281 365 278 800 1,444 2,151 2,313 2,440 2,583 2,840
Ranked by production in 2018 r = revised / e = estimate / f = forecast.
Source: FAS EU Posts based on information in MT and converted to liters (conversion rate of 1 MT = 1,282 liters).
The structure of the EU biodiesel sector is very diverse and plant sizes range from an annual capacity of 2.3 million
litres owned by a group of farmers to 680 million litres owned by a large multi-national company. Biodiesel (FAME)
production facilities exist in every EU member state with the exception of Finland, Luxemburg, and Malta. In contrast,
HVO production is concentrated in only six countries (Table 4). The majority of HVO capacity consists of dedicated
HVO plants, while in Spain HVO is co-processed with conventional fuel in oil refineries.
EU FAME production capacity is expected to decrease by 5 percent in 2018 to 20.3 billion litres, as plants are closed
for good as a result of strong competition. In addition, numerous plants run below capacity or are temporarily shut
down. For example, in Germany three plants announced to either temporarily (at least until the end of June) or until
further notice run at half of their capacity. In France, Saipol announced that it would cut its production by between
400,000 and 600,000 MT in 2018. EU HVO production capacity is forecast to increase to 5.3 million litres in 2018,
when two new facilities will start production in Italy and France [23].
7
traditional first generation biodiesel
8
hydrogenated vegetable oil
208
PRODUCTION AND USE OF BIODIESEL IN SERBIA
The use of available energy sources is very high on the list of energy strategy priorities, primarily to meet energy
needs, as well as to improve the conditions for energy production. By ratifying the Energy Community Treaty, the
Republic of Serbia has accepted the obligation to adopt and implement a plan for the implementation of Directive
2003/30/EC on the promotion of the use of biofuels and other fuels from renewable energy sources in the
transportation sector [21].
In Europe, biodiesel is defined in the standard EN 14214 in 2003, and in Serbia in the standard SRPS EN 14214
(standard for fuels for motor vehicles - fatty acid methyl esters for diesel engines) in 2006 [11]. EU Directive
2003/30/EC defines biofuels and imposes on states the obligation to put 5.75% of biofuels on the market by the end
of 2012 (of the total amount of fuel used in transport). Unfortunately, our country has not met the mentioned target
for biofuels, and in the meantime a new EU directive 2009/28/EC has entered into force, which stipulates that by
2020, the target of 10% of biofuels will reach the market [7]. The proposal for the growth dynamics of RES share in
the period from 2009 to 2020 in the transport sector is presented in Table 5.
In Serbia, over 600,000 hectares are available for the production of oil plants, which indicates that Serbia has the
potential to produce biodiesel at the level of over 200,000 ton per year.
Tabela 5 Dynamics of growth of share Renewable energy source (%) in the sector of energy
consumption [7]
2009 base year for the development of NAPOIE in accordance with the RES Directive 2009/28 / EC
2009 2013 2014 2015 2016 2017 2018 2019 2020
RNE – traffic (%) 0 0 0 2 3 5 7 8 10
NAPOIE – National Action Plan for Renewable Energy Sources
In the European Union, one hectare of rapeseed provides enough grain to produce 1,090 litres of biodiesel fuel [10].
However, in Vojvodina, rapeseed, sunflower and soybean produce significantly lower yields than the European
average (Fig.6). With an average seed yield of 1.69 t/ha, and oil seed content of 36%, 1 ha of rapeseed in Serbia
provides 608 kg of oil or about 690 l of biodiesel fuel. The average yield of sunflower seed in Serbia is 1.79 t/ha, so
with oil content 40% yield of sunflower biodiesel is 716 kg/ha or 816 l/ha. The average yield of soybean in Serbia is
2.25 t/ha, so the content of grain oil is 18%, yield of biodiesel 405 kg ha, ie 460 l/ha [19].
Figure 6 Potential yield of biodiesel per 1 ha of oilseed crops in Serbia and EU depending on the yield of oilseed
crops [19]
The potential area in Serbia for cultivation of oilseeds for processing into biodiesel is estimated at approximately
350,000 ha (668,8009- (271,72210 + 54,00011)) (Table 6).
Theoretically, domestic biodiesel could be substituted by 13 to 16% of domestic fuel consumption (calculated on an
energy basis). In fact, these figures are significantly smaller because of the impossibility of organizing economic
production on crushed and intermittent agricultural holdings in Central Serbia.
For the organized production of large quantities of biodiesel, it is necessary consider the conditions and resources
of raw materials for its production. The most important oilseeds produced in Serbia are: sunflower, soy and rapeseed
rape. In Serbia, the sunflower is mostly cultivated over around 220,000 hectares on an annual basis, whereas the
production of rapeseed is around 22,000 tonnes annually. In Serbia, there are sufficient cultivable areas meeting all
the agro-technical conditions for growing rapeseed as a basic culture for the production of biodiesel. From the aspect
9
Potential area for cultivation of oilseeds, 20% of the orange area,
10
for raw food oil,
11
from seeds and animal husbandry, from which 212,800 to 250,600 t of biodiesel could be provided
209
of resources, this would provide the necessary conditions for Serbia to become an important European biodiesel
producer. However, for the development of the biodiesel industry, it is necessary first of all to take measures at the
state level to support investment in this sector of energy [19].
Table 6 Potential production of biodiesel on the area of 350,000 ha in relation to the

sowing structure of oilseed crops [19]
Potential substitution of fossil diesel with
Potential biodiesel
Variants
production
Sowing structure
of biodiesel
All sectors (%) In agriculture (%)
(t)
I 100 % Oilseed Rape 212,800 13.49 46.67

70% Oilseed Rape
II 224,140 14.21 49.16
30 % Sunflower
50 % Oilseed Rape
III 231,700 14.69 50.81
50 % Sunflower
70 % Sunflower
IV 239,260 15.16 52.47
30 % Oilseed Rape
V 100 % Sunflower 250,600 15.88 54.96
Production economics and ecological motives for the production of

biodiesel
One additional advantage of using biodiesel is that the consumer receives reliable and high-quality fuel at a lower
price. For example, at petrol stations in the EU the difference between diesel and biodiesel is about 10 Eurocents
per liter. The price growth trend in crude oil is much higher than in edible oil, which leads to the expectation that in
the future biodiesel will be even more competitive.
The production of renewable fuels leads to a reduction in imports of fossil fuels, which simultaneously affects exports.
In order for biodiesel to be profitable, it is estimated that it must have the price that is about 8% lower than the price
of fossil fuel.
Research in developed countries has shown that the environment is mainly polluted by motor vehicles (over 50%),
then by thermal power plants and industrial plants. Fuel combustion emits many pollutants in the air, such as lead,
inorganic chlorine and bromine compounds, hydrocarbons, nitrogen oxides, and sulfur oxides. For these reasons,
the international community is increasingly compliant with the regulations that determine the quality of liquid
petroleum fuels, (acceptable for the environment and human health), and increasingly stimulates the use of the fuel
of plant origin (that have considerably reduce harmful emissions).
For the purpose of protecting the environment, a request is made that mineral fuel must contain as little sulfur as
possible. Biodiesel does not damage the diesel engine. On the contrary, its increased lubricity compared to mineral
diesel reduces the wear of vital parts of the engine. A mixture of biodiesel and mineral diesel reduces emissions of
harmful gases, and maximum lubricity is achieved at 10% mixture [19].
Considering that in Serbia biodiesel is not yet widely used, the results of pollution testing using biofuels in Germany
can be listed. The German Environmental Protection Association presented data of comparative testing conducted
on 54 different diesel engines:
 CO2 emissions from biodiesel engines are 10% to 12% smaller than those from diesel engines,
 Hydrocarbon emissions in biodiesel engines are 10% to 35% smaller,
 Emission of solid particles in engines with biodiesel is less than 24% to 36%,
 The soot emission for engines with biodiesel is 50% to 52% smaller.
All of the above results show that biodiesel is much more environmental friendly compared to conventional mineral
fuel, either as a pure fuel or as a mixture with mineral diesel [20].
The ecological effect of using biodiesel is more favorable than when classical diesel is used, so that its effect on the
environmental pollution is significantly smaller, as can be seen in Table 7 [16].
The reduction of fossil CO2 emissions is one of the main motivations behind policies and targets promoting a transition
towards biofuels in the transportation sector. Unsurprisingly, an increasing biofuel share entails a higher emission
reduction potential. This is more significant in scenarios where ethanol plants dominate, than in scenarios where
methanol plants do. The reason is the high conversion efficiency to electricity and heat in ethanol plants, in
210
combination with the in general significantly higher CO2 emission factors of displaced electricity and heat compared
to displaced fossil transport fuels.
Table 7 Emission of harmful gases in biodiesel, and

diesel combustion
100 %
Measuring parameters
Biodiesel Diesel
CO [ppm] 92 179
NO [ppm] 113 129
NO2 [ppm] 26.5 23.4
NOx [ppm] 139 153
SO2 [ppm] 5 102
H2 [ppm] 11 19
O2 [vol%] 17.7 17.5
CO2 [vol%] 2,7 2.84
Particles of the soot [mgm-3] 31.2 98.3
Benzene [mgm-3] <10 825
-3
Toluene [mgm ] <10 398
Xylene [mgm-3] <10 40
CONCLUSIONS
Our country needs to take appropriate measures to encourage the production of this ecological and renewable fuel.
Such measures would ensure a more intensive development of agriculture, improve environmental protection, reduce
crude oil imports, and create new jobs. Moreover, such measures are becoming an obligation, since all EU members,
in proportion to the estimated resources, are obliged to increase the production of biofuels. Given that our country
has significant potential for growing plant cultures for the production of biodiesel, it is expected that the domestic
biodiesel industry will intensify in the forthcoming period.
The advantage of conventional biofuels, especially biodiesel, is that they can be produced in decentralized and
smaller plants, representing an advantage for the development of rural environments. Contrary to that, the production
of advanced biofuels is mainly related to large plants in order to achieve better economic performance. At the same
time, large investments are needed that can only be provided by large companies. Therefore, such companies will
dominate the market of advanced biofuels and control the raw materials prices to a great extent, thereby limiting the
benefits of the production of biofuel for the rural environment.
In 2010, production and use of biodiesel was not recorded in Serbia, although our country has significant land
potential for the production of raw materials for processing into biodiesel. From such a surface it is possible to provide
raw materials for production of over 200,000 tonnes of biodiesel per year.
State policy and production costs of biofuels, as well as requirements for competitiveness, can significantly influence
the prices of biofuels and agricultural raw materials.
Our country has accepted international obligations and has the available potential of biomass, so it is necessary to
take all incentive measures as soon as possible in order to reach the binding target of 10% share of biofuels in
transportation sector by 2020 (EU Directive 2009/28/EC).
REFERENCES
[1] Azad, A.K., Rasul, M.G., Khan, M.M.K., Sharma, S.C., Bhuiya, M.M.K., Mofijur, M.: “A review on socio-economic
aspects of sustainable biofuels”, Int. J. Global Warming, Vol. 10, No. 1/2/3, 2016,
[2] Babić, I., Đurišić, Ž.: “Techno-economic and ecological motives of biodiesel production in Serbia”, Faculty of
Electrical Engineering in Belgrade, March 2008,
[3] Bentsen, N.S., Felby, C.: “Biomass for energy in the European Union–a review of bioenergy resource
assessments”, Biotechnol. Biofuels 5/25, 2012,
[4] Congress US Energy Policy Act of 2005, Washington DC, USA, 2005,
[5] Congress US Energy Independence and Security Act of 2007, Washington DC, USA, 2007,
[6] Crocker, M.: “Thermochemical conversion of biomass to liquid fuels and chemicals”, Royal Society of Chemistry,
London, 2010,
211
[7] Directive 2009/28/EC of the European Parliament and of the Council, 23 April 2009, on the promotion of the use
of energy from renewable sources and amending and subsequently repealing Directives 2001/77/EC and
2003/30/EC, Official Journal of the European Union, L 140/16, 2009,
[8] Dutta, K., Daverey, A., Lin, J-G.: “Evolution retrospective for alternative fuels: First to fourth generation”,
Renewable Energy, Vol. 69, 2014, pp 114–122,
[9] Đurišić-Mladenović, N., Predojević, Z., Škrbić, B.: “Conventional and advanced liquid biofuels”, University of Novi
Sad, Faculty of Technology Novi Sad, Novi Sad, Serbia, Chem. Ind. 70, 2016, pp 225-241,
[10] Enguidanos, M., et al.: “Techno‐economic analysis of Bio‐diesel production in the EU”, a short summary for
decision makers, European Commission, Report EUR 20279 EN, 2002,
[11] Yugoslav Institute for Standardization: “Fuels for motor vehicles fatty acid methyl esters (MEMKs) for diesel
engines, requirements and test methods”, JUS EN14214, 2005,
[12] Lali, A.: “Next generation biofuels-India”, Scenario and areas for cooperation, India-EU and Member State
Partnership for a Strategic Roadmap in Research and Innovation, New Delhi, 12 November 2010,
[13] National action plan for the use of renewable energy sources of the Republic of Serbia - 2013 (In accordance
with the model provided for in Directive 2009/28 / EC - Decision 2009/548 / EC),
[14] Naik, S.N., Goud, V.V., Rout, P.K., Dalai, A.K.: “Production of first and second generation biofuels”, Renewable
and Sustainable Energy Reviews, Vol. 14, 2010, pp 578–597,
[15] Nigam, P.S., Singh, A.: “Production of liquid biofuels from renewable resources”, Progress in Energy and
Combustion Science, Vol. 37, 2011, pp 52–68,
[16] Petrović, M.B., Babić, I.M.: “Perspectives and dilemmas for the production and use of biofuels in Serbia”,
Termotehnika, Vol. 10, No. 1-2, 2013,
[17] Stojiljković, D.: “Alternative Fuels for the SUS Engines in the 21st Century”, Study, Faculty of Mechanical
Engineering, Belgrade, 2007,
[18] Škrbić, B., Đurišić-Mladenović, N.: “Gases of the greenhouse”, Gas 12, pp 11-15, 2007,
[19] Tešić, M., Kiš, F., Janković, V.: “Possibility of production and use of biodiesel in AP Vojvodina”, Monograph,
Vojvodina Academy of Sciences and Arts, Novi Sad, 2008, pp 193,
[20] Tešić, M., Kiš, F., Marinković, R.: “Possibility of production of raw materials for biodiesel in Vojvodina / Serbia
agriculture”, 1450-5029, 2009, pp 31-34,
[21] Decree on the conditions and procedure for acquiring the status of privileged electricity producer, "Official
Gazette RS", no. 8/2013,
[22] BP Statistical Review of World Energy, June 2017; Available from: https://www.bp.com/content/dam/bp/en/
corporate/pdf/energy-economics/statistical-review-2017/bp-statistical-review-of-world-energy-2017-full-
report.pdf, Accessed 07.08.2018.,
[23] EU -28, Biofuels Annual, 2018; Available from: https://gain.fas.usda.gov/Recent%20GAIN%20Publications/
Biofuels%20Annual_The%20Hague_EU-28_7-3-2018.pdf, Accessed 03.07.2018.,
[24] Statista; Available from: https://www.statista.com/statistics/271472/biodiesel-production-in-selected-countries/,
Accessed 07.08.2018.
212
Kragujevac, Serbia
MVM2018-062
Ivan Grujić1
Jovan Dorić2
Nadica Stojanović3
Aleksandar Davinić4
Radivoje Pešić5
Sunny Narayan6
NUMERICAL MODELING OF IC ENGINE COMBUSTION

PROCESS
ABSTRACT: Experimental research generally requires time and money. For this reason, numerical simulations
are increasingly being applied. A numerical modelling of the IC engine working cycle from the moment of intake
valve is closing until the exhaust valve opening is performed in this paper. The maximum pressure value obtained
by the experiment was selected as the control parameter of the simulation accuracy. 3D model of the engine
cylinder is created in the software package Ansys, which according to the dimensions and compression ratio
corresponds to the experimental. The boundary conditions obtained by the experiment were applied to the created
model. Based on the results, it was found that the maximum pressure value is very close to experimental.
KEYWORDS: experimental research, IC engine working cycle, 3D model, maximum pressure value
INTRODUCTION
Internal combustion (IC) engines have a long history followed by technological developments. Until today, the
adequate alternative for IC engines has not been found [1]. During the designing of new or improving existing
engines, the experiment is always considered an integral part of development. However, the application of the
experiment entails a series of inconveniences. Preparing the experiment itself requires a certain amount of time
and involvement of more people. Also, performing an experiment requires certain costs.
For these reasons, more and more methods are being adopted that can accelerate the engine design process, as
well to reduce the cost of the experiment. First, different mathematical models are used to describe the engine
cycle. Lately, as the latest in technology, numerical analyses are increasingly being used. The advantage of
numerical analysis is reflected in the possibility of visualizing of the engine cycle, while mathematical models give
only a numerical values.
1
Ivan Grujić, M.Sc., assistant, University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić STR., 34000
2
Jovan Dorić, Ph. D., assoc. prof., University of Novi Sad, Faculty of technical sciences, Trg Dositeja Obradovića 6
STR., 21101 Novi Sad, Serbia, jovan_d@uns.ac.rs
3
Nadica Stojanović, M.Sc., assistant, University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić STR.,
34000 Kragujevac, Serbia, nadica.stojanovic@kg.ac.rs
4
Aleksandar Davinić, Ph. D., assist. prof., University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić STR.,
34000 Kragujevac, Serbia, davinic@kg.ac.rs
5
Radivoje Pešić, Ph. D., full prof., University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić STR., 34000
Kragujevac, Serbia, pesicr@kg.ac.rs
6
Sunny Narayan, Ph. D., Qassim University, Faculty of Engineering, Saudi Arabia
213
Numerical methods, or more accurate the calculations, are based on the finite element method. The aim of this
paper is to compare the results obtained with by numerical analysis with the results obtained by the experiment, in
order to determine the accuracy of results that are obtained by numerical analysis. As the control parameters, the
pressure maximum and its position will be used.
It is generally known that when constructing any machine assembly, it is necessary to perform mechanical
calculations. In the case of mechanical calculations, the most critical are the maximum load values that occur in the
observed machine assembly. In the case of the IC engine, the greatest load that occurs during operation is
increase of the pressure of the gases in the cylinder of the engine.
Islam et al. [2], in their research, are performed numerical simulation of the gasoline IC engine cycle. As the output
parameters from the simulation, the values of the cylinder pressure, temperature and gas flow speed in the cylinder
were observed. They said that such simulations are of great importance in engine manufacturing and enable
visualization of the cycle in the cylinder.
Mauro et al. [3], in their research, used two approaches to determine the heat release during the combustion. The
first approach is the application of the so-called single-zone (SZ) thermodynamic model. It's a mathematical model.
The second approach is a CFD simulation, similar to simulation that will be performed in this paper. They
emphasized that the mathematical model requires less time. They also said that numerical analysis is enormously
larger then mathematical models. However, in contrast to these good features of this mathematical model, they
said that for the modelling of the heat release, the required input is cylinder pressure that is obtained
experimentally or in some other way. In contrast, the CFD simulation gives the pressure value in the engine
cylinder as well as the heat release.
Rakopoulos et al. [4], in their study, observed the influence of the shape of the intake port and the shape of the
combustion chamber on the gas flow speed. For the numerical simulation they used the geometry of the gasoline
engine. In their work, they also stated how much the emission is concerned today, so instead of gasoline, hydrogen
was used as fuel. The hydrogen injector was mounted in the intake port and the injection is performed during the
intake stroke.
Plengsa-ard and Kaewbumrung [5], in their research, are performed determination of heat transfer through the
cylinder walls. Unlike previous works and similarly to this, they were performed analysis for the diesel engine.
However, how they are observed only the combustion process, they are used only a section of the cylinder. This
approach is usually used to simplify the calculation, and to reduce time of the analysis. In order to determine the
accuracy of the analysis, they compared the calculated pressure and heat release with experimentally obtained. By
comparing these parameters with the experimental, they found that the analysis gives very accurate results.
How is earlier mentioned, the aim of this paper is simulation of the combustion process. However, unlike papers
where only a cylinder section is used, this simulation will be performed for the entire cylinder geometry. Also in
other studies, in most cases, the already existing geometries for simulation of engines are used. In our case, a 3D
model of IC engine geometry is made (Figure 1), which corresponds to the test engine according to the dimensions
of the cylinder and the piston chamber as well as the compression ratio, which is logic because dimensions are the
same as for test engine.
Figure 1 3D model
In order to be able to perform a numerical simulation, the 3D model in Figure 1 is created in the Ansys software
package, where and numerical analysis will be performed. The 3D model in Figure 1 represents the engine when
214
the piston is in the top dead point, while for the piston moving from the top dead point till the bottom dead point,
piston stroke and the connecting rod length are inserted as the input data for the analysis.
EXPERIMENTAL RESEARCH
The experimental research required for this research is based only on cylinder pressure measuring. Cylinder
pressure measuring was achieved by using the modern measurement installation that can be seen in Figure 2. It
can be divided into four parts:
 Analog part that consist piezo pressure sensor AVL QC32D and the signal amplifier KISTLER 5007;
 Digital A/D conversion control, consisting of: the optical marker of the crank angle and the position of the
top dead point AVL 365CC, the optical drive AVL 365.00 & 365.001 and the pulse multiplier AVL
365Z01M.
 Real-time signal processor with 8-channels A/D converter (AVL Indimeter 619) and
 Personal computer with support software AVL IndiCom 1.2.
Figure 2 Measurement installation
As the test engine, one cylinder diesel engine is used. Test engine is shown in Figure 3, and engine specifications
are given in Table 1.
Figure 3 Test engine
215
Table 1 Test engine specifications
Piston bore [mm] 85
Piston stroke [mm] 80
Connecting rod length [mm] 145
Compression ratio [-] 17.5:1
Engine power [kW] 7.4
Engine speed [rpm] 3000
Number of cylinders 1
Fuel Diesel
Injection angle [°CS] 18.5
NUMERICAL SIMULATION
Earlier in the paper it was noted that the numerical simulation of the combustion process will be performed.
However, more precisely, the simulation of the engine operation will be performed from the moment of intake valve
closing until the moment of the exhaust valve opening. More precisely, simulation of compression stroke,
combustion process and expansion stroke will be performed. The intake stroke and exhaust stroke will not be taken
into account because of numerical simulation simplification. After inserting the value of the piston stroke and the
length of the connecting rod, as well for the position of the injector, the 3D model in Figure 1 is decomposed and it
takes shape in Figure 4.
Figure 4 Simulation geometry
It can be notice that the present geometry does not have intake and exhaust port. The reason for this is that they
are not needed, and by their removal, the geometry and the calculation are simplified. Figure 4 also shows the fuel
nozzle, which is defined by the position of the injector. It can also be seen the head of the intake and exhaust
valve. These are necessary because both of these valves are closed during simulated strokes and combustion
process. It has been mentioned earlier that the software is based on the finite element method, so for the given
geometry, the mesh of finite elements is created and shown in Figure 5.
Figure 5 Engine geometry mesh
During the generation of the engine geometry mesh, the special option is used. It is assumed that the shape of
finite elements depends on the geometry shape. Number of mesh elements is 35732 and number of nodes is 7886.
Numerical analysis is performed for one engine working regime. Boundary conditions used for the simulation are
next:
 Engine speed n  2674 rpm
 Cylinder charge v  0.818
 Injection angle   18.5 °CS
216
 Injected fuel per cycle mg  0.0216 g/cycl.
RESULTS AND DISCUSSION

By applying previously mentioned boundary conditions, a numerical analysis was performed for the period from the
moment of the intake valve closing until the moment of the exhaust valve opening. The analysis was performed for
the diesel engine, and therefore in the software is defined that injected fuel is diesel fuel. Also, the ignition timing is
not set, but it is allowed to self-ignition of the fuel-air mixture, which corresponds to diesel engine.
As noted, as control parameters for the accuracy of the analysis, the maximum pressure value and the maximum
value are taken. Figure 6 shows the cylinder pressure, obtained by the numerical simulation.
Figure 6 Cylinder pressure
By analysing the diagram in Figure 6, it was found that the maximum pressure value is 70 bar, and the maximum
pressure position is at 363 °CS. From the experimental data, the maximum pressure value is 71 bar and is at 366
°CS. From these data it can be said that the error in the maximum pressure value is 1.4%, while the error in the
position of the maximum pressure value is 0.8%. Based on the error values, it can be said that the numerical
analysis gives correct results.
CONCLUSIONS
In mechanical engineering, the experiment was always an indispensable part of the research. However, as
experimental research entails time and money, so it is increasingly resorting to some other solutions. Numerical
analyses have become increasingly accepted today in the development of new products.
In this paper, the results of the numerical analysis are compared with the results obtained by the experiment, in
order to determine how accurate numerical results are obtained. Based on the selected control parameters, it was
determined that the errors are small, and as such, the analysis yields satisfactory results. However, it should be
taken into account that for the analysis, an engine model is constructed which by shape and dimension of the
combustion chamber, corresponds to a real engine.
ACKNOWLEDGMENTS
This paper was realized within the researching project “The research of vehicle safety as part of a cybernetic
system: Driver-Vehicle-Environment” ref. no. TR35041, funded by Ministry of Education, Science and
Technological Development of the Republic of Serbia.
217
REFERENCES
[1] Davinić, A.: “Identification of characteristics of multi-process operation of the piston IC engine”, Ph. D. thesis,
University of Kragujevac, Faculty of Engineering, 2013,
[2] Islam, A., Umer Sohail, M., Ali, S.M., Hassan, A., Kalvijn, R.: “Simulation of Four Stroke Internal Combustion
Engine”, International Journal of Scientific & Engineering Research, Vol. 7, Issue 2, 2016, pp 1212-1219,
[3] Mauro, S., Şener, R., Gül, M. Z., Lanzafame, R., Messina, M., Brusca, S.: “Internal combustion engine heat
release calculation using single‑zone and CFD 3D numerical models”, International Journal of Energy and
Environmental Engineering, Vol. 9, 2018, pp 215-226,
[4] Rakopoulos, C.D., Kosmadakis, G.M., Pariotis, E.G.: “Evaluation of a new computational fluid dynamics model
for internal combustion engines using hydrogen under motoring conditions”, Energy, Vol. 34, 2009, pp 2158-
2166,
[5] Plengsa-ard, C., Kaewbumrung, M.: “CFD modelling wall heat transfer inside a combustion chamber using
ANSYS forte”, Proceeding of 8th TSME-International Conference on Mechanical Engineering (TSME-ICoME
2017), IOP Conf. Series: Materials Science and Engineering, Vol. 297, 2018, Bangkok, Thailand.
218
Kragujevac, Serbia
MVM2018-063
Nenad Raspopović1
Jovan Dorić2
Ivan Grujić3
Nebojša Nikolić4
ANALYSIS OF HYBRID TURBOCHARGER IN MOTOR

VEHICLE IC ENGINE
ABSTRACT: Internal combustion engines are dominant motor vehicles propulsion systems. Development of IC
turbochargers is associated with improvement of their efficiency, dimensions, materials and operating characteristics.
In this paper we try to make analysis of working parameters of internal combustion engine with hybrid turbocharger.
To optimally support downsizing of IC engine turbocharger must react as fast as possible and the exhaust gas
volumetric flow to the turbine must be exactly adjustable. Results shown that application of such turbocharger leads
to a better engine performance and increase overall engine efficiency.
KEYWORDS: IC engine, hybrid turbocharger, motor vehicle
INTRODUCTION
A turbocharged engine are widely used today in modern motor vehicle. One of the reasons for massive production
of such engines are better efficiency and higher specific output power. Turbocharger or colloquially turbo is a turbine-
driven forced induction device that increases an internal combustion engine's efficiency and power output by forcing
extra compressed air into the combustion chamber. The engine maximum power is dependent on that amount of fuel
that can be burnt efficiently inside the cylinder. The injected fuel depends on the amount of air that can be available
at each cycle. Thus, by compressing the air to a higher density will allow more air inside the cylinder. Turbocharging
is one of the conventional approaches of increasing the air density by increasing its pressure before entering the
engine cylinder [1]. Besides many advantages over naturally aspirated engines, turbocharged engine also have some
disadvantages. More expensive and more complicated design of the engine and also reduced reliability are one of
them.
In conventional turbocharged engine, to avoid overboosting of the engine, it is necessary to activate waste gate valve
(valve that bypasses the turbo). Other solution is to use Variable-geometry turbochargers (VGTs). VTG are a family
of turbochargers, usually designed to allow the effective aspect ratio (A:R) of the turbo to be altered as conditions
change. This is done because optimum aspect ratio at low engine speeds is very different from that at high engine
speeds. If the aspect ratio is too large, the turbo will fail to create boost at low speeds; if the aspect ratio is too small,
the turbo will choke the engine at high speeds, leading to high exhaust manifold pressures, high pumping losses,
and ultimately lower power output. Both solution in some way decreases potential engine efficiency, and there is
more successful solutions for increasing turbo engine performance [2-5].
1
Nenad Raspopović, Faculty of technical sciences, University of Novi Sad
2
Jovan Dorić, Faculty of technical sciences, University of Novi Sad, jovan_d@uns.ac.rs
3
4
Nebojša Nikolić, Faculty of technical sciences, University of Novi Sad
219
The idea is to improve overall turbocharged IC engine by using electric generator as can be seen from Fig. 1. This
design flexibility leads to further improvements in turbine and compressor efficiency, beyond a conventional
turbocharger. Engine efficiency and power characteristics carry a major importance in the automotive industry as
they incorporate the two major requirements: fuel consumption and drivability. A major inconvenience for
turbochargers constitutes the low response in the low speed region of the engine creating the so called “turbo lag”
affecting drivability. Also, a large portion of the fuel energy is lost through exhaust gas in a SI engine affecting engine
efficiency, fuel consumption implicitly. A hybrid turbocharging assist system is a promising technology for improving
both transient response and engine efficiency [1-5].
Figure 1 One possible solution for hybrid turbocharged [6]
MATERIAL AND METHODS

In this paper analysis of hybrid turbocharged engine was performed with software GT-POWER. GT-POWER is the
industry standard engine performance simulation, used by all major engine manufacturers and vehicle OEMs. GT-
POWER is used to predict engine performance quantities such as power, torque, airflow, volumetric efficiency, fuel
consumption, turbocharger performance and matching, and pumping losses. Beyond basic performance predictions,
GT-POWER includes physical models for extending the predictions to include cylinder and tailpipe-out emissions,
intake and exhaust system acoustic characteristics (level and quality), in-cylinder and pipe/manifold structure
temperature, measured cylinder pressure analysis, and control system modeling. Some basic GT-POWER equations
will be shown.
The first equation ensures the conservation of mass as seen in (1).
(1)
Related to the mass equation is the conservation of momentum relationship shown in (2).
(2)
These two equations are predominately used to determine the trapped air and residual gas masses. They are also
used to determine the fuel dynamics and mass flow rates. GT Power also uses (3) which ensures that energy is
conserved.
(3)
The final equation is an exergy balance as shown in (4) which is implicitly solved.
(4)
Using these equations, the heat transfer from a volume to the walls or another volume can be determined. These
final two equations determine the gas temperatures which also influences the trapped air and residual masses.
220
For our research two engines are analysed. First part of research was focused on F1 hybrid turbocharged engine. In
order to match criteria of fuel consumption of 100kg/h we try to find how much we can get in power increase with
adding a hybrid turbo. Second phase of our research was performed with aim to achieve better fuel consumption in
ordinary motor vehicle engine, in our case 2 litre turbo petrol engine.
In this paper is proposed a turbocharging method [7,8] which is based on the engine characteristics, compressor
performance map, first turbocharger equation and turbine performance map. Engine operating characteristics such
as low-end torque, maximum torque, and maximum power will determine the operating points in the compressor
map. By having the operating points of the turbocharger in the compressor performance map with the calculated
pressure ratio and mass flow rate the first turbocharger equation can be used. The value of the δ contains the
efficiency of the turbocharger, exhaust and intake mass flow rates and the temperatures of the intake air and exhaust
gas. These are taken from the working conditions of the engine and the efficiency can be measured or guessed.
Having δ and compression ratio, πc, the expansion ratio of the turbine can be determined according to the diagram.
From this point on, operating points in the turbine map can be determined. This matching procedure is iterated until
the guessed values, such as the turbine and compressor efficiencies etc., are converged. The whole schematic
process is shown in Fig 2.
Figure 2 Schematic process of turbocharging matching [7]
RESULTS
First engine model was presented on Figure 3. It can be seen from model schematic view that here we have six-
cylinder engine with built-in hybrid turbo equipment. From schematic view it is clear that engine have two motor-
generator (MG) unit. First MG unit is coupled with turbo and second with crankshaft. This system work in such a way
that the surplus energy that is taken on the turbine can be utilized in a better way through the MG units via ECU of
the engine.
221
Figure 3 Schematic view of F1 hybrid turbo engine
Simulation results of power, torque and BMEP curves can be seen on Figure 4, Figure 5, and Figure 6. For the case
when F1 engine was analysed two situations are presented. Red curves refer to the engine with ordinary
turbocharging system and the blue ones are reserved for simulation results of hybrid turbo engine. It can be
concluded from the proposed diagrams that hybrid turbo engine has better examined characteristics.
222
Figure 4 Power curves
Figure 5 Torque curves
223
Figure 6 BMEP curves
Similar simulation is done but in this case on another engine model. As can be seen from Figure 7 four-cylinder motor
vehicle engine was presented in GT-POWER software.
Figure 7 Schematic view of four-cylinder 2.0 hybrid turbo engine
224
In the second phase of our analysis we try to make simulation on the 2.0 turbo petrol engine. In this case the
simulation was done with the condition of giving equal power at the motor output (Figure 8) for two cases. First is
performed with hybrid turbo and the second with conventional turbocharged engine.
Figure 8 Power curves for two investigated cases: blue for hybrid turbo and red for conventional turbocharged
engine
Figure 9 Fuel consumption curves for two investigated cases: blue for hybrid turbo and red for conventional
turbocharged engine
The results show that the mass flow rate of the hybrid turbocharged engine decreasing initially and then starts
increasing with the further increase in the engine speed (Figure 9). The mass flow rates changes with the
performance parameters like volumetric efficiency, in-cylinder temperature and engine speed.The results of the
simulation show that in conventional motors with hybrid turbo we can also expected increased efficiency.
225
CONCLUSIONS
A hybrid turbo, if designed properly, will allow the extra flow and pressure to be achieved at a safe and reliable turbine
speed and at higher compressor efficiencies. The temperature and pressure also gets increased further with the
amount of supercharging. At high temperatures, the dissociation of the gases occurs by absorbing the thermal energy
released during the combustion, thus reducing the thermal efficiency and power produced by the engine. But at
higher boost pressures, the effect of dissociation on power will overcome by the effect of boost pressure. Thus the
power produced by the hybrid turbocharger will be more than that of the conventional turbocharger even though the
thermal efficiency is lesser. In this paper investigation the potential of using the electric motor to provide directly
power to the crankshaft and test the viability of the solution and how it influences the way that the compressor is
operating was done. A study in order to know when the switching between crankshaft assist and turbocharger assist
has to be made in order to obtain optimal performance was also done in this paper. With motor-generator unit added
to turbocharged engine in motor vehicle we can expect better efficiency and higher output power at wide range of
engine speed. Overall, the system brings some benefits but a more detailed analysis has to be made to see if the
actual design is the optimal one.
ACKNOWLEDGMENTS
This research was done as a part of project TR31046 “Improvement of the quality of tractors and mobile systems
with the aim of increasing competitiveness and preserving soil and environment”, supported by Serbian Ministry of
Science and Technological Development.
REFERENCES
[1] Hiereth, H., Prenninger, P.: “Charging the Internal Combustion Engine, Powertrain”, Springer, New York, 2007,
[2] Chadwell and Walls. Analysis of a super turbocharged downsized engine using 1-D CFD simulation, SAE
Technical Paper; 2010-01-1231,
[3] Kusztelan, A., Yao, Y.F., Marchant, D.R., Wang, Y.: “A Review of Novel Turbocharger Concepts for
Enhancements in Energy Efficiency”, Int. J. of Thermal and Environmental Engineering, Vol. 2, No. 2, 2011, pp
75-82,
[4] Shahed S.M.: "The Power of Turbocharging. Industry executives discuss the future of mobility technology", SAE
International, 2005, 86-87,
[5] Seiichi I., Yukio, Y., Kunio, S., Hiroshi, O., Yasuaki, J.: "Development of hybrid turbo, an electrically assisted
turbocharger", Mitsubishi heavy industries Ltd, technical review, Vol. 43 No. 3. 2006,
[6] Jinsha, R., Hans, M.H., Josephb, A., Kiran, T.S., Gireesh, K.T.: "Hybrid turbocharged SI engine with cooled
exhaust gas recirculation for improved performance”, Procedia Technology, Vol 24, 2016, pp 444 – 451,
[7] Alin-Gabriel, D.: "Study on engine efficiency and performance improvements through hybrid turbocharging
assisting”, Master’s thesis in Automotive Engineering, Chalmers University of Technology Gothenburg, Sweden
2014,
[8] Dorić, J.: “Unapređenje efektivnosti motora SUS primenom nekonvencionalnog klipnog mehanizma”, doktorska
disertacija, FTN Novi Sad, 2012.
226
Kragujevac, Serbia
MVM2018-064
Saša Milojević 1
Radivoje Pešić2
Aleksandar Davinić3
Dragan Taranović4
Snežana Petković5
Emil Hnatko6
Radmilo Stefanović7
Stevan Veinović8
INFLUENCE OF VARIABLE COMPRESSION RATIO ON

EMISSION AND VIBE FUNCTION PARAMETERS OF
EXPERIMENTAL ENGINE
ABSTRACT: Compression ratio as a constructive parameter has a very important influence on fuel economy,
emission and other performances of internal combustion engines. Introduction of variable compression ratio has
the number of benefits: limiting maximal in-cylinder pressure, extended field of the optimal operating regime with
minimal fuel consumption, more power, reduced exhaust emission and controlled noise emission. It is particularly
characteristic for multifuel capability of engines with variable compression ratio (VCR), along the control of
maximum temperature and pressure in the cylinders. The paper presents engines and mechanisms for automatic
change of compression ratio value. Modelling of operation process of internal combustion (Otto and Diesel)
engines with direct injection has been performed in parallel with experimental research. The basic problem is the
selection of the parameters in double Vibe function used for modelling of the diesel engine operation process, also
performed for different compression ratio. The optimal compression ratio value was defined regarding the minimal
fuel consumption and exhaust emission. For this purpose, the test bench is brought into operation in the Laboratory
for Engines of the Faculty of Engineering, University of Kragujevac.
KEYWORDS: IC engine, emission, variable compression ratio, Vibe function
1
Dipl.Ing. Saša Milojević, assistant R&D. University of Kragujevac, Faculty of Engineering, Serbia, Kragujevac,
Sestre Janjić 6, sasa.milojevic@kg.ac.rs
2
Dr.Ing. Radivoje Pešić, prof. University of Kragujevac, Faculty of Engineering, Serbia, Kragujevac, Sestre Janjić
6, pesicr@kg.ac.rs
3
Dr.Ing. Aleksandar Davinić,assist. prof. University of Kragujevac, Faculty of Engineering, Serbia, Kragujevac,
Sestre Janjić 6, davinic@kg.ac.rs
4
Dr.Ing. Dragan Taranović, assoc. prof. University of Kragujevac, Faculty of Engineering, Serbia, Kragujevac,
Sestre Janjić 6, tara@kg.ac.rs
5
Dr.Ing. Snežana Petković, prof. University of Banja Luka, Faculty of Mechanical Engineering, Republic of
Srpska, Banja Luka, Bulevar vojvode Petra Bojovića 1A, petkovic1961@gmail.com
6
Dr.Ing, Emil Hnatko, University of Osijek, Mechanical Engineering Faculty in Slavonski Brod, Croatia,
ehnatko@sfsb.hr
7
Dipl.Ing. Radmilo Stefanović, Consultant. Serbia, Belgrade, rad.stefanovic@sbb.rs
8
Dr.Ing. Stevan Veinović, retired prof. University of Kragujevac, Faculty of Engineering, Serbia, Kragujevac,
Sestre Janjić 6, stevan@kg.ac.rs
227
INTRODUCTION
City traffic and traffic flow have the greatest impact on the exhaust emission and air pollution, specifically in the
street canyons, zones of city centres, etc. Moreover, in city area, the traffic can contribute substantially in declining
of exhaust emission, by application of various methods. According to our practical experience, there are currently
two real directions:
1. to switch on alternative fuels, natural gas and hydrogen as driving energy, with parallel introduction of
flexible transport, which enables the decrease in the number of vehicles in urban centres to avoid
congestion [18] and
2. further tribological optimization of conventional internal combustion (IC) engines by lowering internal friction
and mechanical losses, in order to reduce fuel consumption and exhaust emission [911].
Regarding the first proposed direction, nowadays, application of gaseous fuels such as hydrogen and natural gas
as well as their mixtures is more and more interesting in optimized IC engines [1219].
Natural gas is a highquality fuel for motor vehicles. Above all, available reserves equal the known oil reserves;
negative influence upon environment is less than with fuels derived from oil, as well as is the price. Thus, natural
gas as engine fuel has been having a growing application in motor vehicles, which is further proved by the fact that
there are over 26 million vehicles in the world powered by natural gas (1.86 million in Europe). Currently, there are
878 natural gas vehicles (792 passenger and light duty vehicles, 58 buses and 28 medium and heavy duty trucks)
in traffic on the roads in the Republic of Serbia [20].
In terms of engine related criteria, natural gas may be used as the sole fuel for Otto or diesel engines (with certain
engine reconstruction), firstly in engines with automatic variable compression ratio (VCR). In addition, it may be
used in mixture with every kind of oil or proved alternative fuel (lower or higher alcohols, biogas and other) [2124].
In normal ambient conditions, natural gas has a low density of energy per unit volume. Thus, in case of application
of natural gas as a motor fuel, it is necessary to have a certain treatment of natural gas in order to have a higher
concentration of energy by unit volume. Given that it is expected to provide a wider radius of vehicle movement
with one fuel tank supply, it is necessary to have operating pressures of around 200 bar. The simplest way is to
compress natural gas in high-pressure tanks (CNG  compressed natural gas). The second alternative for
increasing the density of energy of natural gas inside tanks on vehicles is its extreme cooling down to 162C, i.e.
conversion to liquefied state (LNG  liquefied natural gas) and storing in cryogenic tanks [25].
In California (California Environment Protection Agency  EPA) and Japan, there are standards for natural gas as
the engine fuel. In the United States federal regulations, the natural gas composition is defined as a reference fuel
for testing exhaust emissions, Table 1 [2527].
Table 1 Specification of the natural gas according to the California EPA standard (01.05.1999.) and the reference
fuel to the Federal EPA characteristics
California EPA Federal EPA Test method
Specification Value
Methane 88 mol.% (min.) 89 mol.% (min.)
Ethane 6.0 mol. % (max.) 4.5 mol. % (max.)
C3 higher HC 3.0 mol. % (max.) 2.3 mol. % (max.)
ASTM D 1945-81
C6 higher HC 0.2 mol. % (max.) 0.2 mol. % (max.)
Oxygen 1.0 mol. % (max.) 0.6 mol. % (max.)
CO2 and N2 1.5-4.5 mol.% 4.0 mol. % (max.)
Hydrogen 0.1 mol. % (max.) -
ASTM 2650-88
CO 0.1 mol. % (max.) -
It should not contain dust, sand, mud, rubber, oil and other substances that could damage
Particulate matter
the equipment or the charging station equipment of the vehicle
Sulphur 16 ppm vol. (max.) - Title 17 CCR Section 94112
Odor Unpleasant and specific in 1/5 lower limit of inflammation
Optimistic studies about natural gas are based on pure methane. Unfortunately, pure methane is nowhere to be
found in nature. It is always in mixture with other impurities in concentration of various percentages. In order to use
natural gas as a fuel, it is necessary to filter it until it reaches more or less the established composition and quality.
There are some national standards that regulate quality of natural gas as engine fuel: SAE J1616, DIN 51624 and
13 CCR § 2292.5 (California regulations). In parallel, there is an international standard ISO 15403, accepted also
by the Republic Serbia. The purpose of this standard is for producers, vehicle users, filling station operators and
other logistics participants in the industry to get informed about quality of fuel for natural gas vehicles.
228
In the second case, the development of vehicle powertrains is increasingly challenged by emission legislation and
by the “end-users” fuel economy demands. In order to meet these requirements it is necessary to continuously
improve existing powertrains and to develop totally new generations of engines, while still giving space for
individual brandspecific features.
According to theoretical studies, the overall effect of IC engines (either Otto or diesel) can be improved, since Otto
engines can be improved up to 40% and the best diesel engines up to 50%. Efficiency of over 50% is achieved
only by two-stroke marine diesel engines (55%) [11, 28].
The current IC engines have serious reservations to improve fuel economy, Figure 1. Average efficiency at the
optimum point is about 30% in automotive engines and about 45% in marine engines [11].
Figure 1 Mercedes-Benz completes vehicle ML 300 CDI and engine rebuild 2.4 l per 100 km.
Percent reduction in fuel consumption at 2.4 l per 100 km compared to the previous model ML 300 CDI:
36% f ourcylinder diesel engine; 16%  eco startstop system; 14%  7GTRONIC PLUS automatic
transmission with fuel economy converter; 12%  reduction of the driving resistance; 7%  electromechanical
steering system; 5%  consumptionoptimized axle drive; 3%  generator management; 3%  regulated fuel pump;
2%  optimized belt drive with decouple; and 2%  clutch and air conditioning compressor
For vehicles, working range loadengine speed at minimum fuel consumption is very close. Its expansion is a
chance for engines with variable systems. A further advantage of such engines is the ability to work with all types of
available liquid and gaseous fuels with the fulfilment of environmental regulations. Energy efficiency propulsion
system will be the number one priority for the future, a model for the comparison of the Airbus A380, whose engine
consumes less then 3 l per 100 km of fuel per passenger (78 passengermiles per US gallon). Criteria for
environmentally friendly drives must be defined in accordance with scientifically based facts [11].
The fourstroke engine known to date, have come from Nikolaus August Otto more than 150 years ago. The
magical limit of about 12.5:1, as the optimum combustion chamber defined by a fuel quality, is above the range
reserved for diesel engines [11].
The starting point is the fact that the mechanical efficiency of the IC engines is dependent on energy losses
associated with friction i.e. the conditions inside tribological system which consists of: piston, piston rings and
cylinder liner. Generally, mechanical efficiency lowers in case of raising the compression ratio (CR) and vice versa.
This leads to a minor mechanical problems, but achieves lower fuel consumption. This is the future of the gasoline
engine - to finally equalize the consumption of diesel engines [2935].
The problem is in the fact that the geometric CR of IC engines is set as a compromise between partial and full-load
requirements. The demand for the highest geometric CR to improve high partload efficiency is opposed by a
knock (gasoline engines) or peak pressure limitations with the Diesel engines.
When CR is raised from 10 to 15:1, efficiency increases by more than 20%! The boundaries of efficiency of heat
engine are about 55%, which results in the consumption of about 160 gkWh1. Currently, most car engines have
the CR around 11:1, efficiency of about 35% and consumption of approximately 250 gkWh1, but in a very narrow
field [36].
229
The increasing load (BMEP  brake mean effective pressure) and power ranges enhanced the compromises to be
taken with deciding on a fixed CR. With gasoline engines, a quite clear correlation between the maximum load and
geometric CR ratio is evident even with a simple statistical observation of actual production engines [37].
The downsizing technology in the recent years has even enhanced this tradeoff. Compared to naturally aspirated
(NA) engines with direct injection (DI) or multi point injection (MPI), the higher specific power and BMEP levels of
turbocharged engines require a lowered CR at full load, while the demand for high CR at partial loads is highly
exacerbated by the fuel consumption requirements. Looking at the distribution of the specific power versus engine
displacement, high performance is no longer restricted to multi cylinder sport cars, Figure 2 [37].
Figure 2 Specific powers of gasoline engines and development trends
According to previous, VCR method is used to determine optimal CR for gasoline and diesel engines under
different loads [38, 39] in combination with other variable systems on engine, such as: flexible intake valve
actuation [40], exhaust gas recirculation (EGR) [41], variable fuel injection time [42, 43], etc. Generally, with
integrated system for VCR, it is possible to increase engine efficiency, power and torque. The engine can operate
under multiple cycles with multifuel capability i.e. well adjust to knock limitations when using various fuels.
Optimization of IC engines is related to equipment through which natural gas (or any other alternative fuel) is
directly injected into the combustion chambers. Compression ratio is self-regulated and it adapts to engine fuel
quality and manages workflows according to the basic requirements: economy, power, torque, ecology, noise and
so on. Today's technical development allows simultaneous management of combustion processes by Otto and/or
Diesel cycle, according to the driving conditions. Such engine freedom offered technological relief for oil industry in
the formulation of a new (and the same for Otto and diesel engines) energy efficient and environmentally friendly
fuel, Figure 3.
Figure 3 New concept of multi-process IC engine
230
The main reason for the growth in engine modelling activities arises from the economic benefits. By using computer
models, large savings are possible in expensive experimental work, when engine modifications are being
considered. Models cannot replace real engine testing but they are able to provide good estimates of performance
changes resulting from possible engine modifications and, thus, can help in selecting the best options for further
development, reducing the amount of hardware development required.
While the more advanced models are extremely large and complex, the basics of an engine thermodynamic model
are quite straightforward and easily understood. The complexity arises later in the refinement of the calculation
methods, the level of detail of sub-system representation and the accommodation of a wide variety of alternative
engine configurations and control systems.
The authors have also successfully studied the influence of CR on combustion process in experimental engine.
Validation of the results received by using the model of the diesel process with double Vibe function was carried
out in parallel by their comparison with relevant results obtained during experimental measurements [1].
The main goal of the paper is to determine the influence of CR and engine working regimes on the combustion
characteristics and shape parameters of double Vibe function, as well as, on the fuel consumption and exhaust
emission. Detailed results of experimental measurements are presented in [1].
OVERVIEW OF VCR MECHANISMS

All companies working on development projects of VCR engine state that applying VCR engines can solve the
problems of economy and ecology. The operating versions are usually systematized according to the similarities.
Numerous construction solutions of VCR engines and systems can be found in the literature. Certain constructions
have been practically realized. All kinematic solutions can be used as combined patent rights, and sometimes as a
serious contribution to constructive thinking. VCR mechanisms can be realized with variable cylinder head volume
or with variable position of piston top dead centre (TDC), Figure 4 [37].
Figure 4 Variants of variable compression ratio:

1) Variable volume in combustion chamber: (a; b) secondary piston for adjust the clearance volume;
2) Variable cylinder height: (a) cylinder moves up by translation; (b) cylinder moves up by rotation;
(c) eccentric crankshaft, effectively adjusts link lengths;
3) Variable compression height: (a) Piston with variable compression height; (b) - eccentric piston pin;
4) Variable connecting rod length: (a; b) - eccentric piston pin and crank pin; (c) - connecting rod with
integrated hydraulic piston; (d) - linkage, engine with three or more connecting rods;
5) Variable crankshaft position; (a; b) - eccentric crank pins.
231
TDC piston position can be regulated by using an unconventional crankshaft in combination with a two-piece
connecting rod, which is controlled by an additional shaft, or by using a rack-and-pinion gear for the transmission of
power from the piston to the crankshaft. Alternatively, retaining a conventional crankshaft drive, the TDC piston
position can be varied by modifying the distance between the crankshaft and the cylinder head or by varying the
kinematic effective lengths of the crankshaft drive. The distance between the crankshaft and the cylinder head can
be changed by tilting the cylinder head together with the cylinder barrel relative to the bearing pedestals, or by
means of a translator mechanism acting on the cylinder head and barrel unit. The required distance change can be
realized also by utilizing an eccentrically supported crankshaft. The variation of cinematically effective lengths
opens up the widest range of constructive possibilities - compression height, connecting rod length and crank
radius can be modified by means of eccentric bearings or by using a linear guidance device.
VCR is an old idea, but still not implemented in the realization of larger series. Thus, if we hear now from Infiniti that
they have found a solution to incorporate it in the series from 2018, we may perceive it with due scepticism. Here,
Atkinson process is used as a way for changing the effective piston stroke.
The approach of Infiniti divides the connecting rod into two parts. However, the crankshaft remains as an
intermediary of the torque to the flywheel. During the run time of the engine, the angle of the two parts can be
changed and, thereby, the connecting rod can be shortened or lengthened. The principle is clearly visible in Figure
5.
This engine mechanism promises higher CR (14:1) for fuel efficiency under low load and lower CR (8:1) to aid
turbocharging, for more power at higher engine speeds, Figure 5 [44].
Figure 5 Infiniti VCT engines with VCR mechanism

Comparison of VCT technology in high (14:1) and low (8:1) compression ratios
Because of its classic construction, IC engine can not provide more complete expansion of combustion products in
the cylinder. Some researchers have tried to solve this problem. The result is a concept of quasi-constant volume
spark ignition engine (QCV SI). In this case, movement of the piston is not linear, because, in this engine, a
mechanism forces the piston to move over curve. Piston is moving along a very complex trajectory and this curve is
very similar to ellipse [45].
As an example, in the area of transport vehicles and heavy-duty engines, Caterpillar has developed VCR engine
for homogeneous charge compression ignition (HCCI) combustion. The result is the 15-liter engine with automatic
compression ratio change, from 8 to 15:1. Crankshaft is mounted eccentrically in the cradle, which sits in the
cylinder block. Rotating cradle moves the crankshaft, which alters CR [4648].
During research an A engine) was developed, with patented mechanism for the continual change of the CR
(CVCR) including two pistons for each cylinder (pistons have relative stroke) for adjusting the cylinder volume,
Figure 6. The idea comes from the two-stroke engines with paired cylinders. The position of the paired cylinders is
in the form of an A letter, and that is the origin of the patents name. Compression ratio changes with the
asynchronous movement of the pistons in the cylinder pair of the engine with two crankshafts. The driving and
driven crankshafts are interconnected by the pair of cylindrical gears. The asynchronicity of the pistons in the
cylinder pair is controlled by the appropriate regulator, which is mounted on one of the crankshafts [3, 4, 36].
232
Developed VCR engine mechanism can be used in all modern IC engines. In the framework of the EUREKA
project led by prof. Pešić, aluminium engine versions were realized [3].
a) b)
Figure 6 a) Cross section of “A” engine and b) Photography of main engine parts which made of aluminium
FUEL INJECTION TECHNOLOGY

For gasoline engines in passenger cars, the most important task is to improve fuel efficiency. Therefore, a big
variety of different technologies potentially can be applied. The system ranges from simple variable charge motion
and low cost variable valve timing devices up to high sophisticated systems like fully variable valve actuation
systems and combustion with auto ignition HCCI. Direct gasoline injection (DGI) systems of generation 1 (wall
guided systems) and even more the systems of generation 2 (spray guided systems) improve fuel efficiency, but
the significant costs for NOx exhaust after treatment have to be taken into consideration [49].
Due to its full load benefits, homogeneous DGI is a preferred solution for high performance engines as well as in
combination with turbocharging for downsizing concepts. The combination of turbocharging, direct injection and
cam phase shifter has proven to be a highly attractive package combining good fuel economy with fun to drive.
Different gasoline engines technologies will have to be applied according to the specific needs of their application
and brand specific requirements. Even keeping high performance characteristics, fuel consumption will
continuously be reduced and future legislative limits can be met. However, system complexity and cost will
increase.
Spray guided DGI combustion systems

The very first DGI concepts, developed even before the wall and the air guided systems, featured a close spacing
between injector and spark plug, Figure 7. However, at that time, such systems were still handicapped by
insufficient injection technology. Now these systems are recovering their importance due to the significant progress
in fuel injection technology. The close spacing results in excellent stratification capability and higher fuel efficiency
improvement compared to alternative concepts [49].
Both stratification injection and mixture formation are largely controlled by the injector itself. Due to the small
distance between the injector and the spark plug, the time available for mixture formation is significantly reduced.
Consequently, the requirements for the injector are higher than with wall or air guided systems.
With conventional spark ignition, the spark has to be arranged in the rim zone of the injection spray to avoid spark
plug fouling. Consequently, the spray shape, especially the spray angle, has to be kept largely constant under all
engine operation conditions and thus independent from injection timing and the respective backpressure. As a
consequence, swirl injectors, as widely used for generation 1 systems, are not a preferred solution for spray guided
systems with conventional spark ignition. Injectors such as piezo injectors currently provide specific advantages in
mixture formation and application.
233
Combustion chamber geometry as well as charge motion can be designed with much less compromises
concerning full load requirements and piston weight than with generation 1 combustion systems.
Exhaust
Injector
Spark plug Intake
Figure 7 Arrangements for spray guided DGI combustion systems
However, the central arrangement of the injector usually requires much more significant modifications of the
existing cylinder head designs than only adapting a base MPFI cylinder head towards a wall guided DGI variant.
The central position of the injector also raises injector tip temperatures and consequently deposits formation risk at
the injector. Both deposit formations as well as spark plug durability are major development issues.
As a future alternative, laser induced ignition might be an attractive alternative, Figure 8 [49].
As the ignition point is only determined by the focus of the respective lens and not by the protrusion of the spark
plug, the ignition point can be directly set within the main spray cone. Thus, from the viewpoint of reliable ignition,
even swirl injectors would be a feasible solution. Currently, such ignitions systems just have proven principal
function at a research level and still require significant further development efforts.
Alternative combustion systems, such as AVL CSI concept, may become option for lean combustion without need
of expensive DeNOx after treatment device, probably in combination and addition to variable valve train systems.
Apart from that, the entire automotive world is looking for other alternatives like Hybrids and Fuel Cells. The future
will tell whether these alternatives will be able to compete with conventional powertrains.
Figure 8 Laser ignition with spray guided DGI
EXPERIMENTAL SETUP
The experiments were carried out in the Laboratory for Engines of the Faculty of Engineering, University of
Kragujevac (FIN). Single cylinder, four-stroke, air cooled diesel engine, DMBLombardini type 3LD450, was used
during experiment (direct fuel injection system, unit pumppipeinjector, injector with four nozzles, 6 kW rated
power, bore 80 mm, stroke 85 mm, stroke volume 454 cm3, valve train DOHC, two valves per cylinder) [1].
The geometric value of CR (ε) is varied from 17.5 to 12.1:1 by replacing the pistons with different piston bowl
volume (realized by changing the piston bowl diameter from 43 mm to 55 mm), Figure 9. It is well known that the
piston bowl geometry design affects the air-fuel mixing and the subsequent combustion and pollutant formation
processes in the DI diesel engines. In this paper, all investigations results were solely related to the CR [1].
During experiments, the engine operated with standard diesel fuel with characteristics specified in Table 2 [1].
234
Table 2 Fuel characteristics
Description Values
Cetane rating (CN) 52
Specific density at 20 °C, (g/cm3) 0.839
Kinematic viscosity at 20 °C, (mm2/s) 3.964
Sulphur content (%) 0.5
The test rig is equipped with the measuring and data acquisition system, Figure 10. Various sensors are mounted
on the engine to measure different parameters. A thermocouple was installed on the surface of high pressure fuel
pipe. A precision crank angle (CA) encoder was coupled with the crankshaft of the engine. The software stores the
data of pressures and volumes corresponding to a particular CA location for plotting the indicator diagram curves.
The cylinder pressure is measured using the AVL QC32D water-cooled piezoelectric transducer. The signal of
pressure is amplified with the Kistler 5007 charge amplifier and it was processed by using the AVL IndiCom
Indicate Software Version 1.2. The software provides the facility for analysing the combustion data, such as the
rate of heat release, ignition delay, combustion timing in degrees and peak pressure and stores them separately for
analysis in the acquisition system [1].
The experimental engine is tested on the SCHENK U116/2 engine dynamometer. Tests are carried out for CR
values of 12.1, 13.8, 15.2 and 17.5:1. Working regimes for fuel consumption and exhaust gas analyses are defined
according to the European Stationary Cycle (ESC) 13 mode cycle. Analysis of the combustion process is
performed for the operating regime shown in Table 3 [1].
Table 3 Working regimes for combustion analysis

Engine speed, Mass of fuel per
BMEP, MPa
rpm cycle (mg/cycle)
0.12 7
0.24 10
1600
0.36 13.5
0.48 18
Figure 9 Pistons before experiments

(the values of the geometric CR for different diameters and Figure 10 Schematic diagram of the experimental
volume of the combustion chamber in the piston head) setup
Exhaust emission is analysed with AVL Dicom 4000 measurement equipment. Particulate matter (PM) emissions
are determined indirectly through the empirical correlation between the measured values of smoke and PM. The
smoke is measured using the AVL 409 equipment according to BOSCH method. Specific emission of exhaust
gases is calculated using the obtained data of exhaust emission and measured engine power at corresponding
working point. The final emission results are expressed in gkW–1h–1 [1].
Modelling the combustion process of diesel engine having VCR

The combustion characteristics can be compared by means of cylinder gas pressure, rate of heat release, ignition
delay, etc. The normalized heat release rate (NHRR) or differential combustion law is an important parameter for
analysis of the combustion process in the cylinder, Figure 11.a. The important combustion process parameters,
such as combustion duration and intensity, can be easily estimated from the rate of the heat release diagram,
Figure 11.b [1].
235
a) b)
Figure 11 a) Normalized rate of heat release during combustion and b) Combustion process in DI diesel engine
In analysed engine regimes, the ignition delay period is too long when the engine works under lower CR value. The
longer ignition delay period results in higher proportion of the injected fuel remaining unburned. The large
accumulation of unburned fuel during the ignition delay period leads to a characteristics sharp peak, Figure 11.a. At
moderate decrease of CR at higher loads (BMEP=0.48 MPa), the decrease of maximal cylinder temperature
occurs, Figure 14. During further decrease of the CR, the maximal temperatures start to increase. Explanation is
found in considerable increase of the combustion law maximum due to the increased ignition delay, Figure 12.a,
induced by lower temperatures at the fuel injection timing. It leads to the increased amount of fuel burned during
premixed combustion [1].
At low loads (BMEP=0.24 MPa), when the CR decreases, maximal values of the combustion law continually
decrease, Figure 11.a, which reflects in the reduction of maximal cylinder temperature values [1].
In diesel engine, cylinder pressure depends on the burnt fuel fraction during the premixed burning phase, that is,
initial stage of combustion. Cylinder pressure characterizes the ability of the fuel to mix well with air and burn
condition. The experimental engine has an old fuel injection system with low injection pressure. Since no new
injection system was available, it was decided to do the testing with the existing fuel injection systems. Therefore,
all the results have been obtained with a low injection pressure.
Vibe function is one of the most famous equations or functions used for modelling the combustion process in IC
engines. Vibe function is often used to approximate the actual heat release characteristics of an engine. Integral of
the Vibe function gives the fraction of the fuel mass that has been burned since the start of combustion.
In this paper, the methodology of modelling combustion process is described, with purpose to see how the working
regimes and CR value affect the double Vibe function parameters. If operation process of diesel engine with DI and
premixed combustion is modelled, the use of double Vibe function is necessary.At DI diesel engines with distinct
explosive combustion and diffusive combustion, a model based on one Vibe function cannot describe the
characteristic form of the combustion process. In that case, it is necessary to use a model based on superposition
of two Vibe functions; one simulating the explosive part (index “1”), the other simulating the diffusive part (index
“2”), Equations (1) and (2), Figure 11.b [1, 50, 51].
m1 1
m1  
   1 
m 1 dx1    C  1 
  z1 
C  1    g  C (m1  1) 1  e

x1   g 1  e   z1       z1 
  d  1  (1)
    z1 
m2 1
m1  
   2 
m 1 dx2   C  2 
 z 2 
C  2   (1   g )  C (m2  1) 2  e
x 2  (1   g ) 1  e     z2
 z 2
 2   (2)
  d  
   z2 
Double Vibe function is expressed as the sum of the first and second part of the Vibe function:
x  x1  x2 and dx dx1 dx2 ,

  (3)
d d d
where:
236
 𝑥  cumulative normalized heat released (mass fraction burned);
 𝛼  crank angle, CA;
 m1, m2  Vibe function shape parameter (for first and second function part);
 1, 2  angle between initial and current time of the first and second Vibe function;
 z1, z2  duration angle of the each simple Vibe function (duration of the heat release);
 C  Vibe function parameter (𝐶 = 6.908); and
 g  share of fuel mass burnt during the first Vibe function (explosive combustion).
In order to establish the influence of the CR on the parameters of double Vibe function, corresponding experimental
and theoretical investigations of the diesel engine with CR change were performed. The research results are
shown in Figure 12. As it can be seen, a good match between the model and experiment is achieved [1].
Variation of the identified combustion process parameters is presented in Figure 13, as function of the CR. The
increase in CR at constant engine speed and load results in significant decrease of the Vibe function coefficient,
Figures 13.a and 13.b [1].
The tests have shown that duration angle of the first explosive part of the combustion is nearly constant,
Φz1=9o CA, and this value remained the same during modelling. The total duration angle of combustion, Φz2=Φ,
increases with the increase of CR, Figure 13.c. It is the consequence of the larger amount of fuel combusting with
diffusive combustion, as shown in Figure 12.a [1].
The amount of the fuel burnt at the first explosive part of first Vibe function, σg, significantly decreases with the
increase of the CR by constant amount of the injected fuel and constant engine speed, Figure 13.d. It is also
connected to a shorter period of ignition delay at higher CR [1].
Figure 12 Experimental and theoretical rate of heat release diagrams under full load, BMEP=0.55 MP and fixed fuel
injection angle (pu). (αps1 and αps2) are angle of initial (start) of combustion (2% of burned mass) for the first and
second Vibe function respectively [1]
237
a) d)
b)
e)
Angle of initial (start) of combustion (αps) under

constant fuel injection timing (αpu=18,5o CA), intensively
increases with the increase of the CR, due to shorter
period of ignition delay, which is a consequence of
higher maximal temperatures at the moment of fuel
injection, Figure 13.e.
All tests were performed with experimental engine and

conventional fuel injection system. Injection pressure
was 20 MPa.
c)
Figure 13 Model parameters variations as functions of the CR [1]
Optimal values of the compression ratio

The maximal pressure in cylinder (pmax) and exhaust gas temperature (Texhaust) as the function of engine load
(Wespecific effective work) for different CR values are shown in Figure 14 [1].
With the increase of CR and engine load, under the same injection timing (18.5o CA) before top dead centre
(BTDC), pmax is increasing, too. This undesirable increase in pmax is followed by a relatively improved atomizing of
larger amount of fuel in cylinder under higher pressure and engine temperature. Because of the improved
conditions for combustion process, the entire working process is improved. Moreover, when the CR is increased,
the Texhaust value is decreased, Figure 14, as revealed by simple isentropic relation at high CR. Similar results can
be seen in [36].
Leaner air-fuel mixture is used in engine operation under low loads. Therefore, the amount of heat released during
the combustion process is decreased. Also, the relatively low fuel injection pressure, which is very poorly
dispersed, leads to the existence of large droplets of fuel. A consequence of this is certain decrease in temperature
of the engine parts and decrease in cylinder temperature in the first phase of fuel injection. There is also an overlap
of Texhaust under middle CR values for all load regimes [1].
Dependence between the brake specific fuel consumption (BSFC) and the CR is presented in Figure 15, for
optimal injection time, in various load conditions [1].
238
Figure 14 Cylinder pressure and exhaust temperature change in experimental diesel engine [1]
The increase of the CR results in less intensive increase of the BSFC at low loads, and then, it more intensively
increases for CRs above 14:1, Figure 15 [1].
At high loads, the BSFC first decreases and reaches the minimum value for CR near 15:1, and then starts to grow
again with further increase of CR, Figure 15. The increase of the engine speed causes the increase in the
mechanical and aerodynamic losses and the increase in BSFC, Figure 16. The combustion process is responsible
for the increase of the fuel consumption. This is obvious in the case of a combined application of low CR and
shortened injection timing, when the delay of ignition becomes longer, and, because of that, the combustion
process is prolonged to expansion stroke. This is followed by a decrease in pmax and temperature within a cylinder,
while temperature during an expansion process shows a tendency to increase. Because of that, the losses become
greater and this may be followed by an increase in the BSFC [1].
Figure 15 Influence of the CR on BSFC [1]
The optimal value of the CR, at which the engine has minimal BSFC, increases with the increase of the load,
Figure 16. At full load, minimal fuel consumption is achieved with ε=17.5, while at low load conditions, minimal fuel
consumption is achieved with ε=12 [1].
239
Figure 16 Selection of the CR optimal value for engine operation with minimal fuel consumption
Under low and medium loads, the emission of NOx increases with the increasing of the CR, Figure 17. At high
loads, the NOx emission firstly decreases and then increases with the increasing of the CR, and reaches its minimal
value with ε=15, Figure 18. Under higher loads, minimal NOx emission is reached with ε=15, Figure 19.
Figure 17 Influence of the CR on NOx emission Figure 18 Selection of the optimal value of the CR for
engine operating with the minimal NOx emission
Figure 19 Influence of the CR on particles emission Figure 20 Selection of the optimal value of the CR for
engine operating with the minimal PM emission
In the case of the largest value of the CR under all loads, the largest temperature occurs inside the engine cylinder.
Large amount of free oxygen under low loads, in spite of relatively low maximal temperature with respect to full
240
load, leads to formation of the largest amount of NOx. NOx emission decreases with load increase, under higher CR
values.
In the case of the lowest value of the CR under low loads, the lowest maximal temperature is achieved within the
working cycle. This leads to formation of the lowest amount of NOx. With an increase in load, temperature also
increases and the amount of free oxygen decreases. Thus, at the beginning of the process, the amount of
produced NOx increases, but, when the amount of free oxygen decreases, a decrease in the amount of produced
NOx would occur with load increase [1].
Particulate matter (PM) emission is the smallest at medium loads and it increases if the engine is running at low or
high loads. At the same time, PM emission increases with the increase of CR at all loads, Figure 19, so optimal CR
is ε=12, Figure 20. Generally, under very low loads, the PM emission is somewhat larger. The major reason for this
is a relatively low injection pressure of the small amount of fuel that does not atomize so well. As the amount of fuel
increases with a load increase, this effect is attenuated and a certain decrease in PM emission occurs, so that,
under large loads, it would begin to increase again. Since the combustion chamber volume increases if the CR
decreases, the amount of air in the cylinder increases, and it is the cause of decreasing of PM emission when the
CR decreases [1].
Analysis of universal diagrams shows that the trends of CR variation are almost equal for the same values of fuel
consumption and emission. Optimal NOx emission is achieved by later injection, while the minimal consumption
and PM are achieved by earlier injection.
The multiple injection strategies in combination with EGR offer the potential to improve the compromise between
engine emissions, noise and fuel economy in diesel engines with lower CR.
ONE ENGINE  ANY FUEL AS FUTURE SCENARIO

According to law of nature, there are no neither pure energy forms, or “pure fuels”, or “clean engines”, nor “clean
vehicles”. All existing reserves such as coal, petroleum and gas are gifts of nature and must be economized not
only as power sources but as unique and sole raw materials. Rational technologies and economical utilization of
these gifts of nature are the only ways to extend life on our planet.
Future vehicles with flexible Otto/diesel engine cycle coupled with the alternative fuels will be key option in
development of future transport in the world as well as in the Republic of Serbia. Future vehicles should be
optimized according to working process software by sensors and with systems of fuel injection, Figure 21.a.
Optimistically defined terms of reference as a symbiosis of superior economy is targeted, which have the best
diesel engines with the best environmental attributes Otto engines.
In a limited development period, fuel mass will be no more of oil origin. On the other hand, well-known technologies
of alternative fuels indicate the economical production of sufficient components as an additive of synthetic origin or
from natural gas from any bio-raw materials. After identical logic, earlier fuels for Otto and diesel engines are
coupled into one universal fuel. Indeed, we have perceived such a futuristic project tasks [6]:
 One engine for all fuels and
 One fuel for all engines.
Multi-process IC engine, Figure 21.b, is running with VCR, as Otto engine, the fuel is added with alcohols (octane
number over 100) as an additive to raise the octane number (premium gasoline having EU 95). Analogously, when
the multi-process engine is running with VCR, as diesel engine, then the additives are bio-diesel oils (cetane
number 80) to raise the cetane quality of fuel (Diesel standard have about 50 cetane)!
Present refineries have their own consumption of 10 to 15% of the total volume of refined oil, depending on the raw
material and on what kind of fuel they produce. If IC engines with VCR (type DieselOtto) are introduced, then
there would no longer be a share in the "gasoline" and "diesel” fuel. Large tankers, which transport fuel, directly use
purified oil. Only a defined energy and environmentally friendly fuel has its thermal power. Then, there are the
disposable complexes as "octane" or "cetane index" numbers and, therefore, a complicated oil refineries, for "high-
octane gasoline", with a "high cetane for the diesel" and the like. Refiners should be less spending and more
producing and providing energy and environmental quality fuel.
241
a) b)
Figure 21 Multi-process multi-fuel IC engine with VCR: a) One engine  any fuel concept and b) Future formulation
of the fuels for VCR engines with Ottodiesel combustion process
CONCLUSIONS
The following generalized conclusions can be written from the study performed during experimental work and
presented modelling:
 VCR engine offers the potential to increase combustion efficiency and decrease emissions under varying
load and speed conditions;
 A good agreement between the model and experimental combustion law of the characteristic shape
developed precisely for direct injection diesel engines is achieved by modelling with double Vibe function;
 Initial combustion angle (αps) intensively increases with the increase of the compression ratio due to
shorter period of ignition delay, which is a consequence of higher temperatures at the of fuel injection
time. This research has shown that the duration angle of the knocking combustion part is nearly constant
and this value remains the same during modelling. Total combustion duration angle increases with the
increase of the compression ratio, because the combustion is of higher quality and equally divided through
all stages;
 The amount of fuel burnt at the first stage significantly decreases with the increase of the compression
ratio, by constant amount of injected fuel and constant engine speed. This is also connected with shorter
duration of ignition delay at higher compression ratios;
 Value of optimal compression ratio, at which the engine runs with minimal fuel consumption, increases
with the increase of load. At full load, the fuel consumption is the smallest at maximal compression ratio of
17.5:1, while at low loads, minimal fuel consumption is achieved for compression ratio 12:1;
 Under lower compression ratio of 12.1:1, the engine has started very difficultly, and operation continued
with higher knocking;
 From the aspect of minimal NOx emission, optimal compression ratio at full load has a value of 15:1.
Particulate matter emission is the smallest for medium loads and it increases if the engine runs at low or
high loads. At the same time, particulate matter emission increases with the increase of the compression
ratio, so the optimal compression ratio is 12:1;
 Future vehicles with flexible Otto/diesel engine cycle coupled with the alternative fuels will be key option in
development of future transport in the world as well as in the Republic Serbia; and
 Optimisticaly defined terms of reference as a symbiosis of superior economy is targeted that has the best
diesel engines with the best environmental attributes Otto engines.
ACKNOWLEDGMENTS
This paper is a result of the research within the project TR35041 financed by the Ministry of Science and
Technological Development of the Republic of Serbia.
242
REFERENCES
[1] Milojević, S.: "Analyzing the Impact of Variable Compression Ratio on Combustion Process in Diesel Engines”,
Mr Thesis, 2005, University of Kragujevac Faculty of Mechanical Engineering,156 p,
[2] Pešić, R., Petković, S., Veinović, S.: “Motor VehiclesEquipment”, University Textbook, 2008, University of
Kragujevac, Faculty of Mechanical Engineering and University of Banja Luka, Faculty of Mechanical
Engineering, 536 p, ISBN 9789993839200, (in Serbian),
[3] Pešić, R.: “ASMATAAutomobile Steel Material Parts Substitution with Aluminum”, Mobility & Vehicle
Mechanics (MVM), Vol. 30, Special Edition, 2004, pp 1165,
[4] Pešić, R.: “The research in the area of IC enginesLaboratory for IC engines on the Faculty of Mechanical
Engineering in Kragujevac”, Monography, 2000, University of Kragujevac, Faculty of Mechanical Engineering,
473 p, (in Serbian),
[5] Pešić, R., Davinić, A., Veinović, S.: “One engine for all fuelsone fuel for all engines”, 10th EAEC European
Automotive Congress, Proceedings, Paper EAEC05YU-EN01, 2005, Belgrade, pp 110,
[6] Pešić, R.: “Automobile SI Engines with minimal fuel consumption”, Monography, Mobility & Vehicle Mechanics
(MVM), Special Edition, 1994, (in Serbian),
[7] Pešić, R., Veinović, S., Orlović, M., Bebić, O.: “Formulation of ecological fuels and design of economical
automotive engines”, Petroleum and Coal, Vol. 37, No. 1, 1995, Slovakia, pp 2228,
[8] Milosavljević, B., Pešić, R., Taranović, D., Davinić, A., Milojević, S.: “Measurements and modeling pollution
from traffic in a street canyon: Assessing and ranking the influences”, Thermal Science, Vol. 19, No. 6, 2015,
pp 20932104. https://doi.org/10.2298/TSCI150402111M,
[9] Veinović, S., Pešić, R., Janićijević, N., Dragović, T.: “Fuel Flexible engines: Two and four stroke engines with
variable compression ratio”, XI International Symposium on Alcohol Fuels, Vol. 1, 1996, Sun City, pp 5257,
[10] Pešić, R., Milojević, S., Veinović, S.: “Benefits and Challenges of Variable Compression Ratio at Diesel
Engines”, Thermal Science, Vol. 14, No. 4, 2010, pp 10631073, https://doi.org/10.2298/TSCI1004063P,
[11] Gruden, D.: “Varovanje okolja v avtomobilski industriji”, Textbook, 2011, Izolit, Ljubljana, 504 pages,
(Translation from the German language: Umweltschutz in der Automobilindustrie, Vieweg + Teubner Verlag,
GmbH. Wiesbaden, 2008.),
[12] Milojević, S., Pešić, R.: “CNG buses for clean and economical city transport”, Mobility & Vehicle Mechanics
(MVM), Vol. 37, No. 4, 2011, pp 5771, http://www.mvm.fink.rs/Journal/Archive/2011/2011V37N4/4_sasa_
milojevic/milojevic_pesic_rad.pdf,
[13] Milojevic, S., Pesic, R.: “Theoretical and experimental analysis of a CNG cylinder rack connection to a bus
roof”, International Journal of Automotive Technology, Vol. 13, No. 3, 2012, pp. 497503,
https://doi.org/10.1007/s12239-012-0047-y,
[14] Milojević, S.: “Reconstruction of existing city buses on diesel fuel for drive on Hydrogen”, Applied Engineering
Letters, Vol. 1, No. 1, 2016, pp 1623,
[15] Milojevic, S.: “Optimization of the Hydrogen System for City Busses with Respect to the Traffic Safety”,
Proceedings of the 20th World Hydrogen Energy Conference WHEC 2014, Vol. 2, 2014, pp 853860,
[16] Milojević, S., Ilić, N., Pešić, R.: “Application of Hydrogen as alternative fuel for propulsion systems in city buses
 Overview”, Mobility & Vehicle Mechanics (MVM), Vol. 40, No. 3, 2011, pp 6984
[17] Milojević, S.: “Sustainable application of natural gas as engine fuel in city buses – Benefit and restrictions”,
Journal of Applied Engineering Science, Vol. 15, No. 1, 2017, pp 8188, 2017, doi: 10.5937/jaes15-12268
[18] Pešić, R., Adžić, M., Petković, S., Hnatko, E., Đokić, D., Veinović, S.: “The future is in the environmental
engagement of energy.”, Tractors and Power Machines, Vol.16, No. 3, 2011, pp 2431,
[19] ACEA European Oil Sequences 2012,
http://www.acea.be/images/uploads/files/2012_ ACEA_Oil_Sequences.pdf, accessed on in October 2013,
[20] NGV Global, Natural Gas Vehicle Knowledge Base: “Natural gas vehicle statistics”, 2018, New Zealand,
http://www.iangv.org/current-ngv-stats/,
[21] Hora, S.T., Agarwal, K.A.: “Effect of varying compression ratio on combustion, performance, and emissions of
a hydrogen enriched compressed natural gas fuelled engine”, Journal of Natural Gas Science and
Engineering, Vol. 31, 2016, pp 819-828, 2016, https://doi.org/10.1016/j.jngse.2016.03.041,
[22] Pešić, R., Petković, S., Hnatko, E., Veinović, S.: “Anthropogenic global warming, Kyoto protocol and transport
ecology”, Tractors and Power Machines, Vol. 13, No. 3, 2008, pp 4350, (in Serbian),
[23] Milojević, S., Pešić, R.: “Logistics of Application Natural Gas on Buses”, CAR 2011 Conference Proceedings,
CAR2011_1105, 2011. pp 110,
[24] Milojević, S., Pešić, R.: “Logistics of natural gas applications as engine fuel”, Tractors and Power Machines,
Vol. 16, No. 2, 2016, pp 6874, (in Serbian),
[25] Pešić, R., Veinović, S., Davinić, A.: “Engine characteristics of gaseous fuel”, Scientific meeting with
international participation, YUNG 4P 2002, Proceedings, 2002, Novi Sad, pp 2733, (in Serbian),
[26] Milojević, S., Pešić, R., Ilić, N.: “The first CNG powered Serbian lowfloor city bus”, 9th International
Conference on accomplishments in electrical and mechanical engineering and information technology,
Proceedings, 1999, Banja Luka, pp 657662, (in Serbian),
243
[27] Milojević, S., Pešić, R.: “Benefit and restrictions related to the application of natural gas as engine fuel for city
buses”, The 13th International Conference on Clean Energy (ICCE2014), 2014, Turkey, pp 902910,
[28] Schommers, J., Scheib, H., Hartweg, M., Bosler, A.: “Minimising friction in combustion engines”, MTZ, Vol. 74,
No. 78, 2013, pp 2835,
[29] Pesic, R., Milojevic, S.: “Efficiency and Ecological Characteristics of a VCR Diesel Engine”, International
Journal of Automotive Technology, Vol. 14, No. 5, 2013, pp 675681, https://doi.org/10.1007/s12239-013-
0073-4,
[30] Gruden, D. et al.: “Traffic and environment”, University Textbook, Springer Verlag, 2003, Berlin Heidelberg,
[31] Pešić, R., Hnatko, E., Golec, K., Stanojević, M., Veinović, S.: “Thermodynamic and construction measures for
radical reduction of fuel consumption in IC engines”, Combustion Engines, Scientific magazine, no. 2007SC2,
2007, Poland, pp 393400,
[32] Milojevic, S., Pesic, R., Taranovic, D.: “Tribological optimization of reciprocating machines according to
improving performance”, Journal of the Balkan Tribological Association, Vol. 21, No. 3, 2015, pp 690699,
[33] Stojanović, B., Milojević, S.: “Characterization, Manufacturing and Application of Metal Matrix Composites. In:
Wythers MC (ed) Advances in Materials Science Research”, 30th edn. Nova Science Publishers, New York,
2017, pp 83133, https://www.novapublishers.com/catalog/product_info.php?products_id=63027,
[34] Milojević, S., Pešić, R., Taranović, D.: “Tribological Principles of Constructing the Reciprocating Machines”,
Tribology in Industry, Vol. 37, No. 1, 2015, pp 1319, http://www.tribology.rs/journals/2015/2015-1/2.pdf,
[35] Milojević, S., Džunić, D., Taranović, D., Pešić, R., Mitrović, S.: “Tribological Reinforcements for Cylinder Liner
of Aluminum  Example Compressors for Brake Systems of Trucks and Buses”, Proceedings of the 15th
International Conference on Tribology SERBIATRIB’17, 2017, Kragujevac, pp 251257,
[36] Pešić, R., Golec, K., Hnatko, E., Kaleli, H., Veinović, S.: “Experimental engine with flexible Otto or Diesel cycle
(VCRvariable compression ratio)”, Innovative Automotive TechnologyIATʼ03, Proceedings, 2003, Koper,
Portorož, pp 281290,
[37] Pischinger, S., Vittek, K., Tiemann, C.: “Twostage Variable Compression Ratio with Eccentric Piston Pin”,
MTZ, Vol. 70, No. 2, 2009, pp 2027,
[38] Mathur, B.Y., Poonia, P.M., Jethoo, S.A., Singh, R.: “Optimization of Compression Ratio of Diesel Fuelled
Variable Compression Ratio Engine”, International Journal of Energy Engineering, Vol. 2, No. 3, 2012, pp
99101,
[39] Reddy, G.A., et al., “Effect of Compression Ratio on the Performance of Diesel Engine at Different Loads”, Int.
Journal of Engineering Research and Applications, Vol. 5, No. 10 (part 2), 2015, pp 6268,
[40] Stricker, K. Kocher, L. Koeberlein, E. Alstine, V.D., Shaver, M.G.: “Estimation of effective compression ratio for
engines utilizing flexible intake valve actuation”, Proceedings of the Institution of Mechanical Engineers, Part
D: Journal of Automobile Engineering, Vol. 226, No. 8, 2012, pp 10011015,
https://doi.org/10.1177/0954407012438024,
[41] Gajarlawara, N., Rao, P.A.G., Murthy, K.M.: “Experimental investigations of effects of cooling/non cooling of
EGR on two level of compression ratio in a common rail diesel engine”, Thermal Science, Vol. 19, No. 6, 2015,
pp 19952002,
[42] Gajarlawar, N., Khetan, A., Rao, P.A.G.: “Investigations of effects of pilot injection with change in level of
compression ratio in a common rail diesel engine”, Thermal Science, Vol. 17, No. 1, 2013, pp 7180,
[43] Ahn, S., Park, J., Won, J., Kim, H., Kang, I., Cho, Y., Shin, S.: “An analytical FEMbased study of the drawing
process of an ultrahighpressure commonrail fuel tube”, Journal of Mechanical Science and Technology,
Vol. 31, No. 7, 2017 pp 33893396, https://doi.org/10.1007/s12206-017-0628-9,
[44] Infiniti VCTurbo: “The world’s first productionready variable compression ratio engine”,
https://kfz-tech.de/Biblio/Benzinmotor/VarVerdichtung3.htm, Accessed on: 31.08.2018.,
[45] Dorić, J., Klinar, I.: “The realization and analysis of a new thermodynamic cycle for internal combustion
engine”, Thermal Science, Vol. 15, No. 4, 2011, pp 961974,
[46] Duffy, K., Kieser, A., Mehresh, P., Frieden, D., Rodman, T., Hardy, B., Hergart, C.: “Heavy Duty HCCI
Development Activities”, DOE DEER Conference, DOE Contract DEFC2605NT42412, 2006, pp 125,
[47] Colle, V., Constensou, C., Dubois, F., Guilly, L.: “Variable Compression Ratio for Future Emission Standards”,
MTZ, Vol. 78, No. 4, 2017, pp 2027,
[48] Helling, S.U., Ghetti, S., Methfessel, P., Marten, C.: “Variable Compression for Large Engine”, MTZ, Vol. 78,
No. 78, 2017, pp 5054,
[49] Fraidl, G.K., Piock, W.F., Fürhapter, A. et al.: “Homogene Selbstzündungdie Zukunft der
BenzinDirekteinspritzung”? MTZ Motortech Z, Vol. 63, No. 10, 2002, pp 828837,
https://doi.org/10.1007/BF03226652
[50] Ghojel., I.J.: “Review of the development and applications of the Wiebe function: Attribute to the contribution of
Ivan Wiebe to engine research”, International Journal of Engine Research, Vol. 11, issue 4, 2010, pp 297–312,
https://doi.org/10.1243/14680874JER06510
[51] Saša Milojević and Radivoje Pešić.: “Determination of Combustion Process Model Parameters in Diesel
Engine with Variable Compression Ratio,” Journal of Combustion, Vol. 2018, Article ID 5292837, 2018, 11
p, https://doi.org/10.1155/2018/5292837.
244
Kragujevac, Serbia
MVM2018-065
Radivoje Pešić1
Saša Milojević2
Dragan Taranović3
Milan Stanojević4
EXPERIENCES RELATED TO THE INSPECTION OF MOTOR

VEHICLES RUNNING ON NATURAL GAS IN THE REPUBLIC
OF SERBIA
ABSTRACT: Natural gas and liquefied petroleum gas represent the most practical, realistic and easiest way to
reduce pollution coming from road transportation. In practice, natural gas vehicles are as safe as any other vehicles
(gasoline, diesel, liquefied petroleum gas, etc.) operating on the roads. However, safety incidents sometimes occur
where a natural gas system, primarily storage cylinders, have failed. Recent incidents involving natural gas vehicles
have shown that cylinders may explode through compliant with current UN ECE 110R regulation. In the paper are
systematized regulations and technical requirements relating to the installation of devices and equipment on
vehicles powered by natural gas. The main objective of the research is to present a good practice related to
inspection of motor vehicles on natural gas. It was also shown the several options for the reconstruction of vehicles
to drive on gas fuel, as well as logistics.
KEYWORDS: logistics, motor vehicles, natural gas, regulations
INTRODUCTION
In energy plan, oil as a conventional fuel is still a very important energy resource. On the other hand, constant
increase in consumption of conventional energy generating products opens the door for all potential substitutes for
oil.
Natural gas is a high-quality fuel for motor vehicles. Above all, available reserves equal the known oil reserves;
negative influence upon environment is less than with fuels derived from oil, as well as the price. Due to that,
natural gas as a motor fuel has been having a growing application in motor vehicles, which is further proved by the
fact that there are over 26 million vehicles in the world powered by natural gas (1.86 million in Europe). As an
example, we could mention the city of Porto whose budget for the current year includes the funds for purchasing
173 new buses powered by natural gas. Currently, in the Republic of Serbia, there are in traffic on the roads 878
natural gas vehicles (792 passenger and light duty vehicles, 58 buses and 28 medium and heavy duty trucks) [1].
1
Dr Radivoje Pešić, prof. University of Kragujevac, Faculty of Engineering, Kragujevac, Sestre Janjić 6,
pesicr@kg.ac.rs
2
Mr Saša Milojević, Assistant R&D. University of Kragujevac, Faculty of Engineering, Kragujevac, Sestre Janjić 6,
sasa.milojevic@kg.ac.rs
3
Dr Dragan Taranović, Associate prof. University of Kragujevac, Faculty of Engineering, Kragujevac, Sestre Janjić
6, tara@kg.ac.rs
4
Mr Milan Stanojević, FCA Serbia d.o.o. Kragujevac, milan.stanojevic@fcagroup.com
245
The growing number of vehicles powered by natural gas has required in parallel new regulations and rulebooks
regulating this field to be adopted. Thus, environmental requirements for vehicles fueled by natural gas have been
prescribed by ECE rulebooks 49 and 83; whereas the requirements related to the aspect of safety and functionality
of installation of gas devices and equipment have been defined under regulations UN ECE 110R and UN ECE
115R. Laws and bylaws applicable in the Republic of Serbia for natural gas vehicles are: Road traffic safety law,
Rulebook on the Classification of Motor Vehicles and Trailers and Technical Conditions for Vehicles in Road
Traffic and vehicle testing rulebook on [26].
For the purpose of managing the process of logistic affairs, from production to periodical checks and maintenance
of vehicles powered by natural gas, it is necessary to systemize the above mentioned regulations and rulebooks.
That would also provide an answer to the question of how safe the vehicles powered by natural gas actually are.
NATURAL GAS AS FUEL FOR DRIVING AND MOBILE SYSTEMS

All gases are good fuels for Otto engines: the mixture is of high-quality and ready for full combustion, engine
operation is economical, pollution is less, oil and engine life is longer etc. In terms of engine related criteria, natural
gas may be used as the sole fuel for Otto or diesel engines (with certain engine reconstruction). In addition to that,
it may be used in mixture with every kind of oil or proved alternative fuel (lower or higher alcohols, biogas, and
other) [7].
In normal ambient conditions, natural gas has a low density of energy per unit of volume. Due to that, in case of
application of natural gas as a motor fuel, it is necessary to have a certain treatment of natural gas in order to have
a higher concentration of energy by unit of volume. Given that it is expected to have a wider radius of vehicle
movement with one fuel tank topping, it is necessary to have operating pressures of around 200 bar which is the
limit that standard bottles for technical gases may endure. The simplest way is to compress natural gas in high-
pressure tanks (CNGcompressed natural gas). The second alternative for increasing the density of energy of
natural gas inside tanks on vehicles is its extreme cooling down to (162 C), i.e. conversion to liquefied state
(LNGliquefied natural gas) and storing in cryogenic tanks.
Optimistic studies about natural gas are based on pure methane. Unfortunately, pure methane is nowhere to be
found in nature. It is always in mixture with other impurities in concentration of various percentages. In order to use
natural gas as fuel, it is necessary to filter it until it reaches more or less the established composition and quality.
There are some national standards that regulate quality of natural gas as engine fuel: SAE J1616, DIN 51624 and
13 CCR § 2292.5 (California regulations). In parallel, there is an international standard ISO 15403, accepted also
by the Republic of Serbia. The purpose of this standard is for producers, vehicle users, filling station operators and
other logistics participants in the industry of (CNG) vehicles to get informed about quality of fuel for natural gas
vehicles [7].
Natural gas has the following characteristics:

 It is non-toxic, non-carcinogenic, non-corrosive, however, it is suffocating, colorless and odorless due to
which it is necessary to add certain odorizes and foul smells,
 Natural gas is lighter than air (density of methane is 0.68 kg∙m3 at 15 C), and when leaking it leaks
upwards, whereas for example propane (1.87 kg∙m3) and butane (2.44 kg∙m3) are heavier and leak
downwards,
 Auto flammability temperature of (CNG) is in the scope of 480650 C, whereas in case of (LPGliquefied
petroleum gas) it is in the scope of 480495 C (methane 595 C, propane 495 C, butane 480 C,
gasoline 260430 C, whereas in diesel it is less than 260 C),
 Flammability scope in volume percent of methane with air is 4.415%, in case of propane it is 2.49.3%,
butane 1.88.8%, (gasoline 1.47.6% and diesel 0.67.5%),
 Natural gas is converted to liquid at ambient pressure at temperature of (162 C); and
 Critical point of methane is defined by temperature of (82.7 C), and pressure of 45.96 bar.
EXAMPLES OF SOLUTIONS FOR NATURAL GAS VEHICLE DESIGN

Natural gas as fuel of vehicle drive systems in road transport is applied in new vehicles with original natural gas
engines or by converting existing engines so as to use gaseous fuels. In practice, there are the following options for
conversion of vehicles so as to use (CNG) [8, 9]:
 Modification of gasoline engines so as to use CNG (Eng. dedicated fuel),
 Modification of gasoline engines so as to use CNG or gasoline (Eng. bifuel),
 Modification of diesel engines so as to use CNG (Eng. dedicated), spark plug ignition; and
 Conversion of diesel engines so as to have parallel function with CNG and diesel fuel (Eng. dual fuel).
246
The concept of vehicles with original factory produced engines designed for use of natural gas as fuel is the only
one that can use all the natural advantages of this fuel [8, 9]. Modifications of old carburetor engines so as to use
natural gas as fuel with use of cheap systems, such as those with injectors, do not contribute to reduction of
emissions and improvement of safety, but quite the contrary, result in their increase and deterioration respectively.
Figure 1 shows the scheme of universal installation for CNG with sequential injection.
Figure 1 Parts of system for sequential injection of CNG to engine
Figure 2 shows position of the installed devices and equipment on the vehicle Volvo S80, with propulsion system
on natural gas or biogas.
Figure 2 Bifuel systems (CNG or biogas) on passenger vehicle Volvo S80
Figure 3 shows parts of gas installation of a city low-floor bus on CNG. Bus propulsion system has been equipped
with original, factory produced gas engine for CNG or LNG. The concept of engine that works on the principle of
combustion of lean mixture meets the Euro IV regulation. With application of engines whose operation has been
based on combustion of stoichiometric mixture of natural gas and air with recirculation of cooled combustion
products and three-component catalyst, the reached level of toxic and hazardous combustion products from vehicle
is within the limits prescribed by Euro VI [9-11].
According to author’s experiences, the best option for reconstruction of existing city buses is the using of
completely new propulsion system with dedicated (CNG) combustion. In the bus, need to be integrated modern
automatic fire protection system [12]. Similar conditions apply if Hydrogen is used as a fuel [13].
247
Figure 3 CNG installations on city low-floor bus
REGULATIONS REGULATING THE ISSUES RELATED TO VEHICLES

POWERED BY NATURAL GAS
International regulations
Environmental requirements that should be met by CNG vehicles have been defined within ECE Rulebook no. 49
for freight carrying vehicles and buses, and within ECE Rulebook no. 83 for passenger vehicles.
From the aspect of safety and proper installation of gas equipment, the requirements to be met by CNG vehicles
have been prescribed within two ECE regulations [5, 6]:
 UN ECE 110R – uniform homologation regulations:
 Specific equipment of motor vehicles powered by CNG; and
 Vehicles equipped with specific equipment for CNG in terms of installation of such equipment.
 UN ECE 115R – uniform homologation regulations:

 Vehicles, previously homologated for basic fuel and subsequently equipped with specific
equipment for LPG drive, in terms of installation of such equipment; and
 Vehicles previously homologated for basic fuel and subsequently equipped with specific equipment
for CNG drive, in terms of installation of such equipment.
National regulations
The following laws and by-laws apply to the vehicles in the Republic of Serbia [24]:
 Law on road traffic safety,
 Rulebook on the Classification of Motor Vehicles and Trailers and Technical Conditions for Vehicles in
Road Traffic; and
 Vehicle testing regulations.
248
TECHNICAL REQUIREMENTS FOR INSTALLATION OF DEVICES AND
EQUIPMENT OF NATURAL GAS VEHICLES
Cylinders for CNG must have the following data permanently inscribed in accordance with provisions of Rulebook
UN ECE 110R, Figure 4:
 Name of manufacturer,
 Factory number,
 Month and year of production (inspection),
 Volume of empty cylinder in liters (l),
 Max. allowable pressure of cylinder filling (MPa/bar),
 Mark of gas to fill the cylinder  "CNG ONLY"; and
 Imprint of cylinder expiry date  "DO NOT USE AFTER XXXXXX" where XXXXXX indicates the month
and year.
a)
xxx
b)
Figure 4 a) Cylinder identification and labels; b) Photography of sticker for CNG cylinder
The cylinders must be fitted to the basic vehicle structure according to the instructions of the manufacturer, i.e.
technical requirements and must not be in direct contact with other metal parts, except in case of cylinders with
special supports designed to be fitted to the metal structure. The technical requirements for cylinder fitting
according to ECE Rulebook no. 115 annex 5 point 1.3 are listed in the table below 1 [6]:
Table 1 Technical requirements for cylinder fitting

Tank Minimum dimensions of Minimum dimensions of fixing metal Minimum dimensions of
volume, l washers, mm straps, mm screws, mm
Up to 85 30x1.5 or 25x2,5 20x3 or 30x1.5 8
85  100 30x1,5 or 25x2.5 30x3 or 20x3 10 or 8
100  150 50x2 or 30x3 50x6 or 50x3 12 or 10
Cylindrical tanks are fixed with at least to insulated straps, except in the following cases:

in this case, three straps are necessary for fixing; and
 in this case four straps are necessary for fixing.
In case of cylinders installed below the vehicle floor, the cylinder must be fixed with at least three suitable straps. The screws
used must be at least in the class 8.8.
249
As for the cylinders of volume over 150 l, it applies that the requirements of UN ECE 110R must be met. Rulebook
UN ECE 110R prescribes that (CNG) cylinders must be fixed in appropriate way, so that they can (when full)
endure deceleration in the moving direction of vehicle and in direction perpendicular to vehicle moving direction,
depending on the vehicle category, as per Table 2. It also requires a certain resistance (absence of any
deformation) of cylinder supports and also of the part of vehicle structure to which the cylinders have been fixed,
which could also be proven by the following calculation [5, 10].
Table 2 Prescribed deceleration as per UN ECE 110R

CATEGORY OF VEHICLES M1 or N1 M2 or N2 M3 or N3
Deceleration  in vehicle moving direction 20∙g 10∙g 6.6∙g
Acceleration  in horizontal plane,
8∙g 5∙g
perpendicular to vehicle moving direction
According to UN ECE 110R, the cylinder for (CNG) must be approved and equipped with the following components
at the least which could be separate or combined [5]:
 Hand valve,
 Automatic valve of cylinder,
 Pressure relief device (PRD) or temperature activated pressure relief device (TPRD); and
 Excess flow valve (EFV).
The cylinder and valves may also be equipped with gas-tight housing with ventilation, if necessary. It is necessary
to protect them from direct sun rays.
High-pressure gas pipes must not be soldered nor welded and instead have to be joined by fixing elements made
of steel. High-pressure gas pipelines must be laid so as not to vibrate, not to have friction among parts especially in
places where the pipes go through openings. The distance between two places of fixing, in case of high-pressure
gas pipes below vehicles, may be at most 500 mm, whereas radiuses of pipe bending must be adjusted to material
and dimensions of pipes.
Gas pipelines are installed in the driver and passenger’s compartment, including the trunk, only if protected in a
special way (impermeable housings, pipe-in-pipe etc.) provided that the protection must be resistant to mechanical
damage, and its openings must be located on the external side of vehicles for ventilation purposes.
Before being connecting to the device (gas valve, pressure reducer, filter, manometer and other), gas pipe must
have a compensation coil that enables expansion. The high-pressure gas pipelines must not have visible damage
and the pipes that have fractures or have undergone mechanical damage due to corrosion process and other must
be replaced immediately.
Low-pressure gas pipelines connecting the pressure reducer with device generating the mixture of fuel and air for
engine must not be in contact with warmed up parts of exhaust system, i.e. with engine, but only with device for
fuel supply to engine. Rubber lines must not be damaged nor cracked. Low-pressure gas lines are joined by
collars. The junctures must be impermeable, collars must be self-blocking, and so there could be no uncontrolled
loosening of joints.
Gas pressure reducer is bracket fixed on the interior part of car body, outside of driver and passenger’s
compartment, including the trunk and it must not be fixed to the parts of vehicle that get warmed up and must be
distanced in a proper way from the parts of exhaust system and vehicle engine.
INSPECTION OF VEHICLES WITH INSTALLED DEVICES AND EQUIPMENT

FOR NATURAL GAS POWERED VEHICLES
Necessary checks
Checking the compliance of devices and equipment with the approval type
Devices and equipment installed in the vehicles fueled by (CNG) must be type approved (homologated) according
to UN ECE 110R, which is to be determined during this check-up.
250
Checking if meeting the requirements related to compliance of devices and equipment with the manner of
combustable mixture preparation and technical implementation of installation
Checking if meeting the requirements related to compliance of devices and equipment with the manner of
combustable mixture preparation
Devices and equipment must be installed in compliance with the instruction manual for installation made by the
manufacturer of devices and equipment for natural gas vehicles, and if there is no such instruction; the minimum
requirements of compliance are as follows:
 Devices and equipment with vacuum regulation system could be applied only in intake carburetor engines;
and
 Engines with a plastic intake manifold and engines with multi point injection system (MPI) in vehicles
produced after January 1st 2001 must have equipment with sequential injection of gas.
Checking if meeting the requirements in terms of technical implementation of installation
Devices and equipment of natural gas vehicles must be installed so as to be within limits of vehicles overall
dimensions, not to reduce possibility, to be protected from mechanical damage that may occur in exploitation, and
to be accessible for the purpose of checking their functionality and impermeability, and checking their identification
marks. At the same time, it is mandatory that installation requirements have been met, as defined under regulation
UN ECE 110R, which is to be determined during this inspection.
The following is to be checked:

 Meeting of requirements for installation of cylinder and its equipment,
 Meeting of requirements for gas lines (high-pressure and low-pressure lines),
 Meeting of requirements for gas pressure reducer; and
 Meeting of requirements for other devices (valves, fuel fill connector, manometers and pressure indicators,
electrical installations, gas dosing devices etc.).
Inspection of tightness of gas installation is done by using a detector sensitive to methane, where each leak of gas
to atmosphere is the reason for making a conclusion that the vehicle is not safe for further exploitation.
Inspection of vehicle mass
Installation of devices and equipment for natural gas fuels changes the vehicle mass, thus requiring during
inspection to determine the mass of vehicle with installed device and equipment. Vehicle mass may be determined
by calculation if you have available information about exact mass of basic vehicle, mass of installed device and
equipment and mass of gas with tank filled by maximum allowed operating pressure in the tank. If the authorized
organization, having checked the relevant technical documentation or in any other way, has any doubt about
accuracy of these data, the mass must be measured. The calculation, i.e. data on grounds of which the mass of
modified vehicle has been determined, must be indicated in the report on testing/inspection of vehicle. In case of
testing a freight vehicle for the purpose of determining the value of increased mass of vehicle ready for driving, it is
necessary to reduce the load bearing capacity of vehicle. In case of buses, if necessary, the total number of places
for standing and seated passengers should be determined by calculation. For that purpose, depending on the bus
body form, you should use the instructions defined by regulations UN ECE 36R and UN ECE 52R.
Periodical inspection
Cylinders are designed and manufactured for a limited design life, which is indicated on the cylinder label. Always
check the label first to ensure that the cylinder has not exceeded its expiration date.
Regulation UN ECE 110R prescribes the time schedule of periodical inspections of CNG cylinders every 48
months, from starting the servicing. In doing so, it is visually checked if there are mechanical, abrasive or chemical
damages. This regulation recommends checking during technical inspection if there is a gas leak [5].
In the Republic of Serbia, periodical inspections of vehicles with CNG devices and equipment are done every four
years.
Periodical inspection of CNG cylinders
Inspection of CNG cylinders is done in accordance with procedure defined by standard ISO 19078, and/or in
accordance with the national standards of the state in which the vehicle and CNG equipment have been exploited.
251
If using during a cylinder periodical inspection the hydraulic pressure test, it is mandatory to use specially treated
water (with inhibitors, without chlorine etc.). Also, it is recommended to have the cylinders exposed to inspection as
little as possible in terms of time, and never to have the cylinders with regular water (but to treated water instead).
The reason for that is in the corrosive effect of water to material out of which the cylinders are made (mostly
aluminum alloy).
Periodical inspections are a safety measure for timely detection of damages of cylinders and cylinder supports.
Inspections must be conducted with use of special tools (measuring tools for measuring length and debt of cracks,
battery lamps and other). There are two types of inspections: general inspection and, as already indicated in the
previous section, inspection by an authorized organization in which case there is also a suitable certificate issued,
Table 3.
Table 3 Overview of procedures for periodical inspections of CNG cylinders and equipment
DESCRIPTION OF
TYPE OF INSPECTION PERIOD
INSPECTION
Inspection is conducted by a
driver, a mechanic or a
technician working in a Recommended to have it
workshop. It includes visual every three months
General inspection inspection of condition of (recommendation of
cylinder and its supports, as well manufacturers of CNG
as of pipelines for the purpose of cylinders and equipment)
determining if they are in good
condition and properly fixed.
Inspection by authorized institutions and Inspection is conducted by a Every four years from starting
issuing of Certificates professional/expert with its service life in accordance
(Certificate and /or attestation of correct
authorization of the relevant with the national regulations.
installation of CNG device and
equipment) institution.
Marking of CNG fueled vehicles

The vehicles of М2 and М3 class fueled by (CNG) must have stickers as a visible indication that they use this type
of natural gas fuel. The form and dimension of stickers have been defined by standard UN ECE 110R. Stickers are
to be place in the front, rear and on the door exterior on the right side of vehicle.
SAFETY MEASURES DURING CHECKUP AND TECHNICAL INSPECTION OF

CNG VEHICLES
Flammability limits of fuel and air mixture are often called explosiveness limits so as to warn about severity of
possible incidents. Below the flammability limits, the mixture is too lean and over it too rich for combustion. The
flammability limits are often called lower and upper limit of flammability/explosiveness.
Gas leaking from the system installed on a vehicle during the check-up and technical inspection of vehicle may
create an explosive mixture in work environment due to which it is necessary to provide the following:
 Take the measures for the purpose of prevention of gas leaking from vehicle system,
 Keep the flame sources far from the place where there is a chance of forming a flammable mixture,
 Surface of heating elements inside the facility must not have the temperature over 400 C, it is forbidden
to use open flame, electrical or gas heaters; and
 It is necessary to adopt and regularly practice the measures of safety at work and fire protection of
facilities, as well as professional training of staff, etc.
Given that in case of leaking from vehicle system, (CNG) leaks upwards it should be observed that explosive
mixture is created in upper zones of the room. On the other hand, the position of the area with explosive
concentration highly depends on the type of the roof structure of facility. In case of no flat roof, the increased
concentration is usually in the highest parts). The position of setting the probes of gas automatic detection and
alarm system also depends on the type of roof structure.
On grounds of applicable U.S. regulations for example ventilation systems in workshops for maintenance of CNG
buses should enable 56 air changes per hour in the room. Also, it requires the air change dynamics of 425 l∙min1
per 1∙m2 for the ventilated areas. Given that the requirements for workshops are similar in case of diesel buses
(must have ventilation system with 46 air changes per hour), it is conclusive that it is not necessary to change the
252
ventilation system, in case of adaptation of existing facility so as to be used to work with CNG vehicles as well, thus
there are no additional costs of maintenance. The relevant regulations also recommend to equip such facilities with
so called roller doors that open faster than the standard ones so that in case of accidental gas leak it enables
additional ventilation of facility in a natural way [14].
Electric equipment of facility for servicing CNG vehicles, such as switches, sockets, lighting devices, motors, should
have a certain degree of explosion protection.
Contrary to the previously given instructions, e.g. when doing the LPG vehicle check-up and technical inspection,
the facilities should be provided with as good as possible ventilation in the floor area of facility, especially in the
zones where pits for repair or inspection of vehicles are, etc.
In general, the basic warnings which apply in cases of gas leakage accidents from vehicles and basic requirements
and safety measures for facility are briefly listed in the Table 4 below.
Table 4 Warnings in case of gas leakage and preventive measures

TYPE OF GAS  FUEL CNG LPG
Warnings in case of gas leakage Gas concentration is up!! Gas concentration is down!!
 Good ventilation of ceiling-  Good ventilation of floor-pit
roof zone zone
Basic requirements and preventive  Elimination of flame source  Elimination of flame source
safety measures for facility above vehicle near floor
 Staff safety training for  Staff safety training for working
working with CNG vehicles with LPG vehicles
General conditions and recommendations for working with cylinders and

equipment for CNG
When handling the cylinders and equipment on vehicles fueled by CNG it is necessary to adhere to the following
measures and instructions for the purpose of improving safety.
1. Ensure that all mounting blocks, brackets and other components are in good condition and properly
secured. If any mountings are loose, re-tightened them by procedures specified by the fuel system
manufacturer. If a mounting is damaged, need to be conducting a re-certification inspection.
2. Use suitable tools:

 Use the tool recommended by manufacturer, if possible; and
 Provide the relevant technical documentation of manufacturer before intervention.
3. Fuel system inspection includes checking all attached components, such as valves, tubing, end plugs,
fittings and pressure-relief devices. During inspection, make sure that each device is securely attached. If
any is loose, tighten it in accordance with the fuel system manufacturer’s instructions:
 Only trained staff may repair and inspect the cylinders and equipment for CNG; and
 It is necessary to be familiar with functionality of parts of system for CNG on vehicles.
4. Eliminate potential flame sources:

 No open flame or other heat sources; and
 Disconnect electrical supply on vehicle; provide good grounding of facility for the purpose of
preventing sparks and similar.
5. Never forget that parts of CNG system are under pressure:

 Loosen the joints slowly due to discharge of gas under pressure, it is mandatory to tighten the
joints with suitable torques according to recommendation of manufacturer (the torques are not the
same for nuts when tightening the pipeline for the first time and in case when tightening the joint
that was previously loosened).
6. Precautionary measures for potentially flammable mixture of natural gas and air:
 Natural gas is flammable in air when the concentration is 4,415%,
 It is forbidden to have service interventions, storing and transport of cylinders and equipment with
potentially flammable mixture. Discharging of gas from cylinder is to be done according to
instructions and technical documentation of manufacturer through a special connector; and
 It is forbidden to connect to electric voltage any parts of device and equipment for CNG in which
there is a potentially flammable mixture.
253
CONCLUSIONS
The issue that primarily refers to emission of particles resulted in reduced use of diesel buses and freight vehicles
in downtown zones of cities in developed countries. It brought about intensive substituting of vehicles of public
utility companies with vehicles powered by natural gas. As a consequence, there are over 26 million vehicles fueled
by natural gas in the world today.
The growth of the number of vehicles fueled by natural gas has been followed in parallel by adopting new
Regulations and Rulebooks regulating the issues in this field. From the aspect of safety and proper installation of
natural gas equipment, the requirements to be met by vehicles fueled by CNG have been prescribed by two
regulations: UN ECE 110R and UN ECE 115R. Thus, devices and equipment for vehicles fueled by gas must be
installed in vehicles in accordance with requirements of applicable national and international regulations.
International and national regulations foresee periodical inspections of devices and equipment for vehicles fueled
by natural gas. In the Republic of Serbia, there is a control check prescribed for the vehicles fueled by CNG that
has to be conducted by an authorized institution every four years followed by a certificate being issued.
During checkups, technical inspections and servicing of CNG vehicles, the facilities should be provided with better
ventilation in the zone of ceiling and roof, and it is mandatory to adhere to measures of safety at work for staff and
fire protection for facilities.
ACKNOWLEDGMENTS
This paper is a result of the researches within the project TR35041 financed by the Ministry of Science and
Technological development of the Republic of Serbia.
REFERENCES
[1] NGV Global, Natural Gas Vehicle Knowledge Base: “Natural gas vehicle statistics”, 2018, New Zealand,
Available from: http://www.iangv.org/current-ngv-stats/, Accessed 15.08.2018.,
[2] Law on Road Traffic Safety, (Official gazette of RS, No. 41/2009, 53/2010, 101/2011, 32/2013, 55/2014,
96/2015, 9/2016 and 24/2018),
[3] Rulebook on the Classification of Motor Vehicles and Trailers and Technical Conditions for Vehicles in Road
Traffic, (Official gazette of RS, No. 40/12, 102/12, 19/13, 41/13, 102/14, 41/15, 78/15, 111/15, 14/16, 108/16,
7/17 and 63/17),
[4] Vehicle testing regulations, (Official gazette of RS, No. 8/12, 13/13, 31/13, 114/13, 40/14, 140/14, 18/15,
82/15, 88/16 and 108/16),
[5] UN ECE Regulation No. 110 Rev.3  CNG and LNG vehicles; Available from: http://www.unece.org/trans/
main/wp29/wp29regs101-120.html, Accessed 25.08.2018.,
[6] UN ECE Regulation No. 115 Rev.1  LPG and CNG retrofit systems; Available from: http://www.unece.org/
trans/main/wp29/wp29regs101-120.html, Accessed 25.08.2018.,
[7] Pešić, R., Davinić, А., Veinović, S.: “Natural gas as raw material and fuel for motor vehicles”, XII International
Scientific Symposium MVM 2002, 2002, Kragujevac, pp 284289,
[8] Milojević, S., Pešić, R.: “CNG Buses for Clean and Economical City Transport”, Mobility & Vehicle Mechanics
(MVM), Vol. 37, No. 4, 2011, pp 5771,
[9] Milojević, S., Pešić, R.: “Logistics of Application Natural Gas on Buses”, The 10th International Automotive
Congress CAR 2011, Conference Proceedings, CAR2011_1105, 2011, University of Pitesti, Romania, pp
5771,
[10] Milojević, S., Pešić, R.: “Theoretical and Experimental Analysis of a CNG Cylinder Rack Connection to a Bus
Roof”, International Journal of Automotive Technology, Vol. 13, No. 3, 2012, pp 497503,
[11] Milojević S.: “Sustainable application of natural gas as engine fuel in city buses – Benefit and restrictions”,
Journal of Applied Engineering Science, Vol. 15, No. 1, 2017, pp 81-88,
[12] Milojević, S., Pešić, R.: “Fire safety of CNG buses  Proper experiences”, Mobility & Vehicle Mechanics
(MVM), Vol. 43, No. 4, 2017, pp 23-37, DOI: 10.24874/mvm.2017.43.04.03
[13] Milojević S.: “Reconstruction of existing city buses on diesel fuel for drive on Hydrogen”, Applied Engineering
Letters, Vol. 1, No. 1, 2016, pp 16-23,
[14] Pešić, R., Milojević, S.: “The issue of control and technical inspection of vehicles at gas plant (Problematika
kontrolisanja i tehničkog pregleda vozila na gasni pogon)”, Technical inspection of vehicles of the Serbian
Republic (Tehnički pregledi vozila Republike Srpske 2012), 2012, Teslić, pp 2542.
254
SECTION
Vehicle Design and

Manufacturing
 Body Structures
 Advanced Materials
 Recycling
 Manufacturing Systems
 Quality Systems in
Automotive Industry
255
256
Kragujevac, Serbia
MVM2018-002
Isak Karabegović1
Ermin Husak2
THE ROLE OF INDUSTRIAL AND SERVICE ROBOTS IN

FOURTH INDUSTRIAL REVOLUTION WITH FOCUS ON
CHINA
ABSTRACT: In the last ten years, digital technology has significantly contributed to the change of people's lives
worldwide, because its application has caused a rapid transformation of all aspects of human life, and especially fast
transformation in the design, production, operation and maintenance of the production system, which caused
anunexpected jump in productivity. It can be said that fourth industrial revolutions ongoing process, which can be
labeledin a variety of ways, such as "intelligent factory", "smart industry" or "advance manufacturing". Development
of the digital technologies in the last twenty years has introduced us from third in the fourth industrial revolution. The
first time the term “Industry 4.0” appears in the Germany in year 2011 whose government promotes automation of
production processes by introducing digital technologies. Germany is one of the most technologically developed
countries in the world and it is logic that this revolution begins there. This example follows the other countries in the
world. Fourth technological revolution depends on a number of new and innovative technological achievements. It is
necessary to integrate production processes in all production phases and further applications by using ICT
technologies for digitalization. The automation of production processes must include advanced sensors and
intelligent robots that can be self-configured to be able to make specific product. It is necessary to collect large
amounts of data to be analysed and used in the production processes. It is also necessary to realize network
communication (include mobile and internet technology) between machines in the production process, the production
system and the operator, as well as suppliers and distributors. Application of previously mentioned leads to the
intelligent manufacturing processes that have a wide range of change of production processes. To intelligent process
we can come only by applying intelligent industrial robots because they represent one of the cornerstones of the
fourth industrial revolution. In this paper an analysis of industrial and service robot application in production
processes. With the implementation of digital technology, robot technology and other advanced technology we, as a
society, strive for intelligent automation, and intelligent factories, and thus create a society in which wealth, created
through the strengthening of global competitiveness, will serve for resolving social issues in society.
KEYWORDS: industry 4.0, digital technology, production processes, M2M, intelligent automation, industrial robot
INTRODUCTION
Automation of production processes in the industry began in sixties of the last century when in the automation
industrial robots have been included and continues to this day. However it must be recognized that it was rigid and
not flexible automation. Reason for this is the fact that if another product is in the same production line, it was
necessary to reprogram each robot, change the tools and end effectors etc. This required a delay in production.It
lasted long as it meant the additional cost in production. We have to mention the fact that industrial robots have been
1
Isak Karabegović, University of Bihać, Faculty of Technical Engineering, Irfana Ljubijankića bb., 77 000 Bihać,
Bosnia and Herzegovina, isak1910@hotmail.com
1
Ermin Husak, University of Bihać, Faculty of Technical Engineering, Irfana Ljubijankića bb., 77 000 Bihać, Bosnia
and Herzegovina, erminhusak@yahoo.com
257
fenced off for reasons that would hurt operators who worked in the production process. In the last 20 years digital
technologies have been developing that are being implemented and in production processes.
Many believe today that we are on the beginning of the new industrial revolution. It is considered that this is for
humanity fourth step which can be entitled fourth industrial revolution or “Industry 4.0”. This term first time have been
emerged in 2011 in Hanover Fair in Germany. This is development concept which is mostly coincides with
development in European countries. Can be called as “Intelligent factories”, “Smart industry” or “Advance
manufacturing”. Fourth industrial revolution is nothing else then set of rapid transformations in design, manufacturing,
work and maintenance of the manufacturing systems, which causes sudden jump in productivity and change of
human life in whole world. Fourth industrial revolution is successor of three earlier industrial revolutions (Figure 1).
As already known, first industrial revolution begins with year 1784 in the middle of the 18th century when steam
engines had been implemented in manufacturing processes. The second industrial revolution begins in year 1870 in
the end of the 19th century when electrical energy had been usedsignificantly. Mass production is implemented on
the moving lines driven by electrical motors. The thread industrial revolution begins in year 1970 in 20th century by
implementation of electronics and informational communication systems and industrial robots which additionally
automatize manufacturing processes. Currently we are in the beginning of fourth industrial revolution which is
characterized by so-called “Cyber-physical systems” (CPS).
Figure 1 Technology industrial revolutions in time periods and development of robotic technology
Potential for transformations of manufacturing processes in Europe and whole world in upcoming period is in using
digital technology, sensor technology, robotic technology and other advanced technologies which leads to
applications of new sensors, expansion of network communications, layout and networking of robots and machines,
increase of computer capacities by lower prices. The fourth industrial revolution gives us greater flexibility in
production processes, maximum adjustment of production to rapid change in the direction of customer requirements,
increasing the speed of production, better quality and increased productivity. Companies to stay on the market and
to be competitive need to use these advantages, invest in new equipment, information and communication
technology (ICT) and perform data analysis that will be on hand throughout the global value chain. The development
of digital and advanced technologies, as well as innovation in production processes represents a challenge for the
development of all technologies including robotic technology [1,12-21,23,24,25,26]. The convergence of digital and
other technologies, primarily referring to the sensor technology have influenced to develop robotic technologies as
indicated in Figure 1.The first generation of industrial robots is the first robotic revolution that occurs in sixties to
seventies of the last century, and it is the industrial automation that is most implemented in the automotive industry,
and many other manufacturing processes. Digital technologies, ICT technologies, sensor technologies and many
advanced technologies change this situation which led to development of second generation robots. Today we are
in second generation of robots where robots have more power to sense environment and there is no need to separate
them with fences. Humans collaborate with robots. Fourth industrial revolution “Industry 4.0”has contributed to the
development of the robotic technology because strategies of the leading courtiers are to go to the full automation of
production processes or “intelligent automation”. This leads to the fourth robotic revolution in robotic technologies
when all robots is going to be intelligent. The second generation of robots has created new requirements for increased
productivity and surpassed the industrial robots of the first generation. The reasons for this are many: making robots
easier to program and use, enhanced the ability to manipulate (does a diverse array of tasks), reduced size and cost
of the robot, robots are working in a wide range of dynamic environment and work together with people. Challenges
for the development of robot technology from the first generation of industrial robots to service robotics, or most
important technology for shaping the future ability of robots lie in the direction of three technical areas as indicated
in Figure 2.The first technical area is knowledge which gives the robot the ability to perceive, understand, and plans
to move into the real world. This feature improves cognitive robot capabilities so that the robot can operate
258
independently in different complex environments. The second technical area is manipulation which gives to robot
precise control and a competence to perform manipulation of objects in his environment . In this way, a significant
improvement exists in the robot control, which gives the possibility that the robot can assume a greater variety of
performing tasks in a variety of manufacturing processes.
Figure 2 Significant technologies in capability shaping of the future robots
Third technical area is interaction. Represents one of the most important areas because it gives the robots to learn
and collaborate with people. Also improving interaction robot - human for the verbal and non-verbal communication.
The robot has the ability to observe and copy human behaving and learning from experience. Safety first is the
absolute prerequisite for the operation of robots with people in the neglected environment. Digital technology made
possible the development of many digital devices (microprocessors that are the brains of digital devices and systems)
tremendous acceleration through the Internet in the industry by: video cameras, RFID readers, mobile phones,
tablets, computers, improving the quality, safety, production, maintenance plants, increase efficiency and
effectiveness in all areas.
IMPLEMENTATION OF INDUSTRIAL ROBOTS IN PRODUCTION PROCESSES

IN THE WORLD AND CHINA WITH REFERENCE TO AUTOMOTIVE INDUSTRY
Today it is impossible to imagine the production process in any industrial branch without industrial robots. In order to
get an actual idea on the modernization and automation of production processes in the industry, and given that we
are in the 21st century, but also that we are currently in the fourth industrial revolution called "Industry 4.0", it is
necessary to make a review of the representation of industrial robots in the past ten years. The statistical data for the
above analysis of the number of industrial robots were taken from the International Federation of Robotics (IFR), the
UN Economic Commission for Europe (UNECE) and the Organization for Economic Cooperation and Development
(OECD) [2-11, 22,24], whereas the data on the number of produced vehicles were taken from the Verband Deutscher
Verkehrsunternehmen (VDV) [11, 12, 22, 24], which are graphically shown in Figure 3a). Having in mind that the
largest companies for vehicle production are located on three continent Asia/Australia as well as the fact that the
highest number of industrial robots has been installed in the automotive industry production processes, we have
conducted an analysis of the representation of industrial robots and vehicle production on the three continents in the
period 2005-2015, as shown in Figure 3b). Analysis of the diagram displaying the representation of industrial robots
in the world in the period 2005-2015 (Figure 3a) brings us to the conclusion that the presence of industrial robots in
the world for the period 2005-2006 was nearly constant and was about 115.000 robot units. In 2009 the application
of industrial robots has dropped significantly to around 60.000 robot units as a result of economic and industrial crisis
in the world. In the period 2009-2015, the representation of industrial robots is growing on annual basis and in 2015
it reached the value of about 254.000 robot units. The diagram of the representation of industrial robots in the world
in the automotive industry and the tendency of application of industrial robots in all industries are completely similar,
except that application is lower, ranging from 32-40% of the total application of robots depending on the year. Thus,
in 2015 industrial robots in the automotive industry are represented with about 98.500 robot units which is about
38.75% of the total application of industrial robots in the world in this year. As can be seen, the largest number of
industrial robots is represented in the automotive industry in the automation and modernization of production
processes, which resulted in the number of vehicles produced in the world, as shown in the third diagram in Figure
3a).
259
Units Units
300.000 180.000
Industrijal Robots 254.000 160.558
160.000
250.000 Industrijal Robots
140.000
x1000 Prod. Vehicles
200.000 120.000
100.000 x1000 Production Vehicles
150.000
Rob.in Automotive Ind.
98.500 80.000
100.000 60.000 44.926.000

40.000
50.000
20.000
91.532.000
0 0
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
a-World b- Asia/Australia
Figigure 3 The tendency of representation of industrial robots and vehicle production in
production processes in World and, Asia/Austarlia in period 2005-2015 [2-11, 24]
When the vehicle production is in question, the tendency is growing annually, so that from 66 million vehicles
produced in 2005, in only ten years the production increased to 91.5 million vehicles. We can conclude that with the
increase of the representation of industrial robots in the automotive industry in the world in ten years, vehicle
production increased for 40%, which seems logical as the automation of production processes in the automotive
industry such as body welding, body painting, assembly and control, lead to decrease of the processing time and
increase in vehicle production. Based on the diagrams in Figure 3b), the representation of industrial robots in
manufacturing processes in all industry branches on all three continent of Asia/Australia in the period 2012-2015, we
see that there is a growing tendency of application of industrial robots on all three continents. In 2009 world recorded
the lowest representation of industrial robots (as shown in diagram 3a), because there was the economic and
industrial crisis in the world that also affected the application of robots on all three continents. We can conclude that
in period 2009-2015 all three continents experienced a growing trend of representation of industrial robots in
production processes, with the exception of a slight decline in the application of industrial robots in 2012 in Asia . In
the period 2005-2009 the tendency of application of industrial robots in Asia/Australia is declining. The highest
representation of industrial robots in the world is on the continent of Asia/Australia for example, in 2015 they applied
160.558 industrial robot units in all production processes. Such high representation of industrial robots is reflected in
the production of vehicles, thus making Asia/Australia the first in the world in the production of vehicles. The tendency
of vehicle production in Asia/Australia in the period 2005-2015 is continuously growing from year to year, so that in
2015 the production of vehicles reached the amount of approximately 44.926.000 units of vehicles. The consequence
of the tendency of increasing vehicle production in Asia/Australia is continuous increase in the representation of
industrial robots during automation of production processes and in the automotive industry. In addition, the fourth
industrial revolution is leading to "smart factories" that are expected in a decade, for which the most responsible is
the application of robotic technology and information technology (IT). Since we came to the conclusion that
Asia/Australia is the first place in representation of industrial robots, as well as the first in the number of produced
vehicles in the world, we need to conduct an analysis of the representation of industrial robots in the country that
produces the most vehicles in Asia.On the continent of Asia, China is in the first place by the representation of
industrial robots in production processes, particularly in the automotive industry, a result of which is that China is the
first in the world in the vehicle production annually. This is not random tendency because China adopted a national
ten-year strategy named "Made in China 2025" which aims to make China the leading country in technology
development in a few years. In order to see the tendency of representation of industrial robots and vehicle production
in China, we used statistical data given in the literature [2-11, 22], while diagramsare shown in Figure 4.
a-Industrial robots b-Predicting the representation robots

Figure 4 The tendency of representation of industrial robots in production processes in China
in period 2005-2015,and the tendency of predicting the representation by 2020 [2-8]
On the continent of Asia, China is in the first place by the representation of industrial robots in production processes,
particularly in the automotive industry, a result of which is that China is the first in the world in the vehicle production
260
annually. This is not random tendency because China adopted a national ten-year strategy named "Made in China
2025" which aims to make China the leading country in technology development in a few years. In order to see the
tendency of representation of industrial robots and vehicle production in China. If we analyze the representation of
industrial robots in production processes in China (Figure 4), we can conclude that China is consistently implementing
its national strategy "Made in China 2025", because the representation of the robots increases annually. The increase
is not per linear function, but rather exponential function, so that in the last six years the representation increased
from 5.525 robot units in 2009 to 68.556 industrial robot units, which is the increase of almost twelve times, which in
the last ten years has not happened, nor recorded in any country in the world.The tendency of application of industrial
robots in China is growing in the period to come. The predictions are that in 2020 China will reach 220.000 industrial
robot units, which presents over 45% of the total representation of industrial robots in the world in that year.
Units Units
80.000 25.000.000
70.000
20.000.000
60.000
50.000 15.000.000
40.000
30.000 10.000.000
20.000
5.000.000
10.000
0 0
a-Industrial robots b-Production vehicle

Figure 5 The representation of industrial robots in top twelve countries in the world in the period 2005-2015, as
well as vehicle production in top twelve countries in the world in 2015 [2-8,22]
Among twelve countries in the world that have the highest representation of industrial robots in production processes
in 2015 are the following: China, North Korea, USA, Japan, Germany, Taiwan, Italy, Spain, India, UK and Brazil. As
we can see based on Figure 4a), the first place is held by China with 68.556 industrial robot units, followed by the
countries in which the automotive industry is highly developed, such as North Korea, USA, Japan and Germany. If
we look at the representation of industrial robots in production processes in China in the last ten years, Figure 5a),
we see that it holds the first place in the last years, the reason being the strategy developed by China named "Made
in China 2025", which aims to make China the leading technology country in the world. Based on the image 5b), we
conclude that China has installed most industrial robots in production processes in the automotive industry because
they are the first in the world in vehicle production [18, 23]. In 2015 China produced close to 25 million vehicle units,
followed by countries that are among top five countries in the representation of industrial robots in 2015: USA, Japan,
Germany and North Korea. In addition to development and increase of application of industrial robots in production
processes, the development and increase of application of service robots in the production process.
DIGITAL TECHNOLOGY AS A DEVELOPMENT GENERATOR IN MACHINE TO

MACHINE COMMUNICATION M2M
As is well known, leading paradigm in every automation of the production process is limiting of human intervention,
and hand tasks to the machine, robot, devices and systems. Digital technology with information and communication
technologies with microelectronics, sensors, actuators and fixed and wireless networks give possibility of creating
communication machine with the machine. This digital interaction between and within the machine and the systems
is nothing but the heart of the fourth industrial revolution. Strategy of industrial development of any country in the
world are moving towards industrial automation of manufacturing processes using digital and advanced technology
and define M2M communication in the context of human and machine.
Companies engaged in robotic technology have developed second-generation of robots that have the ability to work
together with humans, where human during the work are safe. Industrial robots of the first generation had to separate
by fences because of the safety of workers and other plants. Human-Machine Interface (HMI) and Machine-to-
Machine communication is expected to be a key element in the expansion of production automation systems. With
these applications are come to the "intelligent automation" which is the goal of the fourth industrial revolution, i.e. to
improve performance by increasing the total efficiency of installed capacity (equipment) [12-30]. Here we must note
that it is not only the goal communication machine with the machine in a manufacturing process in the industry, but
the goal is the application of digital and other advanced technology for communication of all sorts of devices and
systems. In other words M2M applications can be directed to individuals, companies, communities, organizations in
the public and private sectors, as shown in Figure 5. This is about a pioneering industrial internet which will cover all
production machines, apparatuses, devices and systems which perform certain tasks. This systems can
communicate with each other just by implementing the above digital technology, so that in this communication may,
261
for example exchange the following information:i produced 20 units and I have to stop because my inbox is empty, I
have the ability to work 12% faster in how my inbox is always full, of produced 30 pieces of products two products
have been discarded, i waited for 20 minutes to produce because the inbox was empty ,i am able to reduce its energy
consumption for production as much that my equipment was idle while waiting. Please check? To the right
temperature I worked five minutes, etc.
Figure 6 Concrete example of application M2M in production process in industry
This type of communication between M2M machines, appliances, devices and systems forms the basis for "intelligent
automation" or "intelligent factory", and applications can be in heavy industry, food industry, production of goods, and
in all segments of society and different sectors. Modern M2M applications using micro-electronics and wireless digital
technology, with which these devices can collect and distribute data in real time.In this way, can be accessed at
dozens of billions of connections at will and at any time. M2M applications use sensors and counters for different
events in the range of temperature, through the communication network (fixed, wireless or hybrid), to the application
software that converts the raw data into meaningful information. Telecommunications companies in particular
recognize the opportunity to expand their services and to gain access to operational aspects of their clients. At the
stage of the research are different architecture for M2M systems and technologies that enable the development and
deployment of these systems.M2M communication systems are in development when it comes to its integration and
adaptation of existing technologies and communications systems that are currently used different processes. It is
necessary to train the algorithms to ensure functionality, effectiveness, reliability and safety of the M2M system.
Predictions are that by 2020, in the world, 20 to 50 billion devices will be able to communicate with M2M systems. It
is essential when using M2M system to expectuberisation, and this is one of the sub-category of "disruption". The
term used for the enormous changes which the new companies through technological innovation will cause in every
industry branch. Convergence of digital technologies with other technology created the second generation of
industrial robots. We think that rapidly will come the third generation of industrial robots which will be smaller, cheaper,
more autonomous, flexible and fully cooperative compared to previous generation of robot with simplifying
programming so that they can be programmed by workers. The third generation of industrial robots are intelligent
and autonomous robots whose improvements will be in the direction of: identifying specific objects, manipulation,
knowledge, increase computing performance, numerically controlled remotely, working with miniature and complex
products that require adjustment in the assembly, reliability and precision which exceeds human ability. For these
reasons the industrial and service robots are at the center of automation of production processes today, and in the
future, and it is impossible to make "intelligent automation" and "intelligent factory" without the participation of a new
generation of robots. The second generation of industrial robots can be very easily programmed so they can be
usedby the average workers who are without knowledge of robotics and computer science, in other words we do not
need professional developers that are very expensive. Some companies for development of robots developed
autonomous industrial robots so that they can work together and automatically adjust their activities in making other
products. Also some second generation industrial robots perform tasks by imitating workers. The advantage of the
second generation of robots is that it becomes a tool that improves efficiency. They are very affordable, easy to
install, have a very low rate of investment in them, it is not necessary to reorganize the manufacturing process for its
installation because it is not necessary to separate workers with fences and they are without special safety
equipment.
CONCLUSIONS
The fourth industrial revolution accelerates exponentially. This is the third phase of acceptance of digital technology,
but it was preceded by digital competence and digital use. Digital transformation enables innovation and creativity in
a single domain, not just the application of traditional, but adopted technologies. IoT (Internet of things) represent
new ways in which we communicate with machines, as well as the manner in which the devices that we use at home,
at work, transport, to be connected. From communication M2M is expected to be a key element in the expansion of
262
automation in the manufacturing process, which will, with the participation of sensor and robotic technology lead to
"intelligent automation". Also new way to communicate with devices that are used in the home, transport to work,
etc., is the key to the implementation of the fourth industrial revolution, so it is estimated that by 2020, about 50 billion
devices will be interconnected, of which will be about 10 billion traditional computer devices. In the next five years,
the largest application of digital technology will achieve global companies, hence they will reduce costs, increase
productivity, and allow extension of the implementation to the new areas. With digital design and virtual modeling of
the production process we are able to reduce the time between the design of a product and its delivery to the market.
In this way we come to the great improvements in product quality and a significant reduction of defectsin
manufacturing. Intelligent automation allows greater flexibility in the production, so that different products can be
producedin the same production facility. In the first place in the world in representation of industrial robots in the
automotive industry is China, but also the first in the world in the production of vehicles. It is expected that in the
following period there will be an increase in the representation of industrial robots in production processes worldwide,
with new types of industrial robots that will be able to work together with workers, and in particular in the automotive
industry because the current structures of industrial robots are such that they need to be separated by the
compartments in order not to harm the workers. The predictions are that China will still be the leader in the world in
terms of the representation of industrial robots, as well as the production of vehicles. The fourth industrial revolution
which includes digital and other technologies bring us to the "intelligent production" in the next 10 years. Fourth
industrial revolution provides technology available to everyone. It is assumed that technology in the future will not
provide a competitive advantage, but competitive advantage will be in manner in which we used it.
REFERENCES
[1] Doleček, V., Karabegović, I.: "Robots in the industry", Technical Faculty of Bihac, Bihać, Bosnia and
Herzegovina, 2008,
[2] World Robotics 2015, United Nations, New York and Geneva, 2015,
[9] Verband Deutscher Verkehrsunternehmen VDA: “AutoJahresbericht 2011", Koln, Deutschland, 2011,
[10] Verband Deutscher Verkehrsunternehmen VDA: “AutoJahresbericht 2012",Koln, Deutschland, 2012,
[11] Verband Deutscher Verkehrsunternehmen VDV: “Jahresbericht 2014/2015", Koln, Deutschland, 2015,
[12] BMWi (Hrsg.): “Plattform Industrie 4.0 – Online-Bibliothek“, 2016. URL: http://www.plattform-
i40.de/I40/Navigation/DE/In-der-Praxis/Online-Bibliothek/online-bibliothek.html, Accessed 1.02.2017.,
[13] VDE (Hrsg.): "Arbeitsgruppen und Gremien im Bereich Industrie 4.0", 2016, URL:
https://www.dke.de/de/themen-projekte/excellencecluster-industrie-4-0/arbeitsgruppen-und-gremien-im-
bereich-industrie-4-0, Accessed 12.02.2017.,
[14] The Future of Jobs, Global Challenge Insight Report, World Economic Forum, januar 2016, Geneva,
Switzerland, URL: https:/www.weforum.org, Accessed 14.02.2017.,
[15] Arntz, M., Gregory, T., Jansen, S., Zierahn, U.: "Tätigkeitswandel und Weiterbildungsbedarf in der digitalen
Transformation", Studie des Zentrums für Europäische Wirtschaftsforschung (ZEW) und des Instituts für
Arbeitsmarkt und Berufsforschung (IAB) im Auftrag der Deutschen Akademie der Technikwissenschaften, 2016,
[16] Arntz, M., Gregory, T., Lehmer, F., Matthes, B., Zierahn, U.: “Arbeitswelt 4.0 – Stand der Digitalisierung in
Deutschland: Dienstleister haben die Nase vorn", IAB Kurzbericht 22/2016,
[17] Verl A.: "Robotick & Industrie 4.0", ISW, Stuttgart, Germany, 2016,
[18] Sulavik, C., Portnoy M., Waller T.: “How a new generation of robots is transforming manufacturing”,
Manufacturing Institute USA, September 2014, Gaithersburg, USA, 2016, pp 1-13,
[19] Bunse B., Kagermann H., Wahister W.: "Industrija 4.0. Smart Manufacturing for the Future”, Germany Trade &
Invest, Berlin, Germany, 2015,
[20] Smart industry-a strategy for new industrialisatuon for Sweden, Gevernment Offices of Sweden, Stockholm,
Sweden, 2015,
[21] Pires, N.: "New development on Industrijal Robotics”, Lisabon, Portugal, 2015, https//www.robotics.dem.uc.pt
[22] Karabegović, I.: "Role of Industrial Robots in the Development of Automotive Industry in China", International
Journal of Engineering Works, Vol. 3, Iss. 12, Kambohwell Publisher Enterprises, Multan, Pakistan, 2016, pp 92-
97,
[23] Schlogel, K.: "Hello Industrie 4.0-Smart solutions for smart factories", Kuka, Agsburg, Germany, 2016,
[24] Karabegović, I., Husak, E.: “China as a leading country in the world in automation of automotive industry
manufacturing processes”, Proceedings of the IV International Congress Motor Vehicles & Motors 2016, 06-08
October 2016, Kragujevac, Serbia, 2016,
[25] Jeschke, S.: "Roboter in der Automobilindustrie“, Fachkonferenz, 27 Oktober 2015, Augsburg, Germany, 2015,
263
[26] Richard Kozul-Wright, Robots and industrialization in developing countries, UNCTAD/PRESS/PB/2016/6 (No.
50), United Nations Conference on Trade And Development UNCATD, Geneva, Switzerland, 2015, pp 1-4,
[27] Robotics 2020 Strategic Research Agenda for Robotics in Europe, Produced by euRobotics aisbl, Draft 0v42
11/10/2013, pp 25-43, http://www.eurobotics-project.eu,
[28] Guang-Zhong, Yang: "The Next Robotic Industrial Revolution, Manufacturing Robotics, Robotics and
Autonomous Systems (RAS)”, UK-RAS Manufacturing Review 2015-2016, 2016, ISSN 2398-4422, pp 2-17,
[29] Good Jobs in the Age of Automation, BSR-The Business af a Better World, 2015, New York, USA, pp 1-28,
[30] Fris, D., Editorial: "Universal Robots, Editorial_WR_Industrial_Robots_2016", www.ifr.org/news/ifr-
press.../world-robotics-report-2016-832/.
264
Kragujevac, Serbia
MVM2018-010
Nenad Kostić1
Nenad Petrović2
Vesna Marjanović3
Nenad Marjanović4
AUTOMATED SELECTION OF ANALYTICALLY

CALCULATED GEARBOX CONCEPTS ACCORDING TO
DIMENSIONAL CRITERIA
ABSTRACT: This research allows for a selection of gearbox according to minimal dimensional criteria from
automated analytical calculations. An automation of the calculation process for four different analytical approaches
from literature. The results achieved in the software are compared according to various criteria, minimal length,
height, width, and volume. The automation of the process allows for a selection of gearbox concept depending on
the chosen criteria. For specific input values, a verification of the automated software operation was conducted, and
the resulting values are compared and discussed. Differences in calculated volumes of gearboxes, depending on
choice of calculation, reach over 50%. This approach allows for adequate choice of gearbox for specific cases and
speeds up the calculation process.
KEYWORDS: gearbox volume, comparative analysis, automated selection, analytical calculation
INTRODUCTION
Recent research in the field of development and application of gearboxes is oriented on achieving better working
characteristics, decreasing mass, volume, used material, as well as decreasing the number of different elements
used in the construction. This allows transmissions to be more cost-effective from a production and maintenance
standpoint. Also, these transmissions have useful exploitation characteristics in terms of performance. Improving
gearboxes boils down to an engineer’s vast knowledge and experience in this field and the use of alternate
approaches, in order to allow for progress.
A lot of research is directed towards the development of reducers and their improvement. Marjanovic et al. [1]
developed a practical approach to optimizing gear trains with spur gears based on a selection matrix of optimal
materials, gear ratios and shaft axes positions. Researchers in [2] compared optimization results to analytical
calculations comparing the performance of these processes. Kostic et al. [3] presented a new approach for solving
gearbox optimization using the mimicking of natural processes to achieve layout solutions comparable to optimization
results without using computerized optimization methods. Clearances between cycloid gearbox elements were
optimized in [4] in order to provide a realistic picture of contact and machining tolerances. Golabi et. al [5] presented
1
Nenad Kostić, research assistant, University of Kragujevac, Faculty of Engineering, Sestre Janjić 6, 34000
Kragujevac, nkostic@kg.ac.rs
2
Nenad Petrović, assistant, University of Kragujevac, Faculty of Engineering, Sestre Janjić 6, 34000 Kragujevac,
npetrovic@kg.ac.rs
3
Vesna Marjanović, associate professor, University of Kragujevac, Faculty of Engineering, Sestre Janjić 6, 34000
Kragujevac, vmarjanovic@kg.ac.rs
4
Nenad Marjanović, full professor, University of Kragujevac, Faculty of Engineering, Sestre Janjić 6, 34000
Kragujevac, nesam@kg.ac.rs
265
gear train volume/weight minimization optimizing single and multistage gear trains’ gear ratios. Mendi et al. [6] aimed
to optimize gear train component dimensions to achieve minimal volume comparing GA results to analytic method
parameter volume. Savsani et al. [7] described gear train weight optimization comparing various optimization
methods to genetic algorithm (GA) result values. Gologlu and Zeyveli [8] performed preliminary design automation
through optimization of gear parameters and properties using a GA based approach.
The motivation behind this research is based in the desire to determine the real influential parameters of different
methods and standards for calculating geared speed reducers. The calculation process has been automated, in order
to allow for application on a larger number of reducers. By giving a comparative analysis, based on dimensional
criteria, the best calculation for a specific case can be determined. Choosing the adequate calculation method a
smaller length, width, height or volume of a gearbox can be achieved, resulting in the best possible gearbox
applicable in practice.
PROBLEM DEFINITION
Designing a gearbox presents a complex task due to the large number of parameters which influence its operation.
Design solutions can be achieved analytically or numerically. For the purposes of this research analytical processes
based on ISO standard [9], Petruševič, GOST standard [10] and Kudrijavcec [11] have been automated. Testing
criteria is based on dimensions of a gearbox. Length, width, height, and volume are used for a comparative analysis
of calculation results.
Conditions and constraints of calculations

Input/output values for calculating gearboxes are singular, in order to be comparable. Values for input/output are
given in Table 1.
Table 1 Input parameters and working conditions

Denotation Value
Input power, Pi (kW) 25
Material 34CrAlNi7
Input speed ni (rpm) 2800
Total gear ratio ur 12.5
Usage Electricity motor
Available modules (mm) 1, 1.125, 1.25, 1.375, 1.5, 1.75, 2, 2.25, 2.5, 2.75, 3, 3.5, 4, 4.5, 5, 5.5,
6, 6.5, 7, 8, 9, 10, 11, 12, 14, 18, 20, 22, 25, 28, 30, 32, 36, 40
In order for the gearbox to properly function and to ensure adequate lubrication, certain clearances must be created.
There must be clearances between gears and bearings, between the largest gear diameters and the housing, as well
as between the gears themselves. These values also go into the calculation of the total length, width, height, and
volume of the gearbox. Technical clearances are adopted according to suggestions, and are used for the purposes
of this research as is shown in Table 2.
Table 2 Technical clearances

Denotation Label Value
Distance between gear face and bearing c1 15 mm
Distance between outside gear diameter and gearbox housing c2 15 mm
Distance between outside diameter of the biggest gear and gearbox 50 mm
c3
housing bottom
Distance between gears c5 15 mm
Values and positions of technical clearances are illustrated in Figure 1. The figure also shows the length (L), width
(B), and height (H) which are the output values for comparing calculation results. These values are the basic criteria
for comparing gearboxes.
266
Figure 1 Gearbox dimensions
For the minimal safety factor against pitting and brakeage SHmin = SFmin = 1.2. Testing for the safety factors against
pitting and brakeage was done in Autodesk Inventor, in order to verify the analytically calculated results.
AUTOMATED ANALYTICAL CALCULATION

The conducted research requires a large number of extensive calculations which is why it has been automated. Aside
from this, in order to allow for application on a larger number of problems it is necessary to automate the process so
that the invested time and effort of calculating would not be too large in comparison to the resulting effects.
Automating the previous calculation is conducted in Microsoft Excel and covers the following segments:
 Input of values (material characteristics, input/output of the gearbox, clearances between elements);
 Database of standard modules and constraints of specific values (i.e. integers for number of gear teeth);
 Calculations according to the used suggestions (ISO, Petrusevic, GOST, Kudrijavcev);
 Visualization and result processing ( comparative tables, calculation of length, width, height, and volume)
Values from table 1 are plugged in as input values for calculation in the Excel file and all calculation types are done
based on identical input values. The calculation is automated in such a way as to require manual input of technical
clearance values between the gearbox elements, and are thereby taken into account in the final dimension results
of the gearbox. Technical clearances for this specific case have the same values as are given in table 2. Once these
values are entered, the final results which are used for the analysis are obtained. As output values length, width,
height, and volume are set as the basis for comparing the resulting gearboxes.
Based on defined input values the analytical calculation was automatically done, according to different suggestions.
Stiffness was tested in AD Inventor, and the calculated values were also automatically compared to each other based
on length, width, height and volume criteria. Based on these results an adequate design concept can be adopted
which would be suitable for practical application on real-world problems.
RESULTS
According to calculations based on the various suggestions, parameter values are obtained for the gearboxes. In
values shown in Table 3 these gearboxes are completely defined and can be used. These values show various
dimensions, which means various length, width, height, and volume, therefore it is possible to analyse which
calculation should be used, and when it shouldn’t.
267
Table 3 Gearbox comparison for all design solutions
Denotation ISO Petrusevic GOST Kudrijavcev
1st Stage
Normal module, mn1st (mm) 1.75 1.75 1.75 2
Gear ratio, u12 4.309 4.488 3.15 2.5
Number of teeth on pinion, z1 24 24 24 24
Number of teeth on wheel, z2 103 108 76 60
Face width, b12 (mm) 38 38 38 43
2nd Stage
Normal module, mn2st (mm) 3.75 3.75 3.25 3
Gear ratio, u34 2.901 2.785 4 5
Number of teeth on pinion, z3 18 18 18 19
Number of teeth on wheel, z4 52 50 72 95
Face width, b34 (mm) 61 61 53 51
Based on values from Table 3 the length, width, height, and volume of each gearbox is derived. This approach is
important, as it is not always explicitly required that volume is minimized, but a specific dimension depending on
available space. Therefore it is often more important to adopt a design with a specific value minimized, rather than
the volume. This approach leads to designs of various dimensions and volumes based on the previous calculations.
Calculated lengths are shown in Figure 2 for all calculation cases. The range of values for all the different designs is
from 387.75 (Petrusevic) to 451.5 mm (Kudrijavcev). According to this criteria the best concept is the design
according to Petrusevic’s calculations.
470
451.5
Gearbox length, L [mm]
450
430
410 401.75
390.875 387.75
390
370
350
ISO Petrusevic GOST Kudrijavcev
Calculation type
Figure 2 Gearbox lengths
According to the width criteria values corresponding to each design solution are given in Figure 3. The range of
values is from 151.84 (GOST) to 159.84 mm (ISO and Petrusevic). According to this criteria the most favorable
solution is according to GOST standard.
163
161 159.84 159.84
Gearbox width, B [mm]
159
157 154.96
155
153 151.84
151
149
147
145
Calculation type
Figure 3 Gearbox widths
268
According to height criteria values corresponding to each design solution are given in Figure 4. The range of values
is from 252.6 (Petrusevic) to 350 mm (Kudrijavcev). According to this criteria the most favorable solution is according
to Petrusevic.
370
350
Gearbox heights, H [mm]

350
330
310 299
290
270 260
252.6
250
230
210
Calculation type
Figure 4 Gearbox heights
According to the criteria of overall volume of the gearbox each design solution value is given in Figure 5. The range
of values is from 15.65 (Petrusevic) to 24.49 mm3 (Kudrijavcev). According to this criteria the most favorable solution
is according to Petrusevic.
26
24.49
Gearbox volume, V [dm3]
24
22
20
18.24
18
16.24
15.65
16
14
Calculation type
Figure 5 Gearbox volumes
All calculated values take into account technical clearances. For practical application of gearboxes it is possible to
determine the type of calculation which gives desired results, while all calculated design concepts are practically
applicable.
CONCLUSIONS
For this research an approach to the choice of concept of gearbox was developed based on criteria of length, width,
height, and volume. Gearboxes were designed based on different suggestions for calculation. An automated process
of calculating was developed, which based on input parameters achieves final dimension values. The calculations
take into account clearances between elements of the gearbox, standard modules, as well as other realistic
constraints. A validation of gear stiffness was conducted in AD Inventor. This approach gives practical results and
based on specific situations, an engineer can choose the most suitable calculation for their application.
The calculated values vary greatly, which indicates the large influence of the calculation type on final gearbox design
dimensions. The most suitable value according to length criteria is achieved using Petrusevic’s suggested calculation
(387.75 mm). Compared to the ISO standard the length is greater by 1% which is negligible. The difference for the
GOST standard is around 3.6%, while the difference from Kudrijavcev’s calculation is 16.5%.
According to width criteria, the best value is achieved according to GOST standard (151.84 mm). For ISO and
Petrusevica calculations give larger widths than the GOST for around 5.3%, while Kudrijavcec’s calculations give
around 2% larger width.
269
According to the criteria of height, the best values are achieved using Petrusevica’s calculation (252.5 mm). For ISO
calculation the height is around 3% larger, the GOST standard around 18.4% larger, while the values according to
Kudrijavcec are larger by abour 38.6%.
In terms of volume criteria the best solution is using Petrusevic’s calculations (15.65 dm3), followed by ISO standard
(16.24 dm3), then GOST standard (18.24 dm3), and lastly using Kudrijavcev’s calculations achieving (24.49 dm3).
Compared to the lowest value, ISO standard volume is 3.8% greater, 16.6% greater using GOST standard, and using
Kudrijavcev’s calculations is around 56.5%.
The smallest difference is achieved in terms of volume criteria, then by length, while the greatest difference can be
made in the height of the gearbox. The most common criteria for choice of gearbox concept is volume, however it is
not compulsory. Using the here proposed approach, an alternative is created for engineers to choose adequate
designs depending on the type of chosen calculation. Further research in this field should include a comparative
analysis of achieved values to those achieved through optimization according to defined criteria.
ACKNOWLEDGMENTS
Acknowledgment: This paper is a result of the projects TR33015 of the Technological Development of Republic of
Serbia, titled “Investigation and development of Serbian zero-net energy house”. We would like to thank to the
Ministry of Education, Science and Technological Development of Republic of Serbia for their financial support during
this research.
REFERENCES
[1] Marjanovic, N., Isailovic, B., Marjanovic, V., Milojevic, Z., Blagojevic, M., Bojic, M.: "A practical approach to the
optimization of gear trains with spur gears", Mechanism and Machine Theory, Vol. 53, 2012, pp 1-16,
[2] Marjanovic N., Kostic N., Petrovic N., Blagojevic M., Matejic M., Isailovic B.: "Comparative Analysis of Analytical
Calculation and Optimization on Gearbox Dimensions and Volume", Machine Design, Vol. 6, No. 4, 2014, pp
127-130,
[3] Kostic N., Marjanovic N., Petrovic N.: "A Novel Approach for Solving Gear Train Optimization Problem", Mobility
and Vechicle Mechanics, Vol. 42, No. 2, 2016, pp 67-76,
[4] Kostic N., Blagojevic M., Petrovic N., Matejic M., Marjanovic N.: "Determination of real clearances between
cycloidal speed reducer elements by the application of heuristic optimization", Transactions of FAMENA, Vol.
42, No. 1, 2018, pp 15-26,
[5] Golabi, S. i., Fesharaki, J. J., Yazdipoor, M.: "Gear train optimization based on minimum volume/weight design",
Mechanism and Machine Theory, Vol. 73, 2014, pp 197-217,
[6] Mendi, F., Başkal, T., Boran, K., Boran, F. E.: "Optimization of module, shaft diameter and rolling bearing for
spur gear through genetic algorithm", Expert Systems with Applications, Vol. 37, No. 12, 2010, pp 8058-8064,
[7] Savsani, V., Rao, R. V., Vakharia, D. P.: "Optimal weight design of a gear train using particle swarm optimization
and simulated annealing algorithms", Mechanism and Machine Theory, Vol. 45, No. 3, 2010, pp 531-541,
[8] Gologlu, C., Zeyveli, M.: "A genetic approach to automate preliminary design of gear drives", Computers &
Industrial Engineering, Vol. 57, No. 3, 2009, pp 1043-1051,
[9] Nikolic, V.: "Masinski elementi", Mašinski fakultet u Kragujecu, CIPMES, 2004, Kragujevac,
[10] Trbojevic, M., Jankovic, M., Vukdelija, J., Plavsic, N., Latinovic, V.: "Reduktori", Naučna knjiga, 1977, Beograd,
[11] Kudrijavcev, V. N.: "Detali mašin", Mašinostroenie, 1980, Lenigrad, Russia.
270
Kragujevac, Serbia
MVM2018-027
Saša Mitić1
Ivan Blagojević2
END-OF-LIFE VEHICLE DISPOSAL

AND IT’S INFLUENCE TO THE ENVIRONMENT
ABSTRACT: Motor vehicle industry is nowadays exposed to numerous serious challenges, mostly related to its
influence to environment. Motor vehicles during their life cycles affect the environment in several aspects: through
consumption of energy and other resources, through production of waste in design process and usage, and
through disposal at the end of their life cycle. First two aspects are in consideration for many years and there is
noticeable improvement in reducing the effects, but vehicle disposal at the end of life cycle is relatively new area of
research. This paper presents main items of international legislative related to end-of-life vehicle recycling process,
as well as responsibilities of all participants in this process, from producers to dismantlers and landfills. Also,
current legislation in Serbia is presented, with suggestions and ways for future improvement. At the end, concrete
data of recyclability and reusability rates for some vehicle models are given, approving the positive effects of
applied legislation on vehicle influence to the environment.
KEYWORDS: end-of-life vehicle, recycling, environment, legislative
INTRODUCTION
Vehicle production is continuously growing. Since the end of World War 2, when the global world development and
restructuring of the world industries led to growing of the automotive industry, we have constant increase of vehicle
production in the world. Following the exact data of produced vehicles in the past, the situation is that from year
1950, when the world production amounts 10 million units, we came up to 58 million units in year 2000 [1]. In next
15 years the increase was more than 50%, so we had more than 90 million vehicles produced in 2015. Report for
year 2016 shows 95 million produced vehicles, which indicates further growing of world vehicle production (Figure
1).
End of year 2008 and beginning of 2009 was the only period since 1950 when the drop of vehicle production was
recorded, and it was caused by global financial crisis that shaken all industrial and commercial branches. But very
quickly, yet in year 2010, increase of vehicle production recovered all reductions from previous year, so again we
have constant and stable tendency of vehicle production growing.
Also, all relevant forecasts indicate that this tendency should be kept on in near future.
Regarding the information about vehicles that were sold per year, it was noticeable that their number consistently
follows the number of produced vehicles in each year, with slight reduction comparing to them. The only exception
was year 2009, when the number of vehicles that were sold surpassed the number of produced vehicles. The
reason for that situation was planned reduction of vehicle production and obvious selling of certain number of
vehicles from the stocks.
1
Prof. dr Saša Mitić, University of Belgrade - Faculty of Mechanical Engineering, smitic@mas.bg.ac.rs
2
Prof. dr Ivan Blagojević, University of Belgrade - Faculty of Mechanical Engineering, iblagojevic@mas.bg.ac.rs
271
100.0
95.0
89.8 90.8
87.6
90.0 84.2 93.9
79.9
77.6 88.3 89.7
80.0 73.2 85.6
69.2 70.7 82.1
No. of vehicles in millions 66.7 78.2
70.0 61.7 75.0
71.6
68.4 68.3
60.0 65.9 65.6
50.0 produced
sold
40.0
30.0
20.0
10.0
0.0
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Figure 1 Worlwide vehicle production and selling, period 2005-2016
For further analysis, it is very important to analyze the information about current number of vehicles in use and its
structure. This number also increase from year to year, and approximately have the same tendency as the vehicle
production. The tendency remains the same for passenger and commercial vehicles (Figure 2).
1,400
1,200
335
328
316
1,000 307
289
280
272
265
No. of vehicles in millions
259
800 238
247
commercial vehicles
600 passenger vehicles
947
869 907
400 808 835
748 776
702 727
654 679
200
0
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Figure 2 Worldvide vehicle production and selling, period 2005-2015
All previously mentioned indicators unambiguously indicate the importance of research in this field and creation of
global strategy for managing the end-of-life vehicles (ELV). According to available data, and if we derive
information for year 2015, we come up to the following figures:
 The number of vehicles sold in year 2015 was approximately 90 million units;
 Increase of number of vehicle in use comparing to year 2014 was approximately 47 million units.
The difference between these figures indicates the number of vehicle units in year 2015 that came up to the end of
their life cycle, and that number is 43 million units! Negligence of these indicators and not-setting up of appropriate
measures could lead us to the situation for our planet to become one big landfill.
On the other hand, real situation indicates that the real number of ELVs is probably lower than officially presented.
That fact is related to massive export of used vehicles from EU member countries to the rest of Europe (mainly east
Europe, Russia, etc.) and some North African countries. Obviously, the profit achieved with vehicle export is much
higher, so the problem of ELV treatment EU member countries solve through export. But, the problem of ELV
treatment has not been solved, it just has been postponed and transferred to other countries to deal with it [13].
272
INTERNATIONAL LEGISLATIVE
Considerable national policies and voluntary agreements by major automobile manufacturers have been developed
concerning the environmental impact of vehicles over their lifetimes. At the end of 1999, ten EU member countries
(Austria, Belgium, France, Germany, Italy, the Netherlands, Portugal, Spain, Sweden, and the United Kingdom)
had specific regulations and/or industrial voluntary agreements addressing to ELV. These countries represent
almost 96% of ELV estimated to be in the European Union [6].
The Directive of European Parliament and of the Council 2000/53/EC of September 18th, 2000 [3] organized former
national policies and voluntary agreements. It was aimed to harmonize these existing rules and to push the EU
governments and automobile industry to comply fully with the Directive and to translate its key requirements into
national laws. The ultimate goal of Directive 2000/53/EC is to put only 5% of ELV residues into landfills. It states:
1. Member States shall take the necessary measures to encourage the reuse of components which are
suitable for reuse, the recovery of components which cannot be reused and the giving of preference to
recycling when environmentally viable, without prejudice to requirements regarding the safety of vehicles
and environmental requirements such as air emissions and noise control;
2. Member States shall take the necessary measures to ensure that the following targets are attained by
economic operators:
a. No later than January 1st, 2006, for all ELVs, the reuse and recovery shall be increased to a
minimum of 85% by an average weight per vehicle and year. Within the same time limit the reuse
and recycling shall be increased to a minimum of 80% by an average weight per vehicle and year;
for vehicles produced before January 1st, 1980, Member States may lay down lower targets, but
not lower than 75% for reuse and recovery and not lower than 70% for reuse and recycling.
Member States making use of this subparagraph shall inform the Commission and the other
Member States of the reasons therefore;
b. No later than January 1st, 2015, for all ELVs, the reuse and recovery shall be increased to a
minimum of 95% by an average weight per vehicle and year. Within the same time limit, the reuse
and recycling shall be increased to a minimum of 85% by an average weight per vehicle and year.
Waste prevention, reuse, recycling, and recovery of the ELV constituents so as to reduce automotive shredding
residues (ASR) waste disposal are the objectives of the EC Directive 2000/53/EC. Figure 3 is a schematic
representation of the participants in the ELV chain, according to the EU Directive. The main actor is the producer, a
vehicle manufacturer or professional importer of a vehicle into a EU member state. The producer links the
upstream (supplier) and downstream in the ELV chain (collector, dismantler, and shredder). On the other hand,
collaboration between collector, dismantler, and shredder are necessary to successfully meet the directive goals.
The vehicle produced has to at least meet the following goals: low energy consumption, easy dismantling, suitable
recycling, and less toxic metals (as shown in Figure 3). To fulfill these goals, the producer has to know the technical
and economical facilities, recyclability rate, and efficiencies of the downstream ELV chain. On the other hand, the
producer will provide the dismantling information for each new type of vehicle put on the market. The design of
vehicles appropriate for dismantling, recycling, and reuse, and free of some hazardous substances (Pb, Hg, Cd,
and Cr) will significantly improve the cooperation of the supplier-producer chain [5].
The Directive 2000/53/EC require that the ELV collector and dismantler should be certified (licensed), and as a
result, the number of licensed dismantlers in the EU member countries has increased significantly, exceeding
1,000 licensed enterprises per country in the top five producers of vehicles in the EU. The dismantler’s role is the
removal for sale of reusable parts such as engines, transmissions, gearboxes, and body parts. According to the
Directive 2000/53/EC, removing pollutants from the vehicle becomes an important task of the dismantler business.
This involves the draining of liquids and removing of environmentally harmful constituents such as the battery.
Furthermore, dismantlers are certified to destroy the waste resulting from removing the pollutants (i.e., depollution).
These tasks by the dismantler will facilitate the subsequent hulk shredding and will reduce the ASR generated by
the shredder operators.
Shredding steps include dismantling small parts for recycling, hulk shredding, and ferrous and non-ferrous metal
separation. The separated materials will likely go to automakers for use in the production of the same components
from which they are issued. Energy can be recovered from combustible parts of ELV by using them instead of fossil
fuels in industrial operations, such as cement plants. The remaining part of the vehicles, ELV waste, will go to a
landfill under strict waste control. This will be material for which there is no justification for recovery [8].
273
Figure 3 Main steps in ELV recycling process, according to EC Directive 2005/64/EC
EC Directive 2000/53/EC was used as a base for adoption of another Directive, 2005/64/EC [4], which is related to
the type-approval of motor vehicles with regard to their reusability, recyclability and recoverability. This Directive
strictly prescribe responsibilities of vehicle manufacturers, as well as responsibilities of National Competent
Authorities to ensure the fulfillment the Directive's requirements.
Some of the requests related to vehicle manufacturers:

 The manufacturer shall make available to the Type Approval Authority the detailed technical information
necessary for the purposes of the calculations and checks, relating to the nature of the materials used in
the construction of the vehicle and its component parts;
 The manufacturer and the Type Approval Authority jointly identify the reference vehicle in accordance with
the criteria prescribed by this Directive;
 For the purposes of checks of the materials and masses of component parts, the manufacturer shall make
available vehicles and component parts as deemed necessary by the Type Approval Authority;
 The manufacturer must be in a position to demonstrate that any version within the vehicle type complies
with the requirements of this Directive;
 The manufacturer shall recommend a strategy to ensure dismantling, reuse of component parts, recycling
and recovery of materials. The strategy shall take into account the proven technologies available or in
development at the time of the application for a Vehicle Type Approval;
 The manufacturer declares measures aiming at the reduction of the quantity and the harmfulness for the
environment of end-of life vehicles, their materials and substances.
The competent body shall ensure that the manufacturer has taken the necessary measures to:
 Collect appropriate data through the full chain of supply, in particular the nature and the mass of all
materials used in the construction of the vehicles, in order to perform the calculations required under this
Directive;
 Keep at its disposal all the other appropriate vehicle data required by the calculation process such as the
volume of the fluids, etc.;
 Check adequately the information received from suppliers;
274
 Manage the breakdown of the materials;
 Be able to perform the calculation of the recyclability and recoverability rates in accordance with the
standard ISO 22628: 2002 [7];
 Verify that component part prescribed as not to be reused in the construction of new vehicles is reused;
 Mark the component parts made of polymers and elastomers in accordance with Commission Decision
2003/138/EC of February 27th, 2003 establishing component and material coding standards for vehicles
pursuant to Directive 2000/53/EC of the European Parliament and of the Council on ELV.
Also, EC Directive 2005/64/EC prescribe the component parts of vehicles belonging to category M1 and those
belonging to category N1 which must not be reused in the construction of new vehicles:
 All airbags, including cushions, pyrotechnic actuators, electronic control units and sensors;
 Automatic or non-automatic seat belt assemblies, including webbing, buckles, retractors, pyrotechnic
actuators;
 Seats (only in cases where safety belt anchorages and/or airbags are incorporated in the seat);
 Steering lock assemblies acting on the steering column;
 Immobilizers, including transponders and electronic control units;
 Emission after-treatment systems (e.g. catalytic converters, particulate filters);
 Exhaust silencers.
When the vehicle fulfills all requirements regarding the EC Directive 2005/64/EC, which has to be verified with
appropriate Test Report, competent authority issue an Approval Certificate. Generally, it means that any new
produced vehicle (vehicle of new type) could not be placed into market without valid Approval Certificate.
EC Directive 2005/64 has entered into force in November 2005, starting with application from December 2006,
depending of the level of requirement fulfillment. These cases are covered with Transitional provisions.
The same thing is related to requests of UN Regulation 133 [12], which deals with the approval of motor vehicles
with regard to their reusability, recyclability and recoverability, within member parties of 1958 Agreement at the
World Forum for Harmonization of Vehicle Regulations (WP 29). Since the number of member parties of 1958
Agreement is much higher than EU member countries, the process of harmonization took much more time, so UN
Regulation 133 entered into force from June 2014. Comparing with EC Directive 2005/64/EC, it was noticeable that
all requirements were almost the same, and it was on the line with EU legislative and UN Regulations
harmonization process.
NATIONAL LEGISLATIVE
Recycling is one of the youngest industrial branches in Serbia, and at the same time the only one that has
increasing rate. The best indicator is that in last 3 years this branch of industry engage more than 10,000 people.
Nowadays there are about 2,200 companies dealing with waste collecting and recycling. Comparing with year
2009, when only 200 companies in this field existed, the improvement is enormous.
Along with engineers and environment experts, recycling industry engages waste collectors all around the country
as well. Very often they came from marginalized social groups, so they are covered with social security and
included in legal flows.
The way of ELV management in Serbia has not been systematically covered, although nowadays exist
approximately 2.3 million registered vehicles, average of 16.5 years. With no integrated and system approach to
vehicle recycling, Serbia suffers great loss of resources (materials, energy, employment), and on the other side
there are a lot of negative ecological consequences. Serbian trip for joining to EU and status of candidate imposed
much more serious approach to this matter than it was in the past.
Relevant national legislative related to ELV management is as follows:

1. Waste Management Law [14];
2. Waste Management Strategy for Period 2010-2019 [15];
3. Methods and Procedures of End-of-Life-Vehicles (ELV) Management Regulation [10].
All 3 main documents were enetered into force in 2010, and generally they were designed for thoroughly covering
the whole field of ELV management. Some of main activities scheduled with these documents are [2]:
 Responsibilities of manufacturers, local authorities and all other participants involved into ELV chain;
 Procedures to be taken for ELV, from the owner to the landfill;
 Keeping the evidence about all steps in ELV management;
 Issuing the appropriate documents needed for completing the ELV management procedure;
275
 Short-term and long-term predictions of development process for ELV management;
 Predictions about annual amount of waste by categories;
 Managing the statistic data regarding the recycling process annually;
 To form database of registered and approved vehicle recycling facilities;
 Implementing of EC Directives into national legislative documents;
 Keeping all financial flows through ELV management under control.
The implementation of legislative is another part of the story, unfortunatelly not very successful at the moment. One
of the reasons is previously mentioned fact that this is younges industrial branches in Serbia, faced with numerous
problems. The other reason is very slow and difficult process of “mind changing” and acceptance of changes in
long-term praxis.
Some of the problems noticed through implementation of new legislative are [9], [11]:
 The treatment of ELV is not in accordance with requirements related to environmental protection;
 ELV management system still does not exist as organized activity;
 It is still noticeable the presence of dumped vehicle landfills with possibility for people to come and
dismantle needed used parts for certain financial compensation;
 Domestic manufacturers still do not provide complete information about materials used for vehicle
production, coding of components and general information regarding recycling requirements;
 There is no designed policy of population education regarding the vehicle recycling;
 Penalty policy is still too “soft”, and without stronger implementation of legal part of legislative and higher
penalties there is no further strong implementation of ELV management efficiently.
RECYCLABILITY AND RECOVERABILITY RATES

The method for calculating recyclability and recoverability rates is specified by Standard ISO 22628:2002. It is
based on four main stages inspired by the treatment of ELVs. Recyclability and recoverability rates depend on the
design and material properties of new vehicles, and on the consideration of proven technologies – those
technologies which have been successfully tested, at least on a laboratory scale, in this context.
The calculation method of this Standard cannot reflect the process that will be applied to the road vehicle at the
end of its life.
This Standard specifies a method for calculating the recyclability rate and the recoverability rate of a new road
vehicle, each expressed as a percentage by mass (mass fraction in percent) of the road vehicle, which can
potentially be
 Recycled, reused or both (recyclability rate), or
 Recovered, reused or both (recoverability rate).
The calculation is performed by the vehicle manufacturer when a new vehicle is put on the market.
The calculation of the recyclability and recoverability rates is carried out through the following four steps on a new
vehicle, for which component parts, materials or both can be taken into account at each step:
 Pretreatment;
 Dismantling;
 Metals separation;
 Non-metallic residue treatment.
Recyclability rate, Rcyc, of the vehicle, is calculated as a percentage by mass (mass fraction in percent), using the
following formula:
mp  mD  mM  mTr
Rcyc   100 (1)
mV
Recoverability rate, Rcov, of the vehicle, is calculated as a percentage by mass (mass fraction in percent), using the
following formula:
mp  mD  mM  mTr  mTe
Rcov   100 (2)
mV
276
where:
mp – mass of materials taken into account at the pre-treatment step (all fluids, batteries, oil filters, liquefied
petroleum gas (LPG) tanks, compressed natural gas (CNG) tanks, tyres, catalytic converters);
mD – mass of materials taken into account at the dismantling;
mM – mass of metals taken into account at the metal separation step;
mTr – mass of materials taken into account at the non-metallic residue treatment step and which can be considered
as recyclable;
mTe – mass of materials taken into account at the non-metallic residue treatment step and which can be considered
for energy recovery;
mV – vehicle mass.
For the calculation purposes, a partial mass parameters are determined respectively, at each of these four steps.
At the end, the data for the calculation shall be reported using the formalized table review, either on paper or in
electronic form.
This formalized review has to be presented to the National Type Approval Authority. Also, it presents a part of
Information Document that follows Approval Certificate according to EC Directive 2005/64/EC and/or UN
Regulation 133.
Figure 4 shows completed formalized table for one of the passenger vehicle models, with clear view to all
parameters related to material breakdown and mass rates in recycling process.
Figure 4 Formalized table review for passenger vehicle (Ford Kuga)

277
CONCLUSIONS
ELV Management became very important part of environmental protection. More and more vehicles are in use and
consequently large number of vehicles finishes its life cycle every day, becoming subject of ELV recycling process.
Therefore, strong and distinct legislative is needed in order to protect environment. Current international legislative,
provide reliable and clear way for ELV managing in the future, but only if it is going to be applied strictly.
Regarding Serbian national legislative, year 2010 was turning point, because 3 main acts were entered into force.
Unfortunately, application of legislative is not thorough, hence the effects of possible positive impact are missing.
On the other hand some important parts of legislative, especially in “Methods and Procedures of End-of-Life-
Vehicles (ELV) Management Regulation” were not written distinctly, so we have different interpretations, and in
some cases we have inability to implement them for ELVs because of bad formulations. Therefore, national
legislative need to be revised, and much more adopted to international, especially to EC Directive 2005/64/EC and
UN Regulation 133.
Having all previously mentioned, main conclusion is imposing by itself: without strong and strict legislative
application, it is impossible to keep ELV recycle process under control. Without that control, having in mind
enormous number of very old vehicles on the market, Serbia could become big landfill for ELVs, with no possibility
to recover for many decades. If we want to avoid that scenario, we need to change our way of thinking, to educate
young generations indicating the importance of environment protection and to act “green”.
ACKNOWLEDGMENTS
This paper presents some of the results obtained through the project supported by Serbian Ministry of Education,
Science and Technological Development (Project No. TR 35045 - "Scientific-Technological Support to Enhancing
the Safety of Special Road and Rail Vehicles").
REFERENCES
[1] 2015 Statistics of Road Traffic Accidents in Europe and North America, Volume 53, United Nations Economic
Commission for Europe, 2015, Geneva,
[2] Ćurčić, S., Paunović, L.: “Recycling of the End-of-Life Motor Vehicles - Situation in Serbia (in Serbian)”,
Fakultet tehničkih nauka u Čačku, Informacione tehnologije, obrazovanje i preduzetništvo ITOP17, 2017,
[3] Directive 2000/53/EC of the European Parliament and of the Council of 18 September 2000 on end-of life
vehicles, Consolidated Version, European Parliament and the Council, 2016, Brussels,
[4] Directive 2005/64/EC of the European Parliament and of the Council of 26 October 2005 on the type-approval
of motor vehicles with regard to their reusability, recyclability and recoverability and amending Council
Directive 70/156/EEC, Consolidated Version, 02005L0064-20090203, European Parliament and the Council,
2009, Brussels,
[5] Ex-post Evaluation of Certain Waste Stream Directives – Final Report, European Commission – DG
Environment, 2014,
[6] Frad, A., Revnic, I.: “ProdTect Automotive – Meeting the Requirements of ELV”, International Conference on
Engineering Design ICED ’07, Paper ID: 245, 2007, Paris, pp 28-31,
[7] ISO 22628:2002 – Road vehicles – Recyclability and recoverability – Calculation method, International
Organization for Standardization, 2002, Geneva,
[8] Kanari, N., Pineau, J.-L., Shallari, S.: “End-of-Life Vehicle Recycling in the European Union", The Journal of
The Minerals, Metals & Materials Society (TMS), Vol. 55, No. 8, 2003, pp 15-19,
[9] Kozić, A., Sudarević, D.: “Approach to Vehicle Recycling (in Serbian)”, 32. Nacionalna konferencija o kvalitetu
– Festival kvaliteta 2005,2005, Kragujevac,
[10] Methods and Procedures of End-of-Life-Vehicles (ELV) Management Regulation (in Serbian), Official Journal
98/2010, Belgrade, 2010,
[11] Pešić, R., Babić, S., Milosavljević, B.: “Recycling in Automobile Industry (in Serbian)”, 36. Nacionalna
konferencija o kvalitetu – Festival kvaliteta 2009, 2009, Kragujevac,
[12] Regulation No. 133 - Uniform provisions concerning the approval of motor vehicles with regard to their
reusability, recyclability and recoverability, The United Nations Economic Commission for Europe (UNECE),
2014, Geneva,
[13] Schneider, J., Karigl, B., Neubauer, C., Tesar, M., Oliva, J., Read, B.: End-of-Life Vehicles: Legal aspects,
National Practices and Recommendations for Future Successful Approach, Directorate General for Internal
Policies, European Parliament's Committee on Environment, Public Health and Food Safety, 2010,
[14] Waste Management Law (in Serbian), Official Journal 36/2009, 88/2010, 2010, Belgrade,
[15] Waste Management Strategy for Period 2010-2019 (in Serbian), Official Journal 29/2010, 2010, Belgrade.
278
Kragujevac, Serbia
MVM2018-031
Dimitrios V. Koulocheris1
Clio G. Vossou2
ALTERNATIVE DESIGN FOR A SEMI – TRAILER TANK

VEHICLE
ABSTRACT: The assessment of the design parameters of a semi-trailer tank vehicle is crucial and it is related to
the three motions, namely longitudinal (driving and braking), lateral (guidance and steering), and vertical (suspension
and damping). Since semi-trailer tanks are mainly used for the transportation of dangerous goods, European
Standards specify the minimum requirements for their construction and handling. In the present study, different semi
– trailer tank designs with the same overall length are investigated using the finite element (FE) software ANSYS®
v.18.0 and the calculation method provided in European Directives. The design parameters under consideration
consist on the geometry of cross section and the section of the tank, the number of compartments and the payloads
and the materials used. Different computational models of a semi - trailer tank vehicle have been set up in order to
investigate the influence of each of the above mentioned design parameters. Their boundary values as well as the
suitable computation for structural integrity and handling are defined through the corresponding Standards and the
restrictions posed by the manufacturing procedures of such a semi – trailer tank. The outcomes of this study provide,
among others, a useful insight for tank manufacturers.
KEYWORDS: semi-trailer tank vehicle, structural design, handling, European Directives, Finite Element Method
INTRODUCTION
The typical way of transportation for liquid and granular material by road is with the use of tank vehicles. Common
liquid materials transported by road are fuels which are dangerous goods. Liquid fuels belong in Class 3 of flammable
liquids for which the main danger is this of fire. The safety regulations for vehicles carrying dangerous goods are
outlined by the “European Agreement concerning the International Carriage of Dangerous Goods by Road” (ADR)
and specifically “Part 6 - Requirements for the construction and testing of packagings, intermediate bulk containers
(IBCs), large packagings and tanks” (ADR, 2017). In the aforementioned text the basic requirements for design,
construction, testing, inspection, re-testing, qualification and maintenance of such tanks are thoroughly described.
Tanks, in general, can be divided into categories according to their construction material and their maximum working
pressure. If the tank is metallic and its working pressure is not exceeding 0.5 bar its design and construction is at the
same time related to the European Standard EN 13094 (EN13094, 2015). According to both previously mentioned
Standards the cross-section of tank can be cyclical, elliptical or box-shaped. Furthermore, their section can be
rectangular, wedge-shaped or cone-shaped and they can be compartmented or not. In Annex A of the EN13094 it is
stated that there are four different methods for the verification of the design of such a tank, namely, (a) dynamic
testing, (b) finite element stress analysis, (c) reference design, (d) calculation method or a combination of them. In
order for (a) or (c) to be performed the construction of the tank is a prerequisite, while both (b) and (d) can be
performed right after the preliminary design of the tank, prior to its construction, leaving space for the construction of
an optimized design (Koulocheris, 2017).
In ADR different categories of tanks are mentioned such as fixed tanks, also referred as tank-vehicles, demountable
tanks, tank containers and portable tanks. A tank – vehicle can be a motor vehicle, an articulated vehicle, a trailer or
1
Dimitrios V. Koulocheris, Assistant Professor, Vehicles Laboratory, School of Mechanical Engineering, National
Technical University of Athens, Greece, Zografou Campus, Iroon Polytexneiou 9, 157 80, dbkoulva@mail.ntua.gr
2
Clio G. Vossou, Researcher, Vehicles Laboratory, School of Mechanical Engineering, National Technical
University of Athens, Greece, Zografou Campus, Iroon Polytexneiou 9, 157 80, klvossou@mail.ntua.gr
279
a semi-trailer. Semi–trailers are trailers without front axle. Most European trailers have three axles with single-tire
hubs totaling 6 wheels. A large proportion of the semi-trailer weight is supported by the tractor unit.
According to Eurostat (Eurostat, 2017), 77.6% of the total tonne-kilometres transported in 2015 have been
transported with semi-trailers and road tractors. A common type of semi-trailer is the semi-trailer tank vehicle which
is mainly used for hauling liquids such as gasoline and alcohol, or various types of gases. They are similar in principle
to intermodal trailers but with a different frame intended to be attached to a liquid or gas tank. The maximum
dimensions and weights for international traffic, with respect to road safety reasons and to avoid damaging roads,
bridges and tunnels are set with Directive (EU) 2015/719. According to this Directive two-axle motor vehicles with
three-axle semi-trailers transporting tanks or swap bodies of a length of up to 13.7 m should be allowed in intermodal
transport operations up to a total authorized weight of 42 tonnes.
In this paper guidelines for an alternative design for a semi-trailer tank vehicle are provided. With this respect different
computational models of semi-trailer tank vehicles developed in the Vehicles Laboratory of the National Technical
University of Athens are investigated. Each semi-trailer tank vehicle geometrical model consists of the shell walls,
the front and rear end, different number of partitions or surge plates and top openings. The tank is mounted on the
vehicle with six supports. The front two supports are mounted on one support plate which connects with the motor
vehicle through a fifth wheel coupling while the rest four of them are connected to a support plate each and are
mounted on the semi-trailer. The distance between the two front supports is restricted due to the geometry of the
section of the tank while the distance between the rest of the supports is restricted by the axles of the semi-trailer
thus it is the same for all semi-trailer tank vehicles. With the use of the Finite Element (FE) method the influence of
the cross sectional geometry, the construction material and the compartmentalization of the tank, to the semi-trailer
tank vehicle performance is investigated. For the implementation of the FE method ANSYS® v. 18.0 finite element
(FE) software has been utilized. In total sixteen (16) FE models have been set up and evaluated in the loading cases
provided by the Standard EN13904, for the evaluation of the structural integrity of the tank vehicles. The outcomes
of this study offer a useful insight for tank manufacturers.
FINITE ELEMENT MODELS & DESIGN PARAMETERS

In Figure 1 the side view of a typical semi-trailer tank vehicle is provided.
Figure 1 Semi-trailer tank vehicle
In order to investigate the influence of different design parameters in semi-trailer tank vehicles, the geometrical 3D
models of five (5) main semi-trailer tank vehicle configurations have been constructed. Their geometrical
characteristics are provided in Table 1.
Table 1 Geometrical characteristics of the main semitrailer tank vehicle configurations

Overall Overall Width (mm)
Cross Sectional Number of
Configuration Length
Geometry Front Back Compartments
(mm)
1a Box-shaped 11340 2470/1538 2470/1768 11
2a Circular 11380 1900 2250 11
3a Circular 10280 1900 2250 10
4a Circular 11201 1900 2250 10
5a Circular 12030 1900 2250 11
280
In all configurations the different parts of the semi-trailer tank vehicle are considered to be in fully bonded contact
simulating welded connections. Thus, the shell walls of the each compartment (Ki) are in contact with the front
end/partition (Di) and the shell walls of the final compartment are in contact with both the front partition and the rear
end of the tank. All shell walls are in contact to each other while the top openings are in contact with the corresponding
shell walls. Finally, the tank shell walls are mounted to the semi-trailer with six supports (Si) and their corresponding
support plates (SPi). The first two supports (S1 & S2), which are in contact with SP1, are in contact with the shell
walls of the smaller dimensions (in front of the wedge shaped part of the tank section) while the rest of the supports
(S3 to S6) are in contact with the shell walls of the larger dimensions (behind the wedge shaped part of the tank
section). The rear supports in all cases have a distance of 1320 mm between them, which is equal to the distance of
the axles of the semi-trailer tank vehicle and are placed between them as it is obvious in Figure 1. For every tank
there is a combination of the dimensions da, db and dc (Figure 1) in order for the semi-trailer tank vehicle to host the
designed tank. In Figure 2 the geometrical models of the main semi-trailer tank vehicle configurations are presented.
(a) (b)
(c) (d)
(e)
Figure 2 Geometrical models of the main semi-trailer tank vehicle configurations
In Table 2 information on the construction material, the tare weight and the payload of each configuration is reviewed.
The payload of all semi-trailer tank vehicles has been calculated considering the tanks filled with water up to their
highest point.
Table 2 Material, tare weight and payload of the FEA models

Configuration Material Tare Weight (N) Payload (N)
1a Al 2360 39500
2a Al 1980 40000
3a Al 1820 36000
4a Al 2040 39500
5a Al 1930 42500
Using these semi-trailer tank vehicle configurations as a starting point, new setups are generated in order to evaluate
the influence of different design parameters and a comparative analysis has been performed for all the FEA analyses.
281
The design variables under consideration consist on the geometrical and material characteristics of the tank and the
number of compartments. All the design parameters along with their explored values are presented in Table 3. In
Table 3 also the initial configuration used for the corresponding setups is mentioned.
Table 3 Design parameters under consideration and FEA models

Design Parameter Value Main Configuration
Box shaped Config1
Geometry of cross section
Circular Config2
St
Construction Material Config1 & Config2
Al
10, 8, 7 Config3
Number of compartments for the same overall length
11, 10, 8 Config4
(Setups)
10, 8, 6, 4, 3, 2 Config5
In Figure 3 the partitions for the configurations of different cross sections are presented. In order to construct the
partitions of different geometry, different manufacturing methods are engaged. Hence, the box-shaped partitions are
manufactured through moulding, while the circular ones with press forging.
Isometric view Front View
Box Shaped
Circular
Figure 3 Crossectional geometry of the ends and partitions
In Table 4 the physical and mechanical properties of aluminium alloy (Al) and structural steel (St) are presented.
Table 4 Physical and mechanical properties of aluminium alloy and structural steel
Material
Property Al
(EN14286, St
2008)
Density (kg/m3) 2660 7850
Elastic modulus (MPa) 70300 210000
Poisson’ s ratio 0.3 0.3
Tensile strength (MPa) 125 250
Ultimate strength (MPa) 275 360
Maximum Allowable stress (MPa)
93.75 176.25
(ΕΝ13094, 2015)
The separation of configurations 3, 4 and 5 in compartments lead to different setups generated either from changing
the number and position of the partitions or by replacing the closed geometry of the partitions with an open one, of
the surge plates. Surge plates are non-hermetically closed partitions intended to reduce the effect of surge mounted
at the right angles to the direction of travel. Surge plates have the same geometry with partitions but have holes on
their surface in order to allow the flow of the transported liquid. The area of the surge plate shall be at least 70% of
the cross-sectional area of the tank (ADR, 2017) and the openings are created according to ΕΝ13094 (EN13094). In
Figure 4 the geometrical model of a partition and a surge plate are presented.
282
(a) (b)
Figure 4 3D model of (a) a partition and (b) a surge plate
In Figure 5, the side view of configurations 3a and 4a is presented with respect to the position of the axles of the
semi-trailer. The position of the axles is represented by the three perpendicular centrelines. Support S3 is placed in
such a way with respect to the axles of the semi-trailer that its mid plane is located 700 mm from the first axle.
Name Side view
Config3a
Config4a
Figure 5 Side view of Configuration 3 and 4
For each configuration two setups have been generated without changing the location of the partitions, but with just
replacing them with surge plates. In more details, while in Configuration 3a all partitions are closed, in configuration
3b partition number 5 and 7 are replaced with surge plates leading to a tank of 8 compartments, while in configuration
3c partitions number 4, 6 and 8 are replaced with surge plates leading to a tank of 7 compartments. As far as
configuration 4a is concerned all its partitions are closed and the tank consists of 11 compartments. In configuration
4b partition 8 is replaced with surge plates leading to a semitrailer tank of 10 compartments, while in configuration
4c partitions number 5, 7 and 9 are replaced with surge plates leading to a tank of 8 compartments.
In Figure 6 the alternative set-ups of configuration 5 are also presented with respect to the position of the axles. In
configurations 5d, 5e and 5f the positioning of the partitions does not change, but specific partitions are replaced with
surge plates. In the rest of the setups both the placement and the number of partitions varies.
283
Name Side view
Config5a
Config5b
Config5c
Config5d
Config5e
284
Config5f
Figure 6 Different set-ups of Configuration 5
In more details in configurations 5a and 5b all partitions are closed and they consist of 10 and 8 compartments,
respectively. Configuration 5c has 5 surge plates and 5 partitions and consists of 4 compartments. In configuration
5d, only the second partition is replaced with a surge plate and it consists of 6 compartments. In configuration 5e
additionally partitions number 3, 5 and 7 are replaced with surge plates and it consists of 3 compartments, while
configuration 5f consists of 2 compartments having only the 5th partition closed.
In order investigate the influence of the aforementioned design parameters and evaluate each design, the loading
conditions described in the Standard EN13094 are going to be implemented in each FE model. According to this
Standard, the shell walls, the shell ends, the partitions (or surge plates) and their attachments are designed to
withstand the stresses developed due to the (a) dynamic, (b) pressure and (c) partial vacuum conditions. According
to these conditions, the following loading cases (LC) are implemented.
1. Acceleration of 2g acting on the maximum design mass in the direction of the travel
2. Acceleration of 1g acting on the maximum design mass at right angles of the travel
3. Acceleration of 1g acting on the maximum design mass vertically, upwards
4. Acceleration of 2g acting on the maximum design mass vertically downwards
5. The maximum of
a. the pressure created by a column of water equal to twice the depth of the tank multiplied by the
density of the relative density of the denser substance
b. the pressure created by a column of water equal to twice the depth of the tank
c. 1.3 times the working pressure
6. Test pressure equal to 1.3x(Pts+Pta)
7. Vacuum condition of 3 kPa below atmospheric pressure.
Pts is the pressure of the breathing devices which is equal to 0.012 MPa in all configurations and setup, while Pta
varies among configurations and it is equal to the pressure corresponding to the maximum design mass.
LC 1 – 4 simulate the dynamic conditions and should be applied along with the pressure of the breather device. LC
5 – 6 simulate the pressure conditions, while LC 7 simulates the partial vacuum conditions. LC 6 is applied
considering the payload of each compartment separately, while the rest LC 1 – 5 and 7 are applied to the semi-trailer
tank vehicle considering its total payload. In all LCs the SP were considered fully fixed, simulating welded connections
to the semi-trailer frame.
RESULTS
In total, 16 FE models have been constructed. In Table 5 the number of the FE and the nodes of the FE models of
the five main configurations along with the average quality of the mesh for each configuration is presented. A
convergence study has been performed for all FE models in order to define the mesh and the same mesh has been
used in all setups.
Table 5 Mesh characteristics for the FE models of all configurations

Configuration Number of FE Number of Nodes Average Quality
1a 56341 59362 0.93
2a 13587 14828 0.82
3a 13706 14977 0.82
4a 14623 15949 0.81
5a 15751 17297 0.82
285
The mesh in all configurations consists of SHELL181 surface FE and CONTA174 and TARGE170 surface contact
FE. The criterion for meshing is uniformity, with maximum edge length of 50 mm. SHELL181 is a FE used to model
thin to moderately thick shell structures, such as pressure vessels and tanks (Ansys, 2017, Das, 2105 & Koulocheris,
2016). It consists of 4-nodes with six degrees of freedom at each node, three translational and three rotational ones.
The results of the FE models in terms of maximum stress values and stress contours are presented in this section.
The FE models consist of the 5 main configurations (denoted with the corresponding number of their configuration
and the letter “a”) with the geometrical characteristics presented in Table 1 and the mesh characteristics presented
in Table 3. All main configurations are considered to be constructed out of aluminium alloy. For configurations 1 and
2 a new set up has been created with structural steel as construction material (denoted with the corresponding
number of their configuration and the letters “st”. Finally, new setups have been developed for configurations 3, 4
and 5 changing the compartmentalization of the semi-trailer tank vehicles.
Main configurations
At the beginning the results of the 5 main configurations are going to be presented. The stress results of all the LCs
are going to be presented in groups. Initially, the results of the LCs simulating the dynamic conditions are going to
be presented. In Table 6 the maximum value of equivalent Von Mises stress (SEQV) in MPa for each of the LC1-4
is presented, along with the component where it was observed.
Table 6 Maximum value of equivalent Von Mises stress per FE model for LCs 1-4
LC1 LC2 LC3 LC4
Configuration SEQV SEQV SEQV SEQV
Component Component Component Component
(MPa) (MPa) (MPa) (MPa)
1a 80 S1 55 D4 60 D2 82 D4
2a 34 D1 34 S4 25 K10 34 S3
3a 36 D1 27 S4 34 K10 28 S3
4a 34 D1 40 S2 26 K4 38 S3
5a 38 D1 28 S4 14 K10 24 S4
In Figure 7 the equivalent Von Mises stress contours for LC1-4 for the tank are presented for configurations 1a and
2a. The contours of these two configurations are presented in comparison since they have different cross-section
geometry but almost equal total length and payload.
LC Config1a Config2a
286
3
Figure 7 Equivalent Von Mises stress contours for LC1-4 for configurations 1 and 2
Furthermore, the maximum value of SEQV, in MPa, obtained from LC 5 and 7, which are the pressure conditions
and the vacuum condition, accordingly, are presented in Table 7, along with the component where it appears.
Table 7 Maximum value of equivalent Von Mises stress per FE model for LC 5 & 7
LC5 LC7
Configuration SEQV SEQV
Component Component
(MPa) (MPa)
1a 76 D2 8 D2
2a 57 D12 4 D12
3a 54 D11 4 D11
4a 54 D12 4 D12
5a 51 D11 4 D11
In Figure 8 the equivalent Von Mises stress contours for LC5 & 7 for the tank are presented for configurations 1a and
2a.
287
7
Figure 8 Equivalent Von Mises stress contours for LC5 & 7 for configurations 1 and 2
Finally, in Table 8, the results of the LC 6 which simulates the pressure conditions per compartment are presented.
Table 8 contains the number of the compartment whose pressurization lead to the maximum value of SEQV, in MPa,
and the component where it appeared.
Table 8 Maximum Von Mises stress per FE model

LC6
Configuration SEQV
Compartment Component
(Mpa)
1a 2 83 D3
2a 11 58 D11
3a 9 57 D10
4a 11 55 D12
5a 10 63 D11
In Figure 9 the equivalent Von Mises stress contours for LC6 for the tank are presented for configurations 1a and 2a.
Figure 9 Equivalent Von Mises stress contours for LC6 for configurations 1 and 2
Construction material
For both configurations 1a and 2a new FE models have been set up using structural steel as construction material.
In Table 9 the maximum value of SEQV in MPa for each of LC1-4 is presented, along with the component where it
was observed.
Table 9 Maximum value of SEQV per FE model for LCs 1- 4

LC1 LC2 LC3 LC4
Configuration SEQV SEQV SEQV SEQV
1_St 78 S1 57 D3 67 D2 86 D4
2_St 35 D1 33 S4 36 K10 35 S3
Likewise, in Table 10 the maximum value of SEQV in MPa for each LC5 & 7 is presented, along with the component
where it was observed.
288
Table 10 Maximum value of SEQV per FE model for LC 5 & 7
LC5 LC7
Configuration SEQV SEQV
Component Component
(MPa) (MPa)
1_St 88 D2 10 D2
2_St 58 D12 5 D12
Finally, in Table 11 the maximum value of SEQV in MPa along with the compartment where it appeared.
Table 11 Maximum value of SEQV per FE model for LC 6

LC6
Configuration SEQV
Compartment Component
(MPa)
1_St 1 99 D2
2_St 11 59 D12
Compartmentalization
As mentioned above for configurations 3, 4 and 5 different setups have been constructed and their FE models have
been solved for the same LCs. In the Tables 12 – 14 the results for the setups of these configuration are presented.
For all configurations the results of setup “a” are omitted since they are presented in the previous tables. In more
details, in Table 12 the maximum value of SEQV for LC 1-4 is presented for all setups.
Table 12 Maximum value of SEQV per FE model for LCs 1- 4

LC1 LC2 LC3 LC4
Setup SEQV SEQV SEQV SEQV
3b 35 D1 26 S2 33 K10 25 S3
3c 34 D1 29 D7 34 K9 24 S3
4b 39 D1 53 S2 26 K4 42 S2
4c 34 D1 36 S2 27 K4 37 S3
5b 48 S3 52 S2 21 K8 37 S3
5c 47 S3 53 S5 18 K2 36 S3
5d 48 S3 66 S4 21 K6 37 K4
5e 48 K2 69 S4 18 K2 38 K2
5f 49 S3 85 D6 21 D6 47 D6
In Table 13 the same data is presented for LCs 5 and 7.
Table 13 Maximum value of SEQV per FE model for LC 5 & 7

LC5 LC7
Setup
SEQV(MPa) Component SEQV (MPa) Component
3b 54 D11 4 D11
3c 55 D11 4 D11
4b 54 D12 4 D12
4c 54 D12 4 D12
5b 45 D8 4 D9
5c 45 D10 4 D10
5d 45 D8 4 D8
5e 45 D8 4 D8
5f 43 D8 3 D8
Lastly, in Table 14 the same results are presented for LC 6 with the addition of the compartment where the maximum
value appears.
289
Table 14 Maximum value of SEQV per FE model for LC 6
Setup Compartment SEQV(MPa) Component
3b 7 54 D10
3c 3 55 D4
4b 10 54 D12
4c 8 54 D12
5b 8 44 D9
5c 4 45 D8
5d 6 44 D8
5e 3 44 D8
5f 2 43 D8
DISCUSSION
In the present paper 5 configurations of semi-trailer tank vehicles have been designed and computationally simulated
using the FE method in the loading cases prescribed in the European Standard EN13094 for the construction of
tanks for the transport of dangerous goods. The minimum overall length (Table 1) of these configurations is 10280
mm while the maximum is 12030 mm in order to be mounted on the semi-trailer that it is presented in Figure 1 and
they consist of 10-11 compartments. Furthermore, configuration 1 has a box-shaped cross-section, while the rest
circular ones. The range of their payload (Table 2) is between 36000 to 42500 N and their tare weight ranges from
1800 to 2400 N. It is to be noted that the calculated payload is higher than the real one since all semi-trailer tank
vehicles are constructed for the transportation of liquid fuels with densities less than 800 kg/m3. All semi-trailer tanks
are built with aluminum alloy. In order to lower the overall center of gravity of the semi-trailer tanks all the
configurations have a partially wedge shaped section, as it becomes obvious in Figure 2.
In Figure 10 the ratio of the payload to tare weight, which monitors the material exploitation is provided for all the
basic configurations in terms of bars. Configuration 4 seems to have the most efficient exploitation of construction
material.
Figure 10 Payload to tare weight ratio for all configurations
In Table 5 is obvious that the FE models of all configurations have almost the same number of FE and nodes except
for configuration 1. This happens due to its box-shaped geometry which is more complex than the circular one and
needs more nodes in order for convergence to be achieved.
In Tables 6-8 the results of the FE analysis, in terms of maximum value of SEQV for the main configurations are
summed up. In all the aforementioned tables is obvious that configuration 1, which has a box-shaped cross section,
displays the highest values of SEQV for all loading cases. All the configurations of circular cross-section have stress
values of the same order in all LCs. Furthermore, the location where this value appears remains the same in most
cases. Comparing the LCs corresponding to the dynamic conditions, LC1 which simulates breaking of the semi-trailer
tank vehicle causes the highest Von Mises equivalent stress values. Considering the pressure conditions, LC6, which
simulates the test pressure conditions per compartment, leads to highest values of equivalent Von Mises stress.
290
Observing the stress contours of configuration 1a and 2a is obvious that the way the stresses are distributed on the
semi-trailer tank vehicle varies due to the cross-sectional geometry. The circular cross-section offers a more uniform
distribution, while in the box-shaped cross section is obvious that although the sides of the semi-trailer tank vehicle
are relieved, this is not the case for its top and bottom. Still though, it needs to be stressed out that even if the box-
shaped cross-sectional geometry provides higher stress values they are still below the allowable stress (Table 4).
The same conclusions can be drawn from Tables 9-11 where the results of the setups 1_st and 2_st are presented.
The change of construction material raises the values of SEQV in both configurations and for all loading cases, but
leaves its location unaltered. Regardless construction material, the semi-trailer tank vehicle with the box-shaped
cross-section geometry has higher SEQV values. It is worth mentioning that the highest stress value increase has
been 44% in LC3 for configuration 2 which has circular cross-section. On the other hand the maximum allowable
stress of structural steel is 88% higher than this of the aluminium alloy, so even if the maximum SEQV values
increased the design of the semi-trailer tank vehicle became safer.
Reviewing Tables 12-14 is obvious that the different compartmentalization does not influence particularly the
maximum SEQV value in any LC. On the other hand some alterations on the location where it appears exist. In order
to investigate the role of the supports in Figure 11, that follows, the reaction forces on Z-axis for LC1 for all SPs for
configurations 3 – 5, for all setups are presented.
(a) (b)
(c)
Figure 11 Reaction force on Z-axis on the support plates for LC1
It is obvious that the front support plate (SP1) where S1 and S2 are mounted presents the highest value of reaction
force on Z axis, regardless configuration and setup. The front support plate is used for the fifth wheel coupling and it
sustains 28 – 35% of the total load of the semitrailer tank vehicle in LC1. The load transferred from the tank to the
semi-trailer reduces moving to its rear end from a maximum of 29% to SP2 in configuration 4b to a minimum of 11%
to SP5 in the same configuration. It is worth mentioning that the values of the reaction forces on Z-axis are elevated
since the tank was simulated as fully loaded with water, which is not the real – world case.
In Figure 12, the reaction forces on X-axis for LC2 for all support plates for configurations 3 - 5 are presented. Again
it is obvious that SP1 has the highest value for the reaction force on X-axis regardless configuration or setup. This
support sustains a mean of 30% of the total load of the semitrailer tank vehicle in LC2. The rest 4 SP that are mounted
on the semi-trailer sustain the rest of the load. In configurations 3 and 4 SP4 has the minimum value of reaction force
on X-axis, while SP3 and SP5 have a similar value of reaction force. On the other hand on configuration 5, SP4
displays the minimum value in all configurations, except configuration 5c.
291
(a) (b)
(c)
Figure 12 Reaction force on X-axis on the support plates for LC2
CONCLUSIONS
In the present paper sixteen (16) different configurations and setups of semi-trailer tank vehicles of maximum payload
have been computationally simulated with the use of the FE method and their performance has been monitored in
terms of structural integrity. The influence of the cross-sectional geometry of the tank, its construction material and
its compartmentalization has been emphasized. As far as the cross-sectional geometry is concerned, the circular
one has been found to be more efficient and it is easier to be manufactured. On the other hand since the box –
shaped cross-sectional geometry provides lower centre of gravity their dynamic behaviour remains to be evaluated.
In terms of construction material aluminium alloy provides lighter structures able to sustain all loading conditions and
it is the material of choice of the tank manufacturers. Finally, the compartmentalization does not influence the
structural integrity of the tank in terms of stress distributions, but it does influence the load transfer from the tank to
the semi-trailer vehicle. Alternative placements of the support plates with respect to the axles of the semi-trailer
vehicle remain to be explored.
REFERENCES
[1] ADR, “European Agreement concerning the International Carriage of Dangerous Goods by Road”, 2017,
[2] ANSYS®Academic Research, Release 18. 2017,
[3] Das, S., Sundaresan M.K., Hameed, A.S.: “FEA of Cylindrical Pressure Vessels with Different Radius of
Openings”, Int. J. for Research in Applied Science & Engineering, Vol. 3, Issue IX, 2015, pp 406-414,
[4] Eurostat, and Union européenne. Commission européenne. Energy, transport and environment indicators.
Office for Official Publications of the European Communities, 2017,
[5] EN 13094, “Tanks for the transport of dangerous goods – Metallic tanks with a working pressure not exceeding
0.5 bar – Design and construction”, 2015,
[6] EN 14286, “Aluminium and aluminium alloys Weldable rolled products for tanks for the storage and transportation
of dangerous goods”, European Committee for Standardization, Brussels, 2008,
[7] Koulocheris, D., Vossou, C.: "Computational Assessment of a cylindrical tank vehicle structural integrity, 11th
HSTAM International Congress on Mechanics”, 11th HSTAM International Congress on Mechanics, 2016,
Athens, Greece,
[8] Koulocheris, D., Vossou, C.: "Finite Element Stress Analysis Vs Calculation Method For The Construction Of A
Metallic Tank Used For Dangerous Goods Transportation", 2017 JUMV Automotive Conference, 2017,
Belgrade.
292
Kragujevac, Serbia
MVM2018-033
Ralph Pütz1
Matthias Kampfl2
Thomas Hechfellner3
EXPLOITATION OF THE GOLF BALL EFFECT FOR

FURTHER AERODYNAMIC OPTIMISATION OF RACE CARS
ABSTRACT: The exploitation of the golf ball effect promises further aerodynamic optimization of race cars if the
improvement in air resistance does not significantly affect downforce. Investigations – in the windtunnel and (as a
verification) also by CFD calculations – have been carried out on a Formula 3 rear wing with different dimple setups
on the main wing and the flap such as dimples in front end, in the rear end or all over of both main wing and flap.
Many different combinations have been analysed with different angles of attack. The basis for comparison was a
standard, smooth rear wing with a standard, smooth flap. Furthermore, slight side inclination to simulate crosswinds
has been looked at. The results show that different combinations of dimple setups in main wing and flap allow for
improvements in aerodynamic efficiency over the standard wing only for specific angles of attack, and there is not
"the" optimal combination for all use cases.
KEYWORDS: aerodynamics, golf ball design, dimples, rear wing, race cars, downforce, drag
INTRODUCTION
The physical backgrounds of the golf ball effect to improve aerodynamics by reducing air resistance are largely known
and are increasingly being used outside of the sport of golf. In motorsport, where aerodynamics make up a significant
part of the success, and the improvement of one-hundredth of a second per lap can already be crucial, the golf ball
effect could be of great importance in the future if air resistance improvements do not lead to significant loss of
downforce. The achievable cornering speed can be significantly increased by downforce-increasing measures:
𝐹𝑉,𝐿,𝑧
𝑣𝑉 = √𝜇𝑊,𝑦 · (𝑔 + )·𝑅
𝑚𝑉,𝑡,𝑑𝑟
where:
FV;L;z Downforce by aerodynamic measures, N
The influence of the golf ball effect on the example of the rear wing (with flap) of a Formula 3 (F3) race car is
examined, on the one hand in the model wind tunnel and on the other hand with the aid of a supporting CFD
simulation. Various arrangements of dimples and bumps on the rear wing and as well the flap at different angles of
attack and different flow velocities have been investigated. The original standard F3 rear wing with a smooth surface
on the wing and flap serve as a reference.
1
Prof. Dr.-Ing. Ralph Pütz, Landshut University, BELICON Institute for Vehicle Research – Dept. Motorsport,
Am Lurzenhof 1, D-84036 Landshut, Germany, ralph.puetz@belicon-forschung.de
2
Matthias Kampfl B.Eng., Landshut University
3
Thomas Hechfellner B.Eng., Landshut University
293
Figure 1 Downforce-increasing measures allow a significant increase in cornering speed
(Source: Pütz, R./Serné, T.)
FLUID DYNAMICS FUNDAMENTALS
Types of flow and boundary layer

In principle, a differentiation is made in fluid mechanics between the laminar flow and the turbulent flow. The fluid
particles move in laminar flow in adjacent layers. When the flow velocity is increased, the flow becomes turbulent
from a critical value, with the fluid particles then carrying out irregular movements transversely to the flow direction
in addition to their main flow direction. This leads to significant losses due to a higher momentum exchange. The
Reynolds number Re is a crucial parameter to estimate the type of flow:
𝑣𝑆𝑡𝑟ö𝑚 · 𝐷
𝑅𝑒 =
𝜈𝐿
where:
𝑅𝑒 : Reynolds number [-]
𝑣𝑆𝑡𝑟ö𝑚 : Flow velocity [m/s]
𝐷: Charakteristical length [m]; at wings: chord length c
𝜈𝐿 : Kinematic viscosity oft he air [m2/s] resp. [St = Stokes]; 1 St = 10-4 m2/s
From a transitional range of Re = 0.5 · 105 to 5 · 105, the flow becomes turbulent and lossy. It should be noted that
even with a turbulent flow, the initial flow velocity can be laminar. The flow becomes turbulent only at a transition
point, see Figure 2. From the transition point, the size of the laminar boundary layer continues to increase. If a flow
hits a solid body at a relative velocity, a boundary layer is formed. In this case, the fluid particles which come into
contact with the body, due to the adhesive condition, will have the speed of the body. Further away from the body,
the fluid particles remain at the rate of the initial flow. With larger boundary layer thickness, the friction effects and
thus the flow losses increase significantly.
Figure 2 Example of the flow around an airfoil (Source: Pütz, R./Serné, T.)
294
Air resistance and downforce
On each body in a flow field aerodynamic forces act, which are differentiated into air resistance and lift or downforce
and form a resultant, see Figure 3. The buoyancy that a body (e.g. wing) experiences in a flow field depends on the
pressure differentials between the top and bottom of the body. The origin of this pressure difference is derived from
the laws of energy conservation of the flow described by the Bernoulli equation. The Bernoulli equation shows that
the sum of static, kinetic (dynamic) and potential pressure – that means the total pressure of a (frictionless,
incompressible) flow – is always constant.
𝑝𝑠𝑡𝑎𝑡 + 𝑝𝑑𝑦𝑛 + 𝑝𝑝𝑜𝑡 = 𝑝𝑔𝑒𝑠 = 𝑐𝑜𝑛𝑠𝑡.
1
𝑝𝑠𝑡𝑎𝑡 + 𝜌𝐿 · 𝑣𝑆𝑡𝑟ö𝑚 2 + 𝜌𝐿 · 𝑔 · 𝑧 = 𝑐𝑜𝑛𝑠𝑡.
2
with:
𝑝𝑠𝑡𝑎𝑡 : Static pressure [N/m2] resp. [bar]; 1 bar = 105 N/m2
𝑝𝑑𝑦𝑛 : Dynamic pressure [N/m2] resp. [bar]
𝑝𝑝𝑜𝑡 : Potential pressure [N/m2] resp. [bar]
𝑝𝑔𝑒𝑠 : Total pressure [N/m2] resp. [bar]
𝜌𝐿 : Air density [kg/m3]
𝑧: Height to the reference system [m]
On an airfoil, the flow divides into a lower part and an upper part. At output, the air particles are deflected further
below the wing and accelerated by a compression of the flow strands more than the air particles above the wing. The
flow below the wing is thus faster, which reduces the static pressure below the wing due to the Bernoulli equation.
This then leads to an output force 𝐹𝐿,𝑧 .
For drag 𝐹𝐿,𝑥 and downforce 𝐹𝐿,𝑧 of a race car, the following analogue equations are valid:
𝜌𝐿 𝜌𝐿
𝐹𝐿,𝑥 = 𝑐𝑊 · 𝐴 · · 𝑣𝑆𝑡𝑟ö𝑚 2 = 𝑐𝑊 · 𝐴 · · (𝑣𝑉 ± 𝑣𝐿 )2
2 2
𝜌𝐿 𝜌𝐿
𝐹𝐿,𝑧 = 𝑐𝐴 · 𝐴 · · 𝑣𝑆𝑡𝑟ö𝑚 2 = 𝑐𝐴 · 𝐴 · · (𝑣𝑉 ± 𝑣𝐿 )2
2 2
with:
𝐹𝐿,𝑥 : Drag [N]
𝐹𝐿,𝑧 : Downforce [N]
𝜌𝐿 : Air density [kg/m3]
𝑐𝑊 : Drag coefficiant [-]
𝑐𝐴 : Downforce coefficiant [-]
𝑣𝑉 : Vehicle speed [m/s]
𝑣𝐿 : Wind speed [m/s]
𝐴: Applied vehicle cross-section = projection surface [m2]
The vector sum of vehicle and wind speed corresponds to the flow velocity 𝑣𝑆𝑡𝑟ö𝑚 . In headwind the flow velocity is
𝑣𝑉 + 𝑣𝐿 , in tailwind 𝑣𝑉 − 𝑣𝐿 .
The term 1/2 · 𝜌𝐿 · (𝑣𝑉 ± 𝑣𝐿 )2 is the dynamic pressure 𝑝𝑑𝑦𝑛 oft he flow. Thus, depending on the shape and the cross-
sectional area of the body the flow parameters can be isolated. This has the advantage that these quantities can be
expressed directly in their interaction by measured drag and downforce.
𝐹𝐿,𝑥 𝐹𝐿,𝑥
𝑐𝑊 · 𝐴 = 𝜌 =
𝐿
· (𝑣𝑉 ± 𝑣𝐿 )2 𝑝𝑑𝑦𝑛
2
𝐹𝐿,𝑧 𝐹𝐿,𝑧
𝑐𝐴 · 𝐴 = 𝜌𝐿 =
· (𝑣𝑉 ±𝑣𝐿 )2 𝑝𝑑𝑦𝑛
2
295
Figure 3 Example of drag and downforce at an airfoil (Source: Pütz, R./Serné, T.)
The golf ball effect

To understand how to improve the aerodynamics of a rear wing using small dimples or bumps, it is necessary to
illuminate the principles of the golf ball effect. With a flow of air at the dimples and bumps, small vortices arise, since
the small reductions and / or elevations cause the flow to be deflected at very small intervals and create swirls.
Instead of the laminar boundary layer present in a smooth sphere of the same size and speed, a turbulent boundary
layer is formed by these small irregularities in a golf ball, which separates significantly later and follows the surface
of the ball longer. As a result, the separation area or wake is significantly smaller. A large detachment area behind
the ball would result in a very large flow field with low speed and large, "braking" pressures compared to the pressure
on the front. In the golf ball, however, there is a large area of low pressure behind the ball. Thus, an as small as
possible detachment area is desired, which can be achieved by a turbulent boundary layer, see Figure 4. In
motorsport, therefore, the goal is to reduce the air resistance through an artificially generated, turbulent boundary
layer, without significant loss of downforce.
Figure 4 Comparison of smooth ball and golf ball
INVESTIGATED REAR WING (AIRFOIL) CONFIGURATIONS

Basically, rear wings are used to increase the downforce of a race car. Increasing the contact pressure between the
vehicle tires and the track can significantly improve the driving dynamics characteristics (see Figure 1). Thus, the
cornering speed can be significantly increased by a higher downforce, as more lateral force can be built on the tires.
This leads to a significant improvement in the lap time, also because a later point of deceleration before the curve
and earlier acceleration from the curve and at a higher speed is possible. In addition, the additional downforce also
results in significantly improved traction forces and analogue higher braking forces.
The investigated F3 rear wing consists of two wing groups, the main wing (2) including the flap (1) and the lower wing
(3), see Figure 5. The flap also increases the generated donforce, because the gap between the main wing and flap
296
is flowed through at higher air velocities. The boundary layer at the rear end of the main wing is already large and
thus vulnerable to detachment, so that a further extension of the main wing favored the detachment and thus would
have a negative impact. At the flap, however, a new boundary layer is formed, which can cope with a further increase
in pressure before it comes to a replacement. Often at the back edge of the flap there is a Gurney flap which causes
a speed-dependent "extension" of the airfoil through the formation of opposing vortices. The additionally applied
"winglets" (4) prevent lateral vortex formation on the wings.
Figure 5 Multi-part rear wing of the Formula 3 race car
Figure 6 Model of the F3 rear wing generated by KScan3D
Using a 3D scan (see Figure 6), the contour of the F3 rear wing to be examined was recorded and then transferred
to a CAD program. In the process, about 150 individual photos were assembled using the "KScan3D 1.2" software
in several steps to finally form the 3D model of the rear wing. Thus, various dimple configurations could be designed
on that wing model afterwards. Subsequently, the modeled wing and flap profiles were produced by 3D printing on a
scale of 1:1, so that comparative investigations in the wind tunnel can be carried out. Due to the limited dimensions
of the model wind tunnel (Figure 10), only the width of the wings had to be reduced compared to the original. The
measured parameters are thus not completely transferrable to the original airfoil, but they are very similar, since the
mechanical similarity predominates. Figure 7 shows the investigated basic configurations of the dimples on the main
wing and the flap. In total, 25 different combinations were investigated. The distance between the centers of the
dimples is 10mm in both the x and y directions. The dimples each have a diameter of 3 to 5mm at a depth of 1mm.
In contrast to the original airfoil, a Gurney flap had been omitted in the wind tunnel tests. However, a Gurney flap
was taken into account in the CFD simulation. As a reference serves the conventional rear wing with flap and smooth
surfaces (here: variant 1-1). Figure 8 shows an example of the flaps with the serial numbers 2 and 5.
The evaluation of downforce, air resistance and aerodynamic efficiency was carried out for the various combinations
compared to the reference airfoil (combination of wing 1 with flap 1, designated as variant 1-1). The downforce
generated on the rear wing is mainly dependent on the angle of attack α of the flap and of the chord length c of the
wing profiles at rear wing and flap. The increase in downforce also leads to an increase in air resistance. The optimal
tuning of the ratio of downforce to air resistance is described by the aerodynamic efficiency εaero, which was the focus
of the investigations. The principle is that the higher the aerodynamic efficiency, the better the aerodynamics are
"tuned".
𝑐𝐴 𝐹𝐿,𝑧
𝜀𝑎𝑒𝑟𝑜 = =
𝑐𝑊 𝐹𝐿,𝑥
297
To determine the coefficients cA and cW, the flow velocity must be known which was determined with a pitot tube
anemometer.
Main wing Flap

Serial No. Explanation Serial No. Explanation
1 w/o Dimples ("Reference") 1 w/o Dimples ("Reference")
2 Dimpels over entire top 2 Dimpels over entire top, bottom smooth
3 Dimples only at the front / on nose 3 Dimples only at the front / on nose
4 Dimples only at the rear end 4 Dimples only at the rear end of the profile
5 Dimpels at the front and rear ends 5 Dimpels at the front and rear ends (top and bottom)
Figure 7 Investigated basic configurations for rear wing and flaps
Figure 8 Examples: Flap 5 (left) with dimples also on the bottom, flap 2 (right) with smooth bottom
Figure 9 shows the experimental setup in the wind tunnel. The angle of attack α of the flap is, similar to the original
rear wing, gradually adjustable. The transparent winglets serve as a holding device for wings and flaps. The base
plate is also adjustable in angle to be able to simulate a crosswind influence by a slight inclination of the wing to the
direction of flow can. Due to the side inclination of the wing, in particular by the influence of the winglets, the inflow
area increases significantly. The resulting force from drag and downforce attacks at the pressure point of the airfoil
which is approximately at 1/3 of the chord length c of the wing, see Figure 3. The rear wing construction with the
winglets is built on a support plate, which rests in the downforce direction z on four scales and rests in the drag
direction x of another scale. These measured forces are reduced to a two-dimensional system and converted to the
pressure point.
Figure 9 Experimental setup in the wind tunnel
298
RESULTS OF THE MEASUREMENTS IN THE WIND TUNNEL
For the measurements, a model wind tunnel was used, in which the airfoil (wing) profiles in the scale 1:1, but as
described with reduced width of the wing, can be examined, see Figure 10. Due to the design of this wind tunnel,
flow velocities of almost 100 km / h (27, 8 m / s) can be realized, which are sufficient for the production of turbulent
flows and simulation of the golf ball effects.
A - blower D - measuring section / plenum

B - nozzle E - openable window
C - rectifier F – diffuser
Figure 10 Setup of the modular BELICON model wind tunnel
Firstly, a measurement of the reference airfoil with a smooth surface (variant 1-1) had been carried out. These results
serve as “reference” for the measurements with the alternative concepts with the different dimple arrangements
shown in Figure 7. For the reference measurement, as well as for the concepts with dimples, respectively the angles
of attack 0°, 6°, 8°, 10°, 12°, 15°, 18°, 20° and 25° at a side inclination of 0°, 5° and 10° to the direction of flow had
been examined. The angle of attack is determined by adjusting the inclination of the flap, while the angle of attack of
the main wing profile was not changed as in the original wing. Both the reference airfoil (variant 1-1) and the options
with dimples were measured several times so that statistical mean values for cA, cW and εaero could be formed.
Results with frontal flow (no crosswind)

From all of the measurements, a tolerance range was formed in each case via an error calculation, which extends
from the minium of the measurement minus the measurement error Δ𝑥 and up to the maximum of the measurement
plus the measurement error Δ𝑥. For this reason, in the following the minimum and maximum values derived from the
measurement series which are in the confidence interval are shown in red. Figures 11, 12 and 13 show the curves
of cW, cA and εaero as a function of the angle of attack α of the flap.
299
Figure 11 Drag coefficients as a function of the angle of attack (without crosswind)
Figure 12 Downforce coefficients as a function of the angle of attack (without crosswind)
Figure 13 Aerodynamic efficiency as a function of the angle of attack (without crosswind)
300
It has been revealed that some variants of the alternative concepts with dimples can achieve significant
improvements in air resistance, downforce and aerodynamic efficiency. As expected, strong dependencies on the
angle of attack α were found, so that different dimple variants were identified as effective only for different angles of
attack. However, since it is not practical to always exchange the complete combination of main wing and flap during
a training or qualifying, the mean values of the aerodynamic efficiencies from all angles of attack are calculated below
in order to obtain the "overall" most favorable variant (combination of main wing and flap), see Figure 14. This shows
that variants 5-5, 5-1, 3-2 and 4-1 appear to offer the greatest advantages compared to reference airfoil 1-1 when
the air flow is frontal (without any crosswind), see Figure 15. No improvements were made by dimples outside or
inside on the winglets.
Figure 14 Average aerodynamic efficiency over all angles of attack (without crosswind)
Main wing Flap

5 Dimpels at the front and rear ends 1 w/o Dimples ("Reference")
3 Dimples only at the front / on nose 2 Dimpels over entire top, bottom smooth
4 Dimples only at the rear end 1 w/o Dimples ("Reference")
Figure 15 Configurations of the rear wing and flaps with significant aerodynamic improvement potential
compared to the reference wing 1-1 at frontal flow (without crosswind)
Results with oblique flow (with crosswinds)

In order to investigate the changes that occur under the influence of crosswinds for the different variants with dimples
compared to the reference, the entire wing assembly was rotated by 5° to the direction of flow. As a result, the flow
area and also the position of the pressure point change. It should be noted that here not all previously studied variants
were considered, but only the most promising variants identified at frontal inflow. The aerodynamic efficiency as a
function of the angle of attack and the average aerodynamic efficiency are shown below for selected variants with an
inclination of 5 °, see Figures 16 and 17. Compared to the frontal inflow, the aerodynamic efficiency is sometimes
reduced. It also revealed that other alternative variants with dimples than those with best perfomances at frontal
inflow seem to produce the greatest improvements in oblique inflow. Only the variant 5-1 still offered favorable results,
while now the variants 5-2 and 5-4 offer the greatest advantages compared to the reference airfoil 1-1, see
Figure 18. No improvements were also here found for outside or inside Dimples at the winglets, which are therefore
not recommended. Due to the dimples on the winglets, the aerodynamic properties of the wing were significantly
worsened (cW, cA and εaero) so that it can be assumed that unfavorable turbulence is then generated at the winglets
which prevent exploiting the full potential of the airfoil.
301
Figure 16 Aerodynamic efficiency as a function of the angle of attack (crosswind, 5 ° inclination)
Figure 17 Average aerodynamic efficiency over all angles of attack (crosswind, 5 ° inclination)
Main wing Flap

Serial
Serial No. Explanation No. Explanation
5 Dimpels at the front and rear ends 2 Dimpels over entire top, bottom smooth
5 Dimpels at the front and rear ends 4 Dimples only at the rear end of the profile
3 Dimples only at the front / on nose 4 Dimples only at the rear end of the profile
2 Dimpels over entire top 1 w/o Dimples ("Reference")
5 Dimpels at the front and rear ends 1 w/o Dimples ("Reference")
Figure 18 Rear Wing and Flap configurations with significant aerodynamic improvement potential
over reference airfoil 1-1 with oblique flow (5 °)
From the results it is clearly evident that for different angles of attack α of the flap of the rear wing, for frontal or
oblique flow, different variants with dimples provide the aerodynamic optimum. For the main wing, the dimples version
at the nose and at the end of the profile - variant 5 - seems to be the optimum. In any case, the majority of tested
variants with dimples offer significant improvements over the reference (smooth) wing, which demonstrates the
advantage of selective application of the golf ball effect to the F3 rear wing. Further investigation for application on
other aerodynamic components and their interaction on the overall vehicle is strongly suggested.
When selecting the "right" wing variant for a particular race track, the characteristics for the frontal inflow should
always be considered and weighted more heavily than the results with oblique flow. In most cases, the crosswind
influence is not constant nor can be estimated beforehand. The deciding factor for the application of dimple
302
configurations on the wing is always which aerodynamic target is pursued in priority. On high-speed race tracks, in
addition to the aerodynamic efficiency, the target can also be an optimization of only the cW value. Almost all dimple
concepts attached to the wing profiles have improved the aerodynamic efficiency. This occurs primarily by lowering
the cW value. In the conventional golf ball, the turn from laminar to turbulent flow occurs at a Reynolds number of
about Re = 105. In the measurements, the Reynolds number was also in this range. As soon as turbulent flow is
applied, the air resistance of the golf ball lowers considerably. This effect is therefore also to be assumed in the wings
examined here. The lowest cW values were found for those variants which have dimples at the front and at the rear
of the main profile (variants 5-4 and 5-5). The highest downforce coefficients cA, on the other hand, were found for
variant 4-3 with dimples at the rear of the main wing and at the front of the flap, see Figure 19.
Main Wing Flap

5 Dimpels at the front and rear ends 4 Dimples only at the rear end of the profile
Main Wing Flap

4 Dimples only at the rear end 3 Dimples only at the front / on nose
Figure 19 Configurations for lowest drag (table above) and largest downforce (lower table)
Advantages of aerdynamic converters?

For some time, so-called “aerodynamic converters” have been discussed which are intended to change the material
properties and thereby reduce the aerodynamic properties, e.g. of airfoils. An explanation of the function is based on
the fact that the converters reduce the electrostatic charge, which reduces the surface friction. Other theories also
include the influence of the geomagnetic field and are thereby close to the esoteric border of the sciences. As part
of the tests, two different converters from the company Snipptec were applied to the smooth reference wing
(variant 1-1) and tested for their effect on the aerodynamics. The converters were mounted either at the top or bottom
of the main wing in the area of the front / nose. To check how the shape of the converters affects the aerodynamic
properties, and also as a reference for subsequent measurements with the converters, dummies of identical
dimensions were at first attached to the smooth airfoil variant 1-1 and examined. For the measurement results with
two different converters of the internal designations 31F.o. (black) and 14 (blue, see Figure 20) it was noticeable that
the aerodynamic properties of the wing were in deed affected, see Figure 21. It was found that the full effect of the
converters unfolded only after a certain time after application. The measurements directly after gluing gave
completely different measurement results compared to the measurement after three to four hours or one day / 24
hours. For the converter 31F.o. (black converter), it was found that when mounted on the top of the main wing both
the cW and the cA values were relatively lower compared to the dummy. However, the aerodynamic efficiency
increased. When attached to the bottom, the cA value increased slightly with respect to the dummy, but then the cW
value also increased, reducing the aerodynamic efficiency. Converter 14 (blue converter) mounted at the top
increased cA as well as cW, but overall reduced aerodynamic efficiency. In absolute terms, the converters have not
yet achieved any improvement compared with the smooth airfoil (without dummy), which is certainly due to the fact
that the shape of the converter glued on the surface and therfore the change in the wing contour cause a change in
the aerodynamic properties. Nevertheless, the experiments clearly showed the potential of this interesting approach
to aerodynamic optimization, which should be addressed further. It is therefore recommended to place the converters
in the cavity of the airfoil already during the production of the wings in order not to "disturb" the flow at the surface.
Furthermore, investigations of the wings with the convectors then inside in connection with dimple profiles would be
expedient.
Figure 20 Snipptec® Converter No. 14
303
Figure 21 Average aerodynamic efficiency of Snipptec® converters at all angles of attack
compared to the reference wing with "dummy"
ACCOMPANYING ANALYIS USING CFD

For the supplementary CFD (Computational Fluid Dynamics) calculations, the ANSYS CFX system was used. Since
turbulent flows are strongly dependent on even the smallest disturbances of the flow, a simulation is very complex,
since too many subdivisions would have to be modeled in order to correctly calculate every quantity even for
small(est) vortices. In order to reduce the required computing power and computing time significantly, ANSYS CFX
works with "Reynolds-averaged Navier-Stokes equations". These replace the flow quantities by the sum of their mean
values and their fluctuation quantities. Thus, small turbulences are not completely resolved, but approximately
calculated indirectly via so-called turbulence models. All simulations were performed with two different angles of
attack of the flaps of 10° and 27.5°. Three different flow velocities were chosen: 150, 200 and 250 km/h (41.7, 55.6
and 69.4 m/s).
Investigated basic variants

After test simulations it was noticeable that the existing geometry had to be further simplified in order to achieve a
tolerable simulation time. Finally, it was agreed to investigate the changes by the golf ball effect isolated on only the
angle-adjustable flap, without taking into account the main wing. The task of the winglets to avoid air flows past the
side of the airfoil was realized via settings in the simulation program. Since the wing profile remains constant over
the entire width of the airfoil and no sideways swirl occurs in the flow due to the simulated winglets, a reduced width
of the wing was chosen in the simulation model, also with a view to the reduced width of the 1:1 model investigated
in the wind tunnel. In addition, a Gurney flap was applied to analyze its effects in conjunction with dimples and bumps.
Simplified, the virtual "chord extension" c* is shown for the Gurney flap on the airfoil in Figure 22.
Figure 22 Gurney flaps at the end of the wing create two opposing vortices and thus an extension of the chord
length of the wing with a significant increase in downforce (Source: Pütz, R./Serné, Ton)
304
As a reference for the concepts with dimples served a smooth flap profile (without dimples), which is here referred to
as "variant 1" and corresponds to that of the “flap 1” from the wind tunnel tests. In comparison, three different variants
of dimple concepts were investigated, which are shown in Figure 23.
Serial No. Explanation

1.1 Dimples only in the rear area
1.2 Dimpels only in the front area
1.3 Dimpels all over the top
Figure 23 Investigated dimple configurations on the flap
Results for the basic variants

In Figure 24, the velocity distribution of the fluid particles in the flow field is shown graphically for the reference flap
(Variant 1). It can be seen that, as expected, the flow velocity below the wing (red and yellow arrows) is faster than
above (blue arrows). This leads, as already described, to a higher pressure on the top and a lower pressure on the
underside of the wing and thus to downforce. This is also confirmed by the pressure distribution in Figure 25. The
influence of the Gurney flap on the separation area and thus the virtual "wing extension" is revealed.
Figure 24 Velocity distribution at the reference profile (variant 1) at an angle of attack of 10°
and flow velocity 150 km/h
305
Figure 25 Pressure distribution on the reference profile (variant 1) at an angle of attack of 10°
and flow velocity 150 km/h
Figure 26 shows the three investigated concepts of different dimple arrangements with respect to their aerodynamic
efficiency εaero over the flow velocity. The comparison shows that the characteristics of the different variants are very
similar over the speed. Thus, in all variants (Var. 1.1, 1.2 and 1.3), the efficiency εaero decreases relatively sharply
from 150 km/h to 200 km/h. The observation that εaero of Variant 1.1 slightly increases in the range of 200 km/h to
250 km/h does not contradict, since this is only a consequence of the fact that the kink at 200 km/h in all three variants
is similarly pronounced. Both drag and downforce develop relatively constant over the speed. The largest increase
in aerodynamic efficiency compared to the smooth reference profile shows variant 1.2, those with dimples
exclusively in the front area of the wing, which has significant advantages at any flow velocity of 150 to 250 km/h.
Variant 1.1, with dimples at the rear of the wing, showed only minor improvements in aerodynamic efficiency in the
speed range of 150 to 200 km/h. Variant 1.3 with dimples all over the top surface reveals a deterioration in
aerodynamic efficiency over all speeds.
Figure 26 Change of the aerodynamic efficiency of the concepts with dimples compared to the
reference profile depending on the flow velocity at an angle of 10°
In variant 1.2 (dimples at the front area), an improvement in efficiency from the reference of about 2% to almost
4.2% is evident. In this case, the downforce is increased and the drag is reduced at all flow velocities. Quite
surprising is the result that the variant 1.3 (dimples all over the top surface of the wing) has a deterioration of the
aerodynamic efficiency compared to the reference. This shows how sensitively the golf ball effect responds to minor
changes in the surface.
The significant improvement of the downforce in variant 1.2 compared to the reference (variant 1) is documented in
figure 27. In direct comparison, it becomes clear that with variant 1.2 the maximum value of the pressure (red range:
between +1817 Pa and +2310 Pa) is higher than with variant 1 (red range = between +1811 Pa and +2278 Pa). It
306
can also be seen that the higher pressure above the wing in variant 1.2 also acts over a larger area. In variant 1, on
the other hand, the pressure above the wing temporarily falls below 1817 Pa (dark orange range). In addition, in
variant 1.2, the pressure below the wing is lower than in variant 1 (blue range, -7065 Pa or -6601Pa). It should be
noted that this indicates a relative pressure. Due to the higher pressure difference in variant 1.2, a higher output force
than in variant 1 is also generated there.
Figure 27 Increase of the downforce of variant 1.2 (below) compared to the reference profile (above)
based on the pressure distribution at inflow velocity 150 km/h and 10° angle of attack
However, for 27.5° angle of attack it can be seen that the aerodynamic efficiencies of all Dimple concepts worsen
significantly, from over 8.4% to over 14% compared to the reference flap depending on the speed, see Figure 28.
Both downforce is reduced and drag is increased. The version 1.2, which achieved the best results with an angle of
attack of 10°, even gave the worst results at 27.5°. By contrast, the best aerodynamic values can now be seen in
variant 1.1. Nevertheless, with this steeper angle of attack, the conventional reference wing would be preferable
to all options with dimples.
In order to approve the differences in drag, Figure 29 shows the variants 1.2 (top) and variant 1 (reference, bottom)
compared by the flow lines. It can be seen with the aid of the red bordered control area that the dead space area in
variant 1.2 is significantly larger than in variant 1, which results in larger flow losses, which are more than 5% higher
in variant 1.2. Certainly, the interaction with the Gurney-Flap also plays a role here, which has a downforce-generating
effect.
307
Figure 28 Change in aerodynamic efficiency of the concepts with dimples compared
to the reference profile depending on the flow velocity at an angle of 27.5°
Figure 29 Speeds and streamlines for Variant 1.2 (above) compared to the reference (Variant 1, below)
at 150 km/h and 27.5° angle of attack
308
In summary, it should be noted that the optimal positioning of the dimples strongly depends on the angle of attack.
Thus, no dimple arrangement was found, which allows an improvement in aerodynamic efficiency compared to the
reference flap for all angles of attack. At a shallow angle of 10°, the best results are seen when the dimples are
placed further forward on the wing. However, this variant has great disadvantages at a steep angle of attack. For an
angle of attack of 27.5°, a deterioration compared to the reference flap was found for all Dimple versions.
Results for structural modifications of the basic variants

Further investigations are related to modifications of the dimple structure. Basically, the variabilities shown in
Figure 30 are e.g. the depth hDimple or hBump, the radius RDimple or RBump and the divisions in x (longitudinal) and y
(transverse) directions.
Figure 30 Bumps and Dimples can be arranged in different combinations and zones
of the flap (Source: Pütz, R./Serné, T.)
At first, based on the variant 1.2 (dimples in the front area) a variant 2.1 was created, which has bumps (instead of
dimples) at the same lovations. The bumps have the same size (radius 3.0 mm), but only the center of the basic balls
is offset to effectively get the same expansion. It is expected that with bumps also small turbulences arise, which
should have a positive effect on the aerodynamics. However, it was found that the efficiency of variant 2.1 with bumps
instead of dimples is between 0.3% and 1.2% worse than the reference, at an angle of attack of 10°, see Figure 31.
Thus, dimples (in the front region) of the same size at an angle of attack of 10° bear significant advantages over
bumps.
Both variants show an opposite result at the angle of attack of 27.5°, as the concept with bumps (Variant 2.1) performs
significantly better than the concept with dimples (Variant 1.2), see Figure 32. Nevertheless, the aerodynamic
efficiency of the reference flap is also not matched at an angle of attack of 27.5°. The smooth reference wing offers
the optimum at steep angles of attack.
Figure 31 Change in aerodynamic efficiency of the Variant 1.2 (dimples) and Variant 2.1 (bumps)
compared to the reference profile as a function of the flow velocity at 10° angle of attack
309
Figure 32 Change in aerodynamic efficiency of the Variant 1.2 (dimples) and Variant 2.1 (bumps)
compared to the reference profile as a function of the flow velocity at 27.5° angle of attack
In another simulation, the radius of the dimples was varied, see Figure 33. Instead of a radius R of the dimples of
3.0 mm in variant 1.2, a dimple radius R of 3.5 mm was selected for variant 3.1, while in variant 3.2 the radius was
reduced to 2.5 mm. For all dimple radius modifications, the distance of the sphere center above the surface was
adequately adjusted. Figure 34 shows the result for an angle of attack of 10°. It is obvious that both with a larger
radius (variant 3.1) and a smaller radius (variant 3.2), a deterioration of the aerodynamic efficiency compared to a
radius of 3 mm was determined. It is also noteworthy that in variant 3.1, there is a worsening of around 1.9% to 2.2%
against the reference values. In contrast, variant 3.2 has a comparatively slight deterioration of about only 0.3% to
0.6%, which allows the conclusion that the optimum of the dimple radius for an angle of attack of 10° might be
between 2.5 mm and 3.0 mm.
At an angle of attack of 27.5° and at a speed of 250 km/h, for variant 3.1 (larger dimple radius) a highly significant
deterioration of up to 31.7% compared to the reference can be seen, see Figure 35. This value is derived from a
13.5% increase in drag and a 13.9% reduction in downforce. This example once again shows how sensitive the golf
ball effect is to relative minor changes. In addition, the influence of the depth of the dimples in the profile was also
examined. The optimum depth of the pits was again dependent on the angle of attack α. Thus, the best values for
aerodynamic efficiency were achieved at an angle of attack of 10 ° with a height of the center of the balls of 1.5 mm.
Figure 33 Modification of the Dimple radii: Variant 3.1 (left) und Variant 3.2 (right)
Figure 34 Change in aeodynamic efficiency of the Variant 3.1 (R 3.5) and Variant 3.2 (R 2.5) against
Variant 1.2 (R 3.0) and the reference profile as a function of the flow velocity at 10° angle of attack
310
Figure 35 Change in aeodynamic efficiency of the Variant 3.1 (R 3.5) and Variant 3.2 (R 2.5) against
Variant 1.2 (R 3.0) and the reference profile as a function of the flow velocity at 27,5° angle of attack
Finally, the distance of the dimples in the x and y direction was varied, see Figure 36. In this case, the variant 5.1
was examined, which has a smaller distance between the dimples compared to the variant 1.2 and thus a higher
number of dimples. The calculations show that variant 5.1 (with the higher number of dimples) gave worse efficiency
values than variant 1.2 and even showed a slightly worse aerodynamic efficiency compared to the reference, see
Figure 37. Although, for variant 5.1, an expected improvement in drag compared to variant 1.2 was found, which
however, is due to a significant loss of downforce.
Figure 36 Modification of the dimple spacing: Variant 5.1 (left) und Variant 1.2 (right)
Figure 37 Change in the aerodynamic efficiency of the Variant 5.1 (more dimples) compared to
Variant 2.1 and the reference profile as a function of the flow velocity at 10° angle of attack
CONCLUSIONS
In line with the golf ball effect, attempts are also being made in motorsport to reduce air resistance or improve
aerodynamic efficiency through dimples and bumps in the surface of wings. As a rule, the depth of the dimples or
height of the bumps is about 1 mm with a radius of about 2.5 to 5 mm. While on golf balls even in the low turbulent
speed range (Re ≈0.5·105) there is a reduction in air resistance of around 50% compared to a flat surface, on wings
311
of racing cars the effect is strongly depending on the angle of attack, the arrangement of the dimples/bumps and
their dimensions. It has been demonstrated by wind tunnel tests and accompanying CFD simulations that the golf
ball effect can also be successfully applied to wings, thereby achieving a significant improvement in aerodynamic
efficiency. Depending on the configuration, both the downforce of a rear wing and its air resistance can be reduced.
However, it turned out that this effect is very sensitive to minor changes. Thus, even minor structural modifications
of the dimple configuration, which before had better aerodynamic properties on a rear wing than a smooth reference
wing at a certain angle of attack, can adversely affect or reverse the results. So "the" optimal dimple configuration for
all angles of attack does not exist.
The CFD simulations confirmed the practical test result of the wind tunnel measurements, that the aerodynamic
efficiency depends very much on the wing position. For a larger angle of attack of 27.5° even a significant
deterioration of the aerodynamic efficiency compared to the reference wing was determined by applying the dimple
concepts. On the other hand, for a smaller angle of attack of 10°, with a few dimples variants it was possible to
increase both downforce and drag in comparison to the respective reference. Thus, no arrangement of dimples or
bumps was found which could achieve an improvement to the reference at several angles of attack. Further
investigations by BELICON will be carried out on main wing variant 1 (dimpels at the front and rear ends) and flap
variant 1.2 (dimpels only in the front area plus Gurney flap).
With respect to the flow velocity, it was recognized that the highest aerodynamic efficiency was often found at the
lower speed of 150 km/h, while the efficiency at 250 km/h was always lower. Thus, it is likely that the golf ball effect
has a more favorable impact on wings at a speed of 150 km/h than at a higher speed of 250 km/h. This accommodates
the requirements and average speed profiles of most racing series up to Formula 3.
The widespread belief that bumps can cause the same effect as dimples in certain cases has been confirmed, but
even with identical geometrical arrangements of dimples and bumps at the same angle of attack, different results in
aerodynamic efficiency were found. It is already clear from the theoretical view that the influence of dimples and
bumps on a rear wing with a relatively small dead space even when using Gurney flaps can never be as pronounced
as in an aerodynamically unfavorable form such as a (golf) ball. Nevertheless, as shown, a significant improvement
in the aerodynamics was realized for the examined rear wing, whereby it should be noted that an angle-adjustable
rear wing is quite an "extreme" application for the golf ball effect. Non-adjustable aerodynamic components simplify
the use of the golf ball effect.
Current research results also of other researchers show that due to the high sensitivity of the results depending on
the possible configurations of the dimples and bumps and the angles of attack of the wings and the flow velocities,
the complexity of this application increases highly significantly and results can sometimes not be compared. Based
on the results so far, BELICON Motorsport and the University of Applied Sciences Landshut will carry out further
investigations of the golf ball effect also on other aerodynamic components of race cars and analyze them further in
their interactions.
REFERENCES
[1] Pütz, R., Serné, T.: "Rennwagentechnik – Praxislehrgang Fahrdynamik: Eine praktische Anleitung für Amateure
und Profis“, Springer-Vieweg-Verlag, 2017. [Race car technology – course in vehicle dynamics: a practical guide
for amateurs and professionals],
[2] Hechfellner, T.: "CFD-Rechnungen zu aerodynamischen Bauteilen an Rennwagen zur Untersuchung des
Golfball-Effekts“, Bachelor Thesis, Landshut 2017 [CFD calculations on aerodynamic systems of race cars
analysing the golf ball effect],
[3] Kampfl, M.: Aerodynamische Optimierungen an einem Rennwagenheckflügel unter Anwendung des
Golfballeffekts“, Bachelor Thesis, Landshut 2017 [Aerodynamic optimisation of a race car rear wing exploting
the golf ball effect],
[4] Scott, J.: "Golf Ball Dimples & Drag“, 2005, Available from: http://www.aerospaceweb.org/question/
aerodynamics/q0215.shtml, Accessed 25.06.2018.,
[5] Thomas, M. R.: "Dimple minded“, Racecar Engineering, January 2017,
[6] Anitha, G., Livya, E., Valli, P.: “Aerodynamic Analysis of Dimple Effect on Aircraft Wing”, International Scholarly
and Scientific Research & Innovation, Heft, Vol. 9, No. 2, 2015, pp 350-353.
312
Kragujevac, Serbia
MVM2018-034
Nadica Stojanovic1
Jasna Glisovic2
Ivan Grujic3
Sunny Narayan4
Sasa Vasiljevic5
Bojana Boskovic6
EXPERIMENTAL AND NUMERICAL MODAL ANALYSIS OF

BRAKE SQUEAL NOISE
ABSTRACT: In addition to different kinds of pollutants emitted by the vehicles, noise can also negatively effect
on human health. It does not only threaten drivers, but also people living near major intersections and roads, as well
as roads where traffic-flow is high. One of the biggest problems of the vehicle is the noise that occurs in the braking
process. Despite the large scope of research into the development of brake systems, there are still no reliable
procedures during the development phase to evaluate the robustness of these systems with respect to friction-
induced vibrations. Therefore, the identification of the modal properties by using experimental methods has become
even more important. Experimental and numerical modal analysis of the venting disc with radial ribs was performed
in this paper. This approach enables the determination of the natural frequencies of the brake disc, as well as the
verification of results obtained by the numerical methods. Changes in modal properties-resonance frequencies and
modal damping values due to variation in operating conditions were also analysed.
KEYWORDS: vehicle, braking, noise, disc with radial ribs, natural frequency
INTRODUCTION
Noise affects the heart and cardiovascular system and has a significant effect on various psychosomatic diseases.
Different opinions of the authors have been noted in relation to the noise effect on the cardiovascular system, while
one considers that there is an increased blood pressure, and the others, in contrast to its decrease. Research
conducted for study the prevention of ischemic heart disease in people living in cities or parts of the city, where the
level of public noise exceeds 65 dB, has shown a relative risk of disease from ischemic disease. Today,
cardiovascular diseases are one of the most common causes of death, and these data has a high public health and
social significance. It has been found, by investigating ECG waves and a cardiac rhythm that sounds of higher
intensity have a greater impact on the occurrence of cardiac disorders [1].
1
Nadica Stojanovic, M.Sc., assistant, University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić STR.,
2
Jasna Glisovic, Ph. D., assoc. prof., University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić STR., 34000
Kragujevac, Serbia, jaca@kg.ac.rs
3
4
Sunny Narayan, Ph. D., Qassim University, Faculty of Engineering, Saudi Arabia
5
Sasa Vasiljevic, M.Sc., Mechanical engineer, High technical school of professional studies, Kragujevac,
vasiljevic.sasa036@gmail.com
6
Bojana Boskovic, B.Sc., College of Tehnical and Mehanical Engineering Trstenik, bokaev92@gmail.com
313
It can be said that traffic noise is one of the most prevalent urban polluters in the city environment. Moreover, WHO
(World Health Organization) categorizes the noise as the second worst ecological cause of poor health (PM2.5 is the
first) [2]. This means that road traffic is the most dominant source of environmental noise, it is estimated that 125
million people are exposed to a noise level higher than 55 dB. While noise causes at least 10.000 cases of premature
death in Europe each year, 8 million people have a sleep disorder, hypertension occurs in 900,000 inhabitants, and
only 43.000 are examined by a doctor. A research performed by the Environmental Protection Agency of Ministry of
Environmental Protection for 2016 was carried out in 14 cities in the Republic of Serbia at 195 measuring points.
Statistics included residents exposed to a noise level of 55 dB and more for all sections of the state road network for
which SNM (strategic noise map) was made, where the average annual flow is over 3.000.000 vehicles. During 24
hours, the largest number of inhabitants is exposed to a noise level of 55-59 dB, Figure 1. If it’s a residential area,
the noise level exceeds the allowed values, and in the case of the city centre, the zone along highways, main and
urban roads, then it is below the allowed values [3]. The largest number of people exposed to night noise is 43.132,
and the noise level ranges from 45-49 dB, Figure 2. Results of measurement of night-time noise, which represents
the noise in the period from 22 to 06 hours, vary depending on the area in question. If it is a residential area, then it
exceeds the permissible values, and in the case of a city centre or zone along highways, the main and city roads
then it is below the permitted limits [3].
Figure 1 Population exposed to road traffic noise expressed over the range of total noise indicator Lden [4]
Figure 2 Population exposed to night road traffic noise expressed over the range of average night noise level Lnight
[4]
People primarily use passenger cars for their own personal needs, such as going to work, vacation, shopping, etc.,
simply saying, to travel from one point to another. The biggest problem with today's vehicles is the comfort and noise
produced by the vehicle during its exploitation. The level of noise inside the vehicle has been steadily falling over the
past few years. So the appearance of any sound can lead to disturbance of the driver, as well as to his distrust in his
own vehicle. One of the vehicle components occasionally produces unwanted vibrations, and unpleasant sounds are
their braking systems. Manufacturers of brake and friction materials face very challenging requirements [5]. They
need to know what are the characteristic noise, as well as the cause of the noise generation [6]. Dynamic oscillations
in friction brakes that cause noise, vibration and rigidity can be classified into two groups [7]:
 Dynamic instability which results in a constant resonant frequency independent of rotor and
 Mechanical vibration with a frequency related directly to rotor speed.
314
The brake noise usually occurs in the range from 100 Hz to 20 kHz (which is the upper limit of human hearing) and
in the case of classic brakes, the braking squeal occurs in the range of 1-6 kHz, where the human ear is the most
sensitive. Within the definition of noise and brake judder, there are many subcategories of noise and judder of the
brake [7, 8], Figure 3. A typical estimate of the proportion of warranty claims based on the different categories of
brake NVH (Noise, Vibration and Harshness) is shown in Figure 4.
Figure 3 Frequency range of different vibro-acoustic phenomena generated by vehicle brakes [9]
Figure 4 Typical distributions of warranty issues reported by high performance vehicle manufacturer [7]
There are several basic methods for determining the resonance vibrations of a brake disc such as [10]:
 Using a pulsator with any of the commonly used excitation functions, e.g., sine, random, transient;
 Using a dynamometer and braking system, and
 Impact Hammer.
The most commonly used method is the impact hammer method. The reason for this is the cost of the equipment,
as well as the simplicity of the experiment. As is the case in this study, in a large number of analysed researches,
the response to the excitation is measured by acceleration sensors, while in some studies, the measurement was
performed by a microphone [11]. If the measurement is done by acceleration sensors, the number of sensors used
in the experiment can range from one to four sensors [12-14]. In addition, depending on the sensor, it also depends
on the connection to the brake disc itself. For example, the sensor can be coupled with brake disc by using the
beeswax [15].
Competition on the market is only possible if the experiments are fast, and this is achieved only by the use of numeric
analysis. Of course, the accuracy of the results obtained this way should not be omitted, and this is achieved if the
numeric analysis is calibrated with the experiment [16]. If the obtained numerical results are satisfactory, they can be
used for further analysis [12]. However, an experimental study of structural dynamics has always contributed greatly
in understanding and controlling the vibrational phenomenon that is encountered in practice. Regardless of the fact
that today, there are computers with high performance and capabilities, but the need for an experiment is the same
as before, for several reasons [17]:
 Determining the nature and level of vibration that occurs during operation;
 Verification of the theoretical models and the prediction of various dynamic phenomena which are collectively
referred to as "vibrations", and
 Measurement of the basic properties of materials due to the effects of dynamic loads, such as damping,
friction and fatigue endurance.
The problem of structural vibrations continues to be a huge danger and design constraints for a wide range of
engineering products. These disorders in the system may be unacceptable for the system's planned operation. In
315
this case, the engineer must determine how to minimize or eliminate unwanted vibrations occurring within the system.
However, it can sometimes be very difficult when the cause is unknown [18]. At the end, it is most important that the
vibration level is reduced to permissible level.
One of the aims of the paper is to determine the natural frequency of the passenger car’s brake disc experimentally,
then to modify numeric analysis based on the results obtained by the experiment. So for the future research it is not
necessary to conduct experiment first.
MODAL ANALYSIS
In the past two decades, modal analysis has become one of the main technologies for determining, improving and
optimizing the dynamic characteristics of the different types of structures. Application of modal analysis is very wide,
engineering disciplines where she found the application in engineering, aerospace and civil engineering [19]. The
reason for the application is [20]:
 Revealing the causes of unpredictable material behaviour;
 Construction modification;
 Sensitivity analysis;
 Simplification of mathematical models;
 Predicting the response on disturbance force and
 Detection of structural damage.
Modern constructions are expected to have a small mass and to be a simple as possible, and at the same time
sufficiently strong [20]. The requirement to reduce the mass of the construction is highest in the automotive and
aviation industry. Unless sufficient attention is paid to reduce the mass, unwanted vibrations may occur in certain
exploitation regimes. Modal analysis represents the process of determining the basic dynamic characteristics of the
structure, such as natural frequency, damping and oscillation modes, in order to formulate the dynamic behaviour of
the model [21].
There are two approaches to the modal analysis, the so-called experimental modal analysis, where the modes are
determined by using the acceleration and displacement sensors. The second approach is the theoretical modal
analysis, where the prediction of modes is performed by using the appropriate mathematical methods (FE-based
eigenvalue extraction or MBD analysis of the vibration behaviour of the brake assembly) [9]. Today's software for
modes determination is working according to the principle of mentioned mathematical models. The lack of
experimental research is reflected in the fact that they cannot be examined during exploitation, or more accurate in
road conditions. The vehicle must be examined in the laboratory, on dynamometer rollers, or the brake assembly
must be mounted on the appropriate dynamometer for testing. If a person conducting a test, rather than a robot, the
repeatability of the results is almost impossible. Verification of results obtained by the numeric analysis requires that
they must be approximate to the experimental results or to be a very close to them. So after the experiment, numerical
analysis can be performed.
The research in this paper is performed by using the experiment and numeric analysis. The braking disc that was
research subject in this paper is the vented disc with radial ribs, Figure 5.
Figure 5 3D model of vented brake disc
Experimental analysis
The applied experimental method is based on using impact hammer and with this method, it is necessary to provide
the following:
 Impact hammer with head which can measure input force;
 Acceleration sensor;
 FFT analyser for determining the frequency response function, and
316
 Software for data processing and results showing.
Before starting the experiment, it is necessary to provide the appropriate conditions. This refers to base on which the
measurement object will be placed, in this case the brake disc. It can be noticed, In reviewed literature that the disc
can be located on a sponge, or to hang on elastic ropes that are used for securing cargo during the transportation,
or to be fixed [12, 13, 22-24]. In this research, the chosen way of mounting the brake disc is to place a disc on the
sponge. Measurement installation used for the experiment is shown in Figure 6. Experiment is performed using the
NetdB 12 acquisition device. Excitation of the brake disc is performed by impact hammer (PCB Piezotronics 086C03).
The impact hammer consists of a top, a force sensor, a balancing mass and handles. The top, which is placed on
the impact hammer, is a soft impact cap. The force sensor is a piezoelectric and is embedded in the hammer's head
to capture the impact force. The response is measured by acceleration sensor AC102, marked with a black circle in
Figure 6. The acceleration sensor is connected to brake disc by the magnet. Figure 7 shows the disc with the sensors
used for measurement, as well and disc placement. The experiment results have been stored on a computer, which
was later processed.
Figure 6 Measurement installation scheme
Figure 7 Brake disc with sensors
When performing such an experiment, two conditions must be provided are:

 The location of the acceleration sensor is very important to get the best record of the response. If the
acceleration sensor is placed where impact impulse is very small, it will be very difficult to record FRF.
Acceleration sensor should be placed where high structure responses are expected. By analysis that is
earlier performed on different variants of the disc, it can be concluded that highest modes occur on friction
ring of a brake disc [25].
 The applied top for the impact hammer - the top to be used is the one that activates as many natural
frequencies as possible [15].
The Frequency Response Function (FRF) describes how the system will react to the impulse. If a particular amplitude
and frequency impulse is introduced into the system, the system will start vibrating at the same frequency as the
impulse, but with the phase shift. The response amplitude depends on the amplitude and frequency of the impulse.
The relationship between the amplitude of the response and the excitation impulse is the FRF and is phase-shifted
to any frequency [26].
Numerical analysis
First, the brake disc shown in Figure 5 is created in CATIA software package, and as such is imported into the
analysis software – ANSYS. 3D model of brake disc is created in real size.
The conditions in which the experiment was conducted are also defined in the analysis software. This means that it
is first necessary to define the temperature at which the experiment was performed. The used type of final elements
is tetrahedral; the number of elements is 21.137, while the number of nodes is 38.978. As the disc is tested under
317
static conditions, the degree of freedom is zero, which is also defined in the analysis software. In order to bring the
results of the numerical analysis closer to the values obtained by the experiment, it is necessary to get the accurate
data on the used material characteristics. The results of the modal analysis carried out by using numerical analysis
are most influenced by the Young’s module, which has proven both in this research and in the research of other
authors [15, 27]. The characteristics of the disc brakes’ material are shown in Table 1.
Table 1 Material characteristics of the brake disc

Density, kg/m3 Young’s module, MPa Poison ratio
7600 197,000 0.26
The results obtained by numerical analysis are presented in the form of nodal diameters since they are the most
dominant in the occurrence brake squeal. The reason for this is that the disc tends to vibrate around one or more
node diameters at the same time, Figure 8. The operation of nodes in this case causes the disc to get a waveform.
The number of node diameters is based on the number of nodes and anti-nodes that appeared on the surface of the
disc that is in contact with the brake pads [7]. The number of nodal diameters gives the so-called number of vibration
modes. The result of this phenomenon in the form of a brake disc noise is usually due to one or more nodal diameters.
Figure 8 Typical 8 diametrical mode of vibration—16 antinodes [7]
Results and discussion

The result frequencies of the experiment and numerical analysis are in range up to 11 kHz. This value corresponds
to the upper limit of high-frequency brake noise. The numerical analysis has been modified so that the deviations
between the values of the experiment and the numerical model are as small as can be seen in Figure 9, while the
error values are given in Table 2.
Figure 9 Experimental and numeral response of the free-free disc
318
Table 2 Error between numerical analysis and the experiment
Experimental, Hz 1548 3047 4945 6993 8891 10789

Numerical, Hz 1538 2994 4863 6856 8888 10955
Error, % 0.65 1.74 1.66 1.96 0.03 -1.54
The optimum choice of material's characteristic has resulted in a maximum obtained error value of 1.96%, Table 2.
The error between the results of the experiment and the numerical analysis is small, so for the future research the
characteristics of the materials shown in Table 1 can be used. The number of nodes can be displayed related to the
frequency, which should be approximated by the smooth curve [7], shown in Figure 10. If this condition is satisfied,
the results obtained by the numerical analysis can be accepted completely. For better illustration of the frequency,
the geometry of the disc is observed. This means that the frequency means can be represented related to the angle
between the nodal diameters. Analysing Figure 11, it can be noticed that as the frequency increases, the angle
decreases. The higher the frequency, the higher the number of nodes is, which directly affects the reduction of the
angle between them.
Figure 10 Natural frequency of free-free in function of number of nodes
Figure 11 Frequency in function of antinode angular spacing
319
CONCLUSIONS
Traffic noise is one of the dominant sources of noise in urban areas. Today there is a tendency to reduce noise to a
greater extent. One of the main reasons is that noise has a very negative impact on the health of humans living in
her immediate vicinity. Car manufacturers have, in recent years, been able to reduce the noise generated by the
vehicle during the exploitation. So any sound that occurs during the exploitation can cause distrust from the driver
and that the vehicle spent most of the time on the service instead of what is its primary purpose.
One of the most undesirable and irritating sounds that can occur during the exploitation is the brake noise. The
experimental methods for determining the brake noise are the application of pulsator, dynamometer or impact
hammer method. The method used during research represented in this paper was using the impact hammer. It is
very important to determine the natural frequency of any system and tries to reduce it as much as possible. In this
process, it must be ensured that there are no disturbances of other characteristics that are very important for the
safety of vehicles in traffic.
The modal analysis was performed by the experiment, and then, based on the results of the experiment, the
verification and modification of the numerical analysis results were performed. The numerical results do not deviate
to a large extent, so that for future research, the initial parameters for the performed analysis, or the characteristics
of the material, can be used.
Further research can be carried out on determining the natural frequency for the other components and verification
of the numerical data based on the results obtained by the experiment. The next step would be to determine the
natural frequency for the assembly as well to determine the frequency during braking process.
ACKNOWLEDGMENTS
This paper was realized within the researching project “The research of vehicle safety as part of a cybernetic system:
Driver-Vehicle-Environment” ref. no. TR35041, funded by Ministry of Education, Science and Technological
Development of the Republic of Serbia.
REFERENCES
[1] Budimcic, M., PhD Thesis: “Assessment of the facts of loud music on hearing loss in young adults”, Faculty of
Medical Science, Kragujevac, 2015,
[2] European Environment Agency, Noise in Europe 2014, Luxembourg: Publications Office of the European Union,
2014, ISBN 978-92-9213-505-8,
[3] Decree about noise indicators, limits, methods for assessing indicators of noise, disturbance and adverse effects
of environmental noise, “Official Gazette of the RS “, No. 75/2010,
[4] Republic of Serbia, Ministry of Environmental Protection, Environmental Protection Agency, Available from:
http://indicator.sepa.gov.rs/pretraga/indikatori/svefind/Open View, Accessed 24.06.2018.,
[5] Abd-Rahim, A.B.: “Disc brake squeal – a prediction methodology”, AV Akademikerverlag, ISBN 978-3-639-
45271-6, 2012, Saarbrucken, Germany,
[6] Mauer, G., Haverkamp, M.: “Measurement and assessment of noise caused by vehicle brake systems”, in Inter-
noise, 2007,
[7] Barton, C.D., Fieldhouse, D.J.: “Noise, Vibration and Harshness (NVH)”, Automotive Chassis Engineering,
Springer, pp 255-317, 2018,
[8] Liu, W., Pfeier, J.L.: “Introduction to Brake Noise & Vibration”, Honeywell Friction Materials, Available from:
http://www.sae.org/events/bce/honeywell-liu.pdf, Accessed 28.07.2018.,
[9] Day, A.: “Braking of Road Vehicle”, ISBN 978-0-12-397314-6, Waltham, 2014,
[10] Signalysis, Available from: http://signalysis.com/wp-content/uploads/2015/07/Testing-Brake-Rotor-Natural-
Frequency.pdf, Accessed 25.06.2018.,
[11] Bono, R., Stultz, G.: “Automated resonant inspection to validate resonant frequency characteristics within brake
components for improved NVH performance”, SAE Technical Paper 2010-01-1699, 2010,
[12] Kadu, N.N., Vivekanandan, N.: “Analysis and reduction of squeal noise in a disc brake of an automotive”,
International Engineering Research Journal (IERJ), 2015, pp 1231-1236,
[13] Parra, C., Olazagoitia, J.L., Biera, J.: “Application of testing, modal analysis and numerical methods to the
detection, prediction and avoidance of squeal noise in automotive brake systems”, Proceedings of International
Conference on Noise and Vibration Engineering ISMA2010-USD2010, 2010, Leuven, Belgium, pp 4351-4362,
[14] Williams, J.J.R., Zhang, Z., Oberst, S., Lai, J.C.S.: “Model updating of brake components’ influence on instability
predictions”, Proceedings of the 22nd International Congress on Sound and Vibration, 2015, Florence, Italy,
320
[15] Shodh Ganga, A reservoir of Indian Theses, Available from: http://shodhganga.inflibnet.ac.in/bitstream/
10603/11551/8/08_chapter%203.pdf, Accessed 11.07.2018.,
[16] da Silva, J., Fulco, E., Varante, P., do Nascimento, V., Diesel, F.N., Boniatti, D.L.: "Numerical and Experimental
Evaluation of Brake Squeal," SAE Technical Paper 2013-36-0030, 2013,
[17] Ewins, D.J.: “Modal Testing: Theory, Practice and Application“ second edition, Wiley, 2000, ISBN 978-
0863802188,
[18] Avitabile, P.: “Modal Testing: A Practitioner's Guide”, Wiley, 2017, ISBN 978-1-119-22289-7,
[19] Jimin, H., Zhi-Fang, F.: “Modal Analysis”, ISBN 0 7506 5079 6, Butterworth-Heinemann, Great Britain, 2001,
[20] Veg, Е. PhD Thesis: “Integrity assessment by cross correlated modal identification of steel structures”, Faculty
of Mechanical Engineering, Belgrade, 2015,
[21] Kun-Shan, L., Gene, F., Ray, S.: “The TMS320 Family of Digital Signal Processors”, Texas, 1997,
[22] Nouby, M.G., Sufyan, M., Abd-El-Tawwab, A. M.: “Understanding Mode-Coupling Mechanism of Brake Squeal
Using Finite Element Analysis“, International Journal of Engineering Research and Applications (IJERA), Vol. 2,
2012, pp 241-250,
[23] Nilesh, K.K., Sharad, S.C.: “Investigation of Natural Frequency and Modal Analysis of Brake Rotor Using Fea
and Ema”, International Journal of Innovative Research in Science, Engineering and Technology, Vol. 3, 2014,
pp 16495-16500,
[24] Smith, S., Bilbao-Ludena, J.C., Catalfamo, S., Brake, M.R.W., Reuß, P., Schwingshackl C.W.: “The Effects of
Boundary Conditions, Measurement Techniques, and Excitation Type on Measurements of the Properties of
Mechanical Joints”, Nonlinear Dynamics, Vol. 1, 2016, pp 415-431,
[25] Stojanović, N., Glišović J., Lukić J., Grujić I.: “Influence of vanes shape on the high frequency noise of ventilated
disc brakes in heavy vehicles”, Mobility & Vehicle Mechanics, Vol.43, No.3, pp 1-9, 2017, Doi
10.24874/mvm.2017.43.03.01,
[26] Wheeler, A., Ganji A.: “Introduction to Engineering Experimentation 3th edition”, Pearson/Prentice Hall, ISBN
978-0-13-174276-5, 2010, New Jersey, US,
[27] Abu-Bakar, A.R.: “Disc brake squeal – A prediction Methodology”, AV Akademikerverlag, ISBN 978-3-639-
45271-6, 2012, Stuttgart, Germany.
321
322
Kragujevac, Serbia
MVM2018-036
Sandra Veličković 1
Blaža Stojanović2
Lozica Ivanović3
Slavica Miladinović4
Saša Milojević5
APPLICATION OF NANOCOMPOSITES IN
THE AUTOMOTIVE INDUSTRY
ABSTRACT: The development of new lightweight and strong materials and the design of new products are
among the key elements for the development of new advanced construction and vehicle parts for the automotive
industry. The use of composite materials in the automotive industry has been popular in recent decades due to the
need to reduce vehicle weight, which directly affects fuel consumption and exhaust gases emission. In this way, the
development of improved new materials with improved performance is accomplished. Nanocomposites represent a
new class of materials that has excellent thermal and mechanical properties. The application of nanocomposites for
development of automotive components is reflected in the improvement of the production rate, environmental and
thermal stability, and the reduction in weight in the automotive industry, less wear parts, and indirectly to reduce
CO2 emissions and environmental pollution. This research paper presents a review of the application of
nanocomposites (metal, ceramic and polymeric) in the automotive industry.
KEYWORDS: application, ceramic matrix nanocomposites, metal matrix nanocomposites, polymer matrix
nanocomposites, vehicles
INTRODUCTION
Automotive industry is constantly making efforts to the development and application parts that are light weight, and
which at the same time possess excellent mechanical and tribological properties. Today, the automotive industry
must meet certain requirements in terms of reducing fuel consumption and at the same time to maintain adequate
comfort and safety of the vehicle. Another important factor is the weight of the vehicle, which has a direct impact on
fuel consumption and thus the emission of toxic and harmful products of combustion [13].
The hardest part of the vehicle is the body shell, which makes 40% of its total weight. Therefore, the most common
approach to lightweight design integrated within the structure of the vehicle which uses a combination of different
materials, depending on their engineering properties and functional characteristics. In recent years increased use
of high-strength steel, which in compact European cars makes 50 to 60% of the body. In addition to the use of
1
SandraVeličković, Research assistant, University of Kragujevac, Faculty of Engineering, Serbia, Kragujevac,
Sestre Janjić 6, sandrav@kg.ac.rs
2
Dr Blaža Stojanović, Associate prof., University of Kragujevac, Faculty of Engineering, Serbia, Kragujevac, Sestre
Janjić 6, blaza@kg.ac.rs
3
Dr Lozica Ivanović, prof., University of Kragujevac, Faculty of Engineering, Serbia, Kragujevac, Sestre Janjić 6,
lozica@kg.ac.rs
4
Slavica Miladinović, Research assistant, University of Kragujevac, Faculty of Engineering, Serbia, Kragujevac,
Sestre Janjić 6, slavicam@kg.ac.rs
5
Mr Saša Milojević, Assistant R&D, University of Kragujevac, Faculty of Engineering, Serbia, Kragujevac, Sestre
Janjić 6, sasa.milojevic@kg.ac.rs
323
steel, has increased the use of thermoplastics and thermoset composites used for coating bodywork, bonnet, roof
components and other components of the body.
The main materials used in the automotive industry are composites which are made of high-strength steel,
aluminum, and composites of carbon fiber and plastic. Applying high-strength steel is achieved by lessening the
weight by 20% compared to the conventional steel constructions, use of aluminum of 40% and 50% at the
application of composites of carbon fiber. Therefore, by using light materials, reduces the weight of the vehicle, for
example, the weight of aluminum ties in the suspension is smaller by 15% than the conventional, hollow composite
shaft is 18% lighter than the conventional, and the application of forged aluminum rim saves up to 36% when
compared to steel wheels [4].
Aluminum as a very light material is used as the main matrix element in the production of composite materials
referred to as aluminum metal matrix composites (AMCs). Application of aluminum composite is widespread, from
the military, automotive, flights until the space industry. AMCs compared with primary aluminum alloy have a good
combination of mechanical and tribological properties. In addition, these properties can be tailored to the specific
requests [1,2]. AMCs possess improved physical and mechanical properties such as strength to weight ratio, a
good ductility, a high strength and a high modulus of elasticity, low thermal expansion coefficient, good wear
resistance, good corrosion resistance, high temperature creep resistance, and better fatigue [5]. The use of
lightweight materials for the production of automotive components also contributed to the reduction of air pollution.
Nanocomposites represent a new class of materials that exhibit excellent thermal and mechanical properties. Due
to the high performance are suitable for use in aerospace, automobile, chemical and transport industry [2]. The first
industrial production of nanocomposites in the automotive industry occurred in 1991 with the production of timing
belt covers as part of the engine for Toyota Camry cars [6].
Most research and development based on nanotechnology is being implemented in the automotive sector.
Nanotechnology is applied to parts of the body, chassis and tires, automotive interiors, electrical and electronics, IC
engines and drive systems [7]. In recent years, remarkable progress has been made in various industries which
motivated researchers to work on new structural materials for better performance in engineering applications.
Tiruvenkadam N. et al have made investigations of the use of the aluminum hybrid nanocomposite for the IC
engine cylinder liner (NL) analyzing the performance, combustion and emission characteristics of the engine and
others. The aim was to replace the piston cast-steel composite material with the base of Al 6061 and Nano
reinforcement zirconium dioxide (ZrO2), silicon carbide (SiC) and Solid lubricant graphite (Gr) to the size of
reinforcement 100 nm, 220 μm and 100 microns, respectively. The mass fraction of each reinforcement amounted
to 2.25 wt. %. They concluded that the combination of the selected matrix with reinforcement suitable for NL based
on the density, mechanical and thermal properties and of a homogeneous mixture is carried out based on SEM
analysis. They also found that the application of NL can achieve greater incylinder pressure and higher efficiency.
Applying NL is reduced emissions of CO, HC and smoke quantity output, but the increased NOx emission [8].
Carvalho O. et al have applied a hybrid nanocomposite with an aluminum base for making the piston ring [9]. The
material that was used is a hybrid nanocomposite with aluminum base reinforced with SiC particles, and nano
MWCNT's of the order of 50 nm, produced by hotpressing technique in order to evaluate the different fraction of
CNTs. The best results have shown abrasion in case of the 6 wt. % CNTs, however, because the best ratio
between mechanical and tribological properties of a piston rings are selected from the 2 wt. % of CNTs.
Testing of mechanical and tribological characteristics of the nanocomposite is executive for the production of
connecting rods are Ramachandra M. et al [10]. The application of aluminum alloys with 99% of purity and
reinforcement nZrO2 25 nm in size were made using the nanocomposite powder metallurgy technique for the
production of connecting rods. Testing of wear is made to pin on disc wear testing machine. The study came to the
conclusion that the nanocomposite has a higher value than the hardness of the material base, and also reduces
wear nanocomposites.
The application of the nanocomposite with a base of aluminum A356-reinforced nanoparticles of the CNT is used
for making disc brakes and the drum [11]. Sundaram M. and Mahamane U. produced nanocomposite by the
method of casting with mixing. Tests have shown that an increase in the mass fraction of nano reinforcement CNT
based on the A356 leads to an increase in hardness, while reducing wear, and friction coefficient. Increasing the
hardness of the nanocomposite with an increase in the value of reinforcement in the aluminum base and have
found Ranaa R. et al [12]. They interrogators aluminum nanocomposite reinforcement with SiC is used for the
preparation of camshafts. Stir the nanocomposite is produced casting method with the percentages by weight of
reinforcement of SiC of 1 to 4 wt. % And a particle size of 40 nm in the Al base. To improve the mechanical and
tribological characteristics of aluminum nanocomposite came Asmatulu R. et al [13]. LM6 alloy strengthened
reinforcement nano SiC and Al2O3 in the relationship 0.5, 1.0, 1.5 and 2 wt. % of a styrene casting method. Based
on the results concluded that as such, the composite can be used in the automotive industry for the production of
brake discs.
324
Summarizes the application of a hybrid nanocomposite with aluminum base Al 356 and Al 7075 in the automotive
industry have made Dhanabal S. et al. [14]. As reinforcement nano size is used various kinds of materials and by
methods of making nanocomposites discussed with stirring casting and pressure casting. The study
characterization of nanocomposites came to the result that the piston for making the best show nanocomposites
with reinforcement SiCAl2O3, Al2O3MoS2, TiC, Gr, for the piston ring are nanocomposites reinforced with
Al2O3graphite, drum brakes for the reinforcement TiB2SiC, the shaft reinforcement is SiC, for camshaft
reinforcement SiCCNT, Al2O3TiC and Al2O3 for valves.
The aim of this research is reflected in the review of the application of nanocomposites with metal, polymer and
ceramic base in the automotive industry, and anticipates further growth in the use of these composites. The
increase in the use of nanocomposites for automotive components is expected from year to year because the use
of these composites impact on reducing the weight of vehicles, environmental and thermal stability and promoting
environmental actions.
COMPOSITE MATERIALS
Composite materials are produced by adding basic material of the reinforcement in order to improve certain
characteristics, depending on the purpose of the selected base material and the reinforcement. The composites are
used as a basic structural material in the aerospace and automotive industry is comprised of deformable metal
base and the rigid ceramic as reinforcement. These compositions are known as a composite with a metal base
(MMCMetal Matrix Composite). The first level of classification of composite materials depending on the base
material, in addition to existing MMC polymer matrix composites (PMCPolymer Matrix Composite), a ceramic
matrix when it is about composites with ceramic matrix (CMCCeramic Matrix Composite) [15, 16].
The second level of classification of composites relates to the shape of reinforcement - the particles, short fibers,
laminar composites with continuous fibers and woven composites. In this category are braided and knitted fiber
bundles (short fibers, continuous fiber laminated composites) and woven composites (braided and knitted fiber
architectures are included in this category) [17].
When the composite material has reinforcement in which at least one dimension in the nanometer scale or less
than 100 nm, then the case of nanocomposites [16]. Therefore, the nanocomposites are composites that consist of
the base material and the nano size of the reinforcement which do not dissolve in one another.
Nanocomposites reinforced particles are largely used in the automotive industry due to their ability to withstand
high temperatures and pressures. Several manufacturing methods used to produce them. The main disadvantage
of these materials is a lack of homogeneity in the dispersion of the particles as and weak links between the matrix
and the particles. However, research and testing of nanocomposites are still underway about the process of
obtaining and variations of certain parameters in order to overcome the aforementioned shortcomings.
Metal Matrix Nanocomposites

Nanocomposites reinforced particles are largely used in the automotive industry due to their ability to withstand
high temperatures and pressures. Several manufacturing methods used to produce them. The main disadvantage
of these materials is a lack of homogeneity in the dispersion of the particles as and weak links between the matrix
and the particles. However, research and testing of nanocomposites are still underway about the process of
obtaining and variations of certain parameters in order to overcome the aforementioned shortcomings.
Metal Matrix Nanocomposites (MMNC's) use ductile metallic matrix in which the ceramic reinforcement adds.
Various properties of these composites is improved, including the density, the hardness, abrasion, deformation and
corrosion resistance depending upon the reinforcement (type, type and content) and the matrix as a manufacturing
process. MMNC's become important because composites are used in the industry [19].
Nanoparticles as a reinforcement in metal matrix composites have been progressively replaced other types of
reinforcement such as nanofibers, nanowhiskers or nanoplatelets. The main division is reinforcing the oxides,
carbides, nitrides and borides. The most commonly used types of nanoparticles are: SiC, TiC, WC, TaC, TiB2, AlN,
and Al2O3 [18, 20].
For an easy and highquality use in automobile components as a template for the production of MMNCs used low
density metals such as aluminum (Al), magnesium (Mg), and titanium (Ti) [20]. Aluminum and aluminum alloys
have found many applications in addition to the automotive and aerospace industry in conjunction with
325
carbonreinforced composites, and silicon carbide. Use of aluminum alloy (such as AlSi alloy) for a matrix is
preferred because of its advantages, including low cost and easy workability [17].
The most important production methods that are used for aluminum composites are: stir casting, squeeze casting,
compo casting and infiltration, spray deposition and powder metallurgy [17]. Stir casting ranks first among all the
available techniques because of the low cost, the possibility of using a wide range of materials as well as on
processing conditions, where they can be made with compositions up to 30% in the volume fraction of
reinforcement [18].
Basic classification process for obtaining nanocomposites can be classified into three groups:
1. Solid state processes include different powder metallurgy techniques with modifications in the processing
steps such as, highenergy ball mill, hot pressing, hot isostatic pressing, cold pressing followed by
sintering treatment and extrusion;
2. Liquid state processes include different processes such as casting, stir casting and squeeze casting list;
and
3. Semisolid processes include recasting technique with its variants such as compo casting or in
combination with squeeze casting list.
Figure 1 presents the predictions of application of aluminum in different types of vehicles.
Figure 1 Application of aluminum from 2017 to 2022 [22]
Typical automotive parts which are made of a composite based aluminum alloys as a cylinder head (cylinder head),
the pistons, the wheels, crankcase, cylinder head cover, intake manifold, the intake system and the drive means
such as fork for a changespeed, drive shaft, gears, gear housing and so on. Application of aluminum castings in
the field of automobile industry followed the development of the casting process, the melt treatment and thermal
treatment. Great efforts have been made with the aim of improving mechanical and tribological characteristics, by
controlling the microstructure that is by improving the processes for preparation, treatment of the metal melt and
the heat treatment [15, 16].
Research of composite materials precisely nanocomposites is still in progress. Researchers have been testing the
nanocomposite concluded that they can be applied for the production of the respective parts of cars, such as the
piston rings [8], the pistons [9], the connecting rods [10], the disc brakes and the drum brake [11], the camshaft
[13], the shafts and the valves [14].
Nanoparticles are common byproducts of combustion of oil and other hydrocarbon compounds, combustion and
exhaust gases. As stated above, in the automotive industry uses a lot of materials with nanoparticles, and in recent
years studies have shown that these nanoparticles can be found in water, air, soil, and therefore in the bodies of
animals and humans. These results have led to a debate about toxicity and impact on the environment
nanocomposites. Method maturities of nanoparticles in the human body by inhalation, absorption through the skin
and if swallowed. Diseases that occur accumulation of nanoparticles in the body of the inflammatory airway
disease, bronchitis, asthma, lung cancer, neurodegenerative diseases, cardiovascular disease, liver cancer,
Parkinson's disease, Alzheimer's disease, Crohn's disease, and others. Because of these dangers should be given
a lot of attention to protecting people during manufacture, transport, handling, use, disposal and recycling of parts
and machines which in themselves have nanocomposites [13].
326
Polymer Matrix Nanocomposites
The polymers are excellent in terms of low density, ease of machining and the corrosion resistance, however, their
rigidity and strength are much less than is required for structural applications. The polymers are considered to be
suitable for structural applications matrices, due to their light weight, light processing and longterm stability [23].
Traditional micro reinforcement shall be replaced by nanoreinforcement in a polymer matrix, making a polymeric
nanocomposite. These reinforcement leads to a significant improvement of certain structural and functional
features of the polymer in relation to the polymer microreinforced reinforcement [4].
Polymer matrix nanocomposites (PMNCs) are a combination of materials wherein one dimension of the
reinforcement is less than 100 nm. The main conditions for the development are the low density PMNCs, ecological
and high mechanical properties with a high production speed, good surface finish and with improved corrosion
resistance, as well as better operating temperatures [18]. The preparation of nanocomposites uses a wide range of
polymers, of which the most commonly used elastomers, thermoplastics and thermosets [23].
With the development of nanotechnology, various types of nanoreinforcement such as silicate nanoparticles are
produced and used in the polymer materials to create socalled nanocomposites of polymer matrix (PMNCs).
Uniform dispersion of nanoparticles in the polymer matrix leads to a significant improvement of the properties of the
nanocomposites as compared to those of conventional reinforced composites micro particles [23, 24].
When the polymer matrix is added a small amount of the nanoparticles leads to improve the mechanical properties,
the increase in resistance to cracking and propagation, as well as reducing the coefficient of thermal expansion
(CTE) of the PMNCs. As reinforcement in polymeric materials for making these PMNCs, silica nanoparticles are
used due to relatively high elastic modulus, low density, and good abrasion resistance, low coefficient of thermal
expansion, high thermal stability, and low running costs [24].
There are different processes for the preparation of polymer nanocomposites and each of them has its advantages
and disadvantages. As well as in the manufacture of nanocomposite of metal and this must be taken into account
for selecting the appropriate method, which corresponds to the target application of the material, as well as the
material properties, the ratio of the base and reinforcement, as well as the dispersion of the target performance.
Solvent mixing, melt mixing, and in situ polymerization are some of the most common routes for nanocomposite
synthesis [23, 25].
Automotive market has the largest market share of the polymeric nanocomposites thanks to widespread use such
as a drive mechanism, the inner and outer power train, suspension and braking systems, exhaust systems and
catalytic converters, lubrication, tires and body parts [4, 26].
On the basis of more stringent legal regulations regarding standards of reducing emissions of cars is expected that
the market of polymer nanocomposites recorded a significant increase in applications in the automotive and
defense industries, as well as in aerospace engineering. It is expected that the production of nanocomposites result
in reducing carbon dioxide emissions from cars. In addition to the automotive industry, the use of polymer
composites is increased and the airline industry due to increased production of laminates, sandwich structure,
lightning rods and anti-radar protection. Besides the already mentioned application of polymer composites are also
used in biomedicine at the recording of biological, cell tracking, the magnetic separation of biological,
nanomedicine [26].
Other potential applications PMNCs in the automotive industry are as follows [18]: the cap of belt gear unit, the
bonnet, fuel hoses, fuel valves, door frames, backrests, fuel hoses, and fuel valves, door frames, seat backs, step
assists, heavy duty electrical enclosures, sail panel, box rails, fascia, grills, hood louvers, instrument panels, side
trims, body panels, and fenders, sensors, fuel cells. Figure 2 shows the application of polymer nanocomposites
with the parts of the car.
The automotive industry in Europe invests more than 5% of annual turnover in research and development, and the
main focus is on the development of better coatings and paints and stronger, more robust parts for end use.
Therefore it is expected that the use of polymeric nanocomposites car market growth over the next ten years [6].
Overall, the use of plastics in the global market is promising for passenger cars, light commercial vehicles and
heavy commercial vehicles. It is anticipated that the global plastics market by 2023 jump to 25.2 billion, and is
predicted also a growth of 2.6% in the period from 2018 to 2023.
327
Figure 2 Illustration of the usage of polymer nanocomposites parts [6]
New trends have directly influence the dynamics of the plastic industry, including the development of technologies
for recycling of plastic, and replace PA 66 with cheaper and more advanced PA 6.
The diagrams, Figure 3 shows the prediction of application of polymeric materials from 2018 to 2023.
Figure 3 Histogram display forecasting application of polymeric materials in cars from 2018 to 2023 [22]
Trends, opportunities and market forecasts in the automotive industry until 2023, Figure 4 by type of vehicle:
passenger cars, light-duty commercial vehicles and heavy-duty commercial vehicles.
Based on the diagram in Figure 4 can be concluded that in the next five years, provides for reducing the use of
heavy vehicles, indicating an increase in the application of new materials, i.e. nanocomposites.
328
Figure 4 Histogram display forecasts car market in the world [6]
Ceramic Matrix Nanocomposites
Ceramic composites with dispersed metal particles represent a class of promising material for high performance
applications in harsh environments, such as high temperature. These materials offer the possibility of combining a
resistance to heat, resistance to degradation, and the abrasion resistance of the ceramic phase having a high
mechanical strength and thermal conductivity is provided with metal phase.
Ceramic matrix nanocomposites (CMNCs) is used for making nozzle assemblies, materials stoves, systems for
converting energy, gas turbines, thermal engines, etc. [27]. The ceramic materials are insensitive to extremes of
temperature as opposed to the steel at a high temperature changes the mechanical properties. In particular, they
showed good composite materials reinforced with carbon fibers, which are very durable at high temperatures, have
a high wear resistance, and have substantially lower specific weight compared with steel and aluminum. Carbon
fiber reinforced composites are increasingly being used for car parts that are exposed to high loads, such as e.g.
brake disks, valves, cylinder liners, spark plugs, sensors, isolators, filters, piston and others. Safe driving is
achieved by applying the brake discs made of ceramic composites, because the discs have a great resistance to
deformation and wear, thereby enabling a great savings in weight.
Advantages of using the ceramic composite material in the automotive industry are: the ability to create very
complex shapes, reduce the cost of after-treatment of parts, the possibility of connecting the parts during the
manufacturing process, dimensional stability in extreme working conditions, corrosion resistance, ease of
maintenance, longer life time, possibility to work-up and recyclability [28].
Disadvantages application of ceramic composite materials in the automotive industry are: the high cost of
production, no serial production, the potential for products that are not in use, lack of skilled labor, cruelty (no
deformability), sensitivity to moisture and temperature, inability to repair, toxicity, flammability, etc. Another
downside application of ceramic materials is the tendency to fracture, expensive process of production and
insufficient research [28].
The aforementioned disadvantages are the reason of low presence of micro and nano-ceramic materials in the
automotive industry which can be seen in the diagram in Figure 5.
Compared to conventional CMCs ceramic nanocomposites have improved hardness, high strength, toughness and
creep resistance, thermal shock resistance, flame retardancy, magnetic and optical properties, chemical inertness,
wear resistance and low density. In addition to automotive applications because of their thermal shock resistance
and flame retardancy are used for high-temperature thermal shock zones, such as the fusion energy applications,
in gas turbines, aerospace applications, cutting tools, rock drill tips, tools and dies as well as general wear parts,
and many others [29].
The interest of science in recent decades toward the magnetic nanocomposites because of their functional
properties for which are widely used starting from the field of magnetic fluid catalysis and adsorption,
biotechnology, biomedicine, magnetic resonance imaging, data storage, electronics, magneto-optic and photo
catalysis to supercapacitors, hyperthermia, drug shipments, remediation of water, collecting energy. When
environmental and biomedical and catalytic apply magnetic metalceramic nanocomposites because of its porous
structure. These composites are composed of a dispersion of nanoparticulate Fe0, CO0, or Ni0 nanoparticles
usually in amorphous silica or alumina ceramic matrix [30, 31].
329
STATUS OF APPLICATION OF NANOCOMPOSITES IN THE WORLD IN THE
AUTOMOTIVE INDUSTRIES
Improving the performance of vehicles in terms of quality elements as well as the reduction of greenhouse gas
emissions is the constant tendency of the automotive industry. The main driver of the motor vehicle industry is to
increase fuel efficiency and reducing vehicle weight.
The application of composites in cars is affected by high production as well as stringent market regulations.
According to the International Organization of Motor Vehicle Manufacturers, in 2016, total production of 95 million
cars and commercial vehicles. In the following chart according to research [26] provides an overview of the car
market in the US with revenues by products in recent years and forecasting to 2025, Figure 5.
Figure 5 The car market in the US revenue by product, 2014  2025 (USD Million) [26]
Growing of carbon emissions due to increased number of vehicles in recent years in the world, launched by
governments around the world to react and reconsider the emissions from vehicles. Countries such as China,
Japan, Vietnam and Indonesia have implemented various policies in terms of emission standards.
Car manufacturers and composites are well acquainted with the regulations of the countries and producing
products that meet specified standards. Automakers believe that reducing vehicle weight is one of the most
effective solutions for ensuring compliance and reducing emissions. For every 100 pounds per 45 kg (from 2 to
3%), results in low emissions [26].
According to studies [22] of the global vehicle market is projected to grow by 5.3% in the period from 2018 to 2023
with great possibilities in the areas of passenger cars, commercial vehicles and heavy commercial vehicles
segment. The main growth drivers for this market are increasing production of the vehicle, increasing the need to
improve the ride quality and the growing need for advanced systems of safety and comfort [22].
According to reports and prognosis [32] of the European market of automotive composites there is a possibility of
growth in a variety of applications, exterior, interior, chassis systems, body and others. It is anticipated that the
European market of automotive composites by 2021 will reach $ 4.1 billion and is projected to grow by 5.8% from
2016 to 2021.
The main growth drivers for this market are increasing car production and increasing demand for lightweight
materials. The European Union has set new standards require that passenger cars must meet CO 2 emission target
of 95 g∙km1 (equivalent to 57.9 mpg) by 2021, and for the production of light vehicles is 147 g∙km1 (equivalent to
43.3 mpg) by 2020 years. The European Commission proposes to improve European standards of fuel economy.
New trends that influence the dynamics of the industry include increasing penetration of thermoplastic and carbon
composites. Another emerging trend is in the formation of strategic alliances between the OEMs and the supplier of
carbon fiber and the resin in the automotive industry.
CONCLUSIONS
The following generalized conclusions can be written from the study performed during presented research:
330
 The main initiators of growth in demand nanocomposites are increasing car production and increasing
demand for lightweight materials. It is expected that the implementation of the structural and operating
parts increase the rate because composites act as an excellent substitute for metal parts;
 Increasing importance is attached to reducing the weight of vehicles because of improving fuel economy,
then to improve vehicle performance and reduce emissions. Based on the review of literature predicts an
increase in demand for nanocomposites in the automotive industry in the coming years;
 It is expected that the increase in research and development activities nanocomposites lead to an
increase in their application in structural and operating parts and thus stimulate the growth of the market.
The advantage of using nanocomposites in external components and drive luxury cars have a positive
impact on demand in the market. Nanocomposites based on a variety of materials of metal or plastic
reinforced with metal or ceramic nanoparticles can significantly improve the strength of parts. Different
requirements of the automotive industry, such as improved mechanical, electrical, thermal, corrosion, self-
cleaning and antiwear properties, and sensing abilities correspond nanocomposites reinforced with
nanoparticles, nanofilms, nanoflakes, nanotubes, nanofibers; and
 In the future, attention should be paid, in addition to developing lighter materials, the protection of the
environment through increased recycling rates of the vehicle at end of life. Based on the examination
papers noted that talking about the difficulties in the recycling of composites and can not give up on their
application in the automotive industry, it is necessary to find a proper recycling methods.
ACKNOWLEDGMENTS
This paper presents the research results obtained within the framework of the projects TR35021, TR35033 and
TR35041 financially supported by the Ministry of Education, Science and Technological development of the
Republic of Serbia.
REFERENCES
[1] Sharma, P., Khanduja, D., Sharma, S.: “Dry sliding wear investigation of Al6082/Gr metal matrix composites by
response surface methodology”, Journal of Materials Research and Technology, Vol. 5, No. 1, 2016, pp 29-36,
[2] Suresh, S., Shenbaga Vinayaga Moorthi, N., Vettivel, S.C., Selvakumar, N.: “Mechanical behavior and wear
prediction of stir cast Al–TiB2 composites using response surface methodology”, Materials and Design, Vol. 59,
2014, pp 383-396,
[3] Đorđević, Z., Jovanović, S., Veličković, S., Jović, Đ., Milićević, J.: “Influence of application of new material in
automotive industry on improving quality of life”, 2nd International conference on Quality of Life, Center for
Quality, Faculty of Engineering, University of Kragujevac, 2017, Kragujevac, pp 329-332,
[4] Gutowski, V., Yang, W., Li S., Dean, K., Zhang, X.: “Lightweight Nanocomposite Materials, In book:
Lightweight and Sustainable Materials for Automotive Applications”, 2017, pp 277-302,
[5] Mavhungte, S.T., Akinlabi, E.T., Onitiri, M.A., Varachiа, F.M.: “Aluminum Matrix Composites for Industrial Use:
Advances and Trends”, Procedia Manufacturing, Vol. 7, 2017, pp 178-182,
[6] Nanowerk; Available from: https://www.nanowerk.com/spotlight/spotid=23934.php, Accessed 20.05.2018.,
[7] Mathew, J., Joy, J., George, C.S.: “Potential applications of nanotechnology in transportation: A review”,
Journal of King Saud University - Science, 2018, open access, In Press, Corrected Proof,
[8] Tiruvenkadam, N., Thyla, P.R., Senthilkumar, M., Bharathiraja, M., Murugesan, A.: “Synthesis of new
aluminum nano hybrid composite liner for energy saving in diesel engines”, Energy Conversion and
Management, Vol. 98, No. 2015, pp 440–448,
[9] Carvalho, O., Buciumeanu, M., Madeira, S., Soares, D., Silva, F.S., Miranda, G.: “Optimization of AlSi–CNTs
functionally graded material composites for engine piston rings”, Materials and Design, Vol. 80, 2015, pp 163–
173,
[10] Ramachandra, M., Abhishekb, A., Siddeshwarc, P., Bharathid, V.: “Hardness and Wear Resistance of ZrO2
Nano Particle Reinforced Al Nanocomposites Produced by Powder Metallurgy”, Procedia Materials Science,
Vol. 10, 2015, pp 212–219,
[11] Sundaram, M.U., Mahamani, A.: “Development of Carbon Nanotube Reinforced Aluminum Matrix Composite
Brake Drum for Automotive Applications”, Research and Innovation in Carbon Nanotube-Based Composites,
Dr. Brahim Attaf (Ed.), Chapter, 2015,
[12] Ranaa, R. S., Purohit, R., Sonia, V. K., Das, S.: “Development and Wear Analysis of Al- Nano SiC Composite
Automotive Cam”, Materials Today: Proceedings, Vol. 2, 2015, pp 3586 – 3592,
[13] Asmatulu, R., Nguyen, P., Asmatulu, E.: “Nanotechnology Safety in the Automotive Industry”, Elsevier,
Amsterdam, 2013, pp 57-72,
[14] Dhanabal, S., Vetrivel, S., Vimal, R.M.: “An overview of hybrid metal matrix composites – characterization,
directed applications, and future scope”, International Journal of Scientific Engineering and Applied Science,
Vol. 1, No. 9, 2015, pp 344-350,
331
[15] Stojanović, B.: “Tribological behavior of hybrid composites with A356 matrix”, Doctoral thesis, Faculty of
Engineering University of Kragujevac, 2013,
[16] Džunić, D.: “Friction and wear analisys of Metal Matrix Nanocomposites”, Doctoral thesis, Faculty of
Engineering University of Kragujevac, 2015,
[17] Gazawi, A.A.: “Microstructure and mechanical properties of aluminium based nanocomposites strengthened
with alumina and silicon carbide”, Doctoral thesis, University of Waikato, Hamilton, New Zealand, 2014,
[18] Borgonovo, C.: “Aluminum Nano-composites for Elevated Temperature Applications”, Thesis, Worcester
Polytechnic Institute, 2010,
[19] Saxena, A., Singh, N., Kumar, D., Gupta, P.: “Effect of Ceramic Reinforcement on the Properties of Metal
Matrix Nanocomposites”, Materials Today: Proceedings, Vol. 4, No. 4, 2017, pp 5561-5570,
[20] Ravi, K. R., Nampoothiri, J., Raj B.: “Nanocomposites: A Gaze through Their Applicationsin Transport
Industry”, Chapter in book: Nanotechnology for Energy Sustainability, Vol. 2, 2017, pp 831-856,
[21] Bulei, C., Todor, M.P., Kiss, I.: “Metal matrix composites processing techniques using recycled aluminium
alloy”, IOP Conf. Series: Materials Science and Engineering, Vol. 393, No. 012089, 2018, pp 1-7,
[22] Lucintel – Global Management Consulting & Market Research Firm, Available from: https://www.lucintel.com/,
Accessed 10.06.2018.,
[23] Dubey, K.A., Hassan, P.A., Bhardwaj, Y.K.: “High Performance Polymer Nanocomposites for Structural
Applications”, Chapter 5, Materials Under Extreme Conditions, Elsevier, 2017, pp 159-194,
[24] Hassanzadeh-Aghdam, M.K., Ansari, R., Mahmoodi, M.J., Darvizeh, A.: “Effect of nanoparticle aggregation on
the creep behavior of polymer nanocomposites”, Composites Science and Technology, Vol. 162, 2018, pp 93-
100,
[25] Müller, K., Bugnicourt, E., Latorre, M., Jorda, M., Sanz, E. Y., Lagaron, J. M., Miesbauer, O., Bianchin, A.,
Hankin, S., Bölz, U., Pérez, G., Jesdinszki, M., Lindner, M., Scheuerer, Z., Castelló, S., Schmid, M.: “Review
on the Processing and Properties of Polymer Nanocomposites and Nanocoatings and Their Applications in the
Packaging”, Automotive and Solar Energy Fields Nanomaterials, Vol. 7, No. 4, 2017, pp 1-47,
[26] Nanocomposite Market Analysis, Market Size, Application Analysis, Regional Outlook, Competitive Strategies
And Forecasts, 2015 To 2022”, https://www.grandviewresearch.com/, Accessed 10.06.2018.,
[27] Durowaye, S.I., Sekunowo, O.I., Lawal, A. I., Ojo, O.E.: “Development and characterisation of iron millscale
particle reinforced ceramic matrix composite”, Journal of Taibah University for Science, 2017, Vol. 11, No. 4,
pp 634-644,
[28] Đorđević, Z.: “Composite constructions”, Faculty of Engineering University of Kragujevac, Kragujevac, 2018,
[29] Bose, N.R.: “Thermal shock resistant and flame retardant ceramic nanocomposites”, Woodhead Publishing
Series in Composites Science and Engineering, 2013, India, pp 3-50,
[30] Esposito, S., Dell'Agli, G., Marocco, A., Bonelli, B., Allia, P., Tiberto, P., Barrera, G., Manzoli, M., Arletti, R.,
Pansini, M.: “Magnetic metal-ceramic nanocomposites obtained from cation-exchanged zeolite by heat
treatment in reducing atmosphere”, Microporous and Mesoporous Materials, Vol. 268, 2018, pp 131-143,
[31] Barrera, G., Tiberto, P., Esposito, S., Marocco, A., Bonelli, B., Pansini, M., Manzoli, M., Allia, P.: „Magnetic
clustering of Ni2+ ions in metal-ceramic nanocomposites obtained from Ni-exchanged zeolite precursors”,
Ceramics International, Vol. 44, No. 14, 2018, pp 17240-17250,
[32] Growth Opportunities for the European Automotive Composites Markets, Available from:
https://www.prnewswire.com/, Accessed 10.06.2018.
332
Kragujevac, Serbia
MVM2018_040
Aleksandar Poznić1
Boris Stojić2
SECONDARY CONTRIBUTION TO THE ANALYSIS OF

MAGNETORHEOLOGICAL BRAKE
ABSTRACT: The primary challenge in magnetorheological brakes` design is acquiring the greater braking
torque in constrained volume and weight. This paper represents a part of the research that deals with the feasibility
of increasing the overall braking torque of magnetorheological brake model by varying the shape and the number
of active surfaces in constrained volume. Improved design of the previous magnetorheological brake is modeled
and presented. Magnetic flux density distribution through the rotor of this novel magnetorheological brake`s design
was considered. For the purposes of magnetic flux density distribution analysis a number of simulations were
carried out using commercial finite element method software - COMSOL Multiphysics, AC/DC module. Materials`
magnetic properties, required for simulation process, were previously obtained from manufacturer or were
determined by the measurements and applied to the simulations. Post-processing was utilized to calculate the
magnetic flux density distribution and intensities across the models` specific cross sectional areas. The proposed
magnetorheological brake design shows uniform distribution of the magnetic flux density and great potential for
overall braking torque increase.
(Some figures may appear in color only in an electronic version)
KEYWORDS: magnetorheological brake, simulation, rotor shape, constrained volume, torque increasing
INTRODUCTION
Magnetorheological - MR brake is a type of electromechanical brake that consists of a stator, rotor, working fluid
and one or more excitation coils. The magnetorheological fluid - MRF, is the working fluid of the MR brake - MRB,
and is contained between the stator and the rotor. When excited, by the control current, each coil generates
magnetic field through MRB`s body. Affected by the magnetic field, the MRF`s viscosity changes [2, 4]. This
rheological change leads to change in the MRB`s braking torque value.
There are several MRB types [1]. Regardless of their construction differences, the direction of magnetic flux density
vector is theirs common feature. The magnetic flux density - magnetic flux direction needs to be as perpendicular to
MRF`s flow direction i.e. MRF`s active surfaces as possible and needs to form a closed loop.
From the magnetic and construction point of view, typical MRB is composed of MRF, as (ferromagnetic) working
medium, nonmagnetic and magnetic materials. Magnetic properties of a MRFs` can easily be obtained from theirs
manufacturers. Nonmagnetic materials, such as aluminum, have known magnetic properties. On the other hand,
magnetic properties of magnetic material such as construction steel, usually are not known or are available, and
need to be determined by measurements. The most important material`s magnetic property, in this case, is the
initial magnetization curve, which has a highly nonlinear characteristic, [5].
1
Aleksandar Poznić, PhD assistant, University of Novi Sad, Faculty of Technical Sciences, Trg Dositeja
Obradovića 6, 21000 Novi Sad, alpoznic@uns.ac.rs
2
Boris Stojić, assist. professor, University of Novi Sad, Faculty of Technical Sciences, Trg Dositeja Obradovića 6,
21000 Novi Sad, bstojic@uns.ac.rs
333
The major issue with any MRB`s application, e.g. robotics, automotive, industry etc. is that the overall braking
torque value is still far too small. There are several ways to increase it. One of them is to use MRFs with better
yield characteristics and to reduce MRF`s gap size inside the brake. Second is to increase the applied magnetic
flux acting on the MRF. The last one is to enhance the size of the MRF`s active surface area by multiplying the
number of its layers.
The objective of this work was to simulate magnetic flux distribution through novel MRB`s design MRF layers. MRF
layers vary in height. Particular MRF height and its influence to the magnetic flux distribution were analyzed using a
commercial Finite Element Method - FEM software. The goal was to form a uniform magnetic flux distribution
throughout all layers of the model so consequently each layer contributes equally to the overall braking torque
value, this was not the case in previous model [6]. Obtained magnetic flux values then can easily be converted into
braking torque values for each layer.
NOVEL MAGNETORHEOLOGICAL BRAKE DESIGN

The authors of this paper have used previous, [1], MRB`s type division: drum, inverted drum, disc, T-shape rotor
and multiple discs MRB. There are design variations for each type. The MR disc brake design is very common
MRB type found in literature today [1, 4], but the emphasis of this research was placed on the hybrid design,
combining the drum and T-shape rotor brake design.
Improved design
In Figures 1 a), b) and c) are illustrations of: improved MRB design, its cross-section with emphasis on the MRF
layers and segment the multi-T-rotor assembly respectively. This, improved model, is also a design variant of the
drum and the T-shape rotor design. Opposed to MR drum brake type, that only has one coil, or to the MR T-shaped
rotor brake, that has two separate coils, improved design has several (eight) individual stationary excitation coils,
thus forming a multi-pole structure. Improved MRB design shares excitation coil arrangement with previous [6]
MRB design, Figures 1 d) and e). Here, also, the eight excitation coils are divided into two sets. The first set of coils
is radially arranged on the circumference of a MRB`s stator. The second set of coils is positioned parallel to rotor`s
shaft axis. Each radial coil`s magnetic flux vector - B is directed towards the center of the MRB, thus increasing
the magnetic flux intensity, acting on the MRF contained inside the brake. Both parallel coil`s B are directed
opposite to each other i.e. outwards, away from the center of the MRB, Figure 2 left.
To increase the total MRF active surface area, T-rotor element was concentrically multiplied several times inwards,
thus forming a multi-T-rotor element, Figure 1 c) and f) from [6]. Proposed MRB multi-T-rotor assembly, i.e. shaft
and multi-T element, is composed of both nonmagnetic and magnetic materials, similarly as in previous design [6].
Nonmagnetic shaft also features nonmagnetic disk, designated as multi-T-element inner support. The nonmagnetic
disk diverts magnetic flux lines spreading route through body of the MRB, and splits it into two magnetic flux layers,
Figure 2 right. These two layers act uniformly onto separate but geometrically equal segments of the MRF active
surfaces i.e. MRF layers.
To have a closed magnetic circuit, two six-spoke magnetic flux return bridges were re-applied, Figure 1.
Based on magnetic simulations carried out on previous models, [6], and their results analysis, it was concluded that
the overall magnetic flux intensity inside MRB increases with decrease of the multi-T-rotor radius. Further analysis
lead to conclusion that the induced torques in MRF layers were not equal and were not contributing equally to the
overall torque value, thus stressing unfavorably the entire MRB construction. Also, based on these results and
construction analysis it was concluded that inner layers of MRF are potentially close to MRF saturation. These
problems were tackled in this research and a new i.e. improved T-rotor element was designed, Figure 1 c). The
goal was to compensate the summary influence that the six radial excitation coils have towards the center of the
MRB, by changing the MRF heights across the multi-T-rotor assembly. This change in construction was to lead to
more uniform magnetic flux distribution across T-rotor element.
A brief parameters overview of the proposed MRB design is presented in Table 1. Proposed MRB model is planned
to be manufactured in near future.
334
a) b) c)
d) e) f)
Figure 1 Magnetorheological multi-pole multi-T-rotor brake designs, a) improved design, b) cross-section of the
improved design with emphasis on active surfaces, c) segmental representation of the improved multi-T-rotor
assembly, d) previous design [6], e) cross-section of the previous design with emphasis on active surfaces,
f) segmental representation of the previous design multi-T-rotor assembly
Figure 2 Improved magnetorheological brake design cross-section magnetic flux spreading routes ilustrations
Table1 Multi-pole multi-T-rotor magnetorheological brake parameters
Parameter Value
Magnetorheological brakes` outer diameter (mm) 315
Magnetorheological brake`s length (shaft not included) (mm) 127.45
Multi-T-element outer radius (mm) 100
Magnetorheological fluid`s active area height; from - up to (mm) 20.5 - 90.45
Nonmagnetic disks’ thickness (mm) 2.5
Nonmagnetic disk radius (mm) 100
Shafts’ diameter (mm) 15
Magnetorheological fluid gap (mm) 0.5
Number of coils (-) 8
Maximum control current intensity per coil (A) 1
335
Mathematical modeling
In case of the multi-T-rotor brake the torque generating properties can be described by the same analytical model
used for the MR drum brake model with the results adjustment for additional MRF layers and their specific radiuses
and heights. The maximum field-induced torque, for MR drum brake, is given by:
   
k k
T   2  Ro  2    Ro  l k   y   4     y  l k  Ro2 (1)
k k k
1 1
Similarly, the maximum viscous torque is:

 
k
T   4       Ro3  l k (2)
1 g k
where, k is the number of MRF layers, Ro is a radius of a specific MRF layer, l k is the MRF layer`s height,  y is
k
the yield stress developed in response to the applied magnetic field,  is the viscosity of the MR fluid with no
applied magnetic field,  is the angular velocity of the rotor and g is the thickness of the MR fluid gap.
It was concluded [6] that the overall intensity of the B increases as it progresses toward inner MRF layers,
summarizing the excitation effect of all six coils leading to increase of the  y as the Ro decreases. To exclude
k
 
this effect, the researchers fixed the product of the lk  Ro2 to a specific value. This value was determined at the
k
very far radius and the very least height of the MRF layer. The only variable here was the height of the MRF. This
has lead to a distinctive curvature in T-rotor design.
This combination resulted in a different magnetic flux distribution i.e. equal overall torque contribution from each of
the MRF layers. Magnetic flux intensity simulations results are presented in this paper.
NUMERICAL SIMULATIONS
The proposed MRB design was modeled using commercial FEM software, COMSOL Multiphysics. Due to
presence of the nonmagnetic disk and two six-spoke magnetic flux density return bridges, COMSOL`s 3D space
dimension option was utilized. Magnetic field was considered to be static, so the Stationary Study was used. To
reduce computation power required the model was cut along shaft`s axis, Figure 1 b).
Entire model was surrounded with a sphere shaped air boundary, several times the volume of the model.
Appropriate material nodes and boundary conditions were assigned to every element of the model. In this specific
simulation, materials such as nonmagnetic air and aluminum were selected from COMSOL`s database, but
nonlinear magnetic materials, such as C15E steel and MRF Basonetic 5030, were defined using previously
obtained data, [5]. These data have been loaded to the COMSOL as separate files. Note, presence of elements
such as ball bearings were neglected because of their steel composition and small volume share in overall
construction.
In Magnetic Fields subsection of the model, additional Ampère’s Laws were needed, due to the use of several
different materials. In the same subsection only three Multi-Turn Coil domain nodes were added due to axial
symmetry cut implemented in this model.
Mesh, Figure 3, was generated using the User-controlled mash. The MRF`s layers were meshed using the Free
tetrahedral with custom element size. The minimum tetrahedral element size was at 0.05 mm. Also special
attention was placed on the curvatures and the narrow regions of MRF segments of the brake. The curvature radii
were multiplied by the Curvature factor parameter which in return gives the maximum allowed element size along
specific boundary. The Resolution of narrow regions parameter controls the number of element created in narrow
regions. These parameters greatly improved the mash quality of the models, which is now at the threshold of 0.1,
which is considered satisfactory mesh. The solver was stationary but non-linear.
Main difference between previous and this improved model was the shape of the T-rotor element, Figure 4 a). The
previous model had constant MRF layers height opposite to the improved model where the height of the MRFs
layers gradually raises. Both models had identical excitation coils magnetic flux return bridges arrangement.
336
To determine the overall magnetic flux intensity in a specific MRF layer FEM simulation was carried out. A median
magnetic flux value was determined along three predetermined circular lines, same as in [6]. A 1D Plot Group line
graphs were used to depicture magnetic flux magnitude changes along these three circular lines. Circular lines
were positioned at the very bottom of the MRF layers, at the very top of it and in the middle of them, [6].
Figure 3 Proposed magnetorheological multi-pole multi-T-rotor brake`s mesh
a) b) c)
Figure 4 Proposed magnetorheological multi-pole multi-T-rotor brake`s elements cross section illustrations,
a) MRF layers, b) core and c) nonmagnetic elements of the brake
In the simulation, the hexagonal prism stator, the coils’ cores, the multi-T-rotor element and the six-spoke magnetic
flux return bridges, Figure 4 b), were assigned with the magnetically soft steel C15E. The rest of the MRB assembly
elements, Figure 4 c), were assigned with nonmagnetic materials.
RESULTS AND DISCUSION
Magnetic flux density distribution

Magnetic flux distribution pattern within the proposed MRB was studied and the results are presented. Average
magnetic flux intensity change in each MRF layer is presented in Figure 5 and in Figure 6 a sample of the actual
results is depictured. Average numerical values are presented in Table 2. These values were determined along
three circular lines for each MRF layer. Minimum and maximum values of magnetic flux, along these lines, ranged
from around 0.0004 T up to 0.0009 T in all MRF layers. Good results repeatability is achieved especially in the
outer MRF layers area. Inconsistencies in magnetic flux results are due to exciting coils periodical arrangement and
specific shape of the MRF layers.
Linearity in magnetic flux change through MRF layers is noticeable. This may lead to consistency in the field
induced torque values among MRF layers. This was predicted and considered in this stage of the MRB design.
337
Figure 5 Change in magnetic flux intensity in each magnetorheological fluid`s layer
Figure 6 Magnetic flux simulation sample results for improved magnetorheological multi-pole multi-T-rotor brake
Table 2 Numerical values of magnetic flux intensities at specific radii
Improved magnetorheological brake design

Layer B, [T] Rok, [mm]
1 0,0007 22.75
2 0,0007 33.82
3 0,0007 44.89
4 0,0007 55.96
5 0,0007 67.03
6 0,0007 78.1
7 0,0007 89.17
8 0,0007 100.25
338
CONCLUISION
This study presented a novel viewpoint on magnetorheological brakes. An improved magnetorheological brake
design was presented. By combining magnetic and non-magnetic materials, a new space may have been opened
for magnetorheological technology. The use of non-magnetic material and a specific inner elements shape lead to
magnetic flux uniform distribution through magnetorheological brake’s body.
The goal of this study was to determine magnetic flux intensity change with regard to change in height of
magnetorheological fluid`s layers. For this purpose magnetorheological brake finite element model was made using
commercial finite element model software. This improved model had the same basic geometric properties and
number of excitation coils as previous one. Magnetic flux intensities along specific lines inside magnetorheological
fluid`s layers were obtained. Results are presented graphically and tabular.
Nonlinear relationship between magnetic flux and magnetic field in different materials was applied in the
simulations. Combination of materials shapes may contribute to other magnetorheological applications, where
there is a need for magnetic flux increase in small areas, but where geometric restrictions are present.
The proposed multi-pole multi-T-rotor improved magnetorheological brake design shows big potential. Greater
braking torque, in constrained volume and weight, is now achievable. Future work should be focused on scaling the
magnetic flux intensity through magnetorheological fluid`s layers which was not the case in this study.
ACKNOWLEDGMENTS
This paper presents a part of the researches on the project TR35041 – "Investigation of the safety of the vehicle as
part of cybernetic system: Driver-Vehicle-Environment" and the project TR31046 " Improvement of the quality of
tractors and mobile systems with the aim of increasing competitiveness and preserving soil and environment",
supported by Serbian Ministry of Education, Science and Technological Development.
Magnetorheological fluid data used in research was provided by BASF ® Company. The authors of this paper would
like to express special gratitude to BASF ® Company and especially to its project manager Dr. Christoffer Kieburg
for all its support.
REFERENCES
[1] Avraam, M. T.: "MR-fluid brake design and its application to a portable muscular rehabilitation device", PhD
thesis, Active Structures Laboratory Department of Mechanical Engineering and Robotics, 2009, Bruxelles,
[2] Carlson, J. D., Jolly, M. R.: "MR fluid, foam and elastomer devices", Mechatronics, Vol. 10, No. 4 - 5, 2000, pp
555 - 569,
[3] Comsol Multiphysics, AC/DC Module, User`s Guide. https://www.comsol.com/,
[4] Karakoc, K., Park, E. J., Suleman, A.: "Design considerations for an automotive magnetorheological brake",
Mechatronics, Vol. 18, Issue 8, 2008, pp 434 - 447,
[5] Poznić, A., Miloradović, D., Zelić, A.: "Determination of magnetic characteristics of some steels suitable for
magnetorheological brake construction", Proceedings of the Mechedu 2015, Conference and workshop, 2015,
pp 130 - 133,
[6] Poznić, A., Stojić, B.: "A contribution to the analysis of magnetorheological brake", The 10th International
Symposium Machine and Industrial Design in Mechanical Engineering (KOD 2018), Vol. 10, 2018, 2018, pp 1 -
9.
339
340
Kragujevac, Serbia
MVM2018-045
Sonja Kostić1
Zorica Đorđević2
Milosav Đorđević3
THE INFLUENCE OF THE INTERNAL RADIAL CLEARANCE

ON THE LOAD DISTRIBUTION ON THE ROLLER BODY OF
A SINGLE RADIAL BALL BEARING
ABSTRACT: The distribution of the external radial load on rolling bodies of the radial ball bearing is uneven and
from one ring to the other is transmitted over a few of the total z balls which are located under the meridians flat.
This is a statically undefined problem, so for the analysis of load distribution bedise static equilibrium conditions
need to be introduced and supplemental one which are based on the relationship between the contact deformation
of the contact parts and loads. The unevenness of the distribution is greater considering the influence of the inner
radial clearance of the bearing. Purpose of the paper is to determine as precisely as possible the load transmitted
by the most heavily loaded rolling body as well as the other rolling bodies in the load zone. The paper provides a
complete analytical overview of the clearance impacts on the load distribution. The 2D model was developed by the
FEM whose results are compared with the results obtained by analytics and the results available in the literature.
Detailed study of the influence of the internal radial clearance is very important because it directly influences the
bearing performance characteristics such as bearing radial stiffness and static load.
KEYWORDS: single radial ball bearing, internal radial clearance, load distribution, Finite Element Method,
Ansys
INTRODUCTION
The distribution of the external radial load to the roller bodies of the radial ball bearing is uneven which means that
at the same time all balls are not equal in the transfer of loads from one ring to the other. The total external load is
transmitted from one ring to the other only through a certain number of total balls. The unevenness of the
distribution of the external load is influenced by the intensity of the load, the inner radial clearance, the elastic
deformation of the bearing rings, the internal bearing geometry and other factors. In this paper, a detailed analysis
of the influence of the inner radial bearing clearance on the distribution of the external load on the roller bearing
bodies is given.
DETERMINATION OF EXTERNAL LOAD DISTRIBUTION TO THE BALL

BEARING ROLLING ELEMENTS
It is a statically undefined problem, therefore, for static equilibrium analysis it is necessary to introduce additional
ones which are based on the connection between the contact deformations of the connected parts of bearing
(moving one ring in relation to the other under the influence of the external load) and the load.
1
Sonja Kostić, Visoka tehnička škola strukovnih studija u Kragujevcu, Kosovska 8, sonja25yu@yahoo.com
2
Zorica Đorđević, University of Kragujevac, Faculty of Engineering Kragujevac, Sestre Janjic 6, zoricadj@kg.ac.rs
3
Milosav Đorđević, Visoka tehnička škola strukovnih studija u Kragujevcu, Kosovska 8,
m.djordjevic.kg@gmail.com
341
Consideration of the load distribution on the ball bearing rolling elements is carried out assuming that the bearing
parts are stationary; that all balls of the exact dimensions and shapes and that the diameters and radii of the rolling
path of the rings are of perfectly accurate measures, shape and position.
Radial ball bearing with internal radial clearance

An analysis of the distribution of the external load to the ball bearing rolling elements will take into account that the
housing in which the outer bearing ring is fitted is hard (does not change the shape due to the load) and that the
direction of the external load action passes through the center of one of the ball and is directed to the ball. The
external radial force Fr is distributed unevenly to a number of balls zs in a loaded zone below the meridian plane
which represents the plane of the bearing normal to the direction of the attack line of force (Figure 1). The Meridian
plane divides the bearing into two zones - one is placed unloaded and in the other there are loaded balls. Balls
whose centers lie at the meridian level do not transmit the load.
Fr
e/2
r
M M
0z
F2   Fz-2
2
F2 F1 Fz-1 Fz-2=F2
F0
F1 Fz-1=F1
F0
Figure 1 Load distribution to ball bearing rolling elements of radial bearings with internal radial clearance
From the static equilibrium conditions, the relationship between the external radial load and the force at the place of
the most loaded ball and the force at the place of other balls can be determined. As this is a statically
indeterminate/undefined problem, additional conditions are introduced which connect elastic contact deformations
and loads [1-3]. It is assumed that in the unloaded state at the centric position of the rings the touching of the balls
and the outer ring is achieved and that there is a radial clearance which is equal to half of the total clearance
between the balls and the inner ring, as shown in Figure 1.
The distribution of the external load on the ball bearing rolling elements of radial bearings with radial clearance is
radically uneven compared to the bearing without radial clearance, as the clearance increases with the increase in
the load area and increases the intensity of the force at the point of the most loaded ball. Therefore, the total
contact deformation at the place of the most loaded ball  0 z at the initial touch point is higher than the
corresponding contact deformation of the bearing without a radial clearance  0 .
The total contact deformation of the coupled parts at the initial point of contact at the place of the most loaded ball
in the existence of the radial clearance  0 z is determined by iterative method using the Newton's iteration method
(tangent method) [2], where by determining an unknown parameter  x we obtain:
e
 0z   x (1)
2
When  0 z is determined, the total radial displacement of the bearing axis  r can be determined, and on the basis
of this the total contact deformation at the position of the random i-th ball  iz can be determined. Depending on the
calculated values of the total contact deformations  0 z and  iz , the corresponding forces can also be determined
at the place of the most loaded ball bearing with a radial clearance:
F 0z  C   0z
32
(2)
342
and at the place of random i-th balls:
F iz  C   iz
32
(3)
By knowing the total contact deformation of the coupled parts at the place of the most loaded ball of bearing with a
radial clearance, the factor of unevenness of load distribution on the balls of the given bearing can finally be
determined:
C 
32
k  F 0z   0z
 Fr   Fr  (4)
   
 
z  z 
From the above expression, it can be seen that the factor of unevenness of load distribution on the balls of a given
bearing depends on the size of the external radial load Fr, the number of balls z, the inner bearing geometry and
the elastic modulus and the Poisson coefficients of the bearing parts (constant 𝐶𝛿 ) and the total contact
deformation of the coupled parts at the starting point of the contact  0 z . For a particular bearing and at a given
external load, the factor k is only in function of  0 z . Since  0 z depends on the external load and the internal radial
clearance e, for the given external load factor k is the function of the inner radial clearance. As the higher values of
the radial clearance increase the contact deformation of the coupled parts, the greater is the unbalanced load
distribution factor of the balls of the given bearing k.
An analytical example of the determination of the external load distribution on the ball bearing rolling
elements
Determination of external radial load distribution parameters on the ball bearing rolling elements is performed on a
single ball bearing with radial contact 6206. This bearing has a wide application in industry. In the automotive
industry, it is an inevitable part of the componentes in transmission system (gearboxes, differentials...),
watherpump systems, altenators, wheels etc at commercial and special vehicles. The bearing characteristics
required for calculation are given in the literature [3].
For different values of the radial clearance e whose value ranged from 0 to 20 μm and for different values of the
external radial load whose values ranged from 1000 to 10000 N, there is obtained table of the results of the
calculation (table 1).
Table 1 Results of the calculation for ball bearing 6206

e Results of Fr ,N
the
μm calculation 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
F0 487.2 974.4 1461.7 1948.9 2436.2 2923.4 3410.7 3897.9 4385.2 4872.4
0 F 18 N 326.6 653.3 980.0 1306.7 1633.4 1960.1 2286.7 2613.4 2940.1 3266.8
F 27 35.2 70.51 105.7 141.03 176.2 211.5 246.80 282.0 317.3 352.5
F0 537.6 1042.5 1543.8 2039.7 2534.8 3028.9 3522.2 4014.8 4507.2 4999.0
10 F 18 N 301.7 624.9 950.3 1274.4 1599.2 1924.2 2249.3 2574.6 2899.9 3225.4
F 27 - - 5.7 22.3 43.0 66.1 90.8 116.6 143.3 170.7
F0 581.8 1098.3 1607.8 2114.0 2618.1 3120.9 3622.7 4122.0 4620.1 5117.2
20 F 18 N 272.9 588.4 908.6 1230.9 1554.6 1879.1 2204.3 2528.8 2853.4 3178.1
F 27 - - - - - - 0.962 10.0 23.4 39.4
Based on the obtained results, the load distribution dependency on the rolling balls depends on the intensity of the
external load (Figure 2 - Figure 3) in the function of the radial clearance.
By analyzing the obtained results and the diagrams presented it is concluded that increasing the intensity of the
external radial load linearly increases the load values transmitted by the rolling elements of the bearing. It can be
seen that increasing radial clearance increases the value of force on the site of the most loaded balls and reduces
the loads of other balls involved in the distribution. This increases the unbalance of the load distribution on the
rolling elements as the overall contact deformation increases on the site of the most loaded balls and reduces the
contact deformations of the joined parts at the spots of the other balls. For the sake of clarity, on the diagrams, a
dashed line represents the load value that the rolling elements transpose when ball bearing has no clearance.
343
Theoretically, in the case of large internal radial clearance in the bearing and at certain external loads, the load
distribution on the ball bearing rolling elements is very unfavorable, regarding the expected balls do not come in
contact, so that a large part of the external load is transmitted by the ball at line of attack force. Only after the great
contact deformation of the coupled parts comes the contact of the other rolling elements.
5000
without clearance
4000
F0
3000 F1-8
F , Fi , N
F2-7
2000
0
1000
0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Fr , N
Figure 2 Load distribution on ball bearing rolling elements of 6206 depending on external load, in case of bearing
with internal radial clearance (e=10 μm)
5000
without clearance
4000
F0
3000
F1-8
F0 , Fi , N
F2-7
2000
1000
0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Fr , N
Figure 3 Load distribution on ball bearing rolling elements of 6206 depending on external load, in case of bearing
with internal radial clearance (e=20 μm)
Table 2 gives the results of factor of uneven load distribution in the function of radial clearance e.
Table 2 Factor of uneven load distribution for ball bearing 6206

e Results of Fr ,N
the
μm calculation 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
0 4.385 4.385 4.385 4.385 4.385 4.385 4.385 4.385 4.385 4.385
10 4.838 4.691 4.631 4.589 4.562 4.543 4.528 4.516 4.507 4.499
k
20 5.236 4.942 4.823 4.756 4.712 4.681 4.657 4.637 4.620 4.605
30 5.632 5.193 5.015 4.914 4.849 4.802 4.767 4.739 4.716 4.697
Based on the presented solutions, diagrams of factor changes of balls uneven load distribution in function of the
inner radial clearances are given (Figure 4). The figure shows that the values of the factor k = 5 correspond to the
radial clearance e  29 μm (at a given external load, eg F = 3000 N), and that for other values of the clearance it
344
changes (bigger or smaller). It is concluded that for the different clearances values the same value of factor of
uneven the load distribution on balls k = 5, as it appears in the literature [4], can not be taken but the actual value
(calculated values given in the table). This is especially important when designing bearings and supporting
elements in general when the analysis of load distribution plays an important role. Taking into account the actual
value of factor k it is possible to determine more precisely the static bearing load. For the sake of visibility the
curvature depends only on the external load in the interval Fr  1000  5000 N.
6 F =1000 N
.
.
.
F =5000 N
k
4
0 10 20 30 40
e , m
Figure 4 Change of factor of uneven load distribution to the ball bearing rolling elements in function of the radial
clearance, and for the given external load
THE CHOICE AND MODELLING OF SIMPLIFIED PHYSIC AND FEM MODELS

OF SINGLE ROW DEEP GROOVE RADIAL BALL BEARING FOR ANALYSIS
OF LOAD DISTRIBUTION OF BALL BEARING ROLLING ELEMENTS
In Figure 5, a 2D plane mesh model of finite elements for single row radial ball bearing 6206, which has nine balls,
is given (the worst moment is shown when one of the balls is on the attack line of the given external load). For the
discretization of the 2D physical model, 2D four-node isoparametric finite elements were used. During
generatization of 2D mesh of finite elements a plane state is chosen (it is assumed that in all cross-sections of
bearing the stress distribution is the same) as well as very precise division of finite elements mesh.
Figure 5 The mesh of 2D finite elements for the model of radial bearing 6206 with defined contact finite elements,
boundary conditions (twisted coordinate system) and external load
Model of the bearing is presented in x-y global Descartes' coordinate system with its ground zero in axis of inner
bearing ring. Balls bearing have local coordinate systems. Ball that is in the line of offensive force (line of y-axis)
has the same coordinate system as global model bearing system with its ground zero in the centre of the ball.
Other balls have twisted coordinate systems compared to the global one with its ground zero placed in the center
of them. For the sake of simplified simulation of load and data observing the model is set in such a way that
external load is given in the line of y-axis (radial direction). Thus the total contact deformation of the most stressed
ball (the relative displacement of bearing) could be seen from results obtained from FEM analysis as node
displacement in the direction of y-axis of global coordinate system.
Outer bearing ring is absolutely immobilized (boundary conditions are given so that displacement is not allowed
neither in line of x-axis nor y-axis). Inner bearing ring is limited in line of x-axis that is its displacement is allowed in
345
the line of external load. The ball placed in the line of external load has exactly the same boundary conditions as
inner bearing ring. The limitations of the balls in the line of external offensive line are given in radial twisted
coordinate systems. Boundary conditions are given so that imitate the cage where balls are placed in the real
environment. The displacement is allowed in the radial direction (direction that connects the center of model
bearing to appropriate balls center). In the perpendicular direction in the radial axis of twisted coordinate system,
the displacement equals zero. Without boundary conditions defined that way balls were simply sliding along model
during experiment (by giving them the external load). Nodes placed in expected contact zone of balls with outer
and inner bearing ring have no boundary conditions. External radial load is given in the ground zero of global
coordinate system of bearing model and variates from 1000 up to 5000 N.
The results error obtained by FEM is less than 10% compared to theoretic results.
Figure 6 Numerical results of load distribution on ball bearing 6206 uder the external radial load for the case of ball
bearing without internal radial clearance
From figure 6, the distribution of the given external radial load to the ball bearing rolling elements for the case of
ball bearing without internal radial clearance is clearly visible. The most loaded ball in the bearing is ball in the
direction of the action of the external load then the balls left and right of the most loaded ball. Balls right below the
meridian plane are the least involved in the distribution of loads. It should be emphasized that the load values
distribution by the corresponding balls left and right from the most loaded are equal which is another indication of
the validity of the adopted numerical bearing model.
Figure 7 Numerical results of load distribution on ball bearing 6206 uder the external radial load for the case of ball
bearing with internal radial clearance e  20 m
In Figure 7, the influence of the inner radial clearance on the load distribution is obvious. As obtained from the
analytical calculation and numerical models results have been confirmed that balls located just below the meridian
plane (and farthest from the most loaded ball for the given external loads) do not come into contact until a large
contact deformation occurs.
For the additional verification of the accuracy of numerical solutions, a case in the literature was provided, bearing
6306, the results of which were compared with the available theoretical [4] and numerical [5] solutions. The
appearance of the 2D model of the finite elements of the radial ball bearing 6306 and the boundary conditions are
identical to the bearing 6206 whose words are already detailed.
346
Figure 8 clearly shows that results obtained by using FEM described above are more than satisfactory. Results
error of FEM analysis compared to theoretic data is less than 1% (note: results are obtained by giving the external
load on 6306 bearing). Comparison of the results obtained by both, ANSYS and I-DEAS Supertab softwares,
demonstrate that numeric results of ANSYS are more precise to theoretic data [4] than that of I-DEAS Supertab [5].
40
35 theoretic results
FEM results (1)
30
FEM results (2)
25
r , m
20
15
10
0
0 1000 2000 3000 4000 5000
Fr , N
Figure 8 Graphic demonstration of the results comparison (total radial axis displacement of 6306 bearing and
intensity of external load are compared) according: Theoretic research, FEM analysis by using of I-DEAS Supertab
software -curve (2), [5] and FEM analysis by using of ANSYS software - curve (1)
CONCLUSIONS
Due to the complexity of the problem the numeric model of the ball bearing mentioned above is developed. The
Finite Element Method is the numerical method used. Software used is ANSYS v.5.4. The choice and modelling of
single row radial ball bearing for load distribution on ball bearing elements and stress deformation state analysis is
detailed in literature. Obtained results confirmed the theoretic approach to the problem of load distribution on ball
bearing elements analysis (the error of obtained results compared to the theoretic ones did not excess 10%). The
same approach of ball bearing modelling could be applied to other series and ball bearing types as well. The future
exploration of this problem will be directed to 3D modelling and obtaining the most reliable results possible.
ACKNOWLEDGMENTS
This paper presents the results obtained during research within the framework of the projects TR 33015, supported
by the Ministry of Education, Science and Technological Development of the Republic of Serbia.
REFERENCES
[1] Harris, T.A.: “Rolling Bearing Analysis“, 1984, New York,
[2] Ristivijevic, M., Mitrović, R.: “Load distribution – tooth gearing and rolling bearings”, Faculty of Mechanical
Engineering, 2002, Belgrade,
[3] Kostić (Stefanović), S.: “Analysis of the Influence of Elastic Deformation and Internal Radial Clearance on the
Static Load Rating of the Rolling Ball Bearing“, Master thesis, 2004, Kragujevac,
[4] Mitrovic, R.: “Influence of Elastic Deformation and Internal Radial Clearance Analysis on the External Load
Distribution of the Ball Bearing Rolling Elements“, Master thesis, 1987, Belgrade,
[5] Subic, A.: “Geometry Modelling of Ball Bearing with Numeric Displacement Definition in Bearing Arrangement“,
Master thesis, 1991, Belgrade.
347
348
Kragujevac, Serbia
MVM2018-046
Aleksandar Radaković1
Dragan Milosavljević2
Gordana Bogdanović3
Dragan Čukanović4
Petar Knežević5
ELASTIC WAVES PROPAGATION IN COMPOSITE

LAMINATES USED IN THE AUTOMOTIVE INDUSTRY
ABSTRACT: The paper analyses the propagation of elastic waves in composite laminates which are used in the
automotive industry. The problem is limited to the propagation of Lamb’s waves in laminate plates the surfaces of
which are stress free. The problems of propagation are considered from the aspect of applying high-order shear
deformation theories. Second-order shear deformation theory is applied for obtaining phase velocity diagrams and
dispersion relations. The obtained results are compared with the results of 3D elasticity theory given in the literature.
A code has been developed in Matlab for analyzing the propagation of Lamb’s waves using symbolic variables as
well as for obtaining dispersion relations and phase velocity diagrams in an analytic way. The occurrence of quasi-
modes which result from stacking of plies is also described. Materials whose constitutive matrices are defined through
engineering constants are considered.
KEYWORDS: elastic waves, high-order deformation theories, composite laminates, wave modes
INTRODUCTION
There are two theoretical approaches in the analysis of elastic waves propagation. Theoretical approach based on
3D elasticity theory has been used by authors in the following papers [1-3]. The development of laminate theories
has brought the approach of analyzing the propagation of these waves using high order shear deformation theories
[4]. This approach has been considerably less exploited than the process of stress continuity on interlayers of
laminates which is based on 3D elasticity theory. This approach is more useful for numerical methods, but it is an
approximation in comparison with 3D elasticity theory and it is based on dynamic equations of motion obtained by
using variational methods. In this paper the problem is limited to the propagation of Lamb's waves in laminate plates
the surfaces of which are stress free.
We examine the propagation of the plane wave in the plane xy of the adopted coordinate system. The direction of
wave propagation has been defined by the angle  in relation to the positive side of the x-axis of the adopted
1
Aleksandar Radaković, State University in Novi Pazar, Vuka Karadžića bb, 36300 Novi Pazar,
aradakovic@np.ac.rs
2
Dragan Milosavljević, University of Kragujevac, Faculty of Engineering, Sestre Janjić 6, 34000 Kragujevac,
dmilos@kg.ac.rs
3
Gordana Bogdanović, University of Kragujevac, Faculty of Engineering, Sestre Janjić 6, 34000 Kragujevac ,
gocab@kg.ac.rs
4
Dragan Čukanović, Faculty of techical sciences, University of Prištiai, Knjaza Miloša 7, 38220 Kosovska
Mitrovica, dragan.cukanovic@pr.ac.rs
5
Petar Knežević, Državni univerzitet u Novom Pazaru, Vuka Karadžića bb, 36300 Novi
Pazar, pknezevic@np.ac.rs
349
coordinate system, so that the projections of the wave number к are defined as kx  k cos( ) и ky  k sin( ) . The
obtained results have been compared with the results of 3D elasticity theory given in literature. Constitutive matrices
are defined through engineering constants.
Matlab code has been developed for analyzing the propagation of Lamb’s waves using symbolic variables. Moreover,
a completely clear diagram of phase velocities has been given.
THEORETICAL ASSUMPTIONS
Second order deformation theory belongs to the group of polynomial theories and many authors have been
developing it. Whitney and Sun [5] presented a purified structural second order shear deformation theory for static
dynamic analysis of laminate orthotropic plates. Engblom and Ochoa [6] used this theory to develop a formulation of
finite elements for static analysis of cross-ply laminate plates. Kwon and Akin [7] developed a parabolic shear
deformation theory without a shear correction factor. Fares [8] presented second order shear deformation theory for
analysis of free vibrations of anisotropic plates. Tessler [9] developed a new two-dimensional theory of laminate plate
for linear elastostatic analysis of thick composite plates. According to this author plane displacements are linear,
while transversal displacements of sqare function are thickness coordinates. The shorcoming of these theories is the
use of shear correction factor which occurs in the first order shear deformation theory. One of such examples is the
assumption of displacement fields given by Whitney and Sun [5] with correction factors given by Kane and Mindlin
for homogeneous isotropic plate. According to this theory, displacement field can be represented as follows:
z2
u  x, y , z, t   u0  x, y , t   z x  x, y , t   x  x, y , t  ,
2
z2 (1)
v  x, y , z, t   v 0  x, y , t   z y  x, y , t   y  x, y , t  ,
2
w  x, y , z, t   w 0  x, y , t   z z  x, y , t  .
Pursuant to this theory, deformation tensor components will be:
u u0  x 1 x v v 0  y 1 y

 xx   z  z2   xx
0
 zk xx0  z 2 k 1xx ,  yy   z  z2   yy
0
 zk yy  z 2 k 1yy ,
x x x 2 x y y y 2 y
0
 xx k0 k1 0
 yy k0 k1
xx xx yy yy
w u v u0  x z 2 x v 0  y z 2 y u0 v 0

 zz    z  x, y , t  ,  xy    z   z    
z y x y y 2 y x x 2 x y x
0
 xy
(2)
  x  y  2 1  x
y  u w w 0   
 z      xy  zk xy  z k xy ,  xz   x   z  x  z    xz  zk xz
0 0 2 1 0 0
z ,
 y x  2  y x  z x x  x 
0 0
 xz 0
k xy k1
xy k xz
v w w 0   
 yz   y   z  y  z    yz
0
 zk yz
0
.
z y y  y 
0
 yz k y0z
While the correction factors 1, 2 , 3 , 4 , 5 , are introduced as:
 0 xz  1 0 xz ;  0 yz   2 0 yz ;  0 z   3 0 z
 z    z  z    z  (3)
x    4  x 
 ;           .
x x y  y 
y 5 y
  
According to this theory, apart from the rotation of the normal (to the middle plane), due to bending and transversal
shear, there is the elongation of the normal, as well as bending of the normal which are described by the square
elements of the assumed displacement field. Use of this theory becomes more important in macro-mechanical
analysis of thick and moderately thick laminate plates.
The most used method for obtaining dynamic equations of motion is the well known Hamilton's principle. By defining
of kinetic energy and energy of laminate plate deformation we obtain equations of motion as follows [10]:
350
N x N xy M y M xy I3   y
2
 2u I  2x
 u0 :   I1 20  3 ,  y :   Qy  ,
x y t 2 t 2 y x 2 t 2
N y N xy  v 0 I3  y
2 2
Rx Ry  2 y
 v0 :   I1  ,  z :   N z  I3 ,
y x t 2 2 t 2 x y t 2 (4)
Qx Qy w0 2
Sx Sxy I  2u I  2x
 w0 :   I1 , x :   Rx  1 20  5 ,
x y t 2 x y 2 t 4 t 2
M x M xy I  2 x Sy Sxy I1  2v 0 I5  y
2
 x :   Qx  3 , y :   Ry  
x y 2 t 2 y x 2 t 2 4 t 2
T T
Where F  Nxx , Nyy , Nz , Nxy , Mxx , Myy , Mxy ,Sxx ,Syy ,Sxy  and Fs  Qxz ,Qyz , Rx , Ry  are projections of load vector per
hk
width unit which relate to plane, that is, shear stress, and I1...7    k 1, z, z 2 , z 3 , z 4 , z 5 , z 6  dz , n is the total number
n
k 1 h 
k
of layers in the laminate.
The assumed forms of analytical solutions for this type of problem are given in [11] as:
 
i  k x x  k y y  t   
i  k x x  k y y  t   
i  k x x  k y y t 
u0  U0 e  
; v 0  V0e  
; w 0  W0e  
;
 
i  k x x  k y y  t   
i  k x x  k y y  t   
i  k x x  k y y  t 
 z  ze  
; x   0 e  
; y   y e  
; (5)
 
i  k x x  k y y  t   
i  k x x  k y y  t 
 x  xe  
;  y  y e  
;
where  is angle frequency.
By exchanging the assumed forms of solutions in the equations of motion and using the elementary algebra
transformation of matrices, equations of motion are transformed into a matrix notation as
LU  0; (6)
T
Where L is the matrix of important coefficients, while U  U0 V0 W0 x y z x  y  is vector of
displacement amplitude.
With symmetric laminates there occur a bigger number of zero elements in the matrix L so that the equation (6),
becomes:
 L11 L12 0 0 0 L16 L17 L18   U0 

L  
 12 L22 0 0 0 L26 L18 L28   V0 
0 0 L33 L34 L35 0 0 0  W0 
  
 0 0 L34 L44 L45 0 0 0   x 
0  0 (7)
0 L35 L45 L55 0 0 0   y 
 
L16 L26 0 0 0 L66 L67 L68    z 
L  
L18 0 0 0 L67 L77 L78    x 
 17 
L18 L28 0 0 0 L68 L78 L88    y 
In the last equation the separation can clearly be seen into symmetric mode which refers to extensional waves:
351
 L11 L12 L16 L17 L18   U0 
L  
 12 L22 L26 L18 L28   V0 
L16 L26  
L66 L67 L68   z   0,
 
L78   x 
(8)
L17 L18 L67 L77
 
L L88   y 
 18 L28 L68 L78
Ls
and antisymmetric part which refers to bending waves:
L33 L34 L35  W0 

L  
 34 L44 L45    x   0,
(10)
L35 L45 L55   y 
Las
For the existence of non-trivial solutions of the equations (8) and (9) it is necessary to satisfy the condition:
det LS   0 и det Las   0. (11)
If in the latter equations we exchange expressions for phase velocity cP   / k , and projections of the wave number
kx  k cos( ) and ky  k sin( ) for a fixed propagation direction  we obtain polynomial equations in the function of
k . For the adopted value of fh and the used relation   2 f it is possible to solve polynomial equations in the function
of к, that is, k 2 . The polynomial equation in the function of к for the symmetric mode shall be of the tenth degree,
that is, of the fifth degree in the function of k 2 , while the equation of the anti-symmetric mode shall be polynomial of
the sixth degree in the function of к, that is, of the third degree in the function of k 2 .
If we use some numerical solver for determining the root of the polinomial equations of the fifth, that is, third degree,
we shall obtain five, that is, three different roots for k 2 . When the values of this roots are sorted out and using the
advantages of the Matlab program in the sense of combining symbolic and numerical programming, 5 different
curves shall be obtained relating to symmetric mode, that is, 3 curves relating to antisymmetric mode. If we create
ordered pairs fh and phase velocity values, at a fixed value of the angle of propagation, dispersion curves are
obtained, while ordered pairs of phase velocity and angle of propagation, at a fixed value fh, give diagrams of phase
velocities in the polar coordinate system. Symmetric and anti-symmetric waves are defined as wave modes which
can be symmetric, that is, anti-symmetric in relation to the middle plane of the plate. In one ply, where the direction
of wave propagation matches the material direction, extensional (symmetric) and bending (anti-symmetric) wave
modes separate from shear (symmetric and anti-symmetric) modes. However, in the ply/layer which is a constituent
of the laminate, in which the direction of wave propagation does not match the material direction, extensional and
bending modes stack with shear horisontal modes.
NUMERICAL RESULTS
For diagrams of results we have used the material with the following characteristics:
Е1  168GPa , Е1  9.31GPa , Е1  9.31GPa , G12  3.45GPa , G13  3.45GPa ,

kg
G13  5.17GPa ,  12  0.33,  13  0.33,  23  0.4,  =1610
m3
If the value of the product fh is fixed, in the polar coordinate system, it is possible to obtain the diagram of dependance
of phase velocity c p from the variation of the agle of plane wave propagation. By varying the wave propagation angle
from 00 to 3600 we obtain diagrams fiven in figures 1. and 2. In order to present it easier, here it is necessary to do
normalization, and for this purpose we use longitudinal wave velocity in the direction of fibers cL  C11 /  where
C11 is elasticity constant in fibers direction in the local coordinate system, and  is the density of the selected
material. In the pictures it is possible to clearly see the separation of modes which are obtained by sorting out
polynomial equation roots per k 2 . There are five curves which refer to the symmetric mode of propagation and three
curves which refer to the anti-symmetric mode of propagation. It should be highlighted that in situations when the
352
direction of propagation does not match the material direction of the layer then even in one layer there are no pure
modes but quasi modes occur.
Due to better visibility in the figures we use different colors as follows:
1. In the anti-symmetric mode,
 light blue color presents diagrams of phase velocity obtained for the smallest value of the solution of
polynomial per k 2 . Observed in relation to wave modes, this phase velocity represents quasi-shear
hiorisontal anti-symmetric qSH1 mode where shear mode is dominant,
 purple colour presents phase velocity diagrams obtained for the biggest value of the solution of
polynomial per k 2 . Observed in relation to wave modes, this phase velocity refers to anti-symmetric
quasi-bending mode А0 ,
 red colour presents phase velocity diagrams obtained for values of the solution of polynomial per k 2 ,
which are between the maximal and minimal values of the solution, and it is denoted with qА1, which
clearly indicates the dominant anti-symmetric, that is, bending mode.
2. In the symmetric mode,
 light blue and green colour present phase velocity diagrams obtained for the minimal, that is, maximal
value of the solution of polynomial per k 2 , and refer to quasi-shear horisontal modes qSH2 and qSH0
respectively.
 red, purple and blue colour, respectively, present phase velocity diagrams obtained for sorted values
of solutions (sorted out from the smaller to the bigger solution), which are between the minimal and
maximal values of solution of polynomial per k 2 , and refer to quasi extensional modes which are
denoted with qS2 , qS1, qS0 .
Figure 1 present diagrams of phase velocity in one ply, that is, lamina.
a) Аnti-symmetric mode b) Symmetric mode

Figure 1 Diagrams of phase velocities of wave propagation in a layer at the relation value fh / cL  0.25
In the figure 1. it can be seen that phase velocities of all modes depend on the direction of waves’ propagation, apart
from the fundamental mode where this influence is considerably smaller. It is easy to notice that due to anisotropic
characteristics of the layer, when the direction of propagation matches fibers’ direction (at angles of 180 , that is, 0
), phase velocities reach their maximum values. On the other side, in the situations when the direction of wave
propagation is perpendicular is to the fibers’ direction in the diagram of phase velocities, we can observe occurrence
of their minimum values. Moreover, it should be noted that there are no completely pure modes, apart from
fundamental modes which mostly resemble the pure modes. Also, symmetry can be seen in relation to the directions
which match material directions, that is, in relation to the direction 0  180 , that is, 90  270 .
Diagrams of phase velocities of symmetric and anti-symmetric mode of the laminate 90 / 60 / 90 / 60 / 0s are
given in the figure 2. In the figures 2.a) and 2.b) it can be seen that the propagation of these waves in the laminate
is much more complex than in one ply, that is, it can be observed that there is no symmetry in relation to the horisontal,
that is vertical direction, which is caused by the manner of stacking of plies. It can be noticed that there is polar
symmetry, that is, every point in the phase velocity diagram has own reflection, which can be obtained in relation to
the observed point by the rotation by 180 of the vector which connects coordinate origin and the observed point in
the phase velocity diagram. Also, it can be noticed that there are no pure modes, nor rotations of directions of diagram
353
symmetry, which in this case do not match neither horizontal nor vertical direction. It is possible to also see points in
which diagrams of various modes have similar values. In this case extreme values of phase velocities of all modes
are not on the directions 0  180 , that is, 90  270 , but they coincide with some other directions which define the
direction of polar symmetry. For example, phase velocity diagrams of quasi-bending mode qА1, presented in the
Figure 2b with red color, has polar symmetry in the in the direction which is close to the direction 30  210 .
а) Аnti-symmetric mode b) Symmetric mode

Figure 2 Diagrams of phase velocities of wave propagation in the laminate at the relation value fh / cL  0.25
CONCLUSIONS
This paper examines dynamic behavior of laminate plates at the propagation of Lamb's waves. Second-order
deformation theory with correction factors has been successfully implemented. It has been shown that at the
propagation of these waves, due to non-homogeneity of the material and high degree of anisotropy, there occur
geometric irregularities in the diagrams of phase velocities. Generally, number of curves in the symmetric and anti-
symmetric propagation modes can depend on the implemented deformation theory. By applying second-order
deformation theory it has been shown that there are three curves of phase velocity which refer to the anti-symmetric
propagation mode and five curves which refer to the symmetric propagation mode. In both cases there have not been
pure propagation modes, but quasi modes have occurred, and those are in the form of quasi-bending qА, quasi-
extensional qS and quasi-horizontally polarized modes qSH . Quasi-horizontally polarized modes qSH have
occurred in both clearly separated propagation modes, that is, in the symmetric as well as in the anti-symmetric
propagation mode. In the polar diagrams of phase velocity in one layer, it has clearly been shown that extreme values
occur in situations when direction of propagation of Lamb’s waves matches material direction or when it is
perpendicular to material direction. With laminates consisting of material reinforced by one family of fibers, it has
been demonstrated that extreme values in the phase velocity diagram do not necessarily belong to the same direction
in the polar diagram. It has been noticed that they lie on the directions of polar symmetry. Moreover, it can be
observed that there are points on the diagram where switching from one mode into another can be clearly seen.
Those are not necessarily the same points, but every two neighboring diagrams, apart from the diagrams of the
fundamental and the first higher order, have points where the curves which describe phase velocities approach
mutually. Those points also belong to the direction of polar symmetry.
Based on everything aforesaid it can be easily concluded that using second-order polynomial theories gives the
possibility to analyze the propagation of Lamb’s waved in a fairly simple manner.
ACKNOWLEDGMENTS
Research presented in this paper was supported by Ministry of Education, Science and Technological Development
of Republic of Serbia, Grant TR32036, TR33015 and multidisciplinary project III 44007.
354
REFERENCES
[1] Green, W.A., Milosavljevic, D.: “Extensional Waves in Strongly Anisotropic Elastic Plates", Int. J. Solids &
Structures, Vol. 21, No. 4, 1985, pp 343-353,
[2] Green, W.A., Rogerson, G.A., Milosavljević, D.I.: "Transient Waves in Six-Ply and Eight-Ply Fibre Composite
Plates", Composites Science and Technology, Elsevier Science Publishers Ltd., Vol. 44, 1992, London and New
York, pp 151-158,
[3] Milosavljević, D., Bogdanović, G., Lazić, V., Aleksandrović, S., Lazić, M.: "Bulk waves and dynamical behaviour
in elastic solids reinforced by two families of strong fibres", Journal of Engineering Mathematics, 2014; DOI
10.1007/s10665-014-9747-9, ISSN 0022-0833, Published online 30 December 2014, Springer,
[4] Radaković, A., Bogdanović, G., Milosavljević, D., Veljović, Lj., Čukanović, D.: "Using high-order shear
deformation theory in the analysis of Lamb’s waves propagation in materials reinforced with two families of
fibers”, Acta Mech, Vol. 228, 2017, pp 187–200, DOI 10.1007/s00707-016-1707-1, ISSN: 0001-5970 (Print)
1619-6937,
[5] Whitney, J. M. Sun, C. T.: “A Higher Order Theory for Extensional Motion of Laminated Composites”, Journal of
Sound and Vibration, Vol. 30, No. 1, 1973, pp 85-97,
[6] Engblom, J., Ochoa, O.: "Through-the-Thickness Stress Predictions for Laminated Plates of Advanced
Composite Materials", International Journal for Numerical Methods in Engineering, Vol. 21, No. 10, 1985, pp
1759–1776,
[7] Kwon, Y.W., Akin, J.E.: "Analysis of Layered Composite Plates using a High-Order Deformation Theory",
Computers and Structures, Vol. 27, No. 5, 1987, pp 619–623,
[8] Fares, M.E.: "A Refined Two-Dimensional Theory of Anisotropic Plates, Mechanics Research Communications”,
Vol. 20, No. 4, 1993, pp 319–327,
[9] Tessler, A.: "An Improved Plate Theory for {1,2}-Order for Thick Composite Laminates, International Journal of
Solids and Structures, Vol. 30, No. 7, 1993, pp 981–1000,
[10] Akavci, S. S.: "Two new hyperbolic shear displacement models for orthotropic laminated composite plates“,
Mechanics of Composite Materials, Vol. 46, No. 2, 2010, pp 215-226,
[11] Wang, L., Yuan, F.G.: "Group velocity and characteristic wave curves of Lamb waves in composites: Modeling
and experiments", Composites Science and Technology, Vol. 67, 2007, pp 1370–1384.
355
356
Kragujevac, Serbia
MVM2018-047
Dragan Čukanović1
Aleksandar Radaković 2
THERMAL ANALYSIS OF FUNCTIONALLY GRADED

MATERIAL AS THE SPECIFIC CLASS OF ADVANCED
COMPOSITES USED IN AUTOMOTIVE INDUSTRY
ABSTRACT: Special emphasis is placed on the functional graded materials as advanced material in the family of
engineering composite materials which is made of two or more constituents with continuous variation of chemical
composition through a defined geometric distance. Development, various fabrication techniques as well as the
advantages and disadvantages of modern functional gradient materials compared to conventional composite
materials are described. Structural characteristics of functional gradient plates under different type of loadings are
analysed. This paper presents the method of mathematical idealizations of functionally graded material using the
high order shear deformation theory based on the shape functions. Mechanical and thermal properties of the
functionally graded material, except Poisson’s ratio, are assumed to vary continuously through the thickness of the
plate according to a power-law distribution of the metal and ceramic volume fractions. The effects of power-law index
and temperature gradient on mechanical responses of the plates are discussed and appropriate conclusions are
given.
KEYWORDS: functionally graded plate, power-law distribution, high order shear deformation theory, shape
functions
INTRODUCTION
The use of available materials in different organic and inorganic compounds opened up the path towards the
development of modern materials, advanced polymers, engineering alloys, structural ceramics etc. [1]. The
development of new materials occurred due to the need for materials which are able, in terms of properties, to meet
ever growing needs and requests of new technologies. New technologies often require materials which can perform
multiple functions, or materials possessing the qualities which none of already available materials possess. In those
cases, different materials often have to be combined, provided all the materials keep their own advantages.
Functionally graded materials (FGM) found their place among all those requests.
FGM are modern materials in the family of engineering composite materials, in which there is a continuous and
discontinuous variation of chemical composition through a defined geometric distance. After many years’ research
and development, FGM represent a class of attractive materials that can form a gradient property, which is impossible
to achieve with any other special homogeneous material. Today, it is possible to create FGM with exceptional
mechanical and technical properties that can respond to the most complex challenges of engineering. Mechanical
1
Dragan Čukanović, University of Priština, Faculty of Technical Sciences, Knjaza Miloša 7, 38220 Kosovska
Mitrovica, dragan.cukanovic@pr.ac.rs
2
Aleksandar Radaković, State University of Novi Pazar, Vuka Karadžića bb, 36300 Novi Pazar,
aradakovic@np.ac.rs
3
Gordana Bogdanović, University of Kragujevac, Faculty of Engineering, Sestre Janjić 6, 34000 Kragujevac,
gocab@kg.ac.rs
357
properties, such as: Young’s modulus of elasticity, Poisson’s ratio, shear modulus, thermal expansion coefficient, as
well as material density, are changed through a gradient in recommended directions.
The main aim of FGM is to use the properties of available materials in the best possible way by combining their
potentials. That includes reduction of plane stresses, transverse stresses through plate thickness, improving of
thermal properties, toughness, preventing delamination between two layers, which is the biggest and the most
frequently studied problem concerning conventional composite laminates (Figure 1).
Figure 1 Delamination of composite laminates
Most frequently used FGM is metal/ceramics, where ceramics have a good temperature resistance, fine antioxidant
properties, low thermal conductivity, while metal is superior in terms of mechanical strength, toughness and high
thermal conductivity. FGM, which contain metal and ceramic constituents, improve thermal-mechanical properties
between layers. Due to continuous change between properties of the constituents, delamination between layers
should be avoided (Figure 2). By varying a percentage of volume fraction content of metal/ceramic constituents, FGM
can be formed so that it achieves a desired gradient property in specific directions.
metal
ceramic
a) b)
Figure 2 Traditional composite laminate (a) compared to FGM (b)
The concept of FGMs is applicable to many fields which are subjected to high thermo-mechanical loadings such as
aerospace, chemical plants, electronics, optics, energy conversion, biomaterial etc. In the military industry, FGM is
used for the production of armoured vehicles (thermally resistant doors), armoured vest (materials of a gradient
structure resistant to punching) etc. In the automotive industry, the usage is found in braking systems (brake discs),
where are required FGM with properties such as resistance to high operating temperatures, friction, good thermal
conductivity. FGM as thermal barrier coatings (TBC) are utilized in diesel engines for trucks, buses, locomotive,
marine vehicles, tanks, military transport engines and farm vehicles. Their advantages in this application are
increased power, reduced heat loss and reduced fuel consumption. In addition, TBC have been shown to reduce
exhaust emissions. Thick (2.5 mm) TBC are used on piston crowns and thinner (0.5 mm) ones are used on valve
faces and cylinder heads. Experimental TBCs have been tested on cylinder liners, exhaust valve systems and valve
seats. It has been shown that a 5% reduction in fuel consumption is obtained by insulating the combustion chamber
with 2 mm thick functionally graded TBC [2].
Various fabrication methods of FGM have been developed which enable control of the composition as well as the
structure. The basic classification of the fabrication method of FGM is based on powder metallurgy or casting
metallurgy. If we are talking about powder metallurgy, there are hot and cold pressing, centrifugal mixed-powder
method, sedimentation by gravity force etc. Methods based on molten metal matrix are centrifugal casting and
centrifugal slurry method. Physical and chemistry vapour deposition technique are methods that usually used to
process FGM films with a gradient composition structure at the nanometre level.
Mathematical models are often used to describe certain problems in different engineering fields with the allowed
approximations. Analytical solutions of simple examples are a convenient model to understand the physical essence
of the problem. Based on that, more complex models are developed and commercial software. For the
aforementioned reasons, many authors dealing with problems in the field of FGM.
Thermoelastic behaviour of an FG rectangular ceramic-metal plate was presented by Praveen and Reddy [3] by
using a four-node rectangular finite element based on the first order shear deformation theory (FSDT), including von
Karman’s non-linear effect. This was one of the earliest studies on thermoelastic static and dynamic analysis of FG
plates under pressure and temperature change through plate thickness. Using and expanding the adopted
formulation by Praveen and Reddy, Reddy [4] studied the static analysis of FG rectangular plates using the third
358
order shear deformation theory (TSDT). Javaheri and Eslami derived stability and equilibrium equations for
rectangular FG plates under thermal load, by using classical plate theory (CPT) [5]. Matsunaga [6] presented a two-
dimensional high order shear deformation theory HSDT for thermal buckling of rectangular FG plates. Bodaghi and
Saidi [7] researched thermoelastic buckling behaviour of thick rectangular plates on the basis of Reddy’s theory, by
using the energy method. The authors compared stresses and displacements for ceramic, metal and FG plates and
they concluded that the displacements of the FG plate, even with a small ceramic volume fraction, were significantly
smaller than displacements of the metal plate. While a stress distribution for the isotropic metal or ceramic plate is a
linear function of a coordinate in the thickness direction of the plate, in FG plates it becomes non-linear due to unequal
change of properties through plate thickness. This confirms the fact that FG plates have better properties than plates
made of their constituents alone, namely metal and ceramics. Non-linear bending and buckling of FG plates were
studied by Shen and his associates. Especially Yang and Shen [8] studied non-linear bending of a thin FG rectangular
plate fixed on one pair of opposite edges including the variation of boundary conditions on the second pair of opposite
edges. They used von Karman’s non-linear theory including elastic foundation. The solution was obtained using
semianalytical technique, combining one-dimensional differential quadratic approximation and Galerkin method.
Loads can include transversal pressure, as well as plane pressure. The analysis was further developed by the same
authors [9] into an analysis of geometric nonlinearity of shear deformable plates under thermomechanical loads,
using Reddy’s TSDT. Huang, at al. [10] presented the solutions for 3D elastic theory of thick FG plates based on
Winkler-Pasternak elastic foundation. He assumed that the material properties varied exponentially through the
thickness of the plate. This paper researches the effect of foundation stiffness, loading cases and gradient index of
material properties on mechanical behaviour of plates. The main conclusion of the paper is that mechanical behaviour
of the plate with softer surface supported by elastic foundation differs significantly from the mechanical behaviour of
the plate with the harder surface supported by the same elastic foundation. This especially refers to thick plates.
MATHEMATICAL MODEL OF FGM

FGM rectangular plate of a  b  h dimensions on an elastic foundation where the x-y plane coincides with the
midsurface of the plate while z-axis has a direction of thickness h were studied in this paper (Figure 3). It is assumed
to be rested on a Winkler–Pasternak type elastic foundation, where:
k0 - stiffness of Winkler springs,
k1 - shear coefficient (Pasternak coefficient).
Figure 3 Geometry of a functionally graded plate resting on two parameters elastic foundation:
1- Winkler springs (k0), 2- shear layer (k1)
Young’s modulus of elasticity, thermal expansion coefficient and changes in temperature are defined according to
the power law [11]:
p
1 z
E  z   Em  Ecm    , Ecm  Ec  Em ,
2 h
p
1 z
  z    m   cm    ,  cm   c   m , (1)
2 h
s
1 z
T  z   Tm  Tcr    , Tcr  Tc  Tm ,
2 h
where indexes “c” and” m” refer to ceramics and metal, respectively. This law defines the change of the mechanical
properties as the function of the volume fraction of the FGM constituents in the thickness direction of the plate. As h
represents total thickness of the plate, E(z), (z), T(z) represents a material property in an arbitrary cross-section “z”.
Ec, c, Tc represents a material property at the top of the plate z=h/2 - ceramics, and Em, m, Tm represents a material
359
property at the bottom of the plate z=-h/2 - metal. In the equation (1), index p is the exponent of the equation which
defines the volume fraction of the constituents in FGM (Figure 4). Practically, by varying the index p, homogenous
plates of precisely determined and specific gradient structure could be obtained:
 when p = 0 the plate is homogenous, made of ceramics,
 when 0 < p < ∞ the plate has a gradient structure,
 theoretically, when p = ∞ the plate becomes homogenous again, made of metal, although the plate can be
considered homogenous even when p > 10.
Figure 4 Variation of the volume fraction through the thickness of a plate
Analytical procedure for determining critical buckling temperature for linear (s=1) and nonlinear change temperature
(s>1) across thickness of plate is here developed. It is assumed that the Poisson’s ratio  is constant.
EQUILIBRIUM AND STABILITY EQUATIONS

Disadvantages of the CPT and FSDT theory which require correctional factors, are eliminated by many authors by
introducing shear deformation shape functions. Many of these shape functions are introduced in order to give good
results for specific kinds of loads and specific static or dynamical problems. According to HSDT based on the shape
functions, displacements could be presented in the following way:
w
u  x, y , z, t   u0  x, y , t   z  x, y , t   f  z   x ,
x
w
v  x, y , z, t   v 0  x, y , t   z  x, y , t   f  z   y , (2)
y
w  x, y , z, t   w 0  x, y , t  ,
where f(z) is shape function.
In the reference literature there are many shape functions which can be polynomial, trigonometric, exponential,
hyperbolic etc. Some examples of the shape functions are given in Table 1.
Table 1 Shear deformation shape functions

Number of shape
Name of authors Shape function f(z)
function (SF)
SF 1 Levy, Stein, Touratier [12]  h /   sin  z / h 
SF 2 Soldatos [13] h sinh  z / h   z cosh 1/ 2 
SF 3 Akavci and Tanrikulu [14] 3 / 2 h tanh  z / h  3 / 2 z sec h2 1/ 2
z cos 1/ 2  h sin  z / h 
SF 4 Mechab at al. [15] 
1  cos 1/ 2  1  cos 1/ 2 
This paper proposes a new shape function as follows:
 z 
f  z   z  cosh    1.388  . (3)
  
h 
360
The introduced shape function is an odd function of the thickness coordinate z and satisfies zero stress conditions
for out of plane shear stresses. Observing the shape functions in the Table 1, may see that the proposed function
belongs to the group of simple mathematic functions. This fact makes the integration process easier and thus reduces
considerably the calculation time. Having in mind that the function is analytically integrable, there is no need to switch
to numeric integration, which additionally increases the precision of the obtained results.
In order to define components of unit loads, it is necessary to apply the relations between displacements and strains
in accordance with von Karman’s non-linear theory of elasticity [3]. Using a generalized Hooke’s law as well as
stiffness matrix [11] and taking into consideration the effect of the change in temperature (1) and thermal expansion
which cause a strain  T , the following components of unit loads are obtained:
 u0 1  w 2    2w 0 
      
 x 2  x   x 2 
N xx   xx   Q11 Q12 Q16     Q11 Q12 Q16   2
  h/2   h/2
   v 0 1  w 
2

h/2
   w 0 
Nyy     yy  dz   Q12 Q22 Q26       dz   Q12 Q22 Q26    zdz
  h / 2   y 2  y  y 2 
h / 2
Q16 Q66    h / 2
Q16 Q26 Q66  
 xy 
Q26
N xy   u v w w    2w 
 0
 0
  2 0

 y x x y   x y 
 
  x 
 
x
 Q11 Q12 Q16     Q11  Q12 
h/2
    y  h/2
Q  Q   z T z dz,
  Q12 Q22 Q26     
f z dz    12 22     
h / 2    y  h / 2

 0 

Q16 Q26 C66  
  
 x  y
 y x 
 u0 1  w 2    2w 0 
      
 x 2  x    x 2 
М xx   xx   Q11 Q12 Q16     Q11 Q12 Q16 
  h/2   h/2
Q  v 1  w 
2

h/2   2w 0  2
М
 yy    
 yy  zdz    12 Q22 Q26   0     zdz  
Q
 12 Q22 Q26    z dz
  h / 2   y 2  y  y 2 
h / 2
Q16 Q26 Q66    h / 2
Q16 Q66  
М xy   xy 
Q26
 u v 0 w w    2w 
 0
   2 0

 y x x y   x y 
  (4)
  x 
 
x
 Q11 Q12 Q16     Q11  Q12 
  y 
h/2 h/2
  Q12 Q22 
Q26    
 zf  z  dz   Q12  Q22  z  z T  z  dz,
y
h / 2
Q16 Q26 Q66    h / 2
 0 
   
 x

y

 y x 
 u0 1  w 2    2w 0 
      
 x 2  x   x 2 
Pxx   xx   Q11 Q12 Q16     Q11 Q12 Q16   2
  h/2   h/2
   v 0 1  w 
2

h/2
   w 0 
Pyy     yy  f  z  dz   Q12 Q22 Q26       f  z  dz   Q12 Q22 Q26    zf  z  dz
  h / 2   y 2  y  y 2 
 h / 2 
Q66    h / 2
Q16 Q66  
 xy  Q16 Q26 Q26
Pxy   u v 0 w w    2w 
 0
   2 0

 y x x y   x y 
 
  x 
 
x
 Q11 Q12 Q16     Q11  Q12 
h/2

   
h/2
  Q12 Q22   
  f  z   dz   Q12  Q22  f  z    z T  z  dz,
y 2
Q26  
y
h / 2
Q16 Q26 Q66    h / 2
 0 
   
 x  y

 y x 
361
Ry  h /2  xz  ' h /2
Q44  x 
Q45    '
     f  z  dz    
  f  z  dz. 
2

Rx   h /2  xz   h /2  45
Q  y 
Q55   
In the equation (2), by grouping the terms with the elements of constitutive matrix, it is possible to define new matrices:
 
h/2
 A ,B , D , E , F , G    Qij 1, z, f  z  , z 2 , zf  z  ,  f  z   dz, i , j = 1,2,6 ,

2
ij ij ij ij ij ij
h / 2
h
(5)
 Q f  z   l , r    4,5  .
2
Hlr  lr
'
dz,
h
In order to get an equilibrium equation, it is necessary to define the total potential energy in the following form [11]:
П  U  V, (6)
where strain energy is:
   
h/2
U xx  xx    z T  z     y  yy    z T  z     zz  zz   xy  xy   xz xz   yz yz dAdz (7)
h / 2 A
potential energy of elastic foundation is:
1    w 2  w 2  

V       dA
2
k w k   (8)
 x   x   
0 1
2 A 
Using the principles of minimum potential energy:
 П   U  V    U  V  0, (9)
equilibrium equations become:
 u0 : N xx ,x  N xy ,y  0,  v 0 : Nyy ,y  N xy ,x  0,
 w 0 : M xx ,xx  2M xy ,xy  M yy ,yy  N xx w 0,xx  2N xy w 0,xy  N yy w 0,yy  k 0w 0  k1 w 0,xx  w 0,yy   0, (10)
 x : Pxx,x  Pxy ,y  Rx  0,  y : Pxy ,x  Pyy ,y  Ry  0.
Stability equation for a thick FG plate is derived on the basis of the equilibrium equation (10). The stability equation
of the plate under thermal load can be defined using the displacement components u0 ,v 0 ,w0 ,x 0 and y 0 .
Displacement components of the next stable configuration are the following:
u  u0  u1, v  v 0  v1, w  w 0  w1,

(11)
 x   x 0   x1,  y   y 0   y 1,
where u1,v1,w1,x1 and y1 are the displacement components of arbitrarily small deviation from the stable
configuration. Assuming that temperature is constant in xy-plane of the plate and that it is changing only in the
thickness direction of the plate, the stability equation can be derived by substituting (11) and (4) into (10). In such
obtained equation, terms u0 ,v0 ,w0 ,x 0 и y 0 do not exist because they satisfy the equilibrium condition (10). Therefore,
the stability equations of the functionally graded rectangular plate are the following:
 u0 : Nxx
1
, x  N xy , y  0,
1
 v 0 : Nyy
1
, y  N xy , x  0,
1
 w 0 : M xx , xx  2M xy , xy  M yy , yy  N xx w 1, xx  2N xy w 1, xy  N yy w 1, yy  k 0w1  k1 w1, xx  w1, yy   0,

1 1 1 0 0 0
(12)
 x : P 1
xx , x P1
xy , y  R  0,
1
x  y : P 1
xy , x P1
yy , y  R  0,
1
y
0 0 0
where Nxx , Nyy and Nxy are the resultants of pre-buckling forces:
362
h/2
E  z    z T  z 
 Nyy    0.
0 0 0
N xx dz, N xy (13)
h / 2
1 
It is possible to solve (12) using analytical and numerical methods. In order to obtain analytical solutions, assumed
solution forms and boundary conditions are adopted in accordance to Navier’s methods applied at [16]. Procedure
for obtaining the results by combining symbolic and numerical coefficient values, which occur in these kinds of
problems, is implemented.
Boundary conditions along edges of the simply supported rectangular plate, according to [16], are the following:
v 0  w 0   y  Nxx  M xx  Pxx  0 along edges x  0 and x  a,

(14)
u0  w 0   x  Nyy  Myy  Pyy  0 along edges y  0 and y  b.
Taking into consideration the previously defined boundary conditions, based on [16], it is possible to assume that
Navier’s solution is in the following form:
 
m x n y  
m x n y
u1  x, y    Umn cos sin ,v1  x, y   Vmn sin cos ,
m 1 n 1 a b m 1 n 1 a b
 
m x n y  
m x n y
w1  x, y   Wmn sin sin ,  x1  x, y   Txmn cos sin , (15)
m 1 n 1 a b m 1 n 1 a b
 
m x n y
 y 1  x, y   Tymn sin cos ,
m 1 n 1 a b
where Umn ,Vmn ,Wmn ,Txmn ,Tymn are arbitrary parameters which are to be determined.
Furthermore, using Navier’s solution, equilibrium equation becomes:
L  IU  0, (16)
 
Т
where U  Umn Vmn Wmn Txmn Tymn and  buckling parameter. Coefficients Lij ,  i , j  1  5  are defined in the
following way:
L11   2 A11   2 A66 , L12    A12  A66  , L13  0, L14   2D11   2D66 ,
L15    D12  D16  , L22   2 A66   2 A22 , L23  0, L24    D12  D16  ,
L25   C66   C22 ,
2 2
L33   E11  2  E12  4  E66   4 E22 ,
4 2 2 2 2
(17)
L34   F11   F12  2 E66 ,
3 2 2
L35    F12  2  F66   F22 ,
2 2 3
L44  H 44   G11   G66 ,

2 2
L45   G12  G66  , L55  H55   2G66   2G22 ,
while the matrix Iij ,  i , j  1  5  is defined as:

0,
I 2 0
 
(18)
 Nx   Ny  k0  k1    ,  i , j  3 
2 0 2 2

where   m / a,   n / b . In order to obtain nontrivial solutions, it is necessary that the determinant in (16) is
equal to zero value:
L  I  0. (19)
NUMERICAL RESULTS
The aim of this section is to check the accuracy and the effectiveness of the given theory as well as new introduced
shape function in determining critical buckling temperature of FG plates for linear and nonlinear change temperature
across thickness. In order to do that, different numerical examples were done, and the obtained results were
compared with the results available in the literature. The theory presented in this paper is verified by the examples
of the square plate a/b=1, which was considered in [17] and [11]. Besides the results which are available in the
literature, this paper shows results for other of a/h (10 and 20) and a/b (1 and 2) ratios.
363
This chapter includes tables and diagrams containing results of the research. The numerical values of critical buckling
temperatures of FG plates on an elastic foundation are completely in accordance with the results given in the following
papers [17] and [11]. The authors of this research used material properties of the same material as in the previously
mentioned papers:
Table 1 Material characteristics of FGM constituents

Мaterial Material characteristics
Elasticity modulus Poisson’s ratio Material density Thermal expansion
E[GPa]  [kg / m3 ] coefficient
[C 1 ]
Aluminum ( Al ) Em  70   0.3 m  2702  m  23 106
Alumina ( Al2O3 ) Ec  380   0.3  c  3800 c  7.4 106
In case of gradient change of temperature, temperature at the metal surface is TM  5C .
Table 3 shows comparative results of critical buckling temperatures (∆tcr=∆Tcr.10-3) of square FGM plates resting on
an elastic foundation under a linear increase of temperature across their thickness. Thick and moderately thick square
plates were considered for different length to thickness ratios a/h, power law index p, Winkler and Pasternak
coefficient k0, k1. A comparative review of presented results with the results obtained using different shape functions
(Table 1) and results in [17] and [11] shows that the newly given shape function provides almost identical results.
Analysing results, we can conclude that critical buckling temperatures decrease with increase ratio a/h and power
law index p. Also, critical buckling temperatures increase when the plate rests on an elastic foundation where
Pasternak coefficient k1 has greater impact than Winkler coefficient k0.
Table 3 Comparison of critical buckling temperatures (∆tcr) of square FGM plates resting on an elastic foundation
under a linear increase of temperature across their thickness (a/b=1, m=n=1 and Tm=5°)
k0=0, k1=0 k0=10, k1=0 k0=10, k1=10
p Source
a/h=10 a/h=20 a/h=10 a/h=20 a/h=10 a/h=20
0 [17] 3.2276 0.8330 3.3154 0.8550 5.0479 1.2881
[11] 3.2275 0.8330 3.3152 0.8550 5.0477 1.2881
Present study 3.2274 0.8331 3.3151 0.8550 5.0476 1.2881
SF 1 3.2276 0.8330 3.3154 0.8550 5.0479 1.2881
SF 2 3.2273 0.8330 3.3151 0.8550 5.0476 1.2881
SF 3 3.2275 0.8330 3.3152 0.8550 5.0477 1.2881
SF 4 3.2273 0.8330 3.3151 0.8550 5.0476 1.2881
1 [17] 1.4130 0.3587 1.4897 0.3778 3.0040 0.7564
[11] 1.4130 0.3587 1.4897 0.3778 3.0039 0.7564
Present study 1.4130 0.3587 1.4897 0.3779 3.0039 0.7564
SF 1 1.4130 0.3587 1.4897 0.3778 3.0040 0.7564
SF 2 1.4129 0.3587 1.4896 0.3778 3.0039 0.7564
SF 3 1.4130 0.3587 1.4897 0.3778 3.0039 0.7564
SF 4 1.4129 0.3587 1.4896 0.3778 3.0039 0.7564
5 [17] 1.1600 0.2986 1.2576 0.3230 3.1839 0.8046
[11] 1.1602 0.2986 1.2578 0.3230 3.1840 0.8046
Present study 1.1608 0.2987 1.2584 0.3231 3.1846 0.8047
SF 1 1.1600 0.2986 1.2576 0.3230 3.1839 0.8046
SF 2 1.1607 0.2987 1.2583 0.3231 3.1846 0.8046
SF 3 1.1602 0.2986 1.2578 0.3230 3.1844 0.8046
SF 4 1.1606 0.2987 1.2582 0.3230 3.1844 0.8046
10 [17] 1.2183 0.3156 1.3317 0.3440 3.5699 0.9035
[11] 1.2184 0.3156 1.3318 0.3440 3.5699 0.9035
Present study 1.2187 0.3157 1.3321 0.3440 3.5703 0.9036
SF 1 1.2183 0.3156 1.3320 0.3440 3.5699 0.9035
SF 2 1.2186 0.3156 1.3320 0.3440 3.5702 0.9035
SF 3 1.2184 0.3156 1.3319 0.3440 3.5699 0.9035
SF 4 1.2186 0.3156 1.3319 0.3440 3.5701 0.9035
In Table 4 comparative results of critical buckling temperatures of square FGM plates resting on an elastic foundation
under a nonlinear increase of temperature across their thickness are shown. For fixed parameter s=3, different length
to thickness ratios a/h, power law index p, Winkler and Pasternak coefficient k0, k1 are varied. There was a good
match of the results for all shape functions with the results given in the references [17] and [11]. As in the Table 3,
the same conclusion can be made.
364
Table 4 Comparison of critical buckling temperatures (∆tcr) of square FGM plates resting on an elastic foundation
under a nonlinear increase of temperature across their thickness (a/b=1, m=n=1, s=3 and Tm=5°)
k0=0, k1=0 k0=10, k1=0 k0=10, k1=10
p Source
a/h=10 a/h=20 a/h=10 a/h=20 a/h=10 a/h=20
0 [17] 6.4552 1.6661 6.6308 1.7100 10.0958 2.5763
[11] 6.4550 1.6661 6.6305 1.7100 10.0956 2.5762
Present study 6.4547 1.6661 6.6303 1.7100 10.0953 2.5763
SF 1 6.4552 1.6661 6.6308 1.7100 10.0958 2.5763
SF 2 6.4547 1.6661 6.6302 1.7101 10.0953 2.5762
SF 3 6.4550 1.6661 6.6305 1.7102 10.0956 2.5762
SF 4 6.4547 1.6661 6.6302 1.7100 10.0953 2.5762
1 [17] 2.8269 0.7176 2.9804 0.7560 6.0097 1.5133
[11] 2.8268 0.7176 2.9803 0.7560 6.0096 1.5133
Present study 2.8268 0.7176 2.9802 0.7560 6.0096 1.5133
SF 1 2.8269 0.7176 2.9804 0.7560 6.0097 1.5133
SF 2 2.8267 0.7176 2.9802 0.7559 6.0095 1.5133
SF 3 2.8268 0.7176 2.9803 0.7560 6.0096 1.5133
SF 4 2.8267 0.7176 2.9802 0.7559 6.0095 1.5133
5 [17] 2.0152 0.5188 2.1847 0.5612 5.5309 1.3977
[11] 2.0154 0.5188 2.1849 0.5612 5.5311 1.3977
Present study 2.0165 0.5189 2.1860 0.5613 5.5322 1.3978
SF 1 2.0152 0.5188 2.1847 0.5611 5.5319 1.3977
SF 2 2.0164 0.5189 2.1859 0.5612 5.5321 1.3978
SF 3 2.0154 0.5188 2.1849 0.5612 5.5311 1.3977
SF 4 2.0161 0.5188 2.1856 0.5612 5.5318 1.3978
10 [17] 2.0971 0.5433 2.2923 0.5921 6.1448 1.5552
[11] 2.0972 0.5433 2.2924 0.5921 6.1449 1.5552
Present study 2.0978 0.5434 2.2929 0.5922 6.1455 1.5553
SF 1 2.0971 0.5433 2.2923 0.5921 6.1448 1.5552
SF 2 2.0977 0.5433 2.2928 0.5921 6.1454 1.5553
SF 3 2.0972 0.5433 2.2924 0.5921 6.1449 1.5552
SF 4 2.0975 0.5433 2.2927 0.5921 6.1452 1.5553
Figure 5a and Figure 6a shows a diagram of critical buckling temperature of FG plates under linear and nonlinear
temperature increase in the function of different aspect ratio a/b. Note that for linear increase of temperature
parameter s=1 is set and for nonlinear is s=3. When the ratio a/h=10 is fixed and the values of Winkler-Pasternak
coefficient are exactly specified, it can be noticed in the diagram that the power-law index p has the largest effect.
When the value p>=5, curves corresponding to shape functions SF1, SF2, SF3, SF4 start approaching one another
and finally they merge into one curve. It should be emphasized that the increase of a/b ratio causes the increase of
critical temperature, e.g. curves represent strictly increasing functions of a/b ratio. Also, it can be seen that the critical
buckling temperatures for the same example are larger for nonlinear than the linear increase of temperature across
their thickness.
Figure 5b and Figure 6b shows a diagram representation of critical buckling temperature under linear and nonlinear
temperature increase in the function of length to thickness ratio a/h. For the curves, obtained by using shape functions
SF1, SF2, SF3 and SF4, regardless of the shape function itself, there is an asymptotical fall of critical temperature
diagram. The diagram shows the values of p=0, 1, 5, 10 and it can be clearly seen that the only curve which stands
out is the one corresponding to the value p=0, while all the other curves completely overlap when ratio a/h>5.
The effect of the elastic foundation on critical buckling temperature of FG plates under linear and nonlinear
temperature increase is represented in the Figure 7. The curves are given for the variable value of power law index
p and different values of Winkler-Pasternak coefficients (k0, k1). The curves show that if there is no effect of the elastic
foundation, e.g. when the values are (k0, k1)=(0,0), critical buckling temperature increase with the increase of the
parameter p. However, if the effect of the elastic foundation is considered, that gives an extreme value in the diagram.
Namely, if p=2, there is a minimum value of the function of critical temperature, and after that the diagrams began to
go up and to asymptotically approach the critical value. This phenomenon is even more noticeable when both
coefficients that define the behaviour of the elastic foundation are considered. It should be emphasized that
representative shape function is SF2 as suggested by Soldatos, because it is analytically integrable, and it is not
necessary to use numerical integration. Considering more shape functions is extremely demanding in terms of
hardware and the calculations are very long. The behaviour of a function during integration affects the accuracy of
the given results.
365
а) a/h=10, k0=10, k1=10 b) a/b =1, k0=10, k1=10
Figure 5 Effect of the aspect ratio a/b и a/h on the critical buckling temperature ∆tcr under linear
increase of temperature
а) a/h=10, s=3, k0=10, k1=10 b) a/b =1, s=3, k0=10, k1=10

Figure 6 Effect of the aspect ratio a/b и a/h on the critical buckling temperature ∆tcr under nonlinear
increase of temperature
a) a/h=10, a/b=1 b) a/h=10, a/b=1, s=3

Figure 7 Effect of the power law index p on the critical buckling temperature ∆tcr under linear (a) and
nonlinear (b) increase of temperature
To evaluate the effect of each elastic foundation parameter, k0 and k1 individually, we carried out the numerical
calculations by varying one of them and setting the other as a constant. As in the previous cases, analysis is done
for linear and nonlinear increase of temperature across their thickness and results is shown in Figure 8a and Figure
9a (fixed value k1 and variation of k0) as well as in Figure 8b and Figure 9b (fixed value k0 and variation of k1). The
mentioned figures show that linear curves of critical buckling temperature change have a significantly faster increase
due to the change of coefficient k1 than when coefficient k0 is changed. Practically, it can be concluded that Pasternak
coefficient k1 has greater impact on critical buckling temperatures than Winkler coefficient k0. It can also be noticed
that the curve with the fastest increase is the one corresponding to the value p=10. Its fast increase is confirmed with
the fact that when k1=100, curves which correspond to p=0 and p=10 have an intersection point.
366
a) a/h=10, a/b=1, k1=10 b) a/h=10, a/b =1, k0=10
Figure 8. Effect of the Winkler coefficient k0 and Pasternak coefficient k1 on the critical buckling
temperature ∆tcr under linear increase of temperature
a) a/h=10, a/b=1, s=3, k1=10 b) a/h=10, a/b =1, k0=10

Figure 9. Effect of the Winkler coefficient k0 and Pasternak coefficient k1 on the critical buckling
temperature ∆tcr under nonlinear increase of temperature
CONCLUSIONS
Based on the results given in the tables and diagrams, it can be concluded that the shape functions given in the
Table 1 are acceptable for thermomechanical analysis of the functionally graded plates on an elastic foundation. It
should be emphasized that results obtained through comparative analysis of the given shape functions, have
mistakes within permissible tolerances.
The new shape function has been presented along with the comparative review of it with 4 different shape functions
which were primarily developed by different authors for the analysis of composite laminates but, in this paper, they
have been adjusted and implemented in appropriate relations for the analysis of FGM plates. Based on the obtained
results of the thermal buckling analysis of moderately thick and thick plates, it can be concluded that the newly
developed shape function could be applied in the analysis of FGM plates.
The comparative results obtained by the presented method can be summarized as follows:
 Critical buckling temperatures decrease with increase ratio a/h and power law index p. On other hand,
increasing of parameter s an aspect ratio a/b causes increasing of critical buckling temperatures. It is shown
that the power law index p has the largest effect on the increase of diagram curves. It is also shown that
they increase more slowly with larger p values, and they completely or partly overlap when p  5
 Critical buckling temperatures increase when the plate rests on an elastic foundation. The effect of the
elastic foundation show that critical buckling temperature is more affected by Pasternak coefficient k1 than
Winkler coefficient k0. It is also noticed the occurrence of the function minimum in the diagrams when
coefficients of the elastic foundation are considered. Without the effect of the elastic foundation, the
diagrams do not show any extreme values, but rather they show increasing or strictly increasing functions.
 Critical buckling temperatures for the same example are larger for nonlinear than the linear increase of
temperature across their thickness.
367
ACKNOWLEDGMENTS
Research presented in this paper was supported by Ministry of Education, Science and Technological Development
of Republic of Serbia, Grant TR32036, TR33015 and multidisciplinary project III 44007.
REFERENCES
[1] Jha, D.K., Kant, T., Singh, R.K.: "A critical review of recent research on functionally graded plates", Composite
Structures, Vol. 96, 2013, pp 833-849,
[2] Kohli, G.S., Singh, T.: "Review of funtionally graded materials“, Journal of Production Engineering, Vol. 18, No.
2, 2015, pp 1-4,
[3] Praveen, G. N., Reddy, J. N.: "Nonlinear transient thermoelastic analysis of functionlly graded ceramic–metal
plates", International Journal of Solids and Structures, Vol. 35, No. 33, 1998, pp 4457-4476,
[4] Reddy J. N., Wang, C. M., Kitipornchai, S.: "Axisymmetric bending of functionally graded circular and annular
plates", European Јournal of Mechanics A/Solids, Vol. 18, No.2, 1999, pp 185-199,
[5] Javaheri, R., Eslami, M. R.: "Thermal buckling of functionally graded plates“, AIAA Journals, Vol. 40, No. 1, 2002,
pp 162-169,
[6] Matsunaga, H.: "Thermal buckling of functionally graded plates according to a 2D higher-order deformation
theory", Composite Structures, Vol. 90, No. 1, 2009, pp 76-86,
[7] Bodaghi, M., Saidi, A. R.: "Thermoelastic buckling behavior of thick functionally graded rectangular plates",
Archive of Applied Mechanics, Vol. 18, No. 11, 2011, pp 1555-1572,
[8] Yang, J., Shen, H-S,: "Non-linear analysis of FGM plates under transverse and inplane loads", International
Journal of Non-Linear Mechanics, Vol. 38, No. 4, 2003, pp 467-482,
[9] Yang, J., Shen, H-S.: "Nonlinear bending analysis of shear deformable functionally graded plates subjected to
thermo-mechanical loads under various boundary conditions", Composites Part B: Engineering, Vol. 34, No. 2,
2003, pp 103-115,
[10] Huang, Z. Y., Lu, C. F., Chen, W. Q.: "Benchmark solutions for functionally graded thick plates resting on Winkler
Pasternak elastic foundations", Composite Structures, Vol. 85, No. 2, 2008, pp 95-104,
[11] Akavci, S. S.: "Thermal buckling analysis of functionally graded plates on an elastic foundation according to a
hyperbolic shear deformation theory", Mechanics of Composite Materials, Vol. 50, No. 2, 2014, pp 197-212,
[12] Stein, M.: „Nonlinear theory for plates and shells including the effects of transverse shearing“, AIAA Journals,
Vol. 24, No. 9, 1986, pp 1537-1544,
[13] Soldatos, K.: "A transverse shear deformation theory for homogeneous monoclinic plates", Acta Mechanica, Vol.
94, No. 3, 1992, pp 195-220,
[14] Akavci, S. S.: "Two new hyperbolic shear displacement models for orthotropic laminated composite plates",
Mechanics of Composite Materials, Vol. 46, No. 2, 2010, pp 215-226,
[15] Mechab, B., Mechab, I., Benaissa, S.: "Analysis of thick orthotropic laminated composite plates based on higher
order shear deformation theory by the new function under thermo-mechanical loading", Composites Part B:
Engineering, Vol. 43, No. 3, 2012, pp 1453-1458,
[16] Reddy, J. N.: "Mechanics of Laminated Composite Plates and Shells: Theory and Analysis", CRC Press LLC,
2004, New York, United State of America,
[17] Zenkour, A. M., Shoby, M.: "Thermal buckling of functionally graded plates resting on elastic foundations using
the trigonometric theory", Journal of Thermal Stresses, Vol. 34, No. 11, 2011, pp 1119-1138.
368
Kragujevac, Serbia
MVM2018-056
Dragan Ružić1
AN ANALYTICAL MODEL OF AIR-CONDITIONED MOTOR

VEHICLE CABIN
ABSTRACT: An air-conditioner used to control microclimate in a vehicle cabin works under specific conditions
and microclimate in the vehicle cabin has many differences compared to buildings. In an air-conditioned cabin it is
necessary to obtain good air quality and comfortable environment keeping the power consumption as low as
possible in the same time. In this paper typical features of automotive air-conditioner are modelled and analyzed
using conventional and empirical equations of thermodynamics. Qualitative and quantitative analysis of motor
vehicle cabin's air-conditioning system were carried out. A thermal model of a driver and occupants are included in
the model. The results have good matching with empirical values, making this approach suitable for fast theoretical
analysis of automotive air-conditioning system as well as for educational purpose.
KEYWORDS: motor vehicle cabin, thermal comfort, air-conditioning
INTRODUCTION
An air-conditioner (AC) of motor vehicle cabin is the largest power consumer, excluding the vehicle propulsion
system [10], [12], [16]. The goal of the vehicle's air-conditioning system is to as soon as possible make thermal
state inside the heated cabin comfortable and to maintain the comfortable conditions during the drive. Measures
that can be taken in order to reduce the power consumption are increase of the refrigeration system efficiency,
reduction of the cabin heat gain and efficient use of the air-conditioned air [10], [14], [16]. Apart from experimental
researches of AC performances, many numerical and analytical methods are applicable. The purpose of the
researches is manly to investigate influences of design parameters and operation modes on performance of an air-
conditioning system. The goal is to optimize energy consumption and to obtain thermal comfort and good air quality
for passengers.
The air-conditioned vehicle cabin can be analyzed as an integrated system (AC and a cabin with or without
occupants), or each subsystem can be modelled separately. In analysis of possibilities of reducing vehicle auxiliary
loads, in the project carried out by National Renewable Energy Laboratory (NREL), equation-based modelling was
also used for modelling of thermal processes in air-conditioned vehicle cabin [12]. All the components from the
HVAC system and the thermal envelope are broken down into submodels. Each submodel is described by one or
more equations. In combination with experimental data, fast and reliable simulation model was obtained. More
detailed generic model of entire system is developed in work of Arici et al. [1], using the principles of
thermodynamics. The methodology was validated by comparing the simulation results with the experimental
results. The model allows testing the effects of design, component, and structural changes on the thermal
conditions of the passenger compartment. Ružić [16] presented an analytical model of air-conditioned tractor cab
with proposal of partial air recirculation. It was concluded that the partial air recirculation contribute to reduction of
power consumption for air-conditioning, maintaining the air quality within the required limits. Fayazbakhsh and
Bahrami developed the heat balance method for estimating the heating and cooling loads encountered in a vehicle
cabin [4]. Using a lumped-body approach for the cabin, their model is capable of estimating the thermal loads for
the simulation period in a real-time. In the paper of Kamar et al. [8], results of a parametric study performed on a
passenger car AC in a semi-empirical computer simulation program CARSIM are presented. The effects of the
supply air volumetric flow rate, number of occupants, vehicle speed, and the fractional ventilation air intake, on the
1
PhD Dragan Ružić, assistant professor, University of Novi Sad, Faculty of Technical Sciences, Novi Sad, Trg
Dositeja Obradovića 6, ruzic@uns.ac.rs
369
dry-bulb temperature and relative humidity of the air inside the cabin, and on the evaporator cooling load were
assessed. Modelling of an air-conditioner using general freeware simulation tool CoolPack is presented in work of
Ružić and Sovilj [15]. The results obtained in the simulation tool have good matching with empirical values, making
this software suitable for fast theoretical analysis of automotive air-conditioning system.
This paper deals with qualitative and quantitative analysis of different AC operation modes in a vehicle cabin which
reflects on driver's thermal comfort as well as on AC power consumption. The method used in this research is an
analytical method of modelling thermodynamic processes in an air-conditioned cabin. The problem is treated as a
quasi steady-state lumped system. The reason is the fact that although temperature and velocity fields in the cabin
space are not homogenous, strong mixing of the air caused by ventilation creates very uniform distribution of air
temperature field, which is evident during the cool-down process [17]. The equations of heat and mass transfer and
perfect gas relationships for moist air are solved in MS Excel. The results are moist air states at various segments
of the air-conditioned cabin model as well as other relevant parameters of the system, depending on chosen input
variables.
AIR-CONDITIONED VEHICLE CABIN
Operation modes of an air-conditioner

The operation modes of AC that were considered and analysed are listed in Table 1. The table contains qualitative
estimation of thermal conditions and air quality inside the cabin. An amount of fresh air that is entering the cabin
influences both the air quality and the energy consumption. An air recirculation is typical solution for reduction of
energy consumption in cabin of passenger vehicles, as well as for prevention of ingress of outside contaminants. It
is known that 100% recirculation is the most efficient setting in terms of thermal conditions and particle
concentration, but this could cause elevated CO2 level in the cabin. In transient processes there are three
characteristic cases: closed cabin heated on the sun (hot-soak), beginning of cool-down process and steady-state
conditions when thermal comfort is achieved.
Table 1 The AC modes of operation and estimated influence on the interior air quality
Airflow mode AC Thermal conditions Air quality Comment
recirculation, thermal conditions outside low ingress of
no not applicable
REC = 100% comfort and safety limits contaminants,
recirculation, possible to achieve thermal interior contaminants, allowed for short
yes
REC = 100% comfort moisture and CO2 high time
possible to achieve thermal possible ingress of
outside air,
no comfort under certain contaminants conventional system,
REC = 0%
conditions interior contaminants, filtration of outside
outside air, possible to achieve thermal moisture and CO2 low air is mandatory
yes
REC = 0% comfort
Thermal processes in an air-conditioned vehicle cabin

Analytical model of an air-conditioned vehicle cabin is mathematical model that have a closed form solution. The
solution to the equations used to describe changes in a system can be expressed as a mathematical analytic
function. The model corresponds to the model of air-conditioned room with an external heat gain and an internal
heat and moisture source, i.e. the driver and passengers (Figure 1). The air inside the cabin is assumed to be well
mixed thanks to high air velocities produced by ventilation system, allowing the modelling of the cabin as a lumped
body. The governing equations are conventional equations of processes in moist air and heat and mass transfer
applied on an air-conditioned cabin.
Boundary conditions in a vehicle cabin are in close relation to outside conditions (outside air temperature, air
velocity, intensity and direction of the solar radiation, the vehicle speed). The sum of heat gains for a closed vehicle
cabin in a hot environment, under the steady-state conditions consist of (Figure 1) [5], [14]:
 heat transfer between the environment and the cabin interior air due to the temperature difference Qamb,
 heat gain caused by the solar radiation Qsol,
 heat introduced with the outside air Qair,
 sensible and latent heat released by the driver and passengers QH.
Since only quasi steady-state processes are being analyzed, thermal mass of the cabin materials is neglected in
the model. However, it is possible to extend the model and changes in interior air temperature could be calculated.
370
Figure 1 Schematic presentation of an air-conditioned vehicle cabin with a driver
Heat transfer through the cabin walls
The heat transfer between the ambient and the cabin interior air due to the air temperature difference can be
calculated using:
Qamb = U ·A (taout ˗ tacab) (1)
where
U - total heat transfer coefficient of cabin walls, W/m2K
A - surface area of cabin walls, m2
taout - outside air temperature, °C
tacab - average inside air temperature, °C
Total heat transfer coefficient of cabin's walls is based on empiirical values which are determined experimentally on
motor vehicles [5].
Heat gain due to air infiltration
Due to difference between the enthalpy of outside and inside air, there is a (positive or negative) heat transfer
caused by the air infiltration:
Qair = mout (hout ˗ hcab) (2)
where
mout - outside air mass flow rate, kg/s
hout - outside air enthalpy, kJ/kg
hcab - inside air enthalpy, kJ/kg
Ventilation systems of a motor vehicle cabin generally have an air mass flow rate of around 0.1 kg/s (300 m3/h) and
higher [15].
General expression for moist air enthalpy in dependence of air temperature t, °C, and humidity ratio w, kg/kg, is [2]:
h = 1.006 t + w (2501 + 1.805 t) (3)
Heat gain caused by solar radiation
Solar irradiation on a vehicle cabin surface is variable depending on the sun position and on the orientation of the
cabin surfaces. The total amount of the heat transmitted through the glass caused by solar radiation depends on
the glass solar properties and it is related to the normal projection of the vehicle cabin surfaces in the direction of
the radiation. The absorbed part of the solar irradiation heats the cabin’s outer opaque surfaces (the roof, for
example). The maximum solar irradiation on a horizontal outer surface in the central Europe region on a summer
day may exceed 900 W/m2 [14]. After some exposure to solar radiation, the thermal balance between heat gain
and heat release will be achieved. The heat from the outer surface is released into the environment (mostly to the
sky) by longwave radiation, and into the surrounding air by convection. From the inner surface, the heat from the
heated wall of the cabin is realeased into the air by convection and it is emitted by longwave radiation.
371
Solar absorptivity is also dependent on the colour of the surface. In terms of colour, light colours are preferable.
Under the same conditions, but with a dark coloured roof with solar absorptivity of 0.9, the heat transfer rate
increases three times in comparison with the roof that has the solar absorptivity of 0.3. Also, the temperature of the
roof interior will be 18...30 °C higher, while the radiant heat flux emitted by the roof’s inner surface would in this
case rise more than 50%.
The heat release from the human body
The heat release from the driver is calculated using the equations based on human thermal balance and human
physiology [6], [7], [11]:
QH = QM + QWH = M·ADu + ADu (3,05 10-3(256 tsk - 3373 - pw) + 0,42 (M - 58,15)) (4)
where
QM - sensible heat release
M - the specific metabolic activity, in range from 85 to 110 W/m2 for a driver [7]
ADu - outer surface area of the human body: ADu = 0,202 mb0,425 hb0,725; variables mb and hb are body mass in kg
and body height in m, respectively.
QWH - latent heat release, kJ/kg, depends on skin temperature tsk, metabolic activity level M and partial vapour
pressure pw in Pa [11]
Average skin temperature tsk under the comfort conditions depends upon activity level [11]:
tsk = 35.7 − 0.0275 M (5)
Moisture released by the driver mwH in kg/s can be calculated using:
mwH = QWH ·ADu /hwH (6)
where
hwH - enthalpy of saturated water vapour at skin temperature, kJ/kg.
Model of the AC system
The problem of the cabin air conditioning is determined by the air quantity to be supplied and the air state
necessary to remove given amounts of energy from the cabin at a specified condition of the interior air. The
equations of heat and mass balance in the cabin are [2]:
ma ·h2 + Qtot = ma ·hcab (7)
ma ·w2 + mwH = ma ·wcab (8)
Sum of heat gains Qtot, kW, is equal to:
Qtot = Qamb + Qsol + QH + Qair (9)
The other variables are:

ma - total air mass flow through the evaporator, kg/s, equation (12),
h2 - enthalpy of the moist air leaving the evaporator, kJ/kg,
w2 - humidity ratio of the air leaving the evaporator, kg/kg,
wcab - humidity ratio of the inside air, kg/kg.
In order to achieve steady-state thermal conditions with specified condition of the interior air at given state of
ambient, total air mass flow ma and evaporator cooling load Qevap can be calculated as follows [2]:
ma ·h1 = ma ·h2 + Qevap + mw ·hw2 (10)
ma ·w1 = ma ·w2 + mw (11)
where
mw - moisture removal at the evaporator, kg/s.
If a partial air recirculation takes place with an air mass flow mrec, kg/s, resulting properties of the mixture of the
outside air and the interior air can be calculated using following relations:
372
ma = mout + mrec (12)
mout ·hout + mrec ·hcab = ma ·h1 (13)
mout ·wout + mrec ·wcab = ma · w1 (14)
A recirculation rate REC is in the range from 0 to 100%:
REC = mrec / ma , % (15)
MODEL OF HUMAN THERMAL SENSATION

Value of air temperature inside the cabin is input for mathematical model of the thermal sensation. Human thermal
sensation was calculated using Fanger's equation for PMV (Predicted Mean Vote) calculation [6], [7]. Disadvantage
of the equation for PMV calculation is that it is intended for moderate microclimate conditions near thermal comfort
and does not include effect of sweating due to heat gain but only caused by increased physical activity.
PMV = (0.303  e-0.036M + 0.028) {M - 3.05  10-3 [5733 - 6.99 M - pw] - 0.42 [M - 58.15] -
(16)
1.7  10-5 M (5867 - pw) - 0.0014 M (34 - ta) - 3.96  10-8 fcl [(tcl + 273)4 - (tmr + 273)4] - fcl hc(tcl - ta)}
Clothing surface temperature tcl must be calculated by iteration [6]:
tcl = 35.7 - 0.028 (M - W) - Icl {3.96  10-8 fcl [(tcl + 273)4 - (tmr + 273)4] + fcl hc (tcl - ta)} (17)
Mean radiant temperature tmr is estimated based on the air temperature using empirical formula [9]:
tmr = -0.0078 ta2 + 1.626 ta - 6.880 (18)
Convective heat transfer coefficient hc , W/m2K, is depending on the air velocity va [6]:
hc = 12.1 va0.5 (19)
The air velocity va is estimated according to empirical values [13], [17]. The air velocity is depending on ventilation
air flow rate, but on air distribution and air flow direction too. Therefore, the air velocity is varied in the range from
0.2 to 0.7 m/s, independently of air mass flow in the ventilation system.
Relation between the clothing insulation Icl (Met) and a non-dimensional clothing surface area factor fcl is defined as
follows [6]:
fcl = 1 + 1.290 Icl (20)
Driver's thermal comfort inside the cabin (PMV = 0...0.7) can be achieved for a range of temperatures, air velocities
and relative humidity (Table 2) [6]. It can be noted that relative humidity in the range RHcab = 30...50% does not
affect the PMV value calculated by the Eq. (16).
Table 2 Microclimatic parameters inside the cabin and predicted thermal comfort
tacab, °C 23.0 25.0 27.0
tmrcab, °C 26.4 28.9 31.3
va, m/s 0.22 0.40 0.70
PMV 0.00 0.31 0.70
CALCULATION OF THERMAL PROCESSES

In order to solve the equations listed in the previous sections, equations of heat and mass transfer and perfect gas
relationships for moist air are written in MS Excel. The worksheets present a custom calculator able to determine
moist air states at various points of the air-conditioned cabin model and other system parameters. The calculator
offers a fast way for analysis of a heat exchange at the evaporator in dependency of air mass flow, cabin
characteristics, outer and interior conditions and recirculation rate. Tabular form is very convenient for result
comparison overview.
373
The boundary conditions and inputs for the calculation of thermal processes are:
 the ambient air temperature taout = 35 °C, relative humidity RHout = 30% [3];
 estimated solar heating of the cabin: Qsol = 500 W [14];
 the cabin thermal insulation: U · A = 60 W/K [5];
 sensible heat release from the driver QH = 170 W (metabolic activity M = 85 W/m2 = 1.5 Met [7], outer
body surface area ADu = 2.0 m2);
 driver's clothes insulation including the seat insulation Icl = 0.6 Clo;
 interior steady-state conditions that results with thermal sensation in the range PMV = 0...0.7 (Table 2).
The limitation of the system are maximum cooling power Qevap = 5 kW, maximum total air volume flow rate
ma = 320 m3/h, and minimum air temperature after the evaporator ta2 = 6 °C, as values that correspond to typical
AC of passenger car [5].
Hot-soak condition
The air temperature inside the cabin heated by the solar radiation (HS: hot-soak process, Table 3) is calculated on
the basis of thermal balance between the heat gain and heat loss (i.e. Qtot = 0). The AC is turned off (Qevap = 0) and
it is assumed that there is nobody in the cabin (QH = 0). This approach was verified using the available
experimental data, having a good agreement with outside and inside temperature levels [17].
Qamb = Qsol (21)
Initial average inside air temperature will be (Table 3):
tacab = taout + Qsol /(U ·A) = 43.3°C (22)
Humidity ratio wcab is calculated assuming that vapour partial pressure is to be equal both inside and outside the
cabin. The relative humidity is RH = 19%. From Eq. (21) it can be seen that reflection of the solar radiation is a
direct measure for reduction of air and surface temperatures of a closed cabin. Obviously, the solar absorptivity
and longwave emissivity of the roof material, as well as thermal conductivity, are very important design factors.
Therefore, the outer roof surface should have a low solar absorptivity coefficient and high emissivity, with low
thermal conductivity at the same time. However, in order to reduce thermal radiation towards the operator's body
surface, the inner roof surface should have low thermal emissivity. Unfortunately, this is not the case with the
materials usually used for the underside of the roof [14].
Steady-state conditions
Under the steady-state thermal conditions, the cooling power Qevap is equal to the total heat gain (Qevap = Qtot). The
cooling power in this model is generally controlled by the air mass flow through the evaporator ma. A required air
mass flow through the evaporator is calculated as:
ma = mout + mrec = Qtot / (h1 - h2 - hw2 (w1 - w2)) (23)
Qtot, h1 and h2 are dependent on air mass flow and iterative calculation is necessary.
In the first set of the calculations the influence of the relative humidity inside the cabin was analysed. Two groups of
the simulations were considered, one with RHcab = 30% (cases 1 - 3, Table 3) and another with RHcab = 50% (cases
4 - 6, Table 3).
In the next set of the simulations (cases 7 - 12, Table 3), the evaporator temperature was set to ta2 = 8°C. The
cooling power was the same as in the cases with ta2 = 6°C, by increasing the air mass flow accordingly. State of the
air inside the cabin was kept constant.
Influence of the air recirculation was analysed in the case with the lowest cooling power, cases 13 and 14, Table 3.
374
Table 3 Thermal states of the air inside the cabin and AC operation modes
Case tacab, °C RHcab, % ma, kg/s (m3/h) REC, % ta2, °C Qevap, W Qair, W
HS 43.3 19% 0 0% - 0 0
CD 43.3 19% 320 (0.100) 0% >35 0 -850
1 23 30% 294 (0.092) 0% 6 4272 2828
2 25 30% 218 (0.068) 0% 6 3166 1841
3 27 30% 165 (0.052) 0% 6 2395 1191
4 23 50% 188 (0.059) 0% 6 2731 1287
5 25 50% 144 (0.045) 0% 6 2090 765
6 27 50% 111 (0.035) 0% 6 1615 411
7 23 30% 402 (0.126) 0% 8 5304 3859
8 25 30% 278 (0.087) 0% 8 3673 2348
9 27 30% 201 (0.063) 0% 8 2656 1452
10 23 50% 227 (0.071) 0% 8 2996 1551
11 25 50% 168 (0.053) 0% 8 2217 892
12 27 50% 127 (0.040) 0% 8 1672 467
13 27 50% 111 (0.035) 50% 6 1410 205
14 27 50% 111 (0.035) 100% 6 1204 0
At the beginning of the cooling down (CD, Table 3), the initial average inside air enthalpy is higher than outside air
enthalpy and the heat introduced with the outside air Qair, has a negative value. If the ventilation system is set on a
maximum air flow rate (ma = 320 m3/h), then Qair = -850 W, so the air infiltration contributes to the heat removal.
Solar powered ventilator could contribute to the reduction of the inside temperature of a parked car. If someone
enters the heated cabin, it is recommended to keep the outside air mass flow at maximum but without too much
draft and eyes irritation. Recommended maximum velocity in eyes zone is between 0.05 and 0.30 m/s with
turbulence intensity of up to 70% [6]. Calculated PMV will exceed the upper limit and this condition for a person
inside the cabin could be treated as a heat stress. As soon as the enthalpy of the inside air becomes lower than
enthalpy of outside air during the cooling-down process, a partial recirculation can be activated in order to decrease
power consumption.
Maximum AC cooling power determines the speed of cooling-down and it is higher than cooling power required for
maintaining the steady-state conditions. Apart from initial and ambient conditions (especially the solar radiation),
the cabin material mass and thermal characteristics (heat capacity) are the most influential factors for time required
to achieve comfortable conditions in the cabin.
Although a chosen range of relative humidity has a very low impact on PMV, there is a significant influence on
required cooling power. In terms of energy consumption, it is more preferable to have lower air temperature on the
evaporator because the higher air flow rate results with an increase of cooling power for the same cooling power.
The required cooling power in the case 7 would be higher than the rated AC power, therefore the inside air
conditions would not be achieved.
Partial or full recirculation results with reduction of required power under the same boundary conditions, as
expected. However, there are limitations in the air recirculation due to air quality (see Table 1). Partial recirculation
is not a feature of a conventional automotive HVAC system. Since the driver's breathing zone is within the reach of
airflow from the vents, the ventilation system should enable that only fresh air is directed to the breathing zone. In
this case, if the maximum cooling power is used, heat gain caused by outside air would present around 10% of
total cooling power and a proper local air quality could be obtained. The other part of total airflow, the recirculated
air, has to be used for cooling the driver's body and the cabin, especially the windows and windshield. Cooling the
glass surfaces will reduce thermal radiation to the driver. The model showed that under the steady-state conditions
the cooling power could be at least 13% lower in the cases with 50% partial recirculation (case 13) than in the case
when only outside air is used (case 6).
CONCLUSIONS
The analytical model of air-conditioned vehicle cabin which was treated as a steady-state lumped system is
presented in this paper. The model offers a fast way for analysis of a heat exchange at the evaporator in
dependency of air mass flow, cabin characteristics, ambient and interior conditions and recirculation rate. The
result showed in a quantitative way an influence of the inside air state required for thermal comfort on AC power
consumption, as well as a contribution of the air recirculation to the reduction of cooling power.
Structure of the model in Excel worksheet is suitable for upgrade to transient model or it could be combined with
refrigeration cycle model. However, heat capacity of the cabin and interior materials is not negligible and must be
375
take into account if transient analysis should be carried out. While the calculations of the moist air state or the
refrigerant processes can be carried out with a relatively high precision, assessment of thermal properties of cabin
materials and convective and radiative heat transfer processes is more difficult due to complexity of cabin design.
It can be concluded that despite of many advanced (and expensive) software for simulation of 3D or 4D fluid
dynamics and thermal processes, 1D and 0D models still have important role in research of air-conditioning
processes. Equation-based modelling is suitable for fast solution and should present a first step before more
detailed but time consuming finite element and finite volume methods are to be used.
ACKNOWLEDGMENTS
This research was done as a part of the project TR35041 – "Investigation of the safety of the vehicle as part of
cybernetic system: Driver-Vehicle-Environment", which was supported by the Ministry of Education, Science and
Technological Development of Republic of Serbia.
REFERENCES
[1] Arici, Ö., Yang, S-L., Huang, D., Öker, E.: "Computer Model for Automobile Climate Control System Simulation
and Application", Applied Thermodynamics, Vol. 2, No. 2, 1999, pp 59-68,
[2] ASHRAE, Fundamentals Handbook, 1997, Chapter 6: Psychometrics, American Society of Heating,
Refrigerating and Air Conditioning Engineers, Atlanta, USA,
[3] EN ISO 14505-2, 2008, Ergonomics of the thermal environment - Evaluation of the thermal environment in
vehicles. Part 2: Determination of equivalent temperature. International Organization for Standardization.
Geneva,
[4] Fayazbakhsh, M. A., Bahrami, M.: "Comprehensive Modeling of Vehicle Air Conditioning Loads Using Heat
Balance Method", SAE Paper 2013-01-1507, 2013,
[5] Grossmann, H.: "Pkw-klimatisierung", Springer-Verlag, 2010,
[6] ISO 7730, 1994, Moderate thermal environment- Determination of the PMV and PPD indices and specification
of the conditions for thermal comfort. International Organization for Standardization,
[7] ISO 8996, 2004, Ergonomics of the thermal environment - Determination of metabolic rate International
Organization for Standardization,
[8] Kamar, H. M., Kamsah, N., Senawi, M. Y.: "Computerized Simulation of Automotive Air-Conditioning System:
A Parametric Study", International Journal of Computer Science, Vol. 10, No. 1, 2013, pp 787-794,
[9] Kaynakli, O., Pulat, E., Kilic, M: Thermal comfort during heating and cooling periods in an automobile, Heat
and Mass Transfer, Springer-Verlag, 2004,
[10] Leighton, D., Ruth, J.: "Increasing EDV Range through Intelligent Cabin Air Handling Strategies", Annual
Progress Report, 2016, National Renewable Energy Laboratory, USA,
[11] Parsons, K.: "Human Thermal Environments: The Effects of Hot, Moderate and Cold Environments on Human
Health, Comfort and Performance", 2nd ed. Taylor & Francis, 2003,
[12] Rugh, J., Farrington R.: "Vehicle Ancillary Load Reduction Project" Close-Out Report, Technical Report
NREL/TP-540-42454, 2008, National Renewable Energy Laboratory, USA,
[13] Ružić, D., Časnji, F.: "Personalized ventilation concept in mobile machinery cab", MVM – International Journal
for Vehicle Mechanics, Engines and Transportation Systems, Vol. 37, No. 1, 2011, pp 9-22,
[14] Ružić, D., Časnji, F.: "Thermal interaction between a human body and a vehicle cabin", In book: S. Kazi (ed),
Heat Transfer Phenomena and Applications, 2012, InTech, Rijeka, pp 295-318,
[15] Ružić, D., Sovilj, V.: "Application of Refrigeration Simulation Tools in Study of Automotive Air-Conditioner",
Proceedings of 18th SIMTERM 2017, 2017, Sokobanja, pp 998-1004,
[16] Ružić, D.: "Airflow and thermal management in farm tractor cab", IOP Conference Series: Materials Science
and Engineering 393, IOP Publishing, 2018,
[17] Ružić, D.: "Influence of air-conditioning on thermal comfort in an passenger car", Master Thesis, University of
Novi Sad, Faculty of Technical Sciences, 2006, Novi Sad.
376
Kragujevac, Serbia
MVM2018-059
Boris Stojić1
Dragan Ružić2
Zoran Ivanović3
Evgenije Adžić4
Nikola Ilanković5
LIGHT HYBRID-ELECTRIC VEHICLE AS EDUCATIONAL

AND RESEARCH PLATFORM
ABSTRACT: At the Faculty of Technical Sciences in Novi Sad, there is ongoing student project of developing
and creating light hybrid-electric vehicle as platform for education and research. Main motivation for starting this
project is an improvement of education process through enabling students to have first-hand practical experience
of being involved not only in engineering design process as team members but also into all significant
accompanying activities such as organization, communication, logistics etc. In this paper, main design features of
student vehicle in its current shape are described. Aspects of project contribution to the education process are
considered. Potential of the project for scientific research is also noted.
KEYWORDS: project-based learning, automotive engineering, electric drive, hybrid drive, lightweight design,
student project
INTRODUCTION
Following actual trends of vehicle drivetrain electrification and hybridization, project of design and assembly of light
experimental hybrid-electric vehicle carried out by students was started at the Faculty of Technical Sciences,
University of Novi Sad, in collaboration of Department for Mechanization and Design Engineering and Department
of Power, Electronic and Telecommunication Engineering. This vehicle, besides being realized as a student project
as an example of project-based learning approach, is on the other hand conceived to represent platform for
education and research as well. Main project goal is enabling opportunity for students to gain some hands-on
experience in solving real engineering problems, as well as to study contemporary engineering solutions through
practical application of theoretical knowledge gained during study.
In this paper, vehicle will be described through its main characteristics and design parameters in its current stage of
development. Thereafter, educational and research potentials of this project will be discussed and analyzed.
1
Boris Stojić, Assistant professor, University of Novi Sad, Faculty of Technical Sciences, Novi Sad, Trg Dositeja
Obradovića 6, bstojic@uns.ac.rs
2
Dragan Ružić, Assistant professor, University of Novi Sad, Faculty of Technical Sciences, Novi Sad, Trg Dositeja
Obradovića 6, ruzic@uns.ac.rs
3
Zoran Ivanović, Assistant professor, University of Novi Sad, Faculty of Technical Sciences, Novi Sad, Trg Dositeja
Obradovića 6, zorani@uns.ac.rs
4
Evgenije Adžić, Assistant professor, University of Novi Sad, Faculty of Technical Sciences, Novi Sad, Trg Dositeja
Obradovića 6, evgenije@uns.ac.rs
5
Nikola Ilanković, Teaching associate, University of Novi Sad, Faculty of Technical Sciences, Novi Sad, Trg
Dositeja Obradovića 6, ilankovic@uns.ac.rs
377
BASIC CHARACTERISTICS AND DESIGN PARAMETERS OF THE STUDENT
VEHICLE
Vehicle design process was carried out taking into account general rules of automotive engineering (e.g. [4]), and
considering main features and requirements for hybrid-electric vehicles, e.g. [2, 3]. Practical aspects of project
realization are described on the internet page of the Department for Mechanization and Design Engineering [5].
Informal name of the vehicle given by the student team is “H.E.R.M.E.S.”, which in this case stands for “Hybrid
Educational Recuperative Moto-Electric Special vehicle”. It is conceived that this project will be under long term
development, so that numerous steps of design improvement and optimization may be carried out by next
generation of students, enlarging impact of the project. Subsequently, current stage of vehicle design is described.
Current appearance of the vehicle is shown in Figure 1.
Figure 1 Current appearance of the vehicle
Powertrain
Powertrain is realized as in-parallel hybrid drive, so that the front wheels are driven by two in-wheel electric motors,
and for the rear axle conventional mechanical transmission is provided with the IC-engine as power source.
Thereby front and rear drive can be used together or independently of each other.
IC-engine is one-cylinder, four-stroke 250 cm3-volume gasoline engine originating from all-terrain vehicle “Loncin”.
Fuel system is based on carburetor. Maximum power according to available data [6] is 12 kW and maximum torque
is 17 Nm. Mechanical gearbox is of the same origin, having 4 forward and one reverse gear. Shifting pattern of the
gears is sequential. Differential drive at the rear axle is from passenger vehicle “Zastava Yugo”, slightly modified to
accommodate chain drive by which the power is transmitted from the gearbox to the axle. Chain drive is custom
made for this application.
At the front axle there are two in-wheel 48V BLDC-electric motors, each one for left and right wheel. According to
available data, it was assessed that each electric motor can generate mechanical power of approximately 1.8 kW
[1]. Motors were produced in China. Manufacturer and exact specifications are not known. Power supply is
obtained through 4 12V semi-traction batteries connected in series, via motor controllers.
Appearance of front and rear axle with powertrain components are shown in Figures 2 and 3.
378
Figure 2 Front axle with in-wheel electric Figure 3 Rear axle with conventional
motors and batteries passenger-car final drive, driven by IC engine
Currently there is only the possibility of using either electric or IC-engine power mode, since conditions for
realization of true hybrid drive with simultaneous usage of both kinds of drive are still not fulfilled. In order to
provide for this, several design changes have to be made:
 Electrification of controls for accelerator, clutch and gearbox;
 Automation of clutch and gearbox operation;
 Provision of common controller for simultaneous operation of both drives.
Chassis
Structure of the vehicle is made from steel tubes of Ø 21.3 mm diameter, with 2.6 mm wall thickness. It was
completely developed and realized by the student team. Structure consists of front, middle and rear module that
are connected to each other by using bolt connections. Main idea of using this modular approach is to enlighten
future design modifications, above all regarding powertrain or suspension system. Appearance of the frame in early
stage of development is shown in Figure 4. It was subsequently modified in several stages, including shortening of
the wheelbase. Figure 5 shows 3D-model of the front module, with elements of steering and suspension systems.
Figure 4 Tubular frame consisiting of 3 modules in early stage of development
Suspension system on the front axle is McPherson. It was mostly taken from the small electric vehicle originating
from China, of unknown manufacturer, from which also front powertrain is used (integrated with other components
of the front axle). Steering system is made by combining parts from this donor vehicle and components of the
steering system of “Zastava Yugo” passenger car. 13” diameter wheels are used front.
For the rear axle, almost complete McPherson axle from “Zastava Yugo” is used, though it is originally placed at
the front end of the donor car. The only modification is substitution of steering rods by control arms with inner joints
of fixed position, in order to disable steering movements of rear-axle wheels. One of the future plans is, though, to
utilize this configuration by introducing electronically controlled rear-wheel steering. Rear wheels have diameter of
14”.
Braking system also consists of elements already present on both front and rear axle from donor vehicles,
and the master cylinder is also from “Zastava Yugo”.
379
Figure 5 Model of the front chassis module, with elements of steering and suspension systems
Dimensions and mass properties

Mass of the vehicle: 500 kg.
Mass distribution front/rear: approximately 50/50%.
Wheelbase: 2500 mm.
Future development plans

Using elements of the braking and suspension systems from other vehicles and applying them to new vehicle with
different mass and mass distribution means that braking force distribution and suspension stiffness and damping
properties can generally be expected to be far from their optimal values. Therefore, from the viewpoint of basic
rules of automotive design and vehicle dynamics, one of the most important engineering tasks in order to consider
first stage of vehicle development fully completed, is to redesign i.e. replace elements of these systems according
to fundamental engineering rules for this area.
Other vehicle improvements, modifications and upgrades foreseen to be carried out by following generations of the
students can be classified into two main categories:
 Improvement of design parameters (e.g. making dimensions more compact, reducing vehicle mass,
optimizing suspension setup, modifying powertrain etc.);
 Introduction of automotive electronic and mechatronic systems, installing components enabling their
function.
EDUCATIONAL ASPECT OF THE PROJECT

The motor vehicle is one of the most complex and technologically advanced industrial products. Consequently, the
design process of a modern vehicle encompasses a wide range of technical and technological aspects, with a
particular emphasis on the current trend of the increasing level of automation and growing level of technology in all
systems and assemblies of a motor vehicle. One of the key segments is the transition from conventional forms of
drive systems to alternative concepts, including electric and hybrid electric drives.
The trend of growing number and technological level of electronic and electrical systems present in modern
vehicles requires the improvement of the process of education of experts in various profiles in the field of
automotive engineering such as mechanical engineering, electrical engineering, information technology, etc. In the
framework of student education for various branches of automotive engineering, it is desirable to enable the
acquisition of a wider and more comprehensive insight into the processes and techniques of designing and
developing a motor vehicle or its elements as a final product. For the fundamental adoption and understanding of
certain engineering areas, processes and approaches, as well as the development of the necessary personal skills
within the education process, it is crucial that the theoretical knowledge acquired within the framework of
conventional academic courses is expanded and deepened through their application in practical work.
380
In the scope of the project it is provided that students design and develop certain elements or systems of the
vehicle on their own, while other vehicle components are chosen from available off-the-shelf solutions i.e. parts
from certain series-produced vehicles. Selection from the range of available elements is thereby made through
their evaluation based on design, functional, technological and economic parameters. The finishing process of a
motor vehicle as a final product implies the functional and structural integration of all these components into a
functional unity of the required characteristics, while adhering to the principles and forms used in the automotive
industry.
Learning through practical work and mutual cooperation of students, as well as through the possibility of consulting
with practitioners supported by the teaching staff involved in the project, improves learning outcomes and the
quality of knowledge acquired through participation in the project. An opportunity to participate in a project of
developing its own vehicle awakens the enthusiasm of students, a deeper interest in the engineering process and
motivation for working within the team to achieve a common goal, which in itself further contributes to the
educational goals of the project. Knowledge acquired through participation in the project exceeds the narrowly
specialized frameworks typical for conventional forms of academic courses, contributes to the holistic
understanding of the engineering process, and encourages the development of decision-making skills in a complex
engineering environment. In addition, students are in position to take into account parameters such as market
trends, regulatory requirements, economic and environmental indicators of products and technological processes,
which further contributes to their excellent preparation for taking part in real-world engineering jobs. Innovation
potential and creativity of the student can also be expected to be increased through such kind of work, through
incentive to develop own solutions and approach to problem solution process.
Beside all this, availability of vehicle that is relatively simple in design and with very accessible main components
and systems, on the other hand containing advanced technological solutions such as hybrid-electric drive, strongly
contributes to educational capabilities for academic subjects involved with automotive engineering for very different
educational branches.
POTENTIALS OF THE PROJECT FOR THE RESEARCH WORK

In addition to the main goals of the project, connected to education process and project-based learning, realization
of the project of light hybrid-electric vehicle development provides numerous opportunities for the research
activities. The vehicle can therefore be considered as an educational-research platform for monitoring and
implementation of contemporary and/or forthcoming technologies within the scientific-research work as well as
within the teaching process. Some examples of areas in which scientific research or engineering development
could be carried out are e.g.:
 Study and optimization of vibroacoustic properties of lightweight hybrid-electric vehicles;
 Study and optimization of on-board energy management approaches;
 Improving vehicle lateral dynamic response by using independent left/right drive torque distribution;
 Application of innovative engineering materials to improve vehicle vibroacoustic properties and/or
lightweight design;
 Practical testing of new approaches in engineering design processes, etc.
CONCLUSIONS
In this paper the student project of lightweight hybrid-electric vehicle design and development is introduced. Basic
properties of the vehicle in its current stage are described, and recommendations and expectations for the future
work are given. According to experience with this project gained so far, it was concluded that this project-based
learning activity gives significant contribution to engineering education process, and widens and improves learning
outcomes as well. It was also stated that this project provides numerous opportunities for the scientific and
research work.
ACKNOWLEDGMENTS
This paper was done as a part of the researches on the project TR35041 – "Investigation of the safety of the
vehicle as part of cybernetic system: Driver-Vehicle-Environment" and the project TR31046 " Improvement of the
quality of tractors and mobile systems with the aim of increasing competitiveness and preserving soil and
environment", supported by Serbian Ministry of Education, Science and Technological Development.
381
REFERENCES
[1] Gabrić, T.; “Projektovanje i ispitivanje sinhronog motora sa stalnim magnetima za primenu u električnim
vozilima”, MSc Thesis, 2017, Faculty of technical sciences Novi Sad,
[2] Fuhs, A.: “Hybrid Vehicles and the Future of Personal Transportation”, CRC Press Taylor & Francis Group,
2009,
[3] Hodkinson, R., Fenton, J.: “Lightweight Electric/Hybrid Vehicle Design”, Butterworth-Heinemann 2001,
[4] Weber, J.: "Automotive Development Processes”, Springer 2009,
[5] Departman za mehanizaciju i konstrukciono mašinstvo, Available from: http://mehanizacija.ftn.uns.ac.rs/
index.php/studentski-projekat-hibridnog-vozila-hermes/, Accessed 01.09.2018.,
[6] Lianmei Industrial, Available from: http://www.lianmeigroup.com/product/250CC-ATV-8-INCH.html, Accessed
27.08. 2018.
382
Kragujevac, Serbia
MVM2018-066
Saša Milojević1
Dragan Džunić2
Dragan Taranović3
Radivoje Pešić4
Slobodan Mitrović5
EXPERIMENTAL DETERMINATION OF TRIBOLOGICAL

CHARACTERISTICS OF COMPOSITE MATERIALS IN USE
FOR PARTS IN ALUMINUM AIR COMPRESSOR
(PISTON AND CYLINDER)
ABSTRACT: For reciprocating machines are applicable different materials for the piston skirt and piston rings as
well as for cylinder liner. The friction and wear of the piston group is highly dependent by the material and the
honing of cylinder liner, the coating characteristics, the piston ring dimensioning and tension. If using aluminum
alloys for producing of piston and cylinders in engines and compressors, the results are lower fuel consumption
and exhaust emission, firstly because of lower weight and mechanical losses, too. The problems are poor
tribological characteristics and lower strength of unprotected aluminum comparing with gray cast iron. In this paper,
the tribological properties of ferrous and graphite based reinforcements were analyzed and compared with
aluminum alloy as a base material for cylinder liner and piston skirt in air brake compressor. The ballonplate
CSM nanotribometer was used to carry out these tests under dry sliding conditions and constant sliding distance,
for three different values of sliding speed and normal loads.
KEYWORDS: air compressor, aluminum MMC, reinforcements, tribology
INTRODUCTION
Heavyduty vehicles, trucks and buses are responsible for about a quarter of carbon dioxide (CO2) emissions from
road transport in the EU and for some 6% of total EU emissions. Transport is the only major sector in the EU where
greenhouse gas emissions are still rising. The European Commission has therefore set out a strategy to curb (CO2)
emissions from these vehicles over the coming years. Emissions from transport could be reduced to more than
60% below 1990 levels by 2050 [1].
1
Mr Saša Milojević, assistant R&D. University of Kragujevac, Faculty of Engineering, Kragujevac, Sestre Janjić 6,
sasa.milojevic@kg.ac.rs
2
Dr Dragan Džunić, assistant prof. University of Kragujevac, Faculty of Engineering, Kragujevac, Sestre Janjić 6,
dzuna@kg.ac.rs
3
Dr Dragan Taranović, associate prof. University of Kragujevac, Faculty of Engineering, Kragujevac, Sestre Janjić
6, tara@kg.ac.rs
4
Dr Radivoje Pešić, prof. University of Kragujevac, Faculty of Engineering, Kragujevac, Sestre Janjić 6,
pesicr@kg.ac.rs
5
Dr Slobodan Mitrović, prof. University of Kragujevac, Faculty of Engineering, Kragujevac, Sestre Janjić 6,
boban@kg.ac.rs
383
Figure 1 Share of driving resistances and auxiliary power demand on the total fuel consumption of a semitrailer
truck Euro V
In city buses and trucks a lot of fuel energy is engaging for power of auxiliary units. Specifically, the fuel energy is
engaged for drive of periphery units on engine, as example, for the air compressor, the alternator, the steering
pump, the oil pump, the coolant pump, the fuel high pressure pump and the fuel delivery pump, as well as for air
conditioning (A/C) compressor. The share of auxiliaries on the total power consumption is especially high for city
buses (6.5%) due to (A/C) system and additional consumers of electricity and pressurized air, Figure 1 [1, 2].
Reasons for this are the higher air demand of the wheel brakes, fewer headwinds for the engine cooling or more
steering in curves i.e. the main influence factors on fuel consumption are the engine off heat or the wheel brakes
air demand. As a result, an increasing losses resulting in increase of fuel consumption, that is directly proportional
to emission of (CO2). Generally, city buses are associated with 4.4% of total (CO2) emissions [1]. City buses are
frequently purchased by public institutions and thus they are in the public eye, yet may be the focus of costcutting
measures [28].
In order to reduce exhaust emission, the vehicles equipped with a conventional internal combustion engines must
be optimized by lowering the mechanical losses, specifically internal friction of the mechanical parts in sliding
contacts. Potential actions to reduce friction in vehicles include the use of advanced coatings and surface texturing
technology on engine and transmission components, new lowviscosity and lowshear lubricants and additives,
and tire designs that reduce rolling friction [911].
In accordance with the above mentioned facts, we have realized the research in the field of optimal design of
reciprocating aluminum engines and compressors. Consequently, we investigated new option for increasing
strength and tribological properties of the tribo-system pistoncylinder liner [1215].
The result of researches is patented prototype of aluminum piston and cylinder whose contact surfaces are coated
or modified with reinforcements based on the tribomaterials [16, 17].
METHODS FOR MACHINING OF CYLINDER SURFACE WHICH IS IN SLIDING

CONTACT
As the mechanical efficiency of the aluminum internal combustion (IC) engines i.e. reciprocating machines is
strongly influenced by the tribological situation between piston, piston ring and cylinder surface, the properties of
the cylinder surface become particular significant. Specifically, because the tribological properties of pure aluminum
are poor compare with grey cast iron.
Generally, aluminum alloys have specific disadvantages in the form of higher coefficient of thermal expansion and
inadequate tribological (friction and wear) and mechanical (lower strength) properties.
From the second side, aluminum composites show high stiffness, strength and wear resistance. The elastic
modulus could reach values up to 220 GPa. Conventional aluminum alloys have the elastic modulus of
(7080 GPa). The elastic modulus could be changed by alloying elements, ceramic particles and fibers [1820].
High-silicon-alloys are made mostly via spray forming. For technical application silicon contents up to 36% are in
use for pistons, cylinder liners and other high temperature components. These technical aluminumsiliconalloys
show the elastic modulus of (85100 GPa) and good wear resistance. The machining behavior is good compared
to conventional cast aluminumsiliconalloys [18].
384
a) b)
c) d)
Figure 2 Photography of aluminum cylinder and pistons with tribological reinforcements
a) Cylinder with continual plates; b) Cylinder with discrete plugs in the form of spheres or nodule;
c) The piston with the interchangeable tribological inserts (in area of the piston skirt) and
d) The piston skirt reinforced with discrete plugs in the form of spheres or nodule
Thermal spray coatings are one of option to provide protection of aluminum alloy. Thermal spraying techniques are
coating processes in which melted (or heated) materials are sprayed onto a surface. Today, any makers utilize
thermal spraying as one of most predominant surface treatments, to making fully sprayed cylinder surface
consisting of (FeAl) composite material [20].
Looking from the second side, liquidtype lubricants have effectively served in reducing the friction and wear of
various mechanical devices. However, in compressor components, the liquid-type lubricants have negative effects
on their thermodynamic efficiencies, and also the state of lubrication in these components is usually not known and
is considered to be in the boundary and mixed lubrication regimes. Recently, research interest and efforts are on
oilless compressor conditions to eliminate the adverse effects of liquidtype lubricants and to further improve the
performance of compressors. Consequently, it becomes necessary to develop advanced coating materials that
exhibit lower friction and higher wear resistant under compressor specific conditions [2125].
New concept of aluminum cylinder with tribological reinforcements

Generally, according to real machining conditions, the full contact between piston rings and cylinder is not possible.
This fact leads us to the idea that by casting tribological inserts in the cylinder, we can determine in forward,
contact area between piston rings and cylinder liner.
With the aim to achieving strength as well as tribological characteristics similarly as in case of the application grey
cast iron, we patented the cylinder of composite material for reciprocating air compressor with the reinforcements
consisting of tribological materials, Figure 2 [20].
385
For the purposes of the experiment, internal surface of the aluminum cylinder as base materialmatrix, (alloy
EN AlSi10Mg), was modified by putting tribological reinforcements of cast iron that are arranged in the form of
continuous pads, the plates or like discrete tribological plugs in the form of spheres (nodule), or particles spherical
shape, as reinforcements [20].
By transferring the contact between the piston rings and cylinder made of aluminum on the tribological inserts, we
reduce the wear. This technology extends the service life of cylinder and piston rings.
This optimization can lead to reduce machine weight as well as reduced friction and wear. A reduction of friction
between piston rings and the cylinder running surface is particularly effective, because the majority of frictional
losses in the reciprocating machines are generated inside of this tribological system [9, 24].
EXPERIMENTAL RESEARCHES
Wear is progressive loss of material caused by friction resistance between contact surfaces. The present work
wants to investigate and evaluate the effect of tribological plugs as reinforcements on tribological behavior of
patented aluminum cylinder.
Tribological tests were carried out at CSM nanotribometer with ballonplate contact pair for different normal loads
and sliding speeds in dry conditions. Tribological tests are based on variation of three different normal loads
(0.3, 0.6 and 0.9) N and sliding speeds (3, 9 and 15) mm∙s1. Duration of each test was 500 cycles (distance of
1 m), acquisition rate 100 Hz [26, 27].
a) b)
Figure 3 Optical microscopy for: a) Base aluminum alloy; b) Cast iron reinforcement
Continuous wear monitoring during linear reciprocating sliding was investigated. Coefficient of friction (COF) and
penetration depth (PD) curves were obtained and analyzed. We continually monitor wear process by PD parameter
and we examine effects of changes in applied load and speed.
RESULTS OF TRIBOLOGICAL TESTS AND DISCUSSION
Optical microscopy analysis

Experiments were carried out with the base material for cylinders (aluminum alloy) and with the material for
reinforcements made of cast iron. Figure 3.a presents optical microscopy of base aluminum alloy surface, where
grey phases, which are noted on the surface, presents eutectic silicon. Figure 3.b presents surface of cast iron
nodular discrete pads (reinforcements). Deeper analysis of the presented surface revealed that black lines across
the surface are not micro cracks but graphite inclusions in the cast iron.
More detail analysis of prepared samples, both of aluminum alloy and cast iron inclusions were performed using
Phenom ProX EnergyDispersive Spectroscopy (EDS). EDS analysis results are presented on the Figure 4 and 5.
Figure 4 presents EDS analysis of aluminum alloy and it noticeable that are present three phases, white one iron
particles trapped in aluminum alloy as a result of grinding process. Lighter grey zones represent eutectic silicone,
which is constituent element of the aluminum alloy, while dark grey represent pure aluminum.
386
Figure 4 EDS analysis of base aluminum alloy
Figure 5 represents EDS analysis of cast iron inserts, and it is noticeable that only two phases exist. Black phase
represent graphite inclusions as is it is previously assumed based on optical microscopy. White phase represents
pure iron, although graphite is present, which can be seen on spectrum 2.
Higher percentage of graphite in this phase is, also, result of grinding and polishing process that smears graphite
inclusions all over the surface.
387
Figure 5 EDS analysis of cast iron inserts
Coefficient of friction and penetration depth

Changes of COF and PD (ordinate) during sliding under low load conditions, depending on time, distance and
numbers of cycles (abscissa) are shown in Figure 6.a for basic material (aluminum) and in Figure 6.b for
tribological reinforcement. Due to the reciprocating motion of the needle of CSM nanotribometer between two end
positions, the friction force was also changing direction during the test.
The values of COF for tribological reinforcement were ranging from (0.090.295), and these maximal values are
lower to the results of base material (0.0830.543). The mean value of COF for the reinforcements is also lower
(0.232) relative to the value for the base material (0.350). A steadystate value for COF was reached shortly after
the beginning of the test. PD in material of reinforcements has stable and constant values compared to base
material.
Under higher load conditions and same sliding speed, a lower maximum value of COF of the reinforcement was
also recorded, too, Figure 7.a and Figure 7.b. The obtained COF values range in the range (0.0870.262) for the
reinforcements, and (0.0760.327) for the base material. The mean value of COF of the reinforcements is lower
(0.176) than the value for base material (0.202).
It is noticeable that during wear testing of base material (aluminum alloy) COF sharply rises after a certain period of
sliding as result of material transfer on counter body steel ball. After the material was transferred on the steel ball
contact between transferred aluminum and aluminum as a base material was achieved, which result in increase of
COF value. This process happens regardless the applied load, and it has cyclic nature, which indicates that
transferred material on the counter body surface accumulates until it reaches critical size, that cannot bear
388
tangential loads. Penetration depth plot (Figure 6.a and Figure 7.a) confirms this assumption. After reaching critical
size, transferred material will fall off and process of transferring starts from the beginning.
a) b)
Figure 6 Filtered signal of coefficient of friction and penetration depth for:
a) Base material under (FN = 0,3 N; V = 15 mm ∙ s −1 ); b) Reinforcements under (FN = 0,3 N; V = 15 mm ∙ s −1 )
a) b)
Figure 7 Filtered signal of coefficient of friction and penetration depth for:
a) Base material under (FN = 0,9 N; V = 15 mm ∙ s −1 ); b) Reinforcements under (FN = 0,9 N; V = 15 mm ∙ s −1 )
In case of wear testing of cast iron inserts (reinforcements) no transfer material was observed, which is also
confirmed by COF and penetration depth curves regarding applied load value. Both measured values, COF and
penetration depth has almost constant averaged value during wear testing.
In addition to this conclusion, Figure 8 clearly indicates accumulated transferred material on the counter body steel
ball surface after sliding tests of base material (aluminum alloy) and no transferred material on counter bod surface
after sliding test of reinforcement (cast iron). To be sure that transfer of material took place, counter body steel ball
profile was examined, Figure 9.
a) b)
Figure 8 Optical microscopy of the counter body steel ball profile after sliding test:
a) Base material (aluminum alloy); b) Reinforcement (cast iron inserts)
389
a) b)
Figure 9 Optical microscopy of the counter body steel ball profile a) before and b) after sliding test
Figure 9.b clearly indicates accumulated transferred material on the counter body surface.
Wear analysis
Wear mechanisms based on examinations of worn surfaces by optical microscopy, were analyzed in comparison
with trends of PD curves. Figure 10.a and 10.b present surface micrographs of base material and reinforcements
after test under dry sliding, obtained for the same conditions of applied load and speed. It is noticeable that wear of
aluminum is higher than cast iron sample due to transfer material and increase of COF which occurs for aluminum.
Prior wear tests, mechanical characterization of the prepared samples have been performed. Hardness tests were
performed using CSM NHT2 nanoindenter, using Berkowich three sided diamond tip. Nine measurements for each
tested material were done and averaged hardness values for base material and for reinforcement are 90 and 318
Vickers, respectively. Divergence in hardness values of tested materials indicates in better wear resistance of
reinforcement in comparison to the base material.
Deep grooves in analyzed surfaces, indicates that the abrasive wear is the most dominant mechanism of wear for
both tested materials, although transfer material on the counter body should not be neglected, especially for
aluminum alloy.
Similar conclusions can also be made when testing materials at lower load and at the same sliding speed
(FN = 0,3 N; V = 15 mm ∙ s −1 ) [26].
a) b)
Figure 10 Surface micrographs after dry sliding test for:
a) Base material under (𝐹𝑁 = 0.9 N; 𝑉 = 15 mm ∙ s −1 ); b) Reinforcement under (FN = 0.9 N; V = 15 mm ∙ s −1 )
390
CONCLUSIONS
 In trucks and city buses a lot of fuel energy is engaging for power of auxiliary units of engine, as it is air
compressor inside braking system, reciprocating compressor for air conditioning system, pump of the
steering system, alternator etc.;
 The use of aluminum alloys for making parts of propulsion and mobile systems is important from the
aspect of weight reduction, which directly affects the reduction of fuel consumption. On the other hand, the
problems in application are lower strength and poor tribological characteristics of aluminum as well as
aluminum constructions;
 Some of the available ways to improve the tribological and mechanical properties of the aluminum made
parts are application of coatings with tribological materials, the texture surface making technology, the
application of lower viscosity oils and additives in lubrication oil;
 The problems of lower strength as well as the poor tribological characteristics of the cylinder for
application in reciprocating air compressor for brake system of bus have solved by application aluminum
matrix composite material. One of the constituents is aluminum alloy (EN AlSi10Mg) as the base material
or matrix. Second constituent is being inserted into base material and serves as reinforcement which is
made of cast iron in the form of particles of a spherical shape;
 Presented results of the researches, obtained during tests of materials from which consisting cylinder,
shows that by transferring the contact between the piston rings and cylinder made of aluminum on the
reinforcements, it is possible to reduce the friction and wear. This technology extends the service life of
cylinder and piston rings; and
 Based on the results of tribological tests, more favorable characteristics of the friction and wear of the
reinforcements in relation to the base material were confirmed. The obtained results are useful as
guidelines for further research and selection of the appropriate cylinder construction within the
optimization of the performance of the experimental piston compressor for the production of compressed
air on the heavyduty vehicles (on buses and trucks).
ACKNOWLEDGMENTS
This paper is a result of the researches within the project TR35041 and TR35021 financed by the Ministry of
Science and Technological development of the Republic of Serbia.
REFERENCES
[1] European Commission, Available from: https://ec.europa.eu/clima/policies/transport_en, Accessed 31.08.2018,
[2] Milojević, S., Pešić, R.: “CNG buses for clean and economical city transport”, Mobility & Vehicle Mechanics
(MVM), Vol. 37, No. 4, 2011, pp 5771,
[3] Milojević, S.: “Reconstruction of existing city buses on diesel fuel for drive on Hydrogen”, Applied Engineering
Letters, Vol. 1, No. 1, 2016, pp 1623,
[4] Milojevic, S.: “Optimization of the Hydrogen System for City Busses with Respect to the Traffic Safety”,
Proceedings of the 20th World Hydrogen Energy Conference WHEC 2014, Vol. 2, 2014, pp 853860,
[5] Milojević, S., Gročić, D., Dragojlović, D.: “CNG propulsion system for reducing noise of existing city buses”,
Journal of Applied Engineering Science, Vol. 14, No. 3, 2016, pp 377382,
[6] Milojević, S.: “Sustainable application of natural gas as engine fuel in city buses  Benefit and restrictions”,
Journal of Applied Engineering Science, Vol. 15, No. 1, 2017, pp 8188, 2017, doi: 10.5937/jaes1512268,
[7] Milojevic, S., Pesic, R.: “Theoretical and experimental analysis of a CNG cylinder rack connection to a bus
roof”, International Journal of Automotive Technology, Vol. 13, No. 3, 2012, pp 497503,
https://doi.org/10.1007/s12239-012-0047-y
[8] Milojević, S., Pešić, R.: “Challenges in City TransportAlternative Fuels and Door to Door Model”, Proceedings
of the 4th INTERNATIONAL CONFERENCE MECHANICAL ENGINEERING IN XXI CENTURY, 2018, Niš, pp
387392,
[9] Kennedy, M., Hoppe, S., Esser, J.: “Piston ring coating reduces gasoline engine friction”, MTZ
Motortechnische Zeitschrift, Vol. 73, No. 5, 2012, pp 4043,
[10] Skulić, A., Bukvić, M.: “Tribological properties of pistoncylinder set in internal combustion engines”, Applied
Engineering Letters, Vol. 1, No. 1, 2016, pp 2933,
[11] Mitrović, S., Babić, M., Stojanović, B., Miloradovi,ć N., Pantić, M., Džunić, D., “Tribological Potential of Hybrid
Composites Based on Zinc and Aluminium Alloys Reinforced with SiC and Graphite Particles”, Tribology in
Industry, Vol. 34, No. 4, 2012, pp 177185,
[12] Pešić, R., Milojević, S., Veinović, S.: “Benefits and Challenges of Variable Compression Ratio at Diesel
Engines”, Thermal Science, Vol. 14, No. 4, 2010, pp 10631073, https://doi.org/10.2298/TSCI1004063P
391
[13] Pesic, R., Milojevic, S.: “Efficiency and Ecological Characteristics of a VCR Diesel Engine”, International
Journal of Automotive Technology, Vol. 14, No. 5, 2013, pp 675681, https://doi.org/10.1007/s12239-013-
0073-4,
[14] Saša Milojević and Radivoje Pešić.: “Determination of Combustion Process Model Parameters in Diesel
Engine with Variable Compression Ratio,” Journal of Combustion, Vol. 2018, Article ID 5292837, 2018, 11
pages, https://doi.org/10.1155/2018/5292837,
[15] Milojević, S.: “Analyzing the Impact of Variable Compression Ratio on Combustion Process in Diesel Engines”,
Mr Thesis, 2005, University of Kragujevac Faculty of Mechanical Engineering,156 pages,
[16] Milojević, S., Pešić, R., Taranović, D.: “Tribological Principles of Constructing the Reciprocating Machines”,
Tribology in Industry, Vol. 37, No. 1, 2015, pp 1319, http://www.tribology.rs/journals/2015/2015-1/2.pdf,
[17] Milojevic, S., Pesic, R., Taranovic, D.: “Tribological optimization of reciprocating machines according to
improving performance”, Journal of the Balkan Tribological Association, Vol. 21, No. 3, 2015, pp 690699,
[18] Stojanović, B., Milojević, S.: “Characterization, Manufacturing and Application of Metal Matrix Composites. In:
Wythers MC (ed) Advances in Materials Science Research”, 30th edn. Nova Science Publishers, New York,
2017, pp 83133,
[19] Stojanović, B., Ivanović, L.: “ Application of aluminium hybrid composites in automotive industry”, Technical
Gazette, Vol. 22, No. 1, 2016, pp 247251, https://doi.org/10.17559/TV-20130905094303,
[20] Pešić, R.: “ASMATAAutomobile Steel Material Parts Substitution with Aluminum”, Mobility & Vehicle
Mechanics (MVM), Vol. 30, Special Edition, 2004, pp 116,
[21] Shanmughasundaram, P., Subramanian, R.: “Influence of graphite and machining parameters on the surface
roughness of Alfly ash/graphite hybrid composite: a Taguchi approach”, Journal of Mechanical Science and
Technology, Vol. 27, No. 8, 2013, pp 24452455,
[22] Vencl, A., Rac, A., Bobic, I.: “Tribological behaviour of Albased MMCs and their application in automotive
industry”, Tribology in Industry, Vol. 26, No. (34), 2005, pp 3138,
[23] Vencl, A., Rac, A.: “New wear resistant Al based materials and their application in automotive industry”,
Mobility & Vehicle Mechanics (MVM), vol. 30, Special Edition, 2004, pp 115139,
[24] Gand, B.: “Coating for cylinder surfaces in aluminium engine blocks”, MTZ Motortechnische Zeitschrift, Vol. 72,
No. 1, 2011, pp 3439,
[25] Solzak, T. A., Polycarpou, A. A.: “Tribology of Protective Hard Coatings for use in OilLess PistonType
Compressors”, available at: http://docs.lib.purdue.edu/icec/1790,
[26] Milojević, S., Džunić, D., Taranović, D., Pešić, R., Mitrović, S.: “Tribological Reinforcements for Cylinder Liner
of Aluminum  Example Compressors for Brake Systems of Trucks and Buses”, Proceedings of the 15th
International Conference on Tribology SERBIATRIB’17, 2017, Kragujevac, pp 251257,
[27] Milojević, S., Pešić, R.: “Influence ferrous based reinforcements on tribological parameters of aluminum
cylinder for piston compressor in brake system of bus”, Proceedings of the 2ND INTERNATIONAL
CONFERENCE ON TRIBOLOGY (TURKEYTRIB’18), 2018, Istanbul, pp 78.
392
Kragujevac, Serbia
MVM2018-067
Rodoljub Vujanac1
Nenad Miloradovic2
Milan Djordjevic3
DESIGN OF FORKLIFT SPEED LIMIT CONTROL

SYSTEM FOR WAREHOUSES
ABSTRACT: While the automobile industry of electrically powered cars is still in its infancy stage, the industry
vehicles, such as forklift trucks and other material handling equipment have been in production for several
decades. The new generation forklifts are equipped with modern control and security systems that make them
easier to use. They are designed to operate at low speeds which are not critical in terms of occupational safety. On
the other hand, material handling involves extreme driving manoeuvres, including high steering and fast
acceleration and braking. The older generations of forklifts that do not contain modern control systems are still in
use. These forklifts are involved in many incidents due to their lower stability combined with uncontrolled traffic
areas and speed at workplaces. Also, there are often people injures and damage to buildings, storage systems or
stock. This paper describes a speed limit control system that prevents one of the main causes of accidents -
speeding. The developed system is also related to signalization of speeding of the electrical forklifts. It is designed
to increase safety of transport and storage operations.
KEYWORDS: forklift, speed limit control, warehouse, safety
INTRODUCTION
Materials handling includes the use of many kinds of equipment designed to help in the tasks in warehouses. There
are many kinds of objects and materials to handle, each posing different hazards, and many different vehicles to be
used [1, 2]. At the same time, transport operations should be productive, cost-effective, profitable and safe [3].
Thus, application of preventive measures for reducing the risk is very important.
The forklifts have been frequently used in many industrial and transport operations in recent years. A forklift is a
powerful tool that allows one person to precisely carry, push, pull, lift, stack or tier large heavy loads with little effort.
Using a tool such as a forklift, cart or hand truck instead of lifting and carrying items by hand can reduce the risk of
suffering a back injury. Forklifts vary in size, ranging from 1 t capacity (for general use in warehouses), up to 50 t
capacity (for shipping container). They can be powered by batteries, propane, gasoline or diesel fuel [4].
The forklifts are heavy vehicles that can cause severe injury or even death of unprotected workers in their vicinity.
OSHA estimates that 11% of all forklifts are involved in accidents every year [5]. Speed is a determinant in many
forklift accidents, so it is important that forklift operator observes the speed limits and the warning signs. Even at
low speeds, forklifts can cause serious injuries and fatalities.
1
Rodoljub Vujanac, Ph. D., assist. prof., University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić Str.,
34000 Kragujevac, Serbia, vujanac@kg.ac.rs
2
Nenad Miloradovic, Ph. D., assoc. prof., University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić Str.,
34000 Kragujevac, Serbia, mnenad@kg.ac.rs
3
Milan Djordjevic, Ph. D., prof., The Higher Education Technical School of Professional Studies in Kragujevac, 8
Kosovska Str., 34000 Kragujevac, Serbia, milan.djordjevic@vts.edu.rs
393
All forklifts must be operated in a safe manner. Forklifts should be operated within the manufacturer’s design
parameters, and only in areas for which they have been designed. Generally, systematic method for making
forklifts and their operation as safe as possible involves:
 identifying the hazards (anything with the potential to cause injury or disease),
 assessing the risk,
 determining what control measures to take,
 applying, reviewing and monitoring the control measures.
The warehouse environment is one of the most common places where forklift drivers work. The operator must be
trained and authorised to operate the forklift and also to carry out some required pre-start safety checks (or
procedures). Regarding the forklift operating speeds, the operator should:
 never speed through high speed traffic areas,
 reduce speed throughout heavily trafficked areas to avoid potential collisions,
 always proceed slowly when approaching blind spots,
 reduce speed around pedestrian walkways.
Modern forklifts have the capability of software speed limitation. Thus, when entering the production plant or
warehouse, the operator should presses the button with a maximum speed of 20 km/h in the mode of internal
operation in which the speed is limited to 5-6 km/h. At warehouses, the speed limits should be prominently
displayed, observed and enforced. Corresponding signs must be placed, so that they can be easily seen by forklift
operators. The speed of the forklift together with the weight of its load and road surface determines the distance at
which a forklift can stop in an emergency. For example, at speed of 14 km/h, a forklift will travel four metres in one
second and need at least 10 metres to stop safely and even at 6 km/h (walking pace), a forklift needs at least three
metres to stop.
The older models of forklift trucks do not have the option of changing modes to limit speed, while the modern
forklifts receive a signal from the transmitter through the doors and automatically switch from fast to slow mode and
vice versa. It is important to pay attention to the speed limit of the forklift in order to optimize its use without
sacrificing business productivity.
There are many factors that may influence the selection of forklift speed during operation in the warehouses. The
factors that influence the choice of correct speed limit are:
 forklift size, weight and construction - the higher the weight, the lower the speed should be,
 the state of the corridors through which the forklift circulates (surface resistance and slip capacity),
 visibility - the driving speed must be lower in low visibility areas,
 existence of obstacles like ramps, curves and bumps in the road.
The paper presents a technical solution for speed limit that can be installed in any electric forklift. The idea was to
develop a safe system of work to control risks and help prevent forklift-related incidents.
PROBLEM IDENTIFICATION AND PROJECT TASK

Despite the significant occupational safety problems that the use of forklifts can create, organisations typically have
ill-defined procedures and traffic management policies with respect to forklift truck operations. The most important
rule is that the speed of the forklift must be adjusted in accordance with the characteristics of the roads, the type of
goods being transported, and the possibility of stopping the vehicle. Speed is particularly limited near intersections,
in curves, on slopes, on slippery surfaces, in cases of low visibility, narrow passages, etc. The forklift speed can be
limited by speed bumps (Figure 1), speed limit signs (Figure 2) and low overhead clearances (Figure 3).
Figure 1 Speed bumps Figure 2 Speed limit sign Figure 3 Overhead clearances
394
The forklift speed bumps are used to slow the truck to a complete stop when passing. This reduces the chance for
operator to overspeed. They should be installed in locations such as at the intersections, crosswalks, pathways,
and in other important areas. Speed limit signs provide information about the enforced speed limit in the
warehouse. Although the speed limit signs are placed along the logistics corridors within the warehouse, it was
noticed that drivers often did not respect the limits that are prescribed. The low overhead clearances can also be
helpful in controlling overspeeding of the forklifts in the warehouses.
At the open space, the forklift speed is limited to 10 km/h in unloaded condition or up to 5 km/h in loaded condition.
At the enclosed space, the forklift speed is limited to 5 km/h, regardless of whether the forklift is empty or loaded.
When driving conditions require it, a lower driving speed is appropriate for safe driving.
The excessive speed of the forklift can be detected by using the markings on forklift wheels, Figure 4. When the
driver exceeds the speed of 5 km/h, the side observer can observe an overrun, since at 120 min-1, there is a
"wagon-wheel" effect when the wheel seems to stand in place. The basic problems with this system are:
 the observer must be close to the side of the vehicle to note this effect,
 the effect is achieved at wheel rotation speed of 120 min-1, which, for the wheel radius of 15 cm, amounts
to 6 km/h, while for the wheel radius of 25 cm, the same effect is achieved at speed of 11 km/h,
 the driver has no information about the forklift speed.
Figure 4 Speed detection using the marking on the wheel
All of these problems mentioned above were sufficient reason to approach the development of a system that is
able to signal when the operator of the forklift exceeds the allowed speed of 5 km/h and the possibility of automatic
speed limitation to 10 km/h.
The system should be designed to react to forklift overspeeding by limiting the forklift travel speed without affecting
the lifting power. Design requests for a new speed limit control system were:
 precise speed measurement,
 sound signalization should warn the driver that he exceeded the speed of 5 km/h,
 light signalization - a yellow rotary light that would be noticeable at a distance of 200 m and which would
alert pedestrians in the vicinity of a vehicle moving at a speed higher than permitted,
 physically limit the forklift speed to 10 km/h.
DESCRIPTION OF A NEW SYSTEM

The goal of the design of a new system was to resolve any potential critical situations in warehouse operation using
the forklift speed limitation in order to prevent accidents during handling, Figure 5. At the same time, the maximum
benefits of the operation should be obtained while minimising all risks in warehouse work areas.
The system is designed with a set of six neodymium magnets placed on the wheel hub, rotating together with the
hub [6]. When the wheel rotates, the magnets pass in front of the fixed magnetic encoder (reed relay) at about
5 mm of distance, Figure 6. Every time when a magnet passes near the magnetic encoder, an impulse is induced
and sent to the electronic module for speed limit and signalling (EMSLS). The schematic layout of the system is
shown in Figure 7.
395
Figure 5 Forklift with speed limiter Figure 6 The wheel hub with magnets and magnetic
encoder
EMSLS contains ATMEL microcontroller C8051. It is programmed for each type of vehicle, taking into account the
dynamic radius of the tire which is important for precise measurements. The speed can be generated using the
number of pulses and a time base for a specified range of tires. The measurement error of below 10% meets the
strict requirements. An external programmer programs each microcontroller. The hysteresis loop for „cut off / cut in“
DC power impulse (written in assembler) provided fine or "soft" crossings from „cut in“ to „cut off“ conditions. This
means that the driver would not feel sudden vibrations when breaking the motor [4].
Figure 7 Schematic layout of the system
EMSLS decides which action to take. When the forklift operator exceeds the speed limit of 5 km/h, a 72 dB sound
alarm (piezo buzzer) activates in order to signal the overspeeding. Also, a yellow rotary light turns on, Figure 8, to
warn all persons near the forklift that there is a potential danger.
Figure 8 Basic parts of the system: 1 - The wheel hub with magnets and reed relay (sensor),
2 - EMSLS (Electronic Module for Speed Limit and Signalling), 3 - Rotary light
396
When the forklift operator reaches the second speed limit of 10 km/h, EMSLS gives a signal to switch-off the
power. For electric forklifts, the speed is limited by the cut off of the DC motor current using the NC relay in the
electronic module. In the case of gas forklifts, the speed is limited by the interruption of the fuel supply
(reprogramming the ECU of the newer generation).
In general, the proposed system can automatically limit the forklift speed to required level at the warehouse
entrance or exit, or in specified danger zone. Also, the system allows a full speed driving of the forklift when it is
safe to do. In this way, the safe operation in the warehouse and reduction of possible damage are achieved.
CONCLUSIONS
Manually operated forklifts are widely used in the warehouses today. Controlling the forklift speed limit in working
area is a safety issue of the highest priority. Imposing speed limit can prevent accidents and equipment damage
and also reduce maintenance costs.
Constant demands for the increase of efficiency, reliability and security of logistics processes indicate the need for
research and implementation of new solutions. The new proposed speed limit control system has a simple set up
and low maintenance demands. It contains two kinds of warning: sound alarm and rotary light. It works only when
the forklift is moving, permitting the use of the full lifting power.
The most important advantage of the described system for signalization of speeding and speed limitation is the
improvement of safety conditions for operation. Although designed for electric forklifts, this system can be adjusted
for operation in other drives. The system can be modified to produce different sound warnings for different speed
levels.
REFERENCES
[1] Vujanac, R., Djordjevic, M., Miloradovic, N.: "Mechanical Tugger Trains System for Internal Material Handling",
The Sixth International Conference Transport and Logistic TIL 2017, 2017, Nis, pp 73-76,
[2] Vujanac, R.: "Semi-Automatic Mechanical Transport-Manipulation System for Production and Assembly Lines",
Invention, Response number 2015/5879, 2016, Belgrade,
[3] Vujanac, R., Disic, A., Djordjevic, M.: "Low cost automated system for heavy containers handling",
Proceedings of the XXI International Conference MHCL’15, 2015, Belgrade, pp 189-192,
[4] Djordjevic, M., Vujanac, R., Miloradovic, N: "System for Signalization of Speeding and Limiting the Speed of
the Internal Transport Equipment", Mobility & Vehicle Mechanics (MVM), Vol.43, No.4, 2017, pp 12-21,
[5] Washington State Department of Labor & Industries, Division of Occupational Safety and Health: Forklift
Safety Guide, Publication F417-031-000, 2015,
[6] Djordjevic, M., Jovanovic, S.: "System for Signalization of Speeding as well as Limiting the Speed of the
Internal Transport Equipment", Project developed by company FOX Electronics, 2011.
397
398
SECTION
Vehicle Dynamics and

Intelligent Control
Systems
 Aerodynamics
 Vehicle Ride and Handling
 Intelligent Safety Systems
 Tires/Braking
Systems/Steering
Systems/Suspension
399
400
Kragujevac, Serbia
MVM2018-037
Lazar Stojnić1
Aleksandar Grkić2
Slavko Muždeka3
DEVELOPMENT AND ANALYSIS OF A 4X4 TRUCK

MULTIBODY SIMULATION MODEL
ABSTRACT: The Multibody dynamics simulation solutions today are essential for vehicle development and
testing. Virtual prototypes significantly reduce product development time and costs. Using MSC.Adams/Car software
research teams can reduce their number of prototypes to a minimum number. Given the fact, this paper presents
simulation model of a military 4x4 truck with affection to suspension, chassis and steering systems. Also, simulation
model shows frame stress using FEA method in a couple of dynamic simulation case scenarios. Results presented
in the paper covers vehicle ride and handling characteristics and a comparison of a model with flexible and rigid
frame. Simulation model with this kind of capabilities gives a lot of opportunities to improve prototype phase and
testing of prototypes. Wide variety of MSC.Adams/Car plugins opens up possibilities to develop a simulation model
in extent in which ECU algorithms could be used as input controls and even develop an optimization programs and
durability tests. Given the results that these simulations offer, model should undergo experimental verification that
we set as a task for the future work.
KEYWORDS: multibody dynamics simulation, virtual prototypes, MSC.Adams, FEA analysis
INTRODUCTION
The Multibody dynamics simulation software have taken big role in a vehicle development process. Using this kind
of software not only that we can predict vehicle dynamics and some of the forces imposed on the vehicle, also we
can simulate real life case scenarios and find problems and solutions to the design that we imagined without building
and testing prototypes in a couple of iterations. In any real product-engineering programme, particularly in the ground
vehicle industry, there are always time constraints. The need to introduce a new product to retain market share or to
preserve competitive advantage drives increasingly tight timetabling for product engineering tasks [1]. The most well-
known multibody systems analysis software used today is MSC.Adams. The main use of MSC.Adams within the
automotive industry is to simulate the performance of subsystems independently and test full vehicles in standard
simulation events or even in user defined events with different types of terrain and obstacles. Main advantage of this
type of analysis is quick subsystem substitution and comparative analysis of different vehicle variants. Not only that,
but predictive methods seemingly allow several good things as [1]:
 Improved comprehension and ranking of design variables,
 Rapid experimentation with design configurations,
 Genuine optimization of numerical response variables.
Motion analysis is important because product design frequently requires an understanding of how multiple moving
parts interact with each other and their environment. Also, adding different upgrades to a vehicle and investigating
1
Lazar Stojnić, Research Assistant, Military Technical Institute, Ratka Resanovića 1, 11030 Belgrade, Serbia,
stojnic.lazar123@gmail.com
2
Aleksandar Grkić, Assistant Professor, University of Defense in Belgrade, Military Academy, 33 Generala Pavla
Jurišića Šturma Str., 11000 Belgrade, Serbia, aleksandargrkic@gmail.com
3
Slavko Muždeka, Associate Professor, University of Defense in Belgrade, Military Academy, 33 Generala Pavla
Jurišića Šturma Str., 11000 Belgrade, Serbia, slavko.muzdeka@gmail.com
401
given results is time and money saving but degradation of the ride and handling performance are considerable and
have to be taken into consideration [3]. Although simulations can provide most of the dynamic characteristics, the
simulation model must be validated by experimental analysis.
MODELING THE VEHICLE

Necessity that follows multibody modelling is to outline some of the planning that goes into the development of the
model. This paper deals with the development of a simulation model of a 4x4 truck shown on Figure 1, which is used
in Serbian Army and it represents, by all means, the most appropriate platform for using different types of upgrades.
The first step in MSC.Adams/Car is to sketch out a system topology which would typically illustrate parts, joints,
motions, forces and other elements such as springs, dampers, anti-roll bars, active help systems etc.
Figure 1 MSC.Adams 4x4 truck simulation model
Simulation model, shown in Figure 1 consists of front suspension, front leaf, front tires, front anti-roll bar, steering,
rear suspension, rear leaf, rear tires, rear anti-roll bar and chassis subsystems. The vehicle parameters relevant to
the model are given in the Tables 1 and 2 and Figure 2.
Table 1 Vehicle parameters

Parameter Value
Unloaded mass 4630 kg
Wheelbase 3600 mm
Front track width 1995 mm
Rear track width 2108 mm
Tire dimension 315/80R22.5
Roll moment of
21900 kgm2
inertia
Pitch moment of
352 kgm2
inertia
Yaw moment of
22200 kgm2
inertia
Table 2 Vehicle suspension parameters

Parameter Value
Front leaf stiffness 286 N/mm
Rear spring stiffness 260 N/mm
Front anti-roll bar
8600 Nmm/°
stiffness
Rear anti roll-bar
9000 Nmm/°
stiffness
402
100
80
60
Force (kN)
40
20
-20
-40
-1.5 -1 -0.5 0 0.5 1 1.5
Velocity (m/s)
Figure 2 Front and rear damper diagram
Transmission and brake subsystems are not important for the vehicle dynamic being assessed, so they were not
included in the model. The following assumptions were considered while building the model:
 Sprung mass is a rigid body,
 Used aerodynamic forces parameters were taken from the MSC.Adams truck example,
 Used dampers were taken from the MSC.Adams truck example,
 Bushings parameters are the ones defined in the MSC.Adams/Car,
 Every part of the model is considered as rigid body and
 Tires were modelled using Pacejka 2002 [2] tire model.
The complete model has 829 degrees of freedom, 180 moving parts, 1 translational, 12 revolute, 6 cylindrical, 3
spherical, 20 fixed, 3 Hooke joints, 32 bushings and one movement motion that is defined as steering wheel rotation.
Model topology
Topology of the model is defined by hardpoints (that represents the start and end points of the parts), joints between
parts and communicators. As every subsystem must be made by itself, communicators have to be defined so the full
vehicle assembly can function as one entity. Figure 3 and 4 shown below represents simulation model divided in half,
where every half consists of multiple subsystems. Communicators have a roll to connect joints on both sides of the
subsystems that have to be connected and their purpose is force and displacement transfer.
Front suspension, leaf, anti-roll bar and steering subsystem
This subsystem consists of rigid axle connected to a vehicle body by two leaf springs that also have the role of axle
longitudinal and lateral guidance. Leaf springs are connected to a frame by bushings where the rear part of the leaf
springs has a shackle to absorb big longitudinal deformations during bump or rebound. Connection between the leaf
and shackle is also achieved with bushing. Uprights are connected to the front axle by a revolute joint so it has
rotation around the kingpin axis. At the end of the upright revolute joint provides rotation of the wheel spindle. Both
front and rear suspension systems are fitted with anti-roll bars to help reduce body roll when the vehicle corners.
Anti-roll bar is connected to front axle and to tie rod by bushings and the tie rod is then connected to the vehicle’s
chassis. At the rear end of the leaf, shock absorber can be found and it is connected to both leaf and chassis by a
bushing. Rotation of the steering wheel is provided by revolute joint between the wheel itself and steering column.
Steering column has its own housing that is fixed to the vehicle chassis. Steering column, intermediate and input
shaft are connected by hooke joint so the rotation of the steering wheel isn’t lost during the direction change process.
Revolute joint of the input shaft and pitman arm are connected by a reduction gear that represents steering gearbox.
Pitman arm rotation is transferred to the steering link through spherical joint and to the steer arm through convel joint.
Steer arm is fixed to the upright so the rotation of the steering wheel is transferred to the Ackerman trapezoidal
geometry via steer arm. A leaf spring model is a series of small parts connected by beams based on first principles.
Contact between the leaves is modeled with vector forces. Model requires data of the leaf spring in unloaded
configuration, then a preprocessor performs a static analysis to load the leaf spring to a specified ride height.
403
Rear suspension, leaf and anti-roll bar subsystem
Kinematics of the rear suspension are basically the same with exception to different position of the anti-roll bar and
shock absorbers and of course the upright being put out of the system. Main difference in the rear is that the lower
strut of the shock absorber isn’t connected to the leaf but to the axle itself.
Figure 3 Front suspension and steering kinematic diagram
Figure 4 Rear suspension kinematic diagram
Figure 5 Kinematic diagram legend
404
Characteristics of the tires
The handling performance and directional response of a vehicle are greatly influenced by the mechanical force and
moment generating characteristics of the tires. As mentioned earlier tires were modelled using Pacejka 2002 tire
model and it has characteristics shown on Figure 6.
20 14
10 7
Longitudinal force (kN)
Lateral force (kN)

0 0
-10 -7
-20 -14
-1 -0.5 0 0.5 1 -8 -4 0 4 8
Longitudinal slip (deg) Lateral slip (deg)
220
120
Self aligning torque (Nm)
20
-80
-180
-280
-10 -5 0 5 10
Lateral slip (deg)
Figure 6 Tire characteristics
The figure shown above shows the most important tire characteristics with 7000, 14000 and 21000 N of vertical load
respectively.
After the full vehicle model is built chronological methodology have to be performed, in order to achieve appropriate
simulation results [3]:
 Design appropriate maneuvers to assess the full vehicle response,
 Perform maneuvers on the model,
 Discuss the results and draw conclusions for further researches.
405
SIMULATION ANALYSIS
All the analyses in the MSC.Adams/Car are based on the Driving Machine. The Driving Machine drives your virtual
vehicle according to your instructions much like a test driver would drive an actual vehicle. It has two types of control
during a simulation, open- and closed-loop control or as MSC refers to it Machine Control and SmartDriver. Basic
benefits of this controller are:
 Saving the time previously needed to set proper gains for closed-loop controllers. The Driving Machine
incorporates proprietary control algorithms from MSC Software,
 Shorter simulation times by ending simulations based on targets for lateral acceleration, longitudinal
velocity, and distance travelled,
 Flexibility in how you input vehicle paths and desired velocity profiles, choosing from recorded data stored
in external files or parametrically generated paths profiles based on a few simple inputs.
The most appropriate test scenarios that can show ride and handling dynamics are double lane change course event,
constant radius cornering and maintaining certain velocity while crossing over a bump event. Also, these tests are
relatively easy to perform in the real-life case scenario, so they can be performed in purpose of validating the
simulation model.
0.8
0.6
0.4
Roll angle (deg)
0.2
0
-0.2
-0.4
-0.6
-0.8
0 5 10 15 20 25
Time (s)
Figure 7 Roll angle during the double lane change test
Roll angle, lateral acceleration, side slip angle and vertical loads are calculated at various starting forward velocities.
Yellow line on the Figure 7, Figure 8 and Figure 9 presents result where vehicle starting velocity is 45 km/h, grey line
40km/h, orange line 35 km/h and blue line results with 30 km/h starting velocity. Under cornering conditions, when
forward velocity decreases, tire lateral forces on the tires which the tires must develop to equalize the centrifugal
force will decrease. This also leads to the reduction of the slip angle on the rear tires [6]. This leads to increase of
the vehicle side slip angle as it can be seen on the Figure 9. Low value of the roll angle is consequence of low vertical
value of CoG.
0.4
0.3
Lateral acceleration (g)
0.2
0.1
-0.1
-0.2
-0.3
-0.4
0 5 10 15 20 25
Time (s)
Figure 8 Lateral acceleration during the double lane change test
406
4.5
3.7
2.9
Side slip angle (deg) 2.1
1.3
0.5
-0.3
-1.1
-1.9
-2.7
-3.5
0 5 10 15 20 25
Time (s)
Figure 9 Side slip angle during the double lane change test
19 19
Front right tire vertical load (kN)
Front left tire vertical load (kN)
15.5 15.5
12 12
8.5 8.5
5 5
0 5 10 15 20 25 0 5 10 15 20 25
Time (s) Time (s)
16 16
Rear right tire vertical mode (kN)
Rear left tire vertical load (kN)
14 14
12 12
10 10
8 8
6 6
0 5 10 15 20 25 0 5 10 15 20 25
Time (s) Time (s)
Figure 10 Truck mass distribution in a double lane change test
Vertical load difference between front and rear axle is 52:48 in favor of front axle while the truck remains unloaded.
Figure 10 shows that increase of 15 km/h in starting longitudinal velocity changes vertical load transfer by 111% in
the front of the vehicle and 157% in the rear of the vehicle. With the CoG in given position vehicle is far away from
rolling over with the tested velocities.
0.5 0.9
0.8
Lateral acceleration (g)
0.4 0.7
Roll angle (deg)
0.6
0.3
0.5
0.4
0.2
0.3
0.1 0.2
0.1
0 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)
Figure 11 Lateral acceleration and roll angle during the constant radius cornering test
407
Figures 11, 12, 13 and 14 shows results from the constant radius cornering test. In these figures dark blue line
represents vehicle moving with 30 km/h starting velocity, orange 35 km/h, grey 40 km/h, yellow 45 km/h and light
blue 50 km/h.
Similar results can be seen on the constant radius cornering test. Diagram on figure 12 shows that the driving
machine is performing steering wheel corrections as the starting velocity is increasing.
2.5
1.5
Side slip angle (deg)
0.5
-0.5
-1
-1.5
-2
0 2 4 6 8 10
Time (s)
Figure 12 Side slip angle during the constant radius cornering test
19 12
Front right tire vertical load (kN)
Front left tire vertical load (kN)
18 11
17
10
16
9
15
8
14
13 7
12 6
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)
17 12
Rear right tire vertical load (kN)
Rear left tire vertical load (kN)
16 11
15 10
14 9
13 8
12 7
11 6
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)
Figure 13 Truck mass distribution in a constant radius cornering test
As forward velocity increases, there are bigger oscillations of vertical load at each axle. This is also a consequence
of steering wheel corrections.
408
CONCLUSIONS
This simulation model is a good starting template for a 4x4 truck dynamic simulation. With a little effort model can be
changed to a real-life scenario truck and can be used for assessment of a truck dynamics. Also, implementing a FEM
models into this template doesn’t just improves vehicle assessed dynamics but helps in part optimization in a dynamic
test scenario. This approach to vehicle dynamic analysis potentiate time and money save by reducing the
experimental test scope. On the other hand, it is easy to test different types of a vehicle which is particularly important
with this type of vehicles that can have many different upgrades on the same chassis. After the finished model, results
can be easily analyzed and shared with different development teams which can shorten the development process.
Likewise, use of multibody simulation models is really good for modernization and optimization of already familiar
systems.
REFERENCES
[1] Blundell, M., Harty, D.: "The Multibody Systems Approach to Vehicle Dynamics", Butterworth-Heinemann, 2014,
[2] Pacejka, H. B.: "Tyre and Vehicle Dynamics", Butterworth-Heinemann, 2002,
[3] Khettou, N., Trifković, D., Muždeka, S.: "Modeling and Multibody Simulation of Land Rover Defender 110 Ride
and Handling Dynamics", 7th International Scientific Conference on Defensive Technologies, OTEH 2016,
Belgrade, pp 223-230, ISBN 978-86-81123-82-9,
[4] Khettou, N., Grkić, A., Muždeka, S.: "Modeling and Validation of Ground Vehicle Dynamics Using Multibody
Simulation", 6th International Congress Motor Vehicles & Motors 2016, Kragujevac, Serbia, ISBN 978-86-6335-
037-3,
[5] Stojnić, L., Khettou, N., Grkić, A., Muždeka, S.: "Modelling of vehicle steering system using multibody simulation",
26th JUMV International Automotive Conference Science & Motor Vehicles 2017, Belgrade, Serbia, ISBN 978-
86-80941-41-7,
[6] Hasagasioglu, S., Kilicaslan, K., Atabay, O., Güney, A.: "Vehicle dynamics analysis of a heavy-duty commercial
vehicle by using multibody simulation methods", The International Journal of Advanced Manufacturing
Technology, Vol. 60, 2012, pp 825-839.
409
410
Kragujevac, Serbia
MVM2018-042
Boris Stojić1
Aleksandar Poznić2
Nebojša Nikolić3
EXAMINATION OF SENSITIVITY OF TRACTOR TIRE

FILTERING BEHAVIOR TO CONTACT LENGTH CHANGE
ABSTRACT: Vibrations of agricultural tractors and other off-road vehicles are mainly excited through uneven
ground encountered by such vehicles. Thereby tire acts as geometrical filter that transforms short-wavelength road
undulations into effective excitation profile of milder geometry. Being geometrically conditioned, this filtering
behaviour depends on tyre radial deflection i.e. contact length. In this paper examination of sensitivity of filtering
behaviour to contact length change is described and results are presented.
KEYWORDS: tractor tire, vibration, effective road profile
INTRODUCTION
If tyre rolls over uneven ground whose irregularities have wavelength of the order of magnitude similar to tyre
contact length, wheel centre will not follow exact ground geometry (as would be the case when using point-contact
rigid follower instead of elastic tyre), but will travel along the path of different shape. If tyre motion happens in
quasistatic manner, than tyre response (in view of the wheel centre trajectory shape) will have greater length and
lower height compared to single road irregularity causing this response. This is the result of both tyre geometry and
elasticity, and can be described as geometric low-pass filtering of the original road profile. Wheel centre trajectory
is thereby considered effective road profile, since this motion represents real vehicle vibration excitation. This
mechanism is depicted in Figure 1. This aspect of the tyre behaviour is referred to as „enveloping behaviour“, since
the tyre virtually envelopes i.e. wraps around ground irregularity of the short wavelength. Choosing quasistatic
mode, it is possible to focus analysis purely on tyre enveloping behaviour, not taking into account tyre internal
dynamics and vibration properties.
By using different combinations of tyre load and pressure, it is possible to obtain the same value of the length of
contact area of the tyre with even ground. Furthermore, unambiguous relationship exists between contact length
and radial tyre deflection [3, 4]. If values of these parameters change, the way of establishing contact between tyre
footprint and road irregularities encountered by the tyre during motion also alters. This further leads to alteration of
tyre enveloping behaviour i.e. geometric filtering properties. This phenomenon leads to conclusion that tyre contact
length can be used as parameter of the tyre enveloping model. Such approach has already been used previously,
e.g. [6, 7], though for road vehicle tyres.
1
Boris Stojić, Assistant professor, University of Novi Sad, Faculty of Technical Sciences, Novi Sad, Trg Dositeja
Obradovića 6, bstojic@uns.ac.rs
2
Aleksandar Poznić, Assistant, University of Novi Sad, Faculty of Technical Sciences, Novi Sad, Trg Dositeja
Obradovića 6, alpoznic@uns.ac.rs
3
Nebojša Nikolić, Assistant professor, University of Novi Sad, Faculty of Technical Sciences, Novi Sad, Trg
Dositeja Obradovića 6, nebnik@uns.ac.rs
411
Tractor tyre is characterized by significant differences in some aspects of its behaviour when compared to road
tyres, above all due to different tyre structure and different relation between this physical structure and pneumatic
contribution to tyre behaviour. In [1], results of examination have been published that shown that, although not
completely accurate, approach of using tyre contact length as unique parameter of enveloping behaviour can be
used as satisfying approximation, providing that tyre operational parameters are within usual range. Attaining
higher level of model simplification additionally justifies adoption of this approximation. Primary topic of this paper is
to obtain numerical quantification and qualitative assessment of the level and nature of contact length impact on
tyre enveloping behaviour.
Figure 1 Concept of effective road profile: h(x) - ground profile shape, hEFF(x) - wheel centre trajectory based on
h(x) and transformed by tyre filtering mechanism, x - direction of motion
EXPERIMENTAL SETUP AND METHOD

Tyre enveloping behaviour has been studied by measuring horizontal and vertical displacement of the wheel rolling
very slowly over the singular road obstacle, in nearly quasistatic conditions (i.e. such that all kinds of dynamic
excitation of the tyre structure can be considered negligible). Experimental facility used in these investigations is
shown schematically in the Figure 2. It was described in more details in [2] and [5].
Figure 2 Composition of the test facility: 1-tested wheel, 3-wheel mounting frame, 4-wheel guiding cart, 11-road
irregularity
Experimental measurements were made with the tyre of the size 12.4R28. Shape and dimensions of the road
obstacle used for investigations are shown in the Figure 3.
Figure 3 Shape and dimensions of the road obstacle used for investigations
412
Different levels of contact length were obtained by using constant tyre load of 960 daN and by varying tyre
pressure. Four levels of contact length were used: 30, 35, 40 and 45 cm. Significant level of result fluctuations was
observed, so in order to compensate for this, every measurement was carried out 10 times. Results were averaged
and smoothed subsequently, so as to obtain representative response curves for each case.
Contact length was determined by measuring the distance between two tin plates placed along both front end rear
edges of the tyre contact area, Figure 4. This method was described in more details in [3].
Figure 4 Measuring tyre contact length

In order to quantify impact of contact length change on tyre filtering i.e. enveloping properties, measurements of
tyre response curves have been made for different values of contact length. Geometric and spectral properties of
response curves in different conditions were compared to each other.
In Figure 5, measurement results i.e. enveloping curves obtained are directly compared to each other. Vertical shift
between the curves originates from the different levels of tyre centre due to different values of tyre deflection, at the
same time enabling better visual overview of this comparison. Two main properties that can be clearly seen from
the Figure 5 are:
1. Completely different shapes of individual curves;
2. Significant difference in maximum values of response amplitudes.
Figure 5 Enveloping curves obtained for different values of the contact length (values next to curves)
In the Figure 6, zoomed area from the Figure 5 close to the origin is shown. In this figure it can be noticed that for
greater value of contact length (corresponding to smaller level of radial deflection of the tyre) response curves start
earlier, due to earlier contact with the obstacle. This also leads to the difference of the overall response curve
length.
413
Conclusion that can be made on the basis of these results, corresponding to intuitive comprehension of tyre
behaviour, is that the effect of tyre low-pass filtering will be more pronounced for larger contact length. Therefore
under such conditions amplitude response tends to decrease, while response length grows.
Figure 6 View of response curve beginning point (i.e. therefore overall length) change with the change of the
contact length
Another view of investigation results is shown in Figure 7. In this view all response curves are shifted to the same
referent ground level. For better visibility, only parts of curves left to the vertical symmetry axis are shown. This
view shows properties noticed according to the previous figures, regarding main differences in tyre response for
different contact lengths, more clearly.
70
60
50
40
zEFF [mm]
30 cm
35 cm
30
40 cm
20 45 cm
10
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55
x [m]
Figure 7 Response curves shifted to the same referent ground level
Finally, analysis of response curves spectral content is carried out in term of power spectral density. PSD curve of
ground profile excitation is also shown for comparison. Results are shown in Figure 8. Change of response
frequency content, which is indicator of filtering properties alteration, can be notified from the figure. Curves show
that with increasing contact length, spectral content of the response curves decreases. It can also be seen that this
change in tyre filtering behaviour is especially abrupt in transition of contact length values between 35 and 40 cm.
414
600
500
400
PSD [mm2/m]
30 cm
300 35 cm
40 cm
200 45 cm
Excitation
100
0
0 2 4 6 8 10
W [m-1]
Figure 8 Spectral content of response curves and excitation
CONCLUSIONS
In order to quantify impact of contact length change on tyre filtering i.e. enveloping properties, measurements of
tyre response curves have been made for different values of contact length. Geometric and spectral properties of
response curves in different conditions were compared to each other.
Results of investigation have shown clearly that contact length change lead to significant variation of tyre
enveloping behaviour, in a way that larger contact length means more pronounced low-pass filtering. Analysis of
frequency content of response curves for different values of contact length, which can be regarded as the most
appropriate indicator of tyre filtering properties, suggests that alteration of enveloping behaviour does not happen
evenly with the change of contact length, but that it will be much more significant in certain regions of contact
length gradient.
In this paper tyre contact length was mainly treated as suggested parameter of tyre enveloping model. Further task
that has to be carried out within this scope is to assess level of dynamic fluctuations of tyre contact length, which is
effect of deflection change due to dynamic wheel load. In that way it will be possible to conclude if it is necessary to
include variable contact length into tyre enveloping model. Should it be possible to leave out this part of modelling
approach, it would contribute significantly to the model simplification and therefore improvement of numerical
efficiency.
ACKNOWLEDGMENTS
This paper was done as a part of the researches on the project TR35041 – "Investigation of the safety of the
vehicle as part of cybernetic system: Driver-Vehicle-Environment" and the project TR31046 " Improvement of the
quality of tractors and mobile systems with the aim of increasing competitiveness and preserving soil and
environment", supported by Serbian Ministry of Education, Science and Technological Development.
REFERENCES
[1] Stojić, B.: "Istraživanje mogućnosti korišćenja kontaktne dužine traktorskog pneumatika kao jedinstvenog
eksploatacionog parametra envelope kretanja", Traktori i pogonske mašine Vol. 19, No. 3, 2014, pp 23-28,
[2] Stojić, B., Poznanović, N., Poznić, A.: "Test facility for investigations of quasistatic enveloping behaviour of
tractor tire", Proceedings of 8th International Symposium "Machine and Industrial Design in Mechanical
Engineering", 2014, Balatonfüred, HU, FTN Novi Sad, pp 89-92,
415
[3] Stojić, B., Poznić, A.: "Determination of basic mechanical parameters of the tractor tyre by using universal
approach", Proceedings of 9th International Symposium "Machine and Industrial Design in Mechanical
Engineering", 2016, Balatonfüred, HU, FTN Novi Sad, pp 49-54,
[4] Stojić, B., Poznić, A.: "Research and modelling of tractor tyre radial deflection and ground contact length
response to vertical load", Proceedings of International Congress Motor Vehicles & Motors 2016, Fakultet
inženjerskih nauka Kragujevac, 2016,
[5] Stojić, B., Poznanović N., Poznić A.: "Research and modeling of the tractor tire enveloping behaviour", Journal
of Vibration and Control, Vol. 23, Issue: 2, 2017, pp 290-304,
[6] Zegelaar, P. W. A.: "The Dynamic Response of Tyres to Brake Torque Variations And Road Unevennesses",
Doctoral thesis, Delft University of Technology, 1998,
[7] Хачатуров, А.А.: "Динамика системы дорога - шина - автомобиль – водитель", Машиностроение, 1976.
416
Kragujevac, Serbia
MVM2018-051
Nina Živanović1
OSCILLATIONS OF LIGHT WHEELED VEHICLE UNDER

THE INFLUENCE OF INTEGRATED WEAPON’S RECOIL
FORCE
ABSTRACT: Oscillations of light wheeled vehicle’s centre of gravity, caused by firing weapon integrated on it,
are studied through simulation of its axial and angular displacements. Dynamical model of system with three
degrees of freedom is used and mathematical model of oscillations is defined using the Lagrange’s equations of
second order.
Weapon recoil force and resistance force that causes vehicle’s oscillations is calculated according to the internal
ballistics data of integrated weapon and chosen system construction. System of differential equations is
numerically solved in MATLAB for real firing conditions and vehicle configuration. Coordinates of weapon mount
and elevation and azimuth angles are varied and results are analyzed.
KEYWORDS: oscillations, dynamical model, wheeled vehicle, integrated weapon
INTRODUCTION
Systems composed of light mobile platform with integrated weapon rise interest over the past few years because
modern warfare require light military vehicles in the arsenal, with few crew members, efficient, not too expensive,
intended for scouting, quick actions, neutralization and instant prevention of hostile forces.
Well equipped armies have purpose-designed tactical vehicles, while less developed countries and paramilitary
forces use improvised, already available, adapted vehicles. As a mobile platform, modified light commercial vehicle
on wheels is most commonly used (pickup vehicles, utility vehicles). Besides the basic purpose, it also has to
absorb additional load from weapon’s recoil force. For this paper, Land Rover Defender 110 is selected as a mobile
platform and automatic grenade launcher 30 mm caliber M93 as a weapon for integration.
Mathematical-mechanical model of system’s oscillations is set based on theoretical approach for studying vehicle
dynamics and differential equations of motions are derived. When solved, referring to internal-ballistic data about
weapon and vehicle construction parameters, specific equations of platform oscillations are acquired, which can be
numerically solved in one of contemporary software packages.
Simulation, which is actually solving of previously set equation system, is performed in MATLAB environment,
giving the graphical representation of results, in this case system oscillations. With initial parameters variations,
system precision and alteration of oscillations can be perceived.
1
Nina Živanović, Military technical institute, Ratka Resanovića 1, 11000 Belgrade, Serbia,
nina.zivanovic.kg@gmail.com
2
Gordana Bogdanović, University of Kragujevac Faculty of Engineering, Sestre Janjić 6, 34000 Kragujevac,
Serbia, gocab@kg.ac.rs
417
WEAPON INTEGRATION ON PLATFORM
Weapon can be integrated on several places on the vehicle – it can be mounted on one of the reinforcing bars in
the vehicle trunk, between the front seats, above the engine compartment, in front of the passenger seat or in
some other place where crew members can operate it from. One of the simplest mounts is vertical bar fastened to
vehicle floor. This kind of integration allows specified weapon elevation and azimuth adjustments, while the weapon
remains fixed in the same place. More flexible solution is swing arm mechanism composed out of several axles
(‘arms’), which can cover up to 180° degrees along the horizontal direction. There are also mechanisms as 360-
degrees turret mounts allowing weapon 360° degrees rotation.
Common practice is to mount weapon in the longitudinal vertical symmetry plane, or as close to it as manageable,
so destabilizing moments are the smallest possible. Vehicle has a specific task to carry out when weapon is
mounted on it – beside the loads transferred from the road, it has to withstand and absorb weapon generated
loads. Main problems for this are inadequate stiffness of the vehicle chassis, adverse oscillations, low weapon
firing precision etc. To fulfil these demands and solve occurring problems, the most important thing is to integrate
weapon in optimal place and to provide suitable vehicle suspension system which can absorb loads, reduce
oscillations and prevent resonance from happening.
Force transferred to the platform calculation

Main targets of automatic grenade launchers are unsheltered troops or lightly sheltered targets and lightly armored
vehicles. Those are convenient for mounting on military vehicles, boats and helicopters [1]. Automatic grenade
launcher 30 mm M93 was developed from Russian AGS-17 and the operating principle is blowback. In Table 1
technical data for automatic grenade launcher 30 mm M93 are given.
Table 1 Technical data for AGL 30 mm М93 [2]

Data AGL 30 mm М93
Empty weapon mass 35 kg
Bullet mass 0.360 kg
Effective firing range 1700 m
Vertical field of action (-5°) ÷ (+70°)
No tripod movement - 30°, with tripod
Horizontal field of action
movement - 360°
Muzzle velocity 185 m/s
Maximum pressure in barrel 150·106 Ра
Within the simulation, internal-ballistic calculation [3] based on Drozdov method had been done and acquired
values of maximum pressure in the barrel (𝒑𝒎𝒂𝒙 = 141.7 MPa) and projectile muzzle velocity (𝒗𝟎 = 186.4 m/s) were
used as initial parameters for further computation. Obtained values barely deviate from the real ones.
Working principle of AGL is simple blowback, which obtains energy, used to perform automatic operation of a
weapon, from residual pressure of propellant gasses. Motion of the parts is slowed by the mass of the bolt and the
force required to compress the action spring [4]. In automatic grenade launcher 30 mm M93 internal friction force,
spring resistance force and hydraulic resistance force oppose to the motion of the recoil mass. Those forces
together compose total force of resistance to recoil 𝑹 [N], the force that is transferred to the platform and calculated
according to the next equation:
𝑹 = 𝑭𝒕𝒓 + 𝟐 · 𝑭о𝒑 + 𝑭𝒉𝒌 − 𝒎𝒕 · 𝒈 · 𝒔𝒊𝒏 𝜶 (1)
Figure 1 shows all of the resistance forces which occur during recoil process over time, as well as the total force of
resistance to recoil.
Hydraulic brake absorbs the most of recoil mass’ kinetic energy, ca. 80%, so it has the biggest impact on 𝑹 force
change [5]. Maximum value of total force of resistance to recoil is 𝑹𝒎𝒂𝒙 = 2385.9 N, selected for further calculation.
418
Figure 1 Resistance forces and total force of resistance to recoil
INTEGRATING AGL 30 MM ON SELECTED LIGHT MOBILE WHEELED

PLATFORM
Decision was made to mount automatic grenade launcher on the rooftop of the Land Rover Defender in the area
above the back seat. Created model had to be simplified, but sufficiently accurate replacement for the real and
complex system. Chosen simplified model has three degrees of freedom – translation of mass centrum in the
direction of vertical axis z (Tz) and rotations around longitudinal x and lateral y axis (roll Rx and pitch Ry).
From a constructive point of view, simplifications include integration of weapon on the vehicle’s longitudinal axis,
solid link between them, not deforming construction, assumption that the system behaves as a compact mass,
equal spring stiffness coefficients 𝑐𝑖𝑗 , as well as equal damping coefficients of front and back shock absorbers 𝑘𝑖𝑗 .
When weapon fires, resistance force to recoil 𝑅 transfers to elastically suspended platform and causes system to
oscillate, which makes it an excitation that takes the system out of equilibrium position. With cessation of excitation
system continues to oscillate, damped using shock absorbers, until it stops in equilibrium position.
In Figure 2 system scheme is shown, observed in xz plane, in right-hand oriented, stationary, perpendicular
coordinate system attached to the ground (xyz). Coordinate system x'y'z', attached to the system mass centrum Т,
moves with the system, 𝑎 and 𝑏 are distances from vertical axes of front and back vehicle axles, 𝐻𝑇 is height of the
mass centrum measured from the ground. Driving force operates in the point D, which position is defined by 𝑑 and
ℎ distances from system mass centrum T in z'x' plane.
Figure 2 Simplified system scheme in the zx plane
419
In Figure 3 the same system is shown, observed from the front in the z'y' plane. Labels ѕ are distances from system
mass centrum T to left and right wheel vertical axes.
Figure 3 System model from the front side in zу plane
In z'x' plane (Figure 4) side of the system is observed, translation in the z’ axis direction, one of the front wheels
and one of the back wheels, as well as rotation around y’ axis (pitch) for angular displacement φ, in the arbitrary
point in time.
Figure 4 System model in the arbitrary point in time during the driven oscillation in zx plane
In the z'y' plane (Figure 5) system is observed from the front, vertical translation and rotation around x’ axis (roll) for
angular displacement θ, also in the arbitrary point in time.
Figure 5 System model in the arbitrary point in time during the driven oscillation in yz plane
420
For simulation, weapon was fired from top of the stationary vehicle, elevation and azimuth angles defined in
advance. For simplified model, differential equations as a description of system’s motion were set.
MATHEMATICAL-MECHANICAL MODEL OF MOBILE PLATFORM

OSCILLATIONS UNDER THE IMPACT OF WEAPON’S RECOIL FORCE
Mathematical-mechanical model represents movement of the physical system based on the classical principles of
vehicle dynamics theory and includes all main influencing parameters [6]. Considering adopted simplifications,
system with infinite number of degrees of freedom is reduced to system with thee of them. Chosen coordinates are:
𝑧Т – vertical position of mass centrum Т, 𝜑 – rotation around lateral axis y' (pitch), 𝜃 – rotation around longitudinal
axis х' (roll).
Lagrange equations, second-order partial differential equations are used to describe movement of the modeled
platform with integrated weapon [7], induced by the force generated when fired out of it:
d ∂𝐸𝑘 𝜕𝐸𝑘 𝜕𝐸𝑝 𝜕𝛷

( )− =− − + 𝑄𝑞′ 𝑖 (2)
d𝑡 ∂𝑞𝑖̇ 𝜕𝑞𝑖 𝜕𝑞𝑖 𝜕𝑞𝑖̇
Labeled i = 1, n, and n is the number of degrees of freedom for system motion (in this case n = 3), 𝐸𝑘 – system
𝜕𝐸𝑝
kinetic energy, 𝐸𝑝 – system potential energy, 𝛷 – dissipative function, (− ) – generalized conservative force
𝜕𝑞𝑖
𝜕𝛷
(negative partial derivative of potential energy by generalized coordinate i), (− ) – generalized dissipative force
𝜕𝑞𝑖̇
(negative partial derivative of dissipative function by generalized velocity 𝑖), 𝑄𝑞′ 𝑖 - arbitrary non-conservative
generalized force.
Total system kinetic energy consists of mass centrum translator kinetic energy and rotational kinetic energy, around
the mass centrum:
1 1 1
𝐸𝑘 = · 𝑚 · 𝑧𝑇̇ 2 + · 𝐽𝑦 · 𝜑̇ 2 + · 𝐽𝑥 · 𝜃̇ 2 (3)
2 2 2
𝐽𝑦 is moment of inertia of system mass in regard to longitudinal axis y', аnd 𝐽𝑥 is moment of inertia of system mass
in regard to lateral axis х'.
Formula for potential energy of the system, for equal values of spring stiffness coefficients is:
1 1
𝐸𝑝 = · 𝑐1 · (Δ211 + Δ212 ) + · 𝑐2 · (Δ221 + Δ222 ) (4)
2 2
Labeled Δ𝑖𝑗 is total deformation, or displacement of elastic elements. Dissipative function, with calculated
derivatives of relative displacements and for equal values of damping coefficients, is:
1 1
𝛷= · 𝑘 · (Δ̇211 + Δ̇212 ) + · 𝑘2 · (Δ̇221 + Δ̇222 ) (5)
2 1 2
Generalized non-conservative forces are derived using virtual displacement of the point D in which actuates
external force. For computation maximum value of previously calculated resistance force to recoil 𝑅𝑚𝑎𝑥 is used. It
has three components in the directions of three axes 𝑅𝑥 , 𝑅𝑦 и 𝑅𝑧 , so the formula for virtual work is:
𝛿𝐴 = (𝑅𝑥 (𝑡) · ℎ + 𝑅𝑧 (𝑡) · 𝑑) · 𝛿𝜑 + (𝑅𝑦 (𝑡) · ℎ) · 𝛿𝜃 + (−𝑅𝑧 (𝑡)) · 𝛿𝑧𝑇 (6)
In order to define Lagrange’s equations, derivatives of previous formulas for kinetic energy, potential energy,
dissipative energy and virtual displacements are appointed and their components sorted by the generalized
coordinates. Formed system of differential equations can be noted in matrix form:
[𝐴]{𝑞̈ } + [𝐵]{𝑞̇ } + [𝐶]{𝑞} = {𝑄} (7)
[𝐴] is matrix of inertia or matrix of masses, [𝐵] is damping matrix, [𝐶] stiffness matrix, and {𝑄} is driving force
vector. When substitution of matrixes takes place matrix (7) becomes matrix (8).
421
𝑚 0 0 𝑧𝑇̈ 2 · 𝑘1 + 2 · 𝑘2 2 · 𝑘1 · 𝑎 − 2 · 𝑘2 · 𝑏 0 𝑧𝑇̇
[0 𝐽𝑦 0 ] { 𝜑̈ } + [2 · 𝑘1 · 𝑎 − 2 · 𝑘2 · 𝑏 2 · 𝑘1 · 𝑎2 + 2 · 𝑘2 · 𝑏 2 0 ] { 𝜑̇ }
0 0 𝐽𝑥 𝜃̈ 0 0 2 · 𝑘1 · 𝑠 2 + 2 · 𝑘2 · 𝑠 2 𝜃̇
2 · 𝑐1 + 2 · 𝑐2 2 · 𝑐1 · 𝑎 − 2 · 𝑐2 · 𝑏 0 𝑧𝑇
2 2
+[ 2 · 𝑐1 · 𝑎 − 2 · 𝑐2 · 𝑏 2 · 𝑐1 · 𝑎 + 2 · 𝑐2 · 𝑏 0 ]{𝜑} (8)
0 0 2
2 · 𝑐1 · 𝑠 + 2 · 𝑐2 · 𝑠 2 𝜃
−𝑅𝑧 (𝑡) + 2 · 𝐻𝑇 · (𝑐1 + 𝑐2 )
= {𝑅𝑥 (𝑡) · ℎ + 𝑅𝑧 (𝑡) · 𝑑 − 2 · 𝐻𝑇 · (−𝑐1 · 𝑎 + 𝑐2 · 𝑏)}
𝑅𝑦 (𝑡) · ℎ
Final matrix (8) is valid for time span [0; 0,0522 s] while resistance force to recoil impacts the system and while
oscillations of the platform with integrated weapon have driven-damped character.
After excitation ceases, the system continues to oscillate damped by the absorbers until it stops in equilibrium
position. Matrix that describes free damped oscillating motion of the wheeled platform with integrated weapon and
three degrees of freedom:
𝑚 0 0 𝑧𝑇̈ 2 · 𝑘1 + 2 · 𝑘2 2 · 𝑘1 · 𝑎 − 2 · 𝑘2 · 𝑏 0 𝑧𝑇̇
[0 𝐽𝑦 0 ] { 𝜑̈ } + [2 · 𝑘1 · 𝑎 − 2 · 𝑘2 · 𝑏 2 · 𝑘1 · 𝑎2 + 2 · 𝑘2 · 𝑏 2 0 ] { 𝜑̇ }
0 0 𝐽𝑥 𝜃̈ 0 0 2
2 · 𝑘1 · 𝑠 + 2 · 𝑘2 · 𝑠 2
𝜃̇
2 · 𝑐1 + 2 · 𝑐2 2 · 𝑐1 · 𝑎 − 2 · 𝑐2 · 𝑏 0 𝑧𝑇
+ [2 · 𝑐1 · 𝑎 − 2 · 𝑐2 · 𝑏 2 · 𝑐1 · 𝑎2 + 2 · 𝑐2 · 𝑏 2 0 ]{𝜑} (9)
0 0 2 · 𝑐1 · 𝑠 2 + 2 · 𝑐2 · 𝑠 2 𝜃
2 · 𝐻𝑇 · (𝑐1 + 𝑐2 )
= {−2 · 𝐻𝑇 · (−𝑐1 · 𝑎 + 𝑐2 · 𝑏)}
0
Differential equations are solved for the data defined in advance: firing conditions (elevation angle, azimuth angle,
pressure force in the barrel and resistance force to recoil) and vehicle/platform construction parameters (stiffness
and damping coefficients, mass centrum coordinates and moments of inertia). This way, platform oscillation
equations are obtained, and for them, in regard to time, change of the generalized coordinates and velocities of
those changes.
SIMULATION OF MOBILE PLATFORM OSCILLATIONS AND RESULTS

ANALYSIS
Solving of previously defined system of differential equations represents the main part of the simulation and it was
conducted numerically in MATLAB environment. There are three initial data sets: vehicle data – technical data,
vehicle mass centrum position, vehicle elastic suspension characteristics, moments of inertia; weapon data –
weapon mass, integrating position relative to system mass centrum, elevation and azimuth angles; driven force
data –calculated values of resistance force 𝑅 over time.
For integration point, position on the rooftop above back seat was selected. In Table 2 are shown values of these
data for basic setup of the system, used for computation. Labels 𝑑 and ℎ are horizontal and vertical distances of
weapon integration position from mass centrum of vehicle-weapon system, respectively.
Table 2 Weapon data and its basic position

AGL 30 mm M93
Weapon mass with ammunition, 𝑚𝑜 [kg] 38
Horizontal distance of the weapon from the mass centrum, 𝑑 [m] 0.929
Vertical distance of the weapon from the mass centrum, ℎ [m] 1.033
Elevation angle, 𝛼 [°] 30
Azimuth angle, 𝛽 [°] 45
Code in MATLAB uses these data to calculate system mass centrum and solve defined matrixes so generalized
coordinates over time are determined – translation of mass centrum and angular displacements, as well as their
graphical representation. For analysis purposes of recoil process on platform oscillations when firing integrated
weapon, elevation and azimuth angles are varied, and so is the weapon position on the vehicle.
422
Selected angle values for calculations include boundary conditions for firing automatic grenade launcher mounted
on vehicle of choice and an angle between them. Elevation angles are 𝛼 = [0°, 30°, 60°] and azimuth angles are
𝛽 = [0°, 45°, 90°]. For different combination of these angles changes of generalized coordinates are acquired.
Label 𝑧𝑇 (𝑡) represents translation of system mass centrum relative to basic position over time; 𝜑(𝑡) is angular
displacement of the system around its mass centrum in xz plane over time, relative to basic position, so called pitch
and 𝜃(𝑡) is angular displacement of the system around its mass centrum in yz plane over time, relative to basic
position, so called roll. Initial values of the generalized coordinates are 𝑧𝑇 (0) = 0 m; 𝜑(0) = 0°; 𝜃(0) = 0°.
Vital parameters are segregated out of the results and showed in Table 3 for better review and comparison.
Segregated parameters are generalized coordinates amplitudes, time needed to reach amplitude (𝑡𝐴 ) and time
required for system to stops in equilibrium position (𝑡𝑆 ).
Table 3 Computation results of system oscillations for different elevation and azimuth angles
𝛂 𝛃 Coordinates Amplitude 𝒕𝑨 [s] 𝒕𝑺 [s]
𝑧𝑇 [mm] 1018.11 0.36 1.5
0° 𝜑 [mrad] 5.86 0.31 1.26
𝜃 [mrad] 0 0 0
𝑧𝑇 [mm] 1018.09 0.36 1.4
0° 45° 𝜑 [mrad] 4.14 0.31 1.35
𝜃 [mrad] 14.61 0.31 1.35
𝑧𝑇 [mm] 1018 0 0
90° 𝜑 [mrad] 0 0 0
𝜃 [mrad] 20.66 0.31 1.35
𝑧𝑇 [mm] 1012.1 0.35 1.82
0° 𝜑 [mrad] 7.75 0.31 1.35
𝜃 [mrad] 0 0 0
𝑧𝑇 [mm] 1013.9 0.35 1.72
30° 45° 𝜑 [mrad] 5.48 0.32 1.28
𝜃 [mrad] 14.61 0.32 1,28
𝑧𝑇 [mm] 1018 0 0
90° 𝜑 [mrad] 0 0 0
𝜃 [mrad] 17.89 0.32 1.25
𝑧𝑇 [mm] 1007.7 0.35 1.95
0° 𝜑 [mrad] 7.56 0.31 1.25
𝜃 [mrad] 0 0 0
𝑧𝑇 [mm] 1010.8 0.36 1.65
60° 45° 𝜑 [mrad] 5.34 0.31 1.25
𝜃 [mrad] 7.31 0.31 1.25
𝑧𝑇 [mm] 1018 0 0
90° 𝜑 [mrad] 0 0 0
𝜃 [mrad] 10.33 0.31 1.25
With an increase of the azimuth angle, for constant elevation angle, angular displacement 𝜃 (roll) is increasing,
while angular displacement 𝜑 (pitch) and mass centrum translation 𝑧𝑇 are decreasing. Approximately the same
amount of time is required for generalized coordinates to reach their amplitude. Time required for system to stop in
equilibrium position is longest for the translator displacement 𝑧𝑇 .
Value of the vertical translation 𝑧𝑇 and angular displacement in xz plane 𝜑 notably impacts weapon precision, that
is, it can cause shot deviation on target. Data in Table 3 shows that amplitudes of those displacements are
significant. Rate of fire is tightly dependent on time needed for system to stop in equilibrium position, because
every deviation from basic firing position will not give satisfying results. Now it can be concluded that for this type of
integration AGL onto the vehicle, it is better to use lower rate of fire or single fire. Angular displacement in yz
plane 𝜃 creates rotated shot plane, but does not affect precision. It should be noted that these values are larger
than the real ones, because of the constantly maximum value of the force transferred to the platform.
Besides the varying of the elevation and azimuth angles, position of the weapon on the vehicle is also altered
(Table 4). For further computation, medium values of these angles were selected (α = 30° and β = 45°). For
horizontal (𝑑) and vertical (ℎ) distance of the weapon integration point from the point T chosen values are 𝑑 =
423
[0,929 m; 2,787 m; -1,393 m] and ℎ = [1,033 m; 0,619 m; 0,103 m]. While varying horizontal distance, vertical
distance keeps main value, and vice versa (given in Table 2).
Table 4 Computation results of system oscillations for different position of the weapon on the platform
Distance Coordinates Amplitude 𝒕𝑨 [s] 𝒕𝑺 [s]
𝑧𝑇 [mm] 1013.6 0.35 1.62
𝑑 = 2.787 m 𝜑 [mrad] 9.49 0.31 1.3
𝜃 [mrad] 12.65 0.32 1.28
𝑧𝑇 [mm] 1013.9 0.35 1.62
𝑑 = -1.393 m 𝜑 [mrad] 0.75 0.29 1.12
𝜃 [mrad] 12.65 0.32 1.21
𝑧𝑇 [mm] 1006.6 0.35 1.71
ℎ = 0.619 m
𝜑 [mrad] 4.04 0.31 1.09
(𝐻𝑇 = 1010.9 m)
𝜃 [mrad] 7.58 0.31 1.21
𝑧𝑇 [mm] 997.56 0.35 1.85
ℎ = 0.103 m
𝜑 [mrad] 2.25 0.32 1.11
(𝐻𝑇 = 1001.8 m)
𝜃 [mrad] 1.26 0.32 1.21
Vertical distance alteration changes vertical position of mass centrum which should be taken into accounts while
interpreting results, but it can be concluded that when the integration point is closer to the mass centrum, amplitude
of the angular displacement 𝜃 significantly decreases. Also, amplitude of the angular displacement 𝜑 decreases, as
well as the time required for the system to stop in equilibrium position (for angular displacements).
CONCLUSION
Studied in this paper is simplified platform-armament model in order to analyze impact of firing process on light
mobile wheeled platform oscillations. Automatic grenade launcher 30 mm M93 is mounted on the Land Rover
Defender’s rooftop and the goal is that suspension handle additional loads generated when firing weapon, absorbs
them adequately so precision on target is maintained. Low precision occurs as a consequence of unsuitable
system oscillation evoked by force that transfers to the platform. This paper only pertains recoil process during one
fired shot and its impact on system oscillations.
Defined simplified model has three degrees of freedom, mass centrum translation in vertical z axis direction and
rotations around longitudinal x axis and lateral y axis. Vertical, longitudinal and lateral system dynamics is studied
in two planes, which includes motions vital for oscillation analysis. As a result generalized coordinates over time are
generated.
Vertical translation 𝑧𝑇 changes as much as 10 mm for high elevation angles. Angular displacement 𝜑 has the
greatest impact on precision and goes almost up to 8 mrad for zero elevation and azimuth angles, which could
cause projectile to miss target by 8 meters on 1000 meters distance. Angular displacement 𝜃 has the biggest value
for zero elevation and maximum azimuth angle, but only rotates shots, does not impacts precision. Alterations
made to weapon position in vertical direction have more influence on generalized coordinates then changes in
horizontal direction. For weapon position close to the mass centrum vertical translation 𝑧𝑇 practically does not
change, but angular displacements decrease, 𝜃 far more then 𝜑. For a precise hit, at 1000 meters distances,
displacement of 𝜑 = 3 ÷ 4 mrad is too much. As noted before, it shoud be taken into account that those are
maximum values obtained when maximum recoil force was used.
Outcome accuracy depends on quality of the real system simplification. If experimental results are available
comparison with the theoretical model can be conducted and simplified model assessed. At last, practical tests can
give insight how theoretical model could be updated and, in that way, made more realistic.
First step toward more realistic model is to use variable force transferred to the platform over time, not only its
maximum value, and to include process of the recoil mass restoring to starting position in the computation, as well.
Extension of the paper can be creating more complex mathematical model with more degrees of freedom and
separated masses of integrated components. Elastic suspension system could have more complex geometry or
some existing configuration could be acquired. Also, stiffness and damping coefficients could be altered and their
impact on the system oscillations monitored. With some adjustments, upgraded model could be used for other
weapons and platforms.
424
ACKNOWLEDGMENTS
This investigation is a part of the projects TR 33015. We would like to thank to the Ministry of Education and
Science of Republic of Serbia for the financial support during this investigation.
REFERENCES
[1] Jaramaz, S., Micković, D.: "Primena principa komore visokog/niskog pritiska na optimizaciju pogona municije
za automatski bacač granata", Mašinski fakultet, 2001, Beograd,
[2] Uputstvo za korišćenje i održavanje – 30 mm bacač granata automatski, Zastava oružje,
[3] Cvetković, M.: "Unutrašnja balistika", Vojnoizdavački zavod, 1998, Beograd,
[4] Petrović, M.: "Mehanika automatskog oružja", Vojnoizdavački zavod, 2009, Beograd,
[5] Ristić, Z., Jakovljević, M.: "Osnovi naoružanja – konstrukcija artiljerijskog oruđa", Vojna akademija Beograd,
2001, Beograd,
[6] Simić, D.: "Motorna vozila", treće izdanje, IRO „Naučna knjiga“,1988, Beograd,
[7] Rašković, D.: "Teorija oscilacija", treće izdanje, Građevinska knjiga, 1974, Beograd.
425
426
Kragujevac, Serbia
MVM2018-052
Satoru Furugori1
Takeo Kato2
Yoshiyuki Matsuoka3
AUTOMOTIVE SAFETY CONTROL SYSTEM BASED ON

TIMEAXIS DESIGN
ABSTRACT: Timeaxis design is a design concept that incorporates the concept of time axis into design theory
and methodology. Under this design framework, different methods, such as models those who integrate multiple
timescales, those employ identity mapping to describe non-linear and non-steady phenomena and those employ
genetic network programing to describe phenomena evolving gradually as times pass, have been proposed. We
applied Timeaxis design to develop an automotive safety control system that considers risks at a short timescale
(second/minute), medium timescale (hour/day), and long timescale (month/year). This system displays the driving
state on the basis of the state of the driver’s vehicle, surrounding vehicles, and driver’s physical conditions,
generates the vehicle control algorithm on the basis of the driver’s state, and provides driving advice to the driver in
real time. With this system, it is possible to correspond to future safety issues such as problems caused by mixing
of autonomous vehicles and conventional vehicles, problems caused by reduction in driving skill due to automation
of driving and problems caused by the decline in driving ability with aging.
KEYWORDS: timeaxis design, multi-timescale model, identity mapping model, genetic network programming,
automotive safety control system
INTRODUCTION
To support the maintenance of the global environment and the sustainable development of the society and
economy, the theory and methodology of Timeaxis design has been proposed [1-3]. Under this framework,
changes in different factors operating along different time axes are considered, including different circumstances
and a “sense of value" such as interest in driving are incorporated into design. With respect to the automobile
traffic, it is necessary to explore means of transportation that cause less traffic accidents, are moved efficiently, and
generate with fewer CO2 emissions and less environmental pollution. Hurried driving may increase the probability
of accidents, while too much focus on safety increases travel time, CO2 emissions, and environmental pollution.
Therefore, safety, efficiency, CO2 emissions, and the environmental burden of vehicles are in a relation of the
trade-off each other; the prioritization of these factors depends on the individual's purpose of the movement and on
traffic conditions. However, individual drivers cannot judge whether their driving is appropriate given the situation
and circumstances. Therefore, this study aims to propose a new automotive safety control system having the
following three functions: 1) the vehicle itself objectively displays its current state; 2) the vehicle changes its control
to adapt to the requirement of individual drivers; 3) the vehicle advises appropriate driving in the situation and
circumstances. These functions enable safer driving with minimal environmental impact.
1
Satoru Furugori, Graduate School of Science and Technology, Keio University, 3-14-1 Hiyoshi Kohoku-ku
Yokohama 223-8522, Japan, furusa212@gmail.com; Technical Research Center, Mazda Motor Corporation, 2-5
Moriya-cho Kanagawa-ku Yokohama 221-0022, Japan, furugori.s@mazda.co.jp
2
Takeo Kato, Department of Mechanical Engineering, Keio University, 3-14-1 Hiyoshi Kohoku-ku Yokohama 223-
8522, Japan, kato@mech.keio.ac.jp
3
Yoshiyuki Matsuoka, Department of Mechanical Engineering, Keio University, 3-14-1 Hiyoshi Kohoku-ku
Yokohama 223-8522, Japan, matsuoka@mech.keio.ac.jp
427
In conventional automobile development, the technologies for CO2 reduction (e.g., hybrid vehicles and engines with
high fuel economy), safety (e.g., safety support systems and autonomous vehicles), efficiency (e.g., traffic signal
controls and vehicle platooning) and reducing the environmental burden (e.g., electric vehicles and fuel cell
vehicles) have been individually developed. Each technology has been developed with the goal of clearing
standards measured with fixed modes such as safety and CO2 emission, and it is not necessary to consider the
tradeoff between technologies to acquire the certification. Therefore, no technical framework for optimizing the
automotive control in use situations has been studied. However, there is no technical framework for overall
optimization given the purpose of personal movement, so a driver cannot know whether or not his or her type of
driving is appropriate given the situation. Accordingly, it is impossible to know whether the vehicle is appropriately
controlled in the situation and circumstances and whether the driver is performing appropriate driving behavior for
the vehicle control. In this paper, a framework for an automotive safety control system is proposed based on the
theory and methodology of Timeaxis design. In this system, the vehicle displays its current driving state, and the
vehicle switches its means of control according to the driving state while advising the driver on how to safely drive
and on how to lessen the environmental burden.
MODEL USED FOR TIMEAXIS DESIGN

Timeaxis design is a theory and methodology that incorporates the concept of time axes into design. Explicit
manipulation of time axes cannot be accomplished in conventional optimization methods and systems engineering,
and, so far, design theory has not actively addressed the concept of “time axes.” In some cases, designers have
assumed that time axes change the use stage (as a minor consideration), but design principles have not been
explicitly based on this concept. For this reason, it is not possible to intentionally or effectively respond to changes
in the time axes of circumstances, such as variations in usage environments or type of use. In Timeaxis design,
several models are used to incorporate phenomena operating along distinct time axes. At following, we explain the
three Timeaxis design models that we incorporated into the framework of our proposed automotive safety control
system.
Non-steady model
A non-steady model describes elements of non-steady phenomena and their relationships. Many phenomena
during driving are non-steady. Vehicles are equipped with various sensors, and vehicle data such as speed,
accelerator operation, brake operation, steering wheel operation, etc., are measured in real time. The vehicle data
obtained from sensors are non-steady because roads such as expressways, urban roads, winding roads, etc.,
change irregularly, and conditions of congestion also irregularly vary. Also, it may be necessary to arrive early, for
example, or to change the type of driving. The type of driving may additionally change because of changes in
physical conditions. Furthermore, there is the possibility of encountering an unexpected, non-steady state, such as
a traffic accident. Therefore, a non-steady model is necessary to objectively indicate the state in which a vehicle is
currently in. As an example of a non-steady model, we examined a framework to describe driving states using an
identity mapping model, as shown in Figure 1 [4-7]. The identity mapping model is a hierarchical network model in
which the number of input and output layers is the same and the number of units in the intermediate layer is
smaller than the number of units in the input and output layers. In the identity mapping model, learning is performed
such that the input signals and the output signals are equal. The features of the input layer are expressed as the
output of the intermediate layer, which contains fewer features than the input layer. By mapping and displaying the
intermediate layer, it is possible to monitor the time changes of non-steady phenomena.
Input data
Compression
Feature
x1 x2
expression
Restoration
－
Figure 1 Example of a non-steady model (identity mapping model)
428
Plasticity model
A plasticity model describes properties such as gradually changing structures and functions with respect to time
axes. Driving a vehicle entails plastic phenomena including, for example, the improvement of driving ability or
driving skills to generate fewer CO2 emissions. In the plasticity model, such changes are described based on past
driving ability, and it does not return to the past state. As an example of a plasticity model, we used a framework to
generate engine control algorithms according to changes in driving states based on genetic network programming
(GNP), which is an evolutionary computation, as shown in Figure 2 [8-10]. In nature, organisms change genetic
information by selection, crossover, mutation, etc., and adapt to the surrounding environment. Genetic algorithms
(GAs) apply the evolutionary processes of organisms to express genetic information using a bit array structure that
considers genetic manipulation. GNP is an extension of GAs and is based on the representation of a gene using a
network structure instead of a bit array. As evolution progresses, the connection state gradually changes so as to
maximize the objective function.
Processing node
Judgment node
Crossover Mutation
Simulation
Selection
Repeat
generations
Figure 2 Example of a plasticity model (genetic network programming)
Multi-timescale model
The multi-timescale model describes the experience of multiple time axes at the same time. A driver
instantaneously controls a vehicle so as not to collide with surrounding vehicles and may also perform driving
considering daily CO2 emissions. In addition, current driving behavior may change as a result of the development of
driving skills or long-term changes in driving skills. Therefore, current driving behavior is considered to be the result
of combining these controls. As shown in Figure 3, the multi-timescale model is based on the integration of different
timescales. In addition, the multi-timescale model considers a hierarchy of time-axis scales, wherein the short
timescale corresponds with seconds or minutes, the medium timescale with hours or days, and the long timescale
with months or years. This model is effective for designing each timescale and for integrating the design of
timescales by considering their relationships.
Long timescale Month/year

model
Hour/day
Medium timescale
model
Short timescale Second/minute

model
Figure 3 Example of a multi-timescale model
APPLICATION OF TIMEAXIS DESIGN TO THE AUTOMOTIVE SAFETY

CONTROL SYSTEM
The automotive safety control system proposed in this paper is aimed to provide future transportation for society by
around 2030. In this future transportation society, autonomous vehicles and manual conventional vehicles will
coexist. Electric vehicles, hybrid vehicles, and internal combustion locomotives will also be mixed. Vehicles
429
equipped with the proposed system will change controls according to the desires of individual drivers, such as safe
driving, good fuel economy, or low electricity cost-based driver’s driving history. For drivers who feel a sense of
burden in driving, autonomous driving can be set as the dominant control. Meanwhile, drivers who want to enjoy
driving or to improve driving skill can be given responsive engine control. Furthermore, driving advice can be
provided according to individual driving characteristics. The proposed system can be installed in vehicles as a core
module. Next, the three functions of the core module will be described using analysis data. The driving data used
for this analysis were taken in 300 seconds at intervals of 0.1 seconds using a driving simulator. The data were
used once for model learning and 10 times for verification.
Driving state display based on identity mapping model

The driving state display is a function that instantaneously grasps current driving conditions, including approaching
dangers to the vehicle, etc., at the moment that a driver is operating a vehicle. To instantaneously display the
driving state in an easy-to-understand manner, it is necessary to compress the driving data measured by many
vehicle parameters and to display features with only a small number of parameters. Therefore, the identity
mapping model for compressing non-steady data was used to extract features from the driving data. Figure 4
shows the driving state extracted from features. The inputs for the vehicle parameters are vehicle speed and its
differential, inter-vehicle distance and its differential, and brake pedal angle and its differential. The number of
intermediate layers was set to 2, and the resulting map was displayed. Here, based on the relative relationship with
the preceding vehicle, the driving state was classified into “collision,” “proximity,” “approach,” “leave,” follow,” and
“free.” As time elapses, the driving state value moves within this map; the same driving states are located within the
same vicinity. Several boundaries were described as straight lines, which classify several driving states. In this
example, the features of the driving state with respect to the distance to the preceding vehicle are shown, but it is
also possible to display the state of the behavior of the vehicle itself and the risk state of the driver at the same
time. Thus, a framework for understanding the driving state based on a two-dimensional map was constructed.
Free
Intermediate neuron 2
Approach
Follow
Collision
Leave
Proximity
Intermediate neuron 1
Figure 4 Mapping of the driving state based on an identity mapping model
Vehicle control based on genetic network programming

This vehicle control is a function oriented toward safe, fuel-efficient, and environmentally friendly driving by
switching to a vehicle control that is adapted to individual driving and based on the current driving state. There is a
possibility that the instantaneously acquired driving data may indicate a driving state that has never been
experienced. In such a case, it is necessary to create a new vehicle control algorithm. So, GNP was used to create
a new vehicle control algorithm. Figure 5 shows an example of the vehicle control algorithm that emerged. GNP
derived an optimal engine control algorithm that minimizes travel costs (fuel and battery costs) when new driving
states, such as flat roads, ascending roads, descending roads, or congested roads, are detected, and it was
confirmed that the travel cost was reduced by switching between these control algorithms. Thus, a framework of
vehicle control was constructed that is created for driving states where no control logic is available. In this scenario,
a new control algorithm emerges, and the vehicle control switches to the optimal algorithm.
430
Processing node Judgment node
Delay node Flag node
yes
regT = 40 1s yes
SOCB < 0.5 no start acc > 0.5
WE = 4
yes no
WE = 32 Δacc < -0.2 no acc < 0.2

yes no 3.5s
no 1s
start v > 100 WE = 0 no
WE = 30 no yes
bra = 0
Δbra < -0.8 yes
yes yes
WE = 26 yes no
yes yes Δacc > 0.2 WE = 0 acc > 0.2
acc < 0.2
WE=28 WE = 32 acc < 0.2 yes
no 3s
no
WE=30 SOCB > 0.65
no
no 3s no
yes |i| > 30 WE = 14
yes
2.5s WE = 70 no
3.5s
regT = 0 Δacc < -0.4 SOCC < 0.5
yes
no
Flat road Descending road

start
yes WE = 0
no
I < -300
acc < 0.2 yes
no WE = 0
regT = 400
regT = 300 no
regT = 320 no
WE=26 I > 300 Δacc > 0.4 no
start 0.5s
yes
acc > 0.2
acc > 0.2
yes
I < -900 1.5s no yes yes
no
no
yes no yes 1.5s no
regT = 80
Δbra < -0.8 WE = 30
regT = 300 yes
acc > 0.25 yes no
WE = 44 WE = 0 SOCB > 0.7
yes no acc < 0.2 no
yes Δbra > 0.2
no yes
acc < 0.2 no v > 20 yes
no bra = 0 yes
yes yes
yes acc > 0.25 I > 900 no
no no
Δacc < -0.6
I > 1500
yes
Ascending road Congested road

Figure 5 Development of the engine control algorithm and switching
Timing of advice based on changes in driving state

Driving advice is presented at the time when the driving state changes and is associated with vehicle behavior.
Figure 6 shows an example wherein the timing of driving advice is displayed on a map of inter-vehicle distance and
relative speed. As the vehicle moves, the driving state value moves on the map and crosses the boundary lines
that dictate the timing of advice. The position of the boundary lines varies depending on the individual driver, and
knowledge is acquired via learning from driving data. Ultimately, the boundary lines of the driving state display
shown in Figure 4 indicate the timing of advice.
Driving advice timing

0.15
0.110
0.1
Free
Leave
. Collision
Relative speed L
0.05
0.025
0 Follow
-0.025
-0.05
-0.1 Proximity
-0.15 Approach
-0.190
-0.2
-0.100 0.100 0.580 1.000
-0.25
-0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
Inter-vehicular distance L
Figure 6 Setting of advice timing for the driver
FRAMEWORK FOR PROPOSED AUTOMOTIVE SAFETY CONTROL SYSTEM

The three functions described in the previous chapter were integrated using a multi-timescale model. Figure 7
shows the framework of the proposed vehicle safety control system. At the short timescale, the real-time display of
the driving state, the vehicle control, and the notification of driving advice to the driver at a set time are performed.
First, the driving state is estimated and displayed based on vehicle data and data from the surrounding
environment. When the running state is known, the corresponding control algorithm is selected to control the
vehicle. Furthermore, advice is given to the driver at a set time. If the driving state is unknown, an approximate
driving state is searched for, passing the vehicle data and the surrounding environment data to the middle
timescale while continuing to display the control and advice in real time. At the middle timescale, a control
431
algorithm corresponding to an unknown driving state is developed. Next, the vehicle data, the surrounding
environment data, and the control algorithm are passed along to the long timescale. At this scale, the driver's
desires with respect to the level of safety and the cost of travel are estimated, generating advice so that the driver's
behavior changes to meet his or her desires based on the analysis of long-term records of the driving state.
Short timescale (sec, min)
input Vehicle data

Surround data
Medium timescale (hour, day)
Unknown driving state
Estimation of Emerging algorithm
driving state
Known Genetic network
driving state programing
Selection of vehicle Selection of Decision of vehicle control way
control algorithm approximate
Long timescale (month, year)
driving state
Conversion to Memory
control
Threshold of driving state
Vehicle control Vehicle control algorithm
output
Driving advice Decision of driving advice timing
Display of vehicle state
Figure 7 Framework of the vehicle safety control system based on a multi-timescale model
DISCUSSION
Issues that are important for transportation in future society were considered in the vehicle control system
proposed herein. These issues include risk resulting from the mixture of conventional vehicles and autonomous
vehicles, risk resulting from the decline in driving skills following the introduction of autonomous driving, and risk
resulting from the decline in driving ability with aging. Regarding the first risk, the proposed system will be able to
distinguish whether a vehicle is autonomous or conventional by learning the features of the type of driving. An
autonomous vehicle may obstruct smooth traffic to comply with traffic rules and increase the safety margin. When
such a vehicle is encountered, smooth and safe driving can be achieved based on driver judgment of whether to
interrupt the control and, for example, move ahead of another vehicle or give the right-of-way. Regarding the
second risk, the proposed system is designed to improve according to an individual's growth regardless of whether
a driver prefers autonomous driving or wishes to improve his or her skill and enjoys driving. Therefore, the vehicle
control and driving advice will change according to an individual's growth, compensating for decline in driving skill
as well. Regarding the third risk, the proposed system considers the life cycle of a vehicle from purchase until
disposal as well as past travel history, which can be transferred to a new vehicle. Upon accumulating driving history
from a beginning point in time until a latter point in time as a driver ages, it is possible to cope with changes in
driving ability due to aging.
One future task for completing the automotive safety control system is the consideration of an algorithm for
automatically changing the boundary of the driving state with respect to the driving state display based on the
identity mapping model. Also, in accordance with the growth of individuals, it is necessary to develop artificial
intelligence (AI) that can judge the type of driving that a driver is oriented toward.
CONCLUSIONS
We applied the theory and methodology of Timeaxis design to propose a framework for an automotive safety
control system that may be placed into effect around 2030. The proposed system consists of a display of the
driving state, algorithms for determining vehicle control, and functions for providing driving advice. The use of a
framework based on a multi-timescale model consisting of a short timescale, medium timescale, and long timescale
enables trade-off between problems at different scales, such as efficiency and safety. Furthermore, by designing
the system as a core module to be mounted on the vehicle, even when changing the vehicle, the past travel history
so far is taken over by switching the core module. As a result, it is possible to make safety control and driving
advice in consideration of long-term changes such as a change in driving ability due to aging. In addition, since the
system is designed to improve according to individual’s growth, it is not necessary for the manufacturer to add both
the autonomous vehicle and the safety support vehicle to the product line up. Manufacturers need to provide a
common architecture of safety vehicles that combines both functions and changes with learning.
432
ACKNOWLEDGMENTS
I appreciate the support from Keio University research scholarship.
REFERENCES
[1] Matsuoka, Y.: "Design Science for Product Creation X Product Usage", Kindai-Kagaku-Sha, 2017, pp 43–55,
[2] Arita, M., Kikuchi, M., Kato, T., Hui, P., Matsumoto, A., Sato, K., Kato, T., Matsuoka, Y.: "Development of Basic
Architecture for Health Monitoring Seat System", Journal of the Science of Design, Vol. 1, No. 2, 2017, pp 71-
80,
[3] Kamiya, K., Kito, A., Alvarez, J., Sato, K., Nishimura, H., Matsuoka, Y., Furugori, S.: "Timeaxis Design of a
Service System Growing Values of Mobility Using the M-V Model", Design, User Experience, and Usability.
Experience Design for Everyday Life Applications and Services, Third International Conference, DUXU 2014,
Heraklion, Crete, Greece, June 22-27, 2014, pp 281–292,
[4] Chalodhorn, R., MacDorman, K, Asada, M.: "Humanoid Robot Motion Recognition and Reproduction",
Advanced Robotics, Vol. 23, No. 3, 2009, pp 349-366,
[5] Kramer, M.: "Nonlinear Principal Component Analysis Using Autoassociative Neural Networks", AIChE Journal,
Vol. 37, No. 2, February 1991, pp 233–243,
[6] Cottrell, G., Munro, P., Zipser, D.: "Image Compression by Back Propagation: An Example of Extensional
Programming", Advances in Cognitive Science, Vol. 3, 1988, pp 208-240,
[7] Yoshimura, H., Shimizu, T., Isu, N., Sugata, K.: "Construction of Noise Reduction Filter by Use of Sandglass-
Type Neural Network", IEICE Trans. Fundamentals, Vol. E80-A, No. 8, August 1997, pp 1384-1390,
[8] Mabu, S., Hirasawa, K., Hu, J.: "A Graph-Based Evolutionary Algorithm: Genetic Network Programming and Its
Extension Using Reinforcement Learning", Evolutionary Computation, Vol. 15, No. 3, Fall 2007, pp 369–398,
[9] Mabu, S., Hirasawa, K., Hu, J., Murata, J.: "Online Learning of Genetic Network Programming and Its
Application to Prisoner’s Dilemma Game", Transactions of IEE Japan, Vol. 123-C, No. 3, 2003, pp 535-543,
[10] Katagiri, H., Hirasawa, K., Hu, J.: "Network Structure Oriented Evolutionary Model: Genetic Network
Programming", Transaction of the Society of Instrument and Control Engineers, Vol. 38, No. 5, 2002, pp 485-
494.
433
434
SECTION
Driver/Vehicle Interface,
Information and
Assistance Systems
 Communication & Informatics

 Noise & Vibration
 Human Comfort & Thermal
Management
 Vehicle Security & Occupant
Protection
 Driver Assistance &
Information Systems
435
436
Kragujevac, Serbia
MVM2018-003
Valentina Golubović-Bugarski1
Snežana Petković2
Željko Đurić3
Goran Jotić4
VIBRATION COMFORT OF THE VEHICLE EXPRESSED BY

SEAT EFFECTIVE AMPLITUDE TRANSMISSIBILITY
ABSTRACT: Research of the human body vibrations, carried out under controlled laboratory conditions, shows
that human body is the most sensitive to vibrations in the frequency range that matches the biomechanical resonance.
In the vertical direction, the resonance of the body is approximately 5 Hz, while in the horizontal direction the
resonance occurs at frequencies less than 2 Hz. The vibrations of the vehicle have been transferred to the driver and
passengers over the seats, which have the ability to attenuate or to amplify vibrations which human body is exposed
to while driving. One way to determine the vibration behaviour of the seat is to measure the SEAT (seat effective
amplitude transmissibility) factor, which represents the ratio between the vibrations measured on the seat and
vibration measured directly on the floor under the seat. Measurement of vibrations in these two positions must be
performed simultaneously. If the value of SEAT is less than 1, a seat attenuates vibrations and meets vibrational
comfort, the value of SEAT greater than 1 indicates that a seat amplifies vibration, reducing vibration comfort. This
paper gives results of SEAT factor investigation done on a hybrid vehicle, for different types of road surface and
different modes of driving (electric power and internal combustion engine).
KEYWORDS: vibration comfort, seat effective amplitude transmissibility SEAT, hybrid vehicle
INTRODUCTION
Seat is a primary point of contact between the human body and the vehicle structure which generates the vibrations,
therefore its role in the isolation of the whole-body vibrations is a very important. Proper design of a seat contributes
largely in reduction of the vibration levels. Seat should be designed in such a way that the driver can drive the vehicle
safely and effectively, it should have strength enough to protect the occupant in the event of an accident, it should
have good static and dynamic comfort properties [11]. It is important to point out that the seat which is better in termn
of static comfort may be worse in the case of dynamic comfort. Figure 1 shows Ebe’s model of seat discomfort which
includes a consideration of both static and dynamic factors [3, 11]. This model shows overall comfort characteristics
of two different seats, one being good at rest and the other being good for driving case. In first case, static comfort is
high and dynamic comfort is less. However in second case, initial static comfort is less but its dynamic comfort is
high. Therefore, seat „a“ will be better for environments with large magnitudes of vibration but seat „b“ will be better
for environments with low magnitudes of vibration. It is up to the manufacturers to optimize the comfort for both cases.
1
Prof. dr Valentina Golubović-Bugarski, University of Banja Luka, Faculty of Mechanical Engineering, Stepe
Stepanovića 71, 78000 Banja Luka, BiH, valentina.golubovic-bugarski@mf.unibl.org
2
Prof. dr Snežana Petković, University of Banja Luka, Faculty of Mechanical Engineering, Stepe Stepanovića 71,
78000 Banja Luka, BiH, petkovic1961@gmail.com
3
Željko Đurić, Ma, University of Banja Luka, Faculty of Mechanical Engineering, Stepe Stepanovića 71, 78000
Banja Luka, BiH, zeljko.djuric@mf.unibl.org
4
Goran Jotić, Ma, University of Banja Luka, Faculty of Mechanical Engineering, Stepe Stepanovića 71, 78000
Banja Luka, BiH, goran.jotic@mf.unibl.org
437
Figure 1 Ebe’s model of the overall seat discomfort [3, 11]
Investigation of seat comfort showed that the ideal seat in any situation is a function of the seat static properties, the
seat dynamic properties, the vibration to be controlled, and the human perception characteristics. It can be found in
[11] that investigation of static parameters was done, for example driver posture [14, 13], contact pressures [15, 7],
and thermal properties [12]. When it comes to dynamic comfort, laboratory studies of whole-body vibration have
established a relationship between the magnitude, duration, frequency content, and waveform of the signal. The
nature of the seat dynamic response is much complex than its static performance. The dynamic response of the seat
is influenced by the dynamic properties of the human body, so the real occupant cannot be replaced by an inert mass
as in static investigation. The transmissibility of a seat measured with one occupant will be slightly different when
tested with a different occupant [17]. Also, the transmissibility of the seat measured while driving on a smooth surface
will be slightly different when tested on a rough surface [4]. Present studies are mostly devoted to investigation of
damping properties of the suspension seats in working machines, such as tractors [1, 5] and construction machines
[2, 10].
THE EFFECTS OF VIBRATION ON COMFOR AND PERCEPTION

Thresholds of perception for continuous whole-body vibration vary widely among individuals. Approximately half of
the people in a typical population, no matter standing or seated, can perceive a vertical weighted peak acceleration
of 0.015 (m/s2). A quarter of the people would perceive a vibration of 0.01 (m/s2), but the least sensitive people would
only be able to sense a vibration magnitude of 0.02 (m/s2) or more, as it is described in ISO 2631-1 [8]. Regarding
the vibration comfort in vehicles, a particular vibration condition may be considered to cause unacceptable discomfort.
Many factors combine to determine the degree to which discomfort may be noted or tolerated. Acceptable values of
vibration magnitude for comfort depend on passenger expectations with regard to trip duration, type of passengers
activities during trip (reading, eating, writing) and many other factors (acoustic noise, temperature, etc.). However,
comfort reactions due to vibration amplitudes may be classified as follows [8]:
 Less than 0,315 m/s2 not uncomfortable
 0,315 m/s2 to 0,63 m/s2 a little uncomfortable
 0,5 m/s2 to 1 m/s2 fairly uncomfortable
 0,8 m/s2 to 1,6 m/s2 uncomfortable
 1,25 m/s2 to 2,5 m/s2 vary uncomfortable
 Greater than 2 m/s2 extremely uncomfortable
SEAT EFFECTIVE AMPLITUDE TRANSMISSIBILITY FACTOR

The vibration isolation efficiency of a seat may be expressed by means of The Seat Effective Amplitude
Transmissibility factor (SEAT) [11]. This dimensionless factor is dependent on the vibration spectrum, seat
transmissibility and subject respone frequency weighting. It shows the capability of a seat design to attenuate the
vibrations generated in a vehicle, i.e. to protect the driver from excessive vibrations. The SEAT value is defined as:
ride comfort on seat

SEAT% = 100 × (1)
ride comfort on floor
A seat would improve ride comfort when SEAT factor is smaller than 1 or, expressed in percentage, when SEAT%
is less than 100%. If the value exceeds these limits, the seat actually amplifies vibrations and, thus, worsens ride
comfort.
438
According to ISO 10326-1 Mechanical vibration – laboratory method for evaluating vehicle seat vibration [9], the
measurement of the SEAT involves determination of the vibration magnitudes at two positions, as depicted in Figure
2:
 on the seat pan, which represents the interface between the human body and the seat,
 at the base of the seat, that is at the place of vibration transmission to the seat, Figure 2.
Figure 2 Graphical representation of the SEAT measurement
Measurements at these two points are done simultaneously and the SEAT is computed as the ratio between these
two magnitudes. The magnitude of the vibrations is measured in terms of weighted acceleration root mean square
values (r.m.s) or vibration dose values (VDV). Therefore, the SEAT expression becomes either:
a w,seat
SEATrms % = 100 × (2)
a w,floor
VDVseat
SEATVDV % = 100 × (3)
VDVfloor
The r.m.s. is expressed in (m/s2) and defined as:
1
1 T 2
aw =   a2w  t  dt  (4)
T 0 
where:
aw  t  is the frequency weighted acceleration (using Wk weighting) as a function of time in (m/s2),
T is the duration of measurement in (s).
The VDV is expressed in (m/s1,75) and defined by the relation:
1
T 4
VDV=   a4w  t  dt  (5)
 0 
where:
aw  t  is the frequency weighted acceleration (using Wk weighting) as a function of time in (m/s2),
T is the duration of measurement in (s).
Whether to use r.m.s. or VDV depends on the vibrations encountered during the measurement. If the vibration history
is rather smooth, then RMS vibration magnitude is preferable. If, however, the vibrations included transients and
shocks, it is recommended to compute SEAT based on VDVs.
The SEAT factor measurement belongs to the category of whole-body vibrations measurement. The way in which
humans perceive vibration depends on different factors, including the vibration frequency content and direction.
People are most sensitive to whole-body vibration within the frequency range of 1 to 20 Hz, but there are different
human sensitivities to vibration in different directions of excitation. When measuring whole-body vibration at the seat,
ISO 2631 requires the use of Wk weighting in the Z-direction, whereas Wd is used for the acceleration in the X- and
Y-directions.
439
When assessing a seats ability to attenuate vibrations, it is important to keep in mind that seat and driver must be
seen as one system. The driver will add mass to the seat, which preloads the seat springs, and changes the
resonance behaviour. Further, depending on posture, the seat-driver combination will lead to a more or less stiff
system (e.g., vibrations will be different if the driver sits relaxed or if feet are pressed against the floor). Thus,
depending on the driver’s body and posture, the performance of seats can be very different. What constitutes a
“good” SEAT value depends on the vehicle type [6, 14]. For example, vibration in cars usually has substantial
components at about 10 Hz, which can be easily isolated by conventional seats. Car seats therefore often have
SEAT values in the range of 60 to 80%.
MEASUREMENT METHODOLOGY
The research task in this paper was to evaluate the vibration (dynamic) comfort of a hybrid vehicle by measurement
of the Seat Effective Amplitude Transmissibility factor (SEAT). A hybrid vehicle efficiently combines the internal
combustion engine (ICE) and electric power (EV) from the battery.
In this investigation, Toyota C-HR (1,8 l Hybrid Petrol) was used, Figure 3, having performances as follows: engine
power 98 Hp or 72/5200 kW/rpm, max torque 142/3600 Nm/rpm; electric motor power 53 kW, max torque 163 Nm,
max voltage 600 V; battery type NiMetalHibrid, nominal voltage 201,6 V, number of battery modules 28, battery
capacity 6,5 Ah.
Figure 3 a) test car, Toyota C-HR, b) energy monitor in the car showing driving mode
The final goal of investigation was to find out how vibration comfort depends on the type of vehicle power mode, i.e.
whether driving in electric mode provides better vibration comfort compared to driving in internal combustion engine
mode. Also, our intention was to find out the effect of the road surface quality to the vibration comfort, the influence
of the driving speed, as well as the influence of the drivers weight.
The measurements were made by driving the car on the four roads, with different conditions:
 the new motorway (Laktaši – Gradiška), Figure 4a: straight, well maintained smooth road surface without
any damage; the car was driven at about 50-60 km/h in EV mode and 70-120 km/h in ICE mode; the drivers
weight was about 100 kg.
 the fast road (Banja Luka-Laktaši): straight and little bumpy due to ruts; the car was driven at about 50 km/h
in EV mode and 70-100 km/h in ICE mode; the drivers weight was about 100 kg.
 the city street (in Banja Luka): well-maintained surface of the road, but the driving was with a lot of slowing
down and braking, at about 40-50 km/h in EV and 60 km/h in ICE mode; two drivers, A (100 kg) and B (60
kg), participated in the experiment.
 the suburban street (in Banja Luka), Figure 4b: the road surface is in a poor condition and characterized by
a series of pot-holes and bumps; the car was driven at about 40-50 km/h in EV mode and 60 km/h in ICE
mode; two drivers, A (100 kg) and B (60 kg), participated in the experiment.
Figure 4 Roads where measurements were done. a) motorway, b) suburban street
Each of the measurement sessions durated about 3 minutes. As mentioned before, for evaluation of the SEAT factor
it is necessary to simultaneously measure the vibration signals on the seat and on the vehicle's floor. To perform two
channels measurement, we used following equipment, Figure 5:
440
 Human Vibration Analyser type 4447,
 Seat pad, with built in triaxial accelerometer type 4506,
 Uniaxial accelerometer type 4507 (all by Bruel&Kjaer).
Figure 5 a) The analyzer 4447, b) seat pad strapped to the cushion, c) uniaxial accelerometer on the floor
The Human Vibration Analyser Type 4447 complies with the technical requirements of ISO 8041:2005 Human
response to vibration - Measuring instrumentation, and can perform measurement compliant with the standards
pertaining to human vibration [16]. The instrument possessed the pre-set frequency weightings for hand-arm (Wh)
and whole-body (Wd and Wk) vibrations. Uniaxial and triaxial accelerometers are the piezoelectric vibration sensors
which capture the measured signals and deliver it to the analyser for processing. The uniaxial accelerometer is
mounted on the floor, at the place of the vibration transmission to the seat, Figure 5c. The accelerometer is best
mounted on a rigid part of the floor using glue, a strong magnet or double-sided thin adhesive tape. In this
investigation, we used double-sided adhesive tape. The triaxial accelerometer is mounted in the seat pad which is
located on the seat cushion, Figure 5b. It is necessary to tape or strap it to the cushion in such a way that the
accelerometer is located midway between the ischial tuberosities of the seat occupant. The analyser 4447 is only set
up to evaluate SEAT in the vertical direction, i.e. it compares the vertical vibration at the floor (vibration signal marked
by Auxiliary) with vibrations along the Z-axis of the Seat Pad (vibration signal marked by Z), as depicted in Figure 2.
The whole-body weighting Wk is applied for the Z and the Auxiliary signals when carrying out SEAT measurements.
DISCUSSION OF THE MEASUREMENT RESULTS

The r.m.s. vibration magnitudes (expressed in m/s2) measured along the Z-axis of the Seat Pad (marked by Z) and
at the vehicle’s floor (marked by Aux), together with the values of the SEAT factor are presented in the Table 1.
Table 1 Results of vibration measurement at the seat and at the floor of the vehicle and SEAT factor
Drive r.m.s.
Velocity (m/s 2
) SEAT
Type of road mode/ SEAT
(km/h) %
driver Z Aux
1 55 0,1253 0,1631 0,7684 76,84
2 EV / A 60 0,1208 0,1640 0,7363 73,63
3 60 0,1137 0,1591 0,7149 71,49
new motorway
4 70 0,1221 0,1710 0,7142 71,42
5 ICE / A 90 0,1441 0,1995 0,7225 72,25
6 110 0,1830 0,2392 0,7652 76,52
7 EV / A 60 0,1759 0,2181 0,8066 80,66
8 fast road ICE / A 70 0,2028 0,2606 0,7784 77,84
9 ICE / A 90 0,2768 0,3429 0,8071 80,71
10 EV / A 60 0,2091 0,2888 0,7239 72,39
11 city street ICE / A 50 0,2569 0,3327 0,7721 77,21
12 EV / B 40 0,3116 0,4034 0,7726 77,26
13 EV / A 35 0,4102 0,5593 0,7334 73,34
14 suburban street EV / B 40 0,4826 0,6443 0,7491 74,91
15 ICE / A 45 0,4305 0,5726 0,7518 75,18
Analysing the measurement results, the SEAT factor takes values between 71.42 % and 80.71%. For all
measurement conditions, it can be seen that the increase of the vibration amplitude measured on the vehicle’s floor
results in increase of the vibration measured on the seat, giving the Seat Effective Amplitude Transmissibility average
441
value of 75.44%. This means that 75.44% of vibration generated by driving is transmitted from the vehicle's structure
to the driver's seat.
It is interesting to analyse vibration amplitudes measured at the seat and on the floor of the vehicle for different driving
conditions. For the approximate values of driving speed (40-60 km/h for EV; 45-70 km/h for ICE), vibration amplitudes
are lowest for measurements done along the new motorway. This is expected due to the excellent quality of the road
surface, which is well maintained and smooth, and the road is relatively in a straight line, with minimum curvatures.
The level of vibration is noticeably increasing with the decline in the quality of the driving surface (motorway / fast
road / city streets / suburban streets), for both drive modes, EV (Figure 6a) and ICE (Figure 6b).
a) EV drive mode
1.00
0.81
0.77 0.75
0.80 0.71
0.64
0.60 Z _ rms
m/s2
0.48
0.40 Aux_rms
0.40
0.22 0.31
0.16 0.18 SEAT
0.20 0.11
0.00
motorway fast road city street suburban street
b) ICE drive mode

1
0.77 0.77 0.75
0.8 0.71
0.57
0.6 Z _ rms
m/s2
0.43
0.4 0.33 Aux_rms
0.26
0.17 0.25 SEAT
0.2 0.2
0.12
0
motorway fast road city street suburban street
Figure 6 Vibration amplitude values (Z and Aux) and SEAT factor for different type of roads:
a) EV drive mode, b) ICE drive mode
Comparing the vibration values measured for the same type of road and drive mode, but different driving speeds,
one can conclude that there is no significant difference between vibration values for EV mode (due to insufficient
difference of driving speeds, 55 – 60 km/h). However, regarding ICE drive mode, it can be seen that vibration values
increase as driving speed increases, Figure 7a (driving on the motorway) and 7b (driving on the fast road).
a) motorway / ICE drive mode

0.76
0.8 0.71 0.72
0.6
m/s2
0.4 Z _ rms
0.23
0.17 0.19
0.2 0.18 Aux_rms
0.12 0.14
SEAT
0
70 90 110
driving speed km/h
442
b) fast road / ICE drive mode
0.9 0.8 0.81
0.77
0.8
0.7
0.6
0.5
m/s2
0.4 0.34 Z _ rms
0.26 0.27
0.3 0.21
0.17 0.2 Aux_rms
0.2
0.1
SEAT
0
60 70 90
driving speed km/h
Figure 7 Vibration amplitude values (Z and Aux) and SEAT factor for ICE mode and different driving speeds:
a) driving on the motorway, b) driving on the fast road
Comparing the vibration values measured for the same type of road and approximately equal driving speed, but for
the different drive modes, it can be concluded that EV mode produces lower vibration than driving in ICE mode, as
shown in Figure 8a (driving on the motorway), 8b (driving on the fast road) and 8c (driving on the city street).
a) motorway / driving speed 60-70 km/h

0.8 0.71 0.71
0.6
m/s2
0.4 Z _ rms
0.15 0.17
0.2 0.11 0.12 Aux_rms
0 SEAT
EV ICE
drive mode
b) fast road / driving speed 60-70 km/h

1 0.8 0.77
0.8
0.6
m/s2
Z _ rms
0.4
0.17 0.21 0.2 0.26 Aux_rms
0.2
0 SEAT
EV ICE
drive mode
c) city street / driving speed 60 km/h

1
0.72 0.77
0.8
0.6
m/s2
Z _ rms
0.4 0.33
0.2 0.28 0.25 Aux_rms
0.2
0 SEAT
EV ICE
drive mode
Figure 8 Vibration amplitude values (Z and Aux) and SEAT factor for different drive modes: a) driving on the
motorway, b) driving on the fast road, c) driving on the city street
443
Regarding the influence of the driver’s weight, it can be seen that greater driver’s weight (A - 100 kg, B - 60 kg)
produces lower vibration transmission to the seat pan, Figure 9a (driving on the city street) and 9b (driving on the
suburban street).
1
a) city street / EV / driving speed 50 km/h
0.77 0.77
0.8
0.6
m/s2
Z _ rms
0.4
0.4 0.33 0.31
0.25 Aux_rms
0.2
SEAT
0
A B
driver's weight
b) suburban street / EV / driving speed 40 km/h

0.730 0.740
0.80 0.64
0.56
0.60 0.48 Z _ rms
0.41
m/s2
0.40
Aux_rms
0.20
SEAT_rms
0.00
A B
driver's weight
Figure 9 Vibration amplitude values (Z and Aux) and SEAT factor for different driver’s weight:
a) driving on the city street, b) driving on the suburban street
CONCLUSIONS
An investigation of the vibration comfort of a hybrid vehicle Toyota C-HR was done in this paper. A total of 16
measurements were performed and the measured Seat Effective Amplitude Transmissibility factor, expressing
vibration seat comfort, was in the range between 71.42% (vehicle rides over new motorway with smooth road surface)
and 80.71% (vehicle rides over fast road with a lot of ruts). Although the measured SEAT values varied within 10%,
the average value of the SEAT for all performed measurements is around 75%, which can be considered as a good
dynamic comfort estimation for this category of vehicles (passenger cars). Vibration amplitudes measured on the
vehicle’s floor and at the driver's seat show that the vibration comfort can be rated as "not uncomfortable" (vibration
amplitude less than 0.315 m/s2), except for driving over the suburban streets where the vibration comfort is rated as
"a little uncomfortable" (for vibration amplitude between 0.315 m/s2 to 0.63 m/s2).
Regarding the measured vibration amplitudes on the vehicle's floor and the driver's seat, we can conclude
following:
 The vibration amplitude depends on the type of road the vehicle is riding on and the quality of its surface.
The best vibration comfort is achieved by driving on a motorway where the road surface is smooth, with no
damages. This applies to driving modes, electrical power and internal combustion engine mode. For some
environments, a major contributor to the vibration is the roughness, or tidiness, of road. So, keeping the road
in good condition and repairing the damages would improve the mechanical (vibration) environment for the
vehicle.
 Increasing the driving speed increases the vibration amplitudes, which reduces dynamic comfort. For many
vehicles, lower vibration is received by passengers if the speed of the vehicle is limited.
 At the same driving conditions, i.e. the same type of road and approximately the same speed, EV drive mode
gives a lower vibration level (for about 15 %) than drive in ICE mode. Even though the engine of the vehicle
can be the main contributor to the vibration exposure, we concluded that the lower vibration generated in EV
mode is associated with the lower driving speed, since the hybrid vehicle activates the EV mode at lower
driving speeds.
 At the same driving conditions (the same type of road, the same driving mode and approximate speed), it
was shown that the driver of the lower weight was exposed to higher vibrations. Conventional foam seats,
such as car seats, have a resonance at 4 to 5 Hz, which coincides with the frequency where the human body
is most sensitive to vibration. However, the drivers weight generally has only small effect to vibration
444
response. Several other factors, such as subject size, posture, backrest contact and backrest angle, foot-
support position and support for the arms can have significant effects on seat transmissibility.
Commonly cars are purchased on the basis of comfort evaluation just in the showroom, whereby attention is paid
only to static comfort, neglecting the dynamic and temporal aspects of seating comfort. It is much better for
purchasers to test-drive the vehicle for an extended period of time and evaluate the vehicle's dynamic comfort, which
is especially important for high-mileage business drivers.
REFERENCES
[1] Aisyah, A.S., Nawal, A.A.J.: "Vertical Suspension Seat Transmissibility and SEAT Values for Seated Person
Exposed to Whole-body Vibration in Agricultural Tractor, Preliminary Study", Procedia Engineering, Vol. 170,
2017, pp 435 – 442,
[2] Crenna, F., Rossi, G.: "Measurements for the evaluation of vibration exposure of operators in a ship container
terminal", XIX IMEKO World Congress, Fundamental and Applied Metrology, Lisbon, Portugal, 2009, pp 2537-
2542,
[3] Ebe, K., Griffin, M.J.: "Qualitative models of seat discomfort including static and dynamic factors", Ergonomics,
Vol. 43, No. 6, 2000, pp 771–790,
[4] Fairley, T.E., Griffin, M.J.: "The apparent mass of the seated human body in the fore-and-aft and lateral
directions", Journal of Sound and Vibration, Vol. 139, No. 2, 1990 pp, 299–306,
[5] Gialamas, T., Gravalos, I., Kateris, D., Xyradakis, P., Dimitriadis, C.: "Vibration analysis on driver’s seat of
agricultural tractors during tillage tests", Spanish Journal of Agricultural Research, Vol. 14, No. 4, 2016,
[6] Griffin, M.J.: "Fundamentals of human responses to vibration”, In Fundamentals of Noise and Vibration, London:
E&FN Spon, 1998,
[7] Gyi, D.E., Porter, J.M.: "Interface pressure and the prediction of car seat discomfort", Applied Ergonomics, Vol.
30, No. 2, 1999, pp 99–107,
[8] ISO 2631-1 Mechanical vibration and shock – evaluation of human exposure to whole-body vibration, Part 1:
General requirements, 1997,
[9] ISO 10326-1 Mechanical vibration – laboratory method for evaluating vehicle seat vibration: Part 1: Basic
requirements, 1992,
[10] Maciejewski, I.: "Active control of working machines seat suspension aimed at health protection against
vibration", Proceedings in Applied Mathematics and Mechanic, 2008,
[11] Mansfield, N.: "Human response to vibration", CRC Press, 2005,
[12] Parsons, K.C.: "Human thermal environments: the effects of hot, moderate and cold environments on human
health, comfort and performance", London: Taylor & Francis2003,
[13] Porter, J.M., Gyi, D.E.: "Exploring the optimum posture for driver comfort”, International Journal of Vehicle
Design, Vol. 19, No. 3, 1998, pp 255–267,
[14] Reynolds, H.M.: "Automotive seat design for sitting comfort", In Automotive Ergonomics, London: Taylor &
Francis, 1993,
[15] Shen, W., Parsons, K.C.: "Validity and reliability of rating scales for seated pressure discomfort", International
Journal of Industrial Ergonomics, Vol. 20, 1997, pp 441–461.
[16] User Manual for Human Vibration Analyzer Type 4447, Bruel&Kjaer, 2009,
[17] Wei, L., Griffin, M.J.: "Mathematical models for the apparent mass of the seated human body exposed to vertical
vibration", Journal of Sound and Vibration, Vol. 212, No. 5, 1998, pp 855–874.
445
446
Kragujevac, Serbia
MVM2018-015
Janez Kopač1
Franci Pušavec2
TOUCHSCREEN CAR DASHBOARDS AS SERIOUS

DANGER FOR CAUSING TRAFFIC ACCIDENTS
ABSTRACT: More and more sophisticated cars have also very modern cockpit boards full of information, with
many buttons, touch-screen interactions, commands on steering wheel, radio, A/C, etc. We know how dangerous
alcohol is and its influence on prolonging driver’s reaction time. In recent years, there is also a strong action against
using a phone during driving. Searching through the phone menu is taking attention off the road. Similarly, drivers
with new cars need several months to learn approximately where particular/some commands can be found. We have
measured time periods necessary to find different commands. For example, activating A/C takes 1.0 s -2.0 s. In the
case of car moving with speed of 50 km/h through the city, this means traveling along a path of 14 m – 28 m without
any control (by speed of 120 km/h, this corresponds to 33 m – 66 m). Additional fact is also the existence of a red
triangle emergency button that is becoming smaller and smaller and thus poorly visible in new cars. To find this
button, driver again needs more than one second in most critical situations, when full concentration focused on the
road is needed. This paper presents different car dashboards and measured necessary times for obtaining control.
Results are followed by discussion and the suggestions to the drivers how to deal with the raised problem.
KEYWORDS: cars, dashboard, control buttons, touch screen, lost time, reaction path
INTRODUCTION
The study on the use of mobile phones by the drivers shows problematical situations in using them during driving.
As an answer to the safety issues, there appear some alternatives like: driving mode on the phone allowing just
incoming calls, hands-free installation in the car, voice commands, etc. Even though they have their advantages,
none of those solutions assures a 100% safe usage of the phone during driving. Still, in most cases, this is not serial
equipment of the car - many drivers still hold mobile phones simply in hand, without any hands-free connection with
car radio system. On one side, this is taking their attention from the road to the phone. However, at least they normally
hold the phone between the head and the steering wheel. In this way, they partially see the road and traffic in front
of them. This is not perfect, but still much better than searching/typing on the touchscreen located on the right side
of the driver, thus the driver must physically look away for approximately 35o from the direction of driving. This means
losing the attention from the road for 1.0–2.5 s. Depending on the driving speed, this time corresponds to the path
that car drives without attention of the driver [4].
In addition, the important dilemma has been raised related to the question how many commands it is really necessary
to have on the dashboard and which of those can be really adjusted during driving? However, the trend seems to go
in other direction. The dashboards in new cars are going to be totally digital in parallel with the request from EU
commission to assure higher traffic safety.
The new developed cars are going to be completely autonomous with a lot of electronic support/active supervision
systems: active adaptive cruise control/breaking system, road line assistant, speed assistant, etc. Those systems
are adding a stone to the traffic safety. However, it will take a long time (approx. 20 years) for all cars to be equipped
1
Dr. Janez Kopač, University of Ljubljana, janez.kopac@fs.uni-lj.si
2
Dr. Franci Pusavec, University of Ljubljana, franci.pusavec@fs.uni-lj.si
447
with such technology. Basically, it is going to take the car lifetime. Just after the old cars will reach their lifetime, they
will be changed with the modern ones.
SOME CASES OF COCKPITS AND THEIR SPECIFICATIONS

A multifunction steering wheel, equipped with 4+1 different control handles, offers wide control options. However, the
number of control varieties is also enhancing the complexity and time to adjust lights, gear transmission, cruise
control, windshield wiper, etc. Advantage of a cockpit shown in Figure 1 is a big digital speed display and a big button
for switching the warning emergency lights, in the middle of a dashboard pillar. However, simple search for air
circulation control can then take a lot of time. It can be expected that the driver spends 1-3 s for this action. Figure 1
shows an example of a very transparent speed indicator on top positioned display that offers no need to search
downwards and to the right side, but it is directly in the line of the driving direction.
Figure 1 Cockpit of car A with many options - all functions are activated by mechanical switches/buttons
On the contrary, Figure 2 shows the steering wheel with many commands that is obstructing the view of the
dashboard.
Figure 2 Cockpit of car B with many controls on the steering wheel obstructing the view of the dashboard
In the case shown in Figure 3, during driving with speed limit (for instance, in the city, where speed limit is 50 km/h),
the average time for adjusting the most common settings in the car amounts between 1.5–2.0 s [3]. This confirms
448
the suspicion that the modern cars offer a lot of command possibilities with the drawback in attention that has to be
taken from the road and, in most cases, put on the touchscreen commands for a couple of seconds. One of the
solutions that would help improving this issue could be in the system for track control by video sensing system.
Figure 3 Cockpit of car B with digital dashboard
Some modern cars even have many more systems, equipment, commands, mechanical switches and touchscreen
commands. It is a fact that being able to control all of them takes time to recognize them, remember and scroll their
positions. Especially, searching the position of the desired command during driving can be very dangerous. It is well
known that, in sophisticated cars, drivers need to drive them for 50,000 km or use them for 2 years, before they are
completely comfortable with their usage and are even capable to automatically use some of the commands, without
removing the attention from the road [1, 5].
Even the emergency buttons – triangular buttons for blinking lights can be problematic. In Figure 4, two options are
presented: the conventional command panel (right) and the touchscreen command panel (left). In addition to the
aforementioned differences, a comparison between triangular emergency blinking lights switches is interesting. It
seems that, with modern cars, the size of this switch is decreasing. This should not be the case, because it is clear
that this switch can be used in the most critical situations, where size should not influence the critical moments. In
extreme cases (Figure 5), modern dashboards can even be made so that there is not a single button switch present
any more.
Figure 4 Comparison of B (left) and A (right) car cockpits and ease of access to the triangle warning button
449
Figure 5 Totally digital dashboard as an answer to request for higher traffic safety
CONCLUSIONS
“Principle of Trust” between the drivers on the road is actually stronger then between husband and wife in the family.
Drivers are driving against each other on a narrow roads, with speeds up to 90 km/h. In the case of an accident, this
would present the impact speed that is double, i.e. 180 km/h. In other words, such speed corresponds to the
approaching path of 50 m in every second [2]. In such cases, every tiny part of a second that the attention of the
driver is lost from the traffic by checking instruments on the dashboard can be crucial/critical. On normal roads, traffic
lanes are narrow and they can have 2.5 m of width. In such cases, with attention not on the road, car can reach the
opposite lane in approximately 0.5 s, if the steering wheel deviates 10o from the direction of the road. Thus, in the
case when belief, trust and attention are not present, even luck can hardly prevent occurrence of a car crash. Modern
cars will in those critical situations help a lot with indication of critical moments, with automatic breaking, etc. However,
attention of the driver to the traffic is and will always be the most important parameter rescuing the critical moments
that can happen in the traffic.
REFERENCES
[1] Burg, M., Moser, A.: "Handbuch Verkehrsunfallrekonstruktion. Unfallaufnahme, Fahrdynamik, Simulation",
ATZ/MTZ-Fachbuch, Vieweg+Teubner Verlag, 2007, Wiesbaden,
[2] Kuhelj, A.: "Mehanika, kinematika”, Lecturing handouts, Faculty of Mechanical Engineering, 2010, Ljubljana,
[3] Johnson, K., Krejan, A.: “Fizika: preproste razlage fizikalnih pojavov", Tehniška založba Slovenije, 2007,
Ljubljana,
[4] Kostić, S.: "Ekspertize saobračajnih nezgoda", Fakultet tehničkih nauka, 2016, Novi Sad,
[5] De Smet, A.: "Transportation Accident Analysis and Prevention", Nova Science Publishers, 2008, New York.
450
Kragujevac, Serbia
MVM2018-019
Saša Babić1
Jovanka Lukić2
Nikola Kostić3
Žarko Đorđević4
Jelena Obućina-Erić5
NOISE EMISSION WITH ASPECT OF VEHICLES FLEET

STRUCTURE – THE CASE OF SERBIA
ABSTRACT: Many western countries have developed traffic noise models for forecasting and analysis of traffic
noise. Traffic conditions and traffic characteristics differ from country to country and make that model unreliable for
predicting traffic noise in Serbians roads where traffic characteristic are different. This study is thus aimed at building
a relevant traffic noise model based on local traffic characteristics and vehicle types using more accurate noise level
measures based on an equivalent noise level measurement approach. The method enables calculation of reference
energy mean emission level (the basic noise) for each type of vehicles based on direct measurement of LAeq (20s)
from the real running condition of each type of vehicles. A new basic noise level was developed for vehicle types and
condition in Serbia road for the four vehicles group. This LAeq (20s) technique is used to make model to fit traffic
characteristic and different types of vehicles for local condition.
KEYWORDS: noise emission, traffic noise, noise measurements
INTRODUCTION
As observed in the literature, there are many methods and models for noise assessment and each country tends to
use its own approach among them. Study [1] presents a critical review of this matter and shows the main
characteristics of some methods like the approaches advocated by American FHWA, the English CoTRN, the
German RLS 90, the French NMPB-96 method. There is no overall pattern in these methodologies and those
countries that do not use a national model, usually apply the French NMPB-method, which is a model also suggested
by the EU Directive. Also discussed by [2], even when the same method is applied in different country, some
discrepancies in the results may occur.
This study is thus aimed at building a relevant traffic noise model based on local traffic characteristics and vehicle
types using more accurate noise level measures based on an equivalent noise level measurement approach. The
method enables calculation of reference energy mean emission level (the basic noise) for each type of vehicles
based on direct measurement of LAeq (20s) from the real running condition of each type of vehicles. This technique
is used to make a model fit traffic characteristic and different types of vehicles for local conditions that will be
explained in the next paper.
1
Saša Babić, College of Tehnical and Mehanical Engineering Trstenik, babicsf@yahoo.com
2
Jovanka Lukić, Ph. D., prof., University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić Str., 34000
Kragujevac, Serbia, lukicj@kg.ac.rs
3
Nikola Kostić, College of Technical and Mechanical Engineering Trstenik, kosticn83@gmail.com
4
5
Jelena Erić Obućina, College of Technical and Mechanical Engineering Trstenik, jericobucina@gmail.com
6
Branimir Milosavljević, College of Technical and Mechanical Engineering Trstenik, lenka1809@gmail.com
451
NOISE INDICATORS USED IN MEASURING
Sometimes, the people who make or analyse the measurements understand the complexities of the measurement.
If people are asked to measure the same exact noise at the same exact location, the chances of them getting the
same results are weak. The reasons are many: the type or accuracy of their sound level meters may be different, the
frequency weighting may be different, and how they measure or what, or how they acquire samples, may be different.
Current legislation in Europe is based on noise level of vehicles according to test procedures given in Directive
70/157/EEC. The maximum noise level depends on the category and type of the vehicle and it has been declining
gradually with updates of the Directive. Value of the maximum noise level (Lmax), which is regulated by ECE
Regulation 51.02 (Method A) does not properly represent the noise levels emitted by a vehicle in real driving
conditions. This stems primarily from the analysis of measurement and the fact that noise generated by a single
vehicle in motion cannot be equated with individual noise source indicator which corresponds to the maximum value
of the noise. The vehicle in real conditions behaves as a complex noise source made from multiple noise sources
that alter, depending on different conditions and driving modes. In acoustic terms, vehicles can be described by two
sources, a power unit source and rolling source. For the fluid traffic condition, the power noise can be taken as a
constant. Given traffic growth, however, its reduction has done little to reduce the overall traffic noise. One reason
for this is that the test method does not properly represent the noise levels emitted by a car in real driving conditions.
The fact that the vehicle cannot be accurately approximated as a single source of noise imposes the idea of
introducing another indicator of noise for the individual vehicle driving in real conditions. The task of this indicator is
to describe the vehicle as a complex source of noise in real conditions with energy value. In that case, there would
be no need for an approximation of the vehicle as individual source of noise. Currently, there is no officially defined
procedure and method for measurement of noise of individual vehicle driving in real conditions in our country using
these noise indicators. This type of research of noise measurement techniques is also described in [3].
MEASUREMENT OF EQUIVALENT NOISE LEVEL OF CERTAIN CATEGORIES

OF SINGLE VEHICLES - CHARACTERISTICS OF THE MEASUREMENT
A reference energy mean emission level (the basic noise) for each type of vehicle is developed based on direct
measurement of noise from the real running condition of each type of vehicles. Measuring the value of equivalent
noise level LAeq (10 s) was not done directly, primarily because of the short time interval of sampling, which could
lead to some measurement error, but also because of the possible imprecision of instrument. Device for manual
measurement of noise, Bruel & Kajer type 2230, was used for noise measurement. This type of device is class 1
device for measuring noise with measurement error of 0.1 dB. Noise indicator SEL (SEL - Sound exposure level) is
directly measured on the field. Based on the known transformation, levels of noise exposure are easily transformed
into an equivalent level of noise of the measured values of the interval.
Microphone position is defined at 7.5 m from the centre line of the track, at a height of 1.2 m above the road level.
Speed of each vehicle is measured by radar, oriented in the direction of the coming vehicle. In this way, the
measurement of vehicle speed is done at a small angle from 1 to 3 degrees in relation to the centre line of the moving
vehicle, which provides high accuracy measurements of the speed of the vehicle. The method and location of
equipment for measurement of individual vehicle’s pass-by noise are detailed in Figure 1.
The collection of such data is required for a longer period, primarily due to provision of the necessary conditions for
this type of measurement, so that the measurements are being taken at the different times of a day and in different
localities of road M5. Different measurement sites were chosen, primarily related to the same technical and
exploitation conditions, same ambient conditions, similar meteorological conditions (no strong winds, rain, heat) with
the difference that different speed limits for vehicles were defined at the measurement location. This approach to
measure the noise at different speed limits at certain localities was chosen primarily because of the situation that has
been observed in the field. In fact, the first thing that is noticed is that drivers react to these measurements by selecting
the proper speed and choosing the speed imposed by a traffic sign, which, at least the vast majority of drivers, would
not do in different situation. This situation will arise very quickly at the measuring location, which is most likely the
result of "solidarity" among drivers, causing the speed of passage of individual vehicles to stabilize around a value
slightly below the speed limit for that section of the road and then fluctuate slightly around this value. This behaviour
of the driver in the area of measurement is probably caused by driver’s understanding that the roadside activities are
directed against them, probably believing that it is a measurement of speeding by the police, because of which they
select a special driving mode. For this type of measurement of individual vehicle pass-by noise, it is necessary to
provide special conditions and choose the appropriate equipment, but that is not so clear for the drivers and does
not impact their behaviour very much. It is during this measurement that the given hypothesis that the characteristics
of flow structure and the degree of homogeneity of traffic flow often shows characteristics of the driver is confirmed.
452
Collection zone
A B C D
t=0 s t=10 s
7,5m Speed radar
Position of the
microfone
Popint А - Start measuring point of vehicle speed

Point B - Start measuring point of LAeq (20s)
Point C - End point measurements of vehicle
speed
Point D - End point measurements of LАeq (20s)
Figure 1 Schematic view of measurement of equivalent noise level LAeq (20s)
In order to overcome this problem and get a better measurement results, a different measurement technique is
adopted. The proposed technique and method of measurement are much easier and the results obtained are
sufficiently accurate and free from errors that may occur during the process of measuring the impact on driver
behaviour and judgment of an observer. The idea of this measurement (LAeq (20 s)) is that the measurement starts
at the moment when the vehicle passes directly in front of the microphone and stops after 10 seconds. This technique
of measurement lasts 10 s and starts at t1  0 s and ends at t 2  10 s, compared to a similar measurement technique
that starts at t1  5 s and ends at t 2  5 s (Figure 2). At approaching to LAeq (10 s), an observer may not start the
noise meter exactly at the corect time (5 s before the vehicle passes the meter) and also because of the difference
in a spot speed of each particular vehicle. However, this method of measurement is in fact analogous to measurement
in the interval, which starts at t1  10 s and ends at t2  10 s, because of the time in which the measuring device
records the noise emitted by the energy value of noise for 20 s or LAeq (20s). The transformation of this kind of
measurements of 10 s (LAeq (10 s) into LAeq (20 s)) can be mathematically described by expressions given in [3]. The
average mean energy of LAeq (20 s) ( t1  10 s to t2  10 s) is equal to that of LAeq (10 s) ( t1  0 to t2  10 s). This
technique ensures measurement of the energy contribution of the full pass-by of all vehicles, especially long vehicles
at slow speed.
t = -5 t = 0 t = +5 t = -10 t=0 t = +10
[dB(A)] [dB(A)]
Lmax Lmax Microfon
Microfon
LAeq 10s LAeq 20s
-10 -5 0 +5 +10 [s] -10 -5 0 +5 +10 [s]
Figure 2 Comparison between the distribution curves for noise LAeq (10 s) and LAeq (20 s) of pass-by vehicles
THE ANALYSIS OF NOISE EMISSION LEVELS OF PASSING VEHICLE

GROUP
The level of noise emission depends on the speed of the vehicle, and for this reason, the measured values of
equivalent noise level for individual vehicle categories included different speeds in range of 30 km/h, while the upper
speed limit was not restricted. Such data were used for regression analysis to establish the relationship between LAeq
(20s) and speed for each of the defined group of vehicles. For this study, four categories of vehicles were used.
These types of categories consisted of: light motor vehicles, medium trucks, heavy trucks and buses. Regression
453
curves in different form for each of the defined groups of vehicles in operation speed of the vehicle and are shown in
Figures 3 to 6. In addition to these charts, the functional relationships in the analytical form of these two parameters
and the statistical analysis for proposed model are shown in Table 1.
As can be seen in Table 1, the best fit with the measured results, on the basis of the statistical test for all proposed
vehicle categories, is the logarithmic dependence. This kind of dependence between the speed and the noise
emission level is well known in the literature [1, 3, 4] and confirmed by this analysis under our conditions.
Table 1 Statistical test results for the different categories and speeds of vehicles based on regression analysis
Coefficient Interpreted Residuum
Coefficient of Total Standard
of part of the of the F t
Statistics determination dispersion error
2 correlation 2 dispersion dispersion test test
R sy Sy
R s2yr S2 y
Vehicle category Light motor vehicles
Dependence / Number of samples 130
L Aeq (20s)  40.351  0.319V 0.871 0.933 3706.79 3228.099 478.690 1.919 1.1449 0.0456
LAeq (20s)  17.719 lnV  12.331 0.884 0.940 3706.789 3276.262 430.527 1.820 1.1311 0.0012
LAeq (20s)  34.221  0.54V  0.002V 2 0.864 0.930 3706.79 3204.201 502.588 1.966 1.2650 0.9861
L Aeq (20s)  42.6e 0.005V 0.722 0.850 3706.789 2675.189 1031.600 2.817 1.4156 3.1577
Vehicle category Medium truck
L Aeq (20s)  53.192  0.255V 0.721 0.849 1115.28 803.896 311.384 2.379 1.3908 0.6960
LAeq (20s)  15.522 ln V  5.4523 0.781 0.884 1115.28 870.627 244.652 2.109 1.2808 0.6934
LAeq (20s)  31.204  1.017V  0.006V 2 0.768 0.876 1115.28 856.260 259.020 2.170 1.0551 0.5942
L Aeq (20s)  54.344e 0.004V 0.674 0.821 1115.28 751.718 363.56 2.571 1.1760 0.6089
Vehicle category Heavy truck
L Aeq (20s)  63.87  0.174V 0.731 0.855 312.09 228.118 83.974 1.549 1.3705 0.0145
LAeq (20s)  10.231ln V  32.752 0.762 0.873 312.09 237.861 74.230 1.456 1.3135 0.0038
LAeq (20s)  55.292  0.475V  0.002V 2 0.271 0.521 312.09 84.453 227.639 2,.550 1.3243 2.4499
L Aeq (20s)  64.431e 0.002V 0.381 0.617 312.09 118.828 193.26 2.350 1.9725 2.7111
Vehicle category Bus
LAeq (20s)  60.493  0.22V 0.719 0.848 395.19 284.326 110.861 1.780 1.3856 0.0240
LAeq (20s)  11.615 ln V  26.361 0.698 0.835 395.19 276.035 119.153 1.845 1.4317 0.0019
LAeq (20s)  63.7  0.098V  0.001V 2 0.714 0.845 395.19 282.138 113.049 1.797 1.5165 0.3746
L Aeq (20s)  61.401e 0.003V 0.719 0.848 395.19 284.175 111.01 1.781 1.4209 0.1819
75,00 80,00
70,00 75,00
70,00
LAeq (20 s)
LAeq (20 s)
65,00
65,00
60,00
60,00
55,00
55,00
50,00 50,00
45,00 45,00
Number of measured samples Number o
Measured Exponential dependence Measured Expon

Logarithm dependence Linear dependence Logarithm dependence Linea
Quadratic dependence Quadratic dependence
Figure 3 Functional relationship
Fig. 3ofFunctional
the equivalent noiseoflevel,
relationship LAeq (20s),
the equivalent depending
noise level on
Fig.the light motor
4 Functional vehicle of the equiva
relationship
LAeq (20s) depending onspeed
the light motor vehicle speed LAeq (20s) depending on the medium t
80,00 85,00
85,00
75,00
75,00
70,00 75,00
s) s)
LAeq (20 s)
(20 (20
65,00
LAeq
65,00 65,00
LAeq
60,00
55,00
55,00 55,00
50,00
45,00
45,00 45,00
Number of measured samples Numberof
Number ofmeasured
measuredsamples
samples Number of
Measured Exponential dependence
ed Exponential dependence Measured dependence
Logarithm Exponential
Linear dependence
dependence Measured Expo
hm dependence Linear dependence Logarithmdependence
Quadratic dependence Linear dependence Logarithm dependence Line
ic dependence Quadratic dependence Quadratic dependence
Figure
l relationship of 4 Functional
the equivalent noise relationship
level Fig.of4Functional
Fig.5 the equivalent
Functional noise of
relationship
relationship level,
ofthe L Aeq (20s),
theequivalent
equivalent depending
noise
noise level onFig.
level the 6medium truck
Functional speed of the equiv
relationship
ending on the light motor vehicle speed LLAeq
Aeq (20s)
(20s) depending
depending on on the
the medium truckspeed
heavy truck speed LAeq (20s) depending on the bus
85,00
75,00
454
q (20 s)
65,00
Quadratic dependence Quadratic dependence
Fig. 3 Functional relationship of the equivalent noise level Fig. 4 Functional relationship of the eq
LAeq (20s) depending on the light motor vehicle speed LAeq (20s) depending on the mediu
80,00
75,00 85,00
85,00
70,00
s) (20 s)
75,00
65,00 75,00
LAeq (20LAeq
LAeq (20 s)
60,00
65,00 65,00
55,00
50,00
55,00 55,00
45,00
Number of measured samples Number of measured samples
45,00 45,00
Number of measured samples Numb
Exponential dependence Measured Exponential dependence
pendence Linear dependence Logarithm
Measured dependence Linear dependence
Exponential dependence
pendence Quadratic
Logarithmdependence
dependence Linear dependence Measured
Quadratic dependence Logarithm dependence
lationship of the equivalent noise level Fig. 4 Functional relationship of the equivalent noise level Quadratic dependence
ding on the lightFigure
motor 5
vehicle speed relationship
Functional Lof
Fig.5Aeq (20s)
the depending
equivalent
Functional on the
noise
relationship medium
oflevel,
the LAeq truck speed
(20s),noise
equivalent depending
level on the
Fig. heavy truck speed
6 Functional relationship of the eq
LAeq (20s) depending on the heavy truck speed LAeq (20s) depending on the
85,00
75,00
LAeq (20 s)
65,00
55,00
45,00
Number of measured samples Number of measured samples
Exponential dependence
pendence Linear dependence Measured Exponential dependence
endence Logarithm dependence Linear dependence
Quadratic dependence
Figure 6noise
ationship of the equivalent Functional
level relationship of the equivalent
Fig. 6 Functional relationshipnoise
of thelevel, LAeq (20s),
equivalent noise depending
level on the bus speed
ending on the heavy truck speed LAeq (20s) depending on the bus speed
A comparative analysis of the proposed logarithmic regression curve for our model and the regression curve
proposed by the FHWA model is shown in Figure 7. Comparing the regression curves obtained on the basis of
experimental measurements with regression curves that are recommended by the FHWA, one can clearly see the
difference in terms of the level of noise emission and similarities in terms of trend curves themselves. First, it can be
concluded that all categories of regression curves for our conditions give lower values of noise in the receiving point
than the regression curve defined by the FHWA model. This is in particular due to the different structure of the
vehicles themselves, vehicle technology that in some way belong to the same category.
86
LAeq dBA
84
82
80
78
76
74
72
70
68
66
64
62
60
58
56
54
52
50
48
46
30 35 40 45 50 55 60 65 70 75 80
km/h
Light motor vehicle Medium truck Heavy truck Bus

Passenger car FHWA Medium truck FHWA Heavy truck FHWA
Figure 7 Comparative analysis of regression curves established for the equivalent noise emission depending on
the speed for our conditions and the FHWA
By comparison of these models to the defined speed, it can be seen that the differences in the noise emission levels
of certain categories of vehicles are around some average value in the range of speeds between 30-80 km/h.
Additional reason for this discrepancy lies in the fact that the regression curves for the FHWA model are defined back
in 1970s and they still have not changed, and it is quite obvious that the characteristics of the vehicles have changed.
455
Figure 7 and Table 2 show that the FHWA model for passenger cars of all categories of vehicles may be statistically
different. The comparison for the light motor vehicles and medium and heavy trucks does not show such close
agreement. Aldo they are similar, the regression lines are offset by approximately 7.6 dBA from our proposed model.
For these categories of vehicles, considerable leeway in the interpretation of the vehicles type occurs, as previously
discussed. In this sense, the FHWA regression curve can be utilized with some adjustments for our conditions.
Table 2 The average deviation of noise values for certain categories of vehicles, between the FHWA model and
the model for our conditions, expressed in dB (A)
Category of vehicle in Category of vehicle for local conditions
FHWA model Light motor vehicles Medium truck Heavy truck Bus
Passenger car 5.1
Medium truck 7.8
Heavy truck 7.5 8.4
CONCLUSIONS
Several conclusions can be drawn from this research. First, a new basic energy mean noise emission level was
developed for vehicle types and conditions on Serbian road. It was based on direct measurement of equivalent sound
level over a period of 20 s (LAeq (20s)) in the running condition on a road M5, with a speed range of 30-100 km/h for
four classifications of vehicles. This technique can provide a more accurate measurement of individual vehicle’s noise
with better representation of the contributions to overall energy emission. The goodness of the fit between the
predicted values from the developed models and the measured field data was tasted using the statistical analysis. It
is shown that the best fit between noise emission level and the running speed for all categories of vehicles is
logarithmic dependence.
Another conclusion is that the FHWA curves show large variance for all categories of vehicles, most likely due to the
very broad definition of vehicle type. This could lead to significant errors in noise prediction.
The proposed regression curve can be a useful tool for determination of the noise emission level for individual vehicle
in real running conditions. More work is needed to determine the national reference noise emission level and to
improve the accuracy of the proposed mathematical models.
REFERENCES
[1] Steele, C.: "A critical review of some traffic noise prediction models", Applied Acoustics, Vol. 62, 2001, pp 271-
287,
[2] Javier, A., Ian, H. F., Andrew, J. B.: "Variability in road traffic noise levels", Applied Acoustics, Vol. 66, 2005, pp
1180-1195,
[3] Pamanikabud, P., Tansatcha, M., Brown, A. L.: "Development of a highway noise prediction model using an Leq
20 s measure of basic vehicular noise", Journal of Sound and Vibration, Vol. 316, 2005, pp 317-330,
[4] Roger, L. W, Timothy, O., Win, L.: "Development of reference energy mean emission levels for highway traffic
noise in Florida", Transportation research record 1416, 2008, pp 82-91.
456
Kragujevac, Serbia
MVM2018-021
Jelena Trifunović1
Aleksandar Manojlović2
Snežana Kaplanović3
Nebojša Bojović4
METHODOLOGY FOR DETERMINING THE AMOUNT OF

ECOLOGICAL FEE FOR THE USE OF PASSENGER CARS
ABSTRACT: The number of private passenger cars, the degree of motorization and transport requirements is
constantly increasing, which significantly affects the dependence of Europe on oil and it's products and the
emission of greenhouse gases. According Annual European Union greenhouse gas inventory 1990–2015 and
inventory report 2017, CO2 emissions from road transport increased by 642% in the period from 1990 to 2015.
Therefore, efforts are being made to change this trend. In addition to reducing the CO2 emissions of newly
produced passenger cars by increasing energy efficiency, additional measures are needed to reduce pollutant
emissions at an annual level. In order to reduce emissions of pollutants, internalization of external CO2 costs
through annual fees for the possession and use of passenger cars is being carried out. Today in the world there
are different methodologies that are used for the calculation of the annual fee for the use and registration of
passenger cars. The paper proposes a methodology for calculating the annual ecological fee for the use of petrol
and diesel passenger cars. The methodology covers greenhouse pollutants - CO2, CH4, N2O.
KEYWORDS: ecological fee, passenger cars, greenhouse gases, energy efficiency
INTRODUCTION
The negative effects of road transport are subject to the study of different scientific fields both due to global
consequences, and of individual, local adverse effects. Unwanted road transport effects are congestion, traffic
accidents, environmental pollution, noise, an inadequate road network of urban areas in terms of limitations
concerning the period of construction and the increase in the degree of motorization. In addition to the mentioned,
characteristic for road transport are dependence on oil and oil derivatives and high energy consumption.
On the other hand, the development of society leads to economic development, without which, in modern times,
transport is unthinkable. Challenges that pose completely opposing views, increasing transport requirements and
reducing the negative effects of transport on the environment, reducing energy consumption, need to be reconcile.
Emissions of pollutants affect climate change, atmospheric composition, air quality, etc. Thus, it contributes to the
creation of a greenhouse effect [11]. It is therefore necessary that the development of road transport goes in the
direction that is economically, socially and in terms of environmental protection sustainable.
1
Jelena Trifunović, University of Belgrade, Faculty of Transport and Traffic Engineering, Vojvode Stepe 305, 11000
Belgrade, Serbia, j.trifunovic@sf.bg.ac.rs
2
Aleksandar Manojlović, University of Belgrade, Faculty of Transport and Traffic Engineering, Vojvode Stepe 305,
11000 Belgrade, Serbia, a.manojlovic@sf.bg.ac.rs
3
Snežana Kaplanović, University of Belgrade, Faculty of Transport and Traffic Engineering, Vojvode Stepe 305,
11000 Belgrade, Serbia, s.kaplanovic@sf.bg.ac.rs
4
Nebojša Bojović, University of Belgrade, Faculty of Transport and Traffic Engineering, Vojvode Stepe 305, 11000
Belgrade, Serbia, nb.bojovic@sf.bg.ac.rs
457
In 2011, 299.31 million passenger cars were registered in the territory of the European Union, and by 2015 this
number was increased to 323.58 million while the motorization level in 2011 was 383, and in 2015 406 passenger
cars per 1000 inhabitants [3]. Thus, the number of passenger cars, the degree of motorization and transport
demand is constantly increasing, which significantly affects the dependence of Europe on oil and petroleum
products and the emission of greenhouse gases. Also, the total annual CO2 emissions from road transport at the
state level are constantly increasing. In the period from 1990 to 2015, CO2 emissions from road transport increased
almost 6.42 times [4]. Therefore, efforts are being made to change this trend.
With the introduction of binding Euro-standard engine technologies in the European Union, CO2 emissions of new
passenger cars have been reduced from 123.4g/km to 119.5g/km between 2014 and 2015. The average CO2
emission of new cars in 2016 was 118.1 g CO2 / km [5]. The target of 130 g CO2/km by 2015 was achieved two
years earlier, but efforts continue to be made to achieve the set target of 95g CO2/km by 2020 [5]. Research on the
impact of measures on reducing the negative impact of transport on the environment and energy consumption has
been done, including engine and fuel technology, intelligent transport systems (ITS) and mobility management [9]
In addition to these measures, in order to reduce emissions of pollutants, the internalization of external CO2 costs
through annual fees for the possession and use of passenger cars is carried out. Today there are different
methodologies used in the world for calculating the annual fee for the use and registration of passenger cars [6]. A
large number of countries are investing efforts to internalize and reduce external emissions costs by introducing
taxes based on CO2 emissions and / or fuel consumption. Nevertheless, the internalization of external costs of CO2
and other pollutants remains a challenge [6].
In this paper, the methodology for calculating the annual ecological fee for the use of privately owned passenger
cars (or annual usage fee) is proposed. The results of the calculation of the annual environmental fee for petrol and
diesel cars, displacement from 1400 to 2000 cm3, of the passenger car fleet in private ownership of Serbia.
METHODOLOGY FOR DETERMINING THE AMOUNT OF ECOLOGICAL FEE

FOR THE USE OF PASSENGER CARS
The procedure for determining the amount of annual ecological tax fee includes the calculation of the annual total
passenger car emissions at the state level and the allocation to individual passenger cars, so that everyone is
charged the costs of polluting the environment to the extent that it contributes to their activities. The emission of
pollutants in the private passenger car fleet is realized according to the type of fuel (gasoline and diesel), engine
volume, engine technology. The allocation of the amount of pollutants emitted to individual privately owned
passenger cars is realized by calculating the emissions per kilometer after determining the total amount of emitted
pollutants of each category of passenger cars (type of propulsion fuel, engine displacement, engine technology).
Thus, by adopting the principle that "polluter pays" for activities that have a negative impact on the environment, it
is proposed to calculate the annual environmental tax per kilometer for privately owned passenger cars. This
calculation includes the greenhouse gases pollutants CO2, CH4, N2O.
Annual ecological fee for the use of passenger cars

By adopting the principle that "polluter pays" for activities that have a negative impact on the environment, it is
proposed calculation of the amount of annual ecological fee for privately owned passenger cars.
The annual ecological fee per kilometer, for the category of passenger cars, is calculated as follows:
(1)
Where is:
Gtax,km,k – annual ecological fee per kilometer, passenger car category k,
k – category of passenger cars by type of fuel, engine volume and engine technology
eCO2,k, eCH4,k, eN2O,k – annual emitted CO2, CH4, N20, passenger car category k (t),
sk – average annual mileage of passenger cars category k (km),
NPA,k – number of passenger cars category k and
cCO2, cCH4, cN2O – emission unit costs CO2, CH4, N20 (€/t).
The value of the cost of the emissions was adopted from the project [19] and amounts 40 €/t CO2, 540 €/t CH4,
21400 €/t N2O.
458
This procedure is used to allocate the amount of pollutants emitted to privately owned passenger cars. The total
emission of CO2, CH4, N2O of each category of passenger cars is determined. The total activity of passenger cars
is used to reduce the emissions to a unit rate, after which each passenger car, depending on the annual mileage
travelled, is entitled to a financial compensation.
Emission calculation CO2 CH4, N2O

For the calculation of emissions of pollutants from privately owned passenger cars, the software tool COPERT
(Computer Programme to calculate Emissions from Road Transport) is used. COPERT software is used by more
than 20 EU countries to obtain the results of the assessment of the emission of pollutants derived from road
transport in order to form annual national reports [1, 2, 7, 10]. The data needed to estimate pollutant emissions
include: number of vehicles by vehicle category and technology, average millage travelled by category of vehicles
and total kilometers, average vehicle speed per category of vehicles and types of road, structure of the route
travelled by type of road (urban, rural, highway) expressed in percentages, average length of travel and climatic
conditions [8].
The following table shows data from Ministry of Interior of the Republic of Serbia for 2016, the number of
passenger cars with gasoline and diesel engine capacity from 1400 to 2000 cm3 according to engine technology
(Table 1). The largest number of passenger cars with gasoline and diesel engines are the Euro 3 engine
technology cars, 95049 petrol and 364700 diesel (data for 2016). Also, a significant number of conventional diesel-
powered vehicles make the structure of Serbia's fleet of vehicles (98773).
Table 1 Number of passenger cars with petrol and diesel engine capacity from1400 do 2000 cm3
Engine technology, Engine technology,
Number of Number of
petrol diesel
passenger cars passenger cars
(1400-2000cm3) (1400-2000cm3)
pre-euro 565
ECE 15/00-01 629
- -
ECE 15/02 920
ECE 15/03 6801
ECE 15/04 40102 Conventional 98773
Euro 1 18202 Euro 1 27950
Euro 2 32710 Euro 2 54935
Euro 3 95049 Euro 3 364700
Euro 4 21078 Euro 4 104435
Euro 5 13994 Euro 5 38904
Euro 6 1521 Euro 6 4539
Information on annual mileage, speed and road type
Data on the estimated annual mileage, speed, percentage participation of the travelled route on the urban, rural
and highway network are also input data for the estimation of emissions of pollutants from passenger cars. These
data are not available as statistical data, nor are they recorded in a single vehicle registry. Below is a list of driving
conditions, a procedure for assessing data in order to obtain a picture of the particular characteristics of the
vehicle's exploitation.
Driving conditions
Vehicle speed according to the type of road requires a complex assessment procedure. In order to simplify the
procedure, the speed limits on each road type were reduced by 10% when determining the speed of the vehicle on
each type of route. The following hypothesis was set up:
„The average speed of the vehicles is the same regardless of the type of vehicle, the total mass, the volume of the
engine, and differ according to driving conditions (urban – 45km/h, rural – 72 km/h, highway – 108 km/h)“
Average annual mileage and mileage travelled by road type
The average annual mileage was determined by conducting a survey. The survey included a survey of privately
owned passenger cars by conducting on-line surveys.
459
The following data were entered in the survey form: the brand and type of vehicle, the year of production of the
vehicle, the engine volume (cm3), the type of fuel used by the vehicle, the average annual mileage (km), the total
vehicle mileage (km) and the annual mileage on urban, rural and highways (in% or km).
The result of the data processing of the survey is the average annual mileage of passenger cars, as well as the
structure of the route according to the types of road (urban, urban, highway) expressed in percentages. The
following table shows the average annual mileage travelled on the basis of the conducted survey (Table 2).
Table 2 Average annual mileage of Serbian passenger car fleet, petrol and diesel (1400 - 2000cm3)
Engine technology, Engine technology,
Average annual Average annual
petrol diesel
3 mileage 3 mileage
(1400-2000cm ) (1400-2000cm )
pre-euro 13000
ECE 15/00-01 6391
- -
ECE 15/02 6705
ECE 15/03 7685
ECE 15/04 10336 Conventional 17458
Euro 1 5600 Euro 1 65000
Euro 2 10133 Euro 2 20000
Euro 3 13904 Euro 3 27222
Euro 4 13788 Euro 4 24950
Euro 5 16408 Euro 5 31050
Euro 6 23358 Euro 6 33385
The following table shows the distribution of distance travelled to the type of road obtained based on the research
(Table 3).
Table 3 Distribution of distance travelled according to the type of road, (%)

Engine technology,
petrol Urban share (%) Rural share (%) Highway share (%)
(1400-2000cm3)
pre-euro 45.0% 47.5% 7.5%
ECE 15/00-01 45.0% 47.5% 7.5%
ECE 15/02 45.0% 47.5% 7.5%
ECE 15/03 45.0% 47.5% 7.5%
ECE 15/04 45.0% 47.5% 7.5%
Euro 1 26.4% 54.2% 19.4%
Euro 2 59.6% 25.4% 15.0%
Euro 3 54.2% 24.8% 21.0%
Euro 4 54.7% 27.7% 17.6%
Euro 5 53.0% 23.2% 23.9%
Euro 6 55.9% 22.8% 21.3%
Engine technology,
diesel Urban share (%) Rural share (%) Highway share (%)
(1400-2000cm3)
Conventional 75.0% 20.0% 5.0%
Euro 1 5.0% 1.0% 94.0%
Euro 2 50.0% 15.0% 35.0%
Euro 3 49.8% 28.9% 21.3%
Euro 4 48.8% 21.9% 29.4%
Euro 5 44.3% 20.1% 35.6%
Euro 6 33.0% 21.4% 45.6%
The following table shows the values of the emitted CO2, CH4 and N2O emissions per km, per fuel type and
displacement from 1400 to 2000 cm3 using the COPERT model (Table 4).
460
Table 4 Amount of emitted CO2, CH4 and N2O per km according to the type of engine fuel consumption from 1400
to 2000 cm3, g/km
Engine technology,
petrol CO2 (g/km) CH4 (g/km) N2O (g/km)
(1400-2000cm3)
pre-euro 216.14 0.06725 0.00536
ECE 15/00-01 176.44 0.06725 0.00536
ECE 15/02 167.93 0.06725 0.00536
ECE 15/03 167.93 0.06725 0.00536
ECE 15/04 166.07 0.06725 0.00536
Euro 1 158.88 0.01140 0.00657
Euro 2 159.84 0.01046 0.00578
Euro 3 167.48 0.00215 0.00052
Euro 4 174.77 0.00243 0.00066
Euro 5 175.34 0.00258 0.00091
Euro 6 175.47 0.00252 0.00116
Engine technology,
diesel CO2 (g/km) CH4 (g/km) N2O (g/km)
(1400-2000cm3)
Conventional 160.34 0.01619 0
Euro 1 160.93 0.00282 0.00376
Euro 2 153.47 0.00275 0.00391
Euro 3 145.59 0.00068 0.00404
Euro 4 146.52 0.00024 0.00399
Euro 5 146.80 0.00019 0.00377
Euro 6 146.73 0.00006 0.00332
Annual ecological fee for passenger cars with petrol and diesel engines,
1400-2000 cm3
After calculating the emission of CO2, CH4, N2O, from privately owned passenger cars in Serbia, with a petrol
engine and diesel engine (displacement from 1400 to 2000 cm3) per kilometer, it was approached to determination
of the annual ecological fee according to the defined classification of passenger cars.
The fee for the annual ecological tax per mileage ranges from 0.008668 € / km to 0.006377 € / km for gasoline
vehicles, and from 0.006435 € / km to 0.005891 € / km for diesel vehicles depending on engine technology (Table
5). The following table shows the fees for the annual ecological tax in case the vehicle travels the annual average
mileage from 2016 and the fee for the travelled 10000 km per year. The highest amount of tax fee in 2016 has a
passenger car with a diesel engine Euro 1 standard and amounts 419.86 € for a yearly 65000 km travelled, then
Euro 6 196.67 (33385 km) and Euro 5 183 (31050 km). According to the data for 2016, gasoline-powered vehicles
travel less mileage than diesel-powered vehicles. Thus, the highest amount of ecological fee for gasoline
passenger cars for the mentioned year is for the Euro 6 engine technology - 196.67 €.
Table 5 The annual ecological fee per kilometer, for the annual mileage travelled in 2016, and 10000 km per year
I II
Engine unit annual The annual Amount of The annual Amount of
technology, petrol environmental distance annual distance annual
(1400-2000cm3) tax (€/km) traveled ecological tax traveled ecological tax
(km) (€) 10000 km (€)
pre-euro 0.008668 13000 112.68 10000 86.68
ECE 15/00-01 0.007080 6391 45.25 10000 70.80
ECE 15/02 0.006739 6705 45.19 10000 67.39
ECE 15/03 0.006739 7685 51.79 10000 67.39
ECE 15/04 0.006665 10336 68.89 10000 66.65
Euro 1 0.006377 5600 35.71 10000 63.77
Euro 2 0.006416 10133 65.01 10000 64.16
Euro 3 0.006721 13904 93.46 10000 67.21
Euro 4 0.007013 13788 96.69 10000 70.13
Euro 5 0.007036 16408 115.44 10000 70.36
Euro 6 0.007041 23358 164.47 10000 70.41
461
Table 6 (continued)
The annual Amount of The annual Amount of
Engine unit annual
distance annual distance annual
technology, diesel environmental
traveled ecological tax traveled ecological tax
(1400-2000cm3) tax (€/km)
(km) (€) 10000 km (€)
Conventional 0.006435 17458 112.35 10000 64.35
Euro 1 0.006459 65000 419.86 10000 64.59
Euro 2 0.006161 20000 123.22 10000 61.61
Euro 3 0.005845 27222 159.12 10000 58.45
Euro 4 0.005883 24950 146.77 10000 58.83
Euro 5 0.005894 31050 183.00 10000 58.94
Euro 6 0.005891 33385 196.67 10000 58.91
CONCLUSIONS
The proposed methodology for determining the annual ecological fee also includes passenger cars that do not
meet the prescribed Euro standards and still are part of the structure of the Serbian car fleet. Within the framework
of the methodology, the COPET methodology for the emission calculation is based on a large set of classified input
data including exploitation conditions (speed, road type, climate conditions). Exploitation conditions related to
climate conditions, speed and percentage distribution of kilometers travelled according to the road type, according
to the calculation, in the case of gasoline vehicles, lead to a slight increase in CO2 and CH4 emissions per kilometer
of the latest euro engine technology standards, and hence to amount of fees. Thus, the annual ecological fee,
which includes pollutants that have an impact on the greenhouse effect, can only initiate a reduction in the average
annual mileage of privately owned passenger cars. This is especially true for diesel-powered vehicles that travel
more kilometers yearly than gasoline-powered vehicles. Changing the age structure of the vehicle fleet and / or
changing the structure of the vehicle fleet according to the type of fuel must also include pollutants that otherwise
have a negative impact on the environment. Further research would include pollutants whose reduction or
mitigation of their emission is prescribed by the NEDC Directive. In addition, further research would include all
passenger cars displacement with petrol and diesel engines
The complexity of the problem of environmental pollution can not be solved by applying only one measure that
would reduce or mitigate the amount of greenhouse pollutant emissions (carbon-dioxide, methane and nitrogen-
suboxide).
ACKNOWLEDGMENTS
The research presented in this paper has been realized within the project Development of the Model for Managing
the Vehicle Technical Condition in Order to Increase its Energy Efficiency and Reduce Exhaust Emissions (No.
36010), supported by the Ministry of Education, Science and Technological Development of the Republic of Serbia.
REFERENCES
[1] Bellasio, R., Bianconi, R., Corda, G., Cucca, P.: "Emission inventory for the road transport sector in Sardinia
(Italy) ". Atmospheric Environment, Vol. 41, No. 4, 2007, pp 677-691,
[2] Cai, H., Xie, S.: "Estimation of vehicular emission inventories in China from 1980 to 2005". Atmospheric
Environment, Vol. 41, No. 39, 2007, pp 8963-8979,
[3] European Automobile Manufacturers’ Association, "Vehicles in use Europe 2017". Available
from: https://www.acea.be/uploads/statistic_documents/ACEA_Report_Vehicles_in_use-Europe_2017.pdf,
Accessed 30.08.2018.,
[4] European Environment Agency, "Annual European Union greenhouse gas inventory 1990–2015 and inventory
report 2017", Available from: https://www.eea.europa.eu/publications/european-union-greenhouse-gas-
inventory-2017, Accessed 05.09.2018.,
[5] European Environment Agency Report, “Monitoring CO2 emissions from new passenger cars and vans in
2015”. EEA, Copenhagen, 2016. Available from: https://www.eea.europa.eu/publications/co2-emissions-new-
cars-and-vans-2016, Accessed 05.09.2018.,
[6] Kaplanović S., Ivković I, Manojlović A.: "Passenger cars taxation in european countries for the purpose of
solving environmental problems", Proceedings of the Fifth Scientific Conference "Road and the Environment",
Belgrade: Serbian Road Society "Via Vita", 2017, pp 177-181,
462
[7] Kelly, J. A., Ryan, L., Casey, E., O’Riordan, N.: "Profiling road transport activity: emissions from 2000 to 2005
in Ireland using national car test data", Transport Policy, Vol. 16, No. 4, 2009, pp 183-192,
[8] Papić V., Manojlović A., Momčilović V. Trifunović J.: "Određivanje količina emitovanih zagađujućih materija
poreklom od drumskog saobraćaja primenom copert iv modela evropske agencije za životnu sredinu", Institute
of Faculty of Transport and Traffic Engineering, Belgrade, 2010, Available
from: http://www.sepa.gov.rs/download/COPERT.pdf,
[9] Shaheen, S. A., Lipman, T. E.: "Reducing greenhouse emissions and fuel consumption: sustainable
approaches for surface transportation", Iatss Research, Vol. 31, No. 1, 2007, pp 6-20,
[10] Soylu, S.: "Estimation of Turkish road transport emissions", Energy Policy, Vol. 35, No. 8, 2007, pp 4088-4094,
[11] Uherek, E., Halenka, T., Borken-Kleefeld, J., Balkanski, Y., Berntsen, T., Borrego, C., Melas, D.: "Transport
impacts on atmosphere and climate: Land transport", Atmospheric Environment, Vol. 44, Issue 37, 2010, pp
4772-4816.
463
464
Kragujevac, Serbia
MVM2018-035
Tijana Ivanišević1
Dragan Taranović2
Radivoje Pešić3
Sreten Simović4
Vedran Vukšić5
ROLE OF VEHICLES ON ALTERNATIVE PROPULSION

SYSTEMS IN LOGISTICS AND TRANSPORT
ABSTRACT: Cities are places where large concentrations of people are present, and as such, they generate large
quantities of goods. In Europe, 80% of people live in urban areas. Distribution of goods is a fundamental prerequisite
for the maintenance of urban life and business activity that generates income and development of urban areas. The
distribution of goods within the cities leads to the emergence of "diseases in urban areas", such as traffic congestion,
noise, adverse environmental impact, road safety, etc. With a view to sustainable development, increasing attention
is paid to the development of environmentally friendly systems and goods transport technologies in the distribution
of goods. This paper analyzes the role of vehicles on alternative propulsion systems in logistics and transport.
KEYWORDS: transport, logistics, vehicles, alternative propulsion systems
INTRODUCTION
In Europe, 80% of people live in urban areas. Cities represent places of great concentration of people and generate
large quantities of goods [13]. With that in mind, cities represent the core of trade and in this sense can not function
without an adequate system of realization of commodity flows.
Research shows that 60% of cities have significant difficulties in managing CITY logistics, 55% of emissions are
caused by distribution of goods, while 40% of goods are delivered to city centers [18]. Every year, about 57 million
people are disturbed by the noise of road transport, with 42% having serious consequences [2]. At least a million
years of healthy life is lost each year due to traffic noise in Europe [17]. According to estimates, traffic noise costs
amount to about 40 billion euros a year, of which 90% is from road transport [2]. The energy of the transport sector
is made up of oil, in 97%, natural gas 2%, electricity 1% and renewable energy less than 0.5%. In the European
Union, 30% of total energy consumption is lost on transport, 43% of which are used by freight transport. The
development of urban areas results in the loss of land resources. Land loss in European capitals varies between
23% and 78% [5], Of which 30% belongs to transport, mainly to road infrastructure [7]. Transport in urban areas
makes up about 25% of emissions of harmful gases responsible for climate change, and almost all are attributed to
1
M. Sc. Tijana Ivanišević, University of Kragujevac, Faculty of Engineering Sciences, Sestre Janjic no. 6,
tijana.ivanisevic@mail.com
2
Associate prof. Dragan Taranović, University of Kragujevac, Faculty of Engineering Sciences, Sestre Janjić no. 6,
tara@kg.ac.rs
3
Prof. Radivoje Pešić, University of Kragujevac, Faculty of Engineering Sciences, Sestre Janjić br. 6,
pesicr@kg.ac.rs
4
Doc. Phd Sreten Simović, University of Montenegro, Faculty of Mechanical Engineering, Dzordza Vasingtona bb,
sretens@ac.me
5
B. Sc. Vedran Vukšić, Road Safety Center, Kumanicka no. 20e, vedran.vuksic@centarbs.com
465
road transport [6]. More than 1.2 million people die every year on the roads in the world [16]. About 50 million people
are injured annually in traffic accidents [16].
Distribution of goods is a fundamental prerequisite for the maintenance of urban life and business activity that
generates income and development of urban areas. Cities, today, represent the "core" of the trade and can not
function without the realization of commodity flows.
Initial - finishing operations in the supply chain is usually implemented by road transport [13]. Type and characteristics
of goods, influence the choice of mode of transport technology and logistics systems. Road transport is dominant in
the transport of high-value goods, easily perishable goods and/or goods with a regional character of production [13].
Bearing in mind that the largest number of commodity flows begins and/or ends in urban areas, it is easy to explain
the dominant role of road transport in the logistics of urban transport. The volume of road transport between 1995.
year and 2009. year in the EU increased by 46% [3], and according to forecasts, the trend of growth will continue (in
the EU by 40% by 2030. year, or by 80% by 2050. year compared to the 2005. year level) [4].
Freight transport has a constant growth. The functioning of cities is unimaginable without urban freight transport
[1,10]. In European developed areas over 80% of road freight is carried out at distances up to 80 km, which is defined
as urban or urban-regional transport [12]. A survey conducted by IVECO in nine European countries, more than 20
years ago, has shown that about 48% of vehicles go within urban and 32% of vehicles within suburban zones. In
Italy, for 70% of the transported goods final destination is located within the region of the provider. Over 50% of goods
transported is at a distance of 50 km, and 25% of goods within the city.
Logistic activities, primarily urban freight transport, are not desirable in urban areas. In order to maintain the economic
growth of cities and meet the demands of the generator of goods flows for more frequent deliveries of smaller volume,
there is an increase in the number of cargo and "small" delivery vehicles. In addition to the increased number of
cargo and "small" delivery vehicles intended for distribution of goods within cities, the distribution of goods within
cities leads to emissions, noise, energy consumption, occupation of land, climate change, traffic congestion, damage
to natural habitats and species, road safety, etc.
Conducted studies indicate that 60% of cities have significant difficulties in managing CITY logistics, 55% of
emissions are caused by distribution of goods, while 40% of goods are delivered to city centers [18].
The concept of sustainability should answer the question: How can a certain part of the cargo be transported, in an
efficient and environmentally friendly way, which transports goods at the shortest distance, without wasting time, with
minimal costs, occupying the least land and causing minimal environmental pollution? [11].
With a view to sustainable development, increasing attention is paid to the development of environmentally friendly
systems and road transport technologies, such as electric vehicles, hybrid vehicles, CARGO trams and other rail and
water transport systems in the distribution of goods [15].
ROLE OF VEHICLES ON ALTERNATIVE PROPULSION SYSTEMS IN

LOGISTICS AND TRANSPORT
The concept of sustainability should answer the question: How can a certain part of the cargo be transported, in an
efficient and environmentally friendly way, which transports goods at the shortest distance, without wasting time, with
minimal costs, occupying the least land and causing minimal environmental pollution? [11].
Several European cities have tried to implement ecological logistics solutions in their areas. The common solution
that came about was that the last kilometer of delivery would be achieved with the help of vehicles on alternative
propulsion systems [8].
According to Grainic [9], two-stage city logistics is intended for large cities. The use of two-stage city logistics reduces
the number of vehicles in the city center, but the number and length of empty trips is reduced. Two-stage city logistics
consists of two levels, the outer zone and the city zone, and has a main terminal, satellite terminal, eco vehicles, city
vehicles and empty vehicles (Figure 1).
466
Figure 1 Two-stage city logistics [9]
A model for city logistics usually has a distribution center near the city center where goods are produced and/or
where goods are collected from primary suppliers, and then delivered to retailers in urban areas [8].
The effects of city logistics, which is organized as shown in Figure 2 are: vehicles used for large volumes/weight are
not suitable for urban transport, the use of the vehicle may be low, especially when the number of customers in the
same city is not high, the number of vehicles moving in the city center for delivery of goods may be large, the distance
at which the delivery takes place may be high as the effect of a low level of transport efficiency [8].
Figure 2 Schematic representation of city logistics [8]
The application of vehicles on alternative propulsion systems for the delivery of goods in the last kilometer has several
advantages: reduction of traffic congestion as an effect of reducing vehicle traffic, noise reduction, pollution and
vibration, reduction in shipments per trader, and increased efficiency in transport [8].
Figure 3 Innovative logistic system [8]
467
The main constraints of the innovative logistics system are: every city has a transport center which shows that
solutions are applicable to medium-sized cities, the position of a transport center near the city center can negatively
affect traffic congestion, harmful emissions and noise in urban areas, and companies often are not interested in
leaving the last "step" of distribution to one of the cooperators for several reasons: revenues are limited, their brand
is not visible to the retailer, and can not sell door to the door [8].
Tadic, based on the existing problems, world experiences, demands for the modification of urban plans, etc. defines
four concepts of CITY logistics, which clearly and precisely define the need for vehicles on alternative propulsion
systems, which are [13]:
 CL1: Decentralized, satellite system with the dominant role of road transport.
 CL2: Centralized-decentralized system with the use of cargo trams.
 CL3: The core of the network with the use of cargo trams and electric vehicles.
 CL4: Network system with intermodal transport..
Concept CL1 implies a system of storage of goods on the perimeter of the city. For the needs of supplying the central
zone, the development of satellite terminals with access road infrastructure is envisaged. The goal of satellite
terminals is to transfer goods from larger to smaller delivery vehicles and consolidate flows. The function of CITY
logistics terminals would be the storage and distribution of goods to vehicles on alternative propulsion systems or
electric vehicles. In addition, CITY logistics terminals are envisaged to offer a return logistics service and delivery to
home address [13].
The CL2 concept implies the development of several CITY logistics terminals on the periphery of the central city zone.
Apart from having the role of storing goods and consolidating the delivery, these centers would also develop different
services in the form of logistics of return flows, delivery to home addresses, delivery to special areas for the takeover
of goods. Delivery of goods from distant locations of freight and transport centers or warehouses on the periphery of
the city to the nearest CITY logistics terminal would be realized by road transport, and between the CITY logistics
terminals would circulate "CARGO" tram. Distribution of goods from CITY logistics terminals to generators in the
gravity zone would be realized by the application of small delivery vehicles and vehicles on alternative propulsion
systems [13].
Figure 4 Concept CL1 [13] Figure 5 Concept CL2 [13]
The CL3 concept is the beginning of a complex CITY logistics network with two conveyance centers on the perimeter
of the city and several CITY logistics terminals on the periphery of the central city zone. This concept implies the
application of rail transport between the freight transport centers and the "CARGO" tram between the goods transport
centers and the associated CITY logistics terminals. The system of "CARGO" trams is also developed within the
central city zone, but in the function of supply, delivery to special areas for the takeover of goods and logistics of
return flows. Between the stations of "CARGO" trams and generators flows would be realized with not motorized
traffic and/or using a vehicle on alternative propulsion systems [13].
The CL4 concept represents the formation of a network of different categories of logistics centers and a significant
involvement of the railway in the realization of flows. In the area of the central city zone, CITY logistical terminal
would be developed to consolidate the supply of generators in the gravity zone, as well as the intermodal terminal.
These two systems have the option of connecting rail intermodal terminals in other locations. Goods cargo terminals
located on the periphery of the city would distribute goods using shuttle carts. A circular "CARGO" tram line would
conduct flows between the CITY logistics terminals. Distribution of goods in the zone of CITY logistics terminals
would be realized by the use of delivery vehicles on alternative propulsion systems [13].
468
Figure 6 Concept CL3 [13] Figure 7 Concept CL4 [13]
Zecevic [14] defines three scenarios for the development of the CBDD (Central Business Danube District) logistics
system:
 S1: Minimum infrastructure changes scenario. Introducing new technologies into existing logistic systems for
retaining the functions of the port and the railway freight station.
 S2: Significant changes scenario. The removal of a part of logistics and accompanying systems from the
observed area, the modernization of the intermodal terminal, the development of CITY logistics terminals
and the application of vehicles on alternative propulsion systems.
 S3: A complete change scenario. Dislocation of port and railway freight station, development of CLT minimal
configuration, application of cargo trams and vehicles on alternative propulsion systems.
Scenario S1 implies retention and modernization of existing structures and subsystems in the observed area. The
harbor would retain certain functions, and above all the function of intermodal transport. Bearing this in mind, further
development and modernization of intermodal terminals should be expected. Existing storage and distribution
systems could increase efficiency through the use of new technologies [14].
Scenario S2 involves the reduction of distribution and warehouse systems whose presence is not necessary in the
CBDD area, as well as freight forwarding, customs and other activities that are not necessary for the supply of the
CBD. This scenario involves IT modernization, as well as three modal node and CLT development for consolidated
supply of generators in the gravity zone. This scenario would result in the reduction of railway installations, but the
role of the railway in the efficient connection of this space would be strengthened. Through the CLT, CBD would be
supplied with small delivery vehicles on alternative propulsion systems [14].
Figure 8 Scenario S1 of CBDD logistics system [13] Figure 9 Scenario S2 of CBDD logistics system [13]
Scenario S3 implies the relocation of all existing facilities of the port complex and the railway freight station, and the
CBDD observation area remains a business-shopping center. Delivery of goods to the CITY terminal would be
realized from the logistic center at another location in the city, using "CARGO" trams, and the distribution to the
generator in the area of the central city zone and the CBDD would be carried out by vehicles on alternative propulsion
systems.
469
Figure 10 Scenario S3 of CBDD logistics system [13]
Scenarios can be "delimited" based on a set of criteria [13], according to: degree of traffic congestion by heavy goods
vehicles at access points and traffic infrastructures in the CBDD area, degree of space occupancy by logistics
systems that are not necessary in the CBDD area, investments for the development of the system, the costs of
delivery of goods, time losses in the transport of goods and services, the quality of logistics services, environmental
and energy aspects, security aspects, the complexity of logistics chains, technological and visual integration of
logistics systems in the urban environment.
CONCLUSIONS
Logistic systems should enable fast and reliable activities related to the delivery and collection of goods, which are
essential for the maintenance of life in urban areas and for economic prosperity.
Freight transport has a constant growth. The functioning of cities is unimaginable without urban freight transport [1,
10]. In European developed areas over 80% of road freight is carried out at distances up to 80 km, which is defined
as urban or urban-regional transport [12]. In Italy, for 70% of the transported goods final destination is located within
the region of the provider. The distance of transport of over 50% of goods is up to 50 km, and for 25% within the city
itself.
With the growth of the population and the economic development of urban areas, the problems of logistics flows are
also growing [13]. Realization of commodity flows has negative effects on the environment, traffic safety and quality
of life [13].
Several European cities have tried to implement ecological logistics solutions in their areas. The common solution
that came about was that the last kilometer of delivery would be achieved with the help of vehicles on alternative
propulsion systems.
The application of vehicles on alternative propulsion systems for the delivery of goods in the last kilometer has several
advantages: reduction of traffic congestion as an effect of reducing vehicle traffic, noise reduction, pollution and
vibration, reduction in shipments per trader, and increased efficiency in transport.
ACKNOWLEDGMENTS
This work was created as one of the results of the research carried out within the project TR 35041 supported by the
Ministry of Education, Science and Technological Development of the Republic of Serbia.
REFERENCES
[1] Allen, J., Tanner, G., Browne, M., Anderson, S., Christodoulou, G., Jones, P.: "Modelling policy measures and
company initiatives for sustainable urban distribution", Final Technical Report, 2003, London,
[2] Den Boer, L.C., Schroten, A.: "Traffic noise reduction in Europe/Health effects, social costs and technical and
policy options to reduce road and rail traffic noise", Report, Delft, 2007,
470
[3] EC.: "EU transport in figures: Statistical pocketbook 2011", Office for Official Publications of the European
Communities, 2011a, Luxemburg,
[4] EC.: "Accompanying the White Paper - Roadmap to a Single European Transport Area – Towards a competitive
and resource efficient transport system", Office for Official Publications of the European Communities, 2011b,
Luxemburg,
[5] EEA.: “The European Environment State and Outlook 2010-Land Use, State of the environment”, Report., No
1/2010., 2010c, European Environment Agency,
[6] EEA.: "Annual European Union greenhouse gas inventory 1990–2011 and inventory report 2013", Technical
report., No 8/2013., 2013c, European Environment Agency,
[7] EEA.: "A closer look at urban transport, TERM 2013: transport indicators tracking progress towards
environmental targets in Europe", 2013d, European Environment Agency,
[8] Faccio, M., Gamberi, M.: "New City Logistics Paradigm: From the "Last Mile" to the "Last 50 Miles", 7, 14873-
14894; doi:10.3390/su71114873, 2015,
[9] Grainic, G, T., Errico, F., Rei, W., Ricciardi, N.: "Integrating c2e and c2c traffic into city logistics planning", 2012,
The Seventh International Conference on City Logistics, https://ac.els-cdn.com/S1877042812005575/1-s2.0-
S1877042812005575-main.pdf?_tid=1e5fb5d3-325f-4314-9a8f-342b0b90c0c3&acdnat=1532007709_db9b8e1
4d640b3e2272639e216a4be5d, Accessed 19.07.2018.,
[10] Munuzuri, J., Larraneta, J., Onieva, L., Cortes, P.: "Solutions applicable by local administrations for urban
logistics improvement cities", Vol. 22., No. 1., 2005, pp 15-28,
[11] Melo, S.: "Evaluation of Urban Goods Distribution Initiatives towards Mobility and Sustainability: Indicators,
Stakeholders and Assessment Tools”, Ph.D. Thesis, Faculdade de Engenharia da Universidade do Porto, 2010,
Portugal,
[12] Ruesch, M., Petz, C.: "E-Commerce and urban freight distribution (home shopping)", Best Practice Update
(2008), Updated Handbook from Year 2001, 2008, BESTUFS Consortium,
[13] Tadic, S., Zecevic, S., Krstić, M. "Ranking of logistics system scenarios for central business district”, Promet -
Traffic&Transportation, 2014, Vol. 26, No. 2, pp 159-167,
[14] Zecevic, S., "The development concept of the logistics system in the Ada Huja area", Expert research, Urban
Planning Institute of Belgrade, 2006, Belgrade,
[15] Zecevic, S., Tadic, S.: "CITY logistics", Faculty of transport and traffic engineering, University of Belgrade, 2013,
Belgrade,
[16] World Health Organization: "The global burden of disease: 2004 Update", 2008,
[17] WHO/JRC (World Health Organization/Joint Research Centre): "Burden of disease from environmental noise",
Quantification of healthy life years lost in Europe, World Health Organization, 2011, Geneva,
[18] Wang, J., Chi, L., Hu, X., Zhou, H.: "Urban traffic congestion pricing model with the consideration of carbon
emissions cost", Sustainability 2014, Vol. 6, 2014, pp 676–691.
471
472
Kragujevac, Serbia
MVM2018-053
Slavica Mačužić1
Jovanka Lukić2
APPLICATION OF DIGITAL HUMAN MODEL IN ORDER TO

DETERMINE THE HUMAN BODY DISCOMFORT
ABSTRACT: Modern car buyers, in addition to style and safety, require and seating comfort. In the design stage
and car seat development, a posture of car drivers is a critical factor that has to be considered carefully. There is an
optimal seating position with accurately body angles for comfortable driving, and the seating position is more or less
the same for all vehicles, with small variations in the parameters of seating. Anthropometrics characteristics of the
driver has significant role in car design. The aim of this research is to analyze the driver's comfort, in the case of a
small city car, for different populations. Software package Ramsis was used to perform posture analysis of
“mannequin” in driving condition. Also, different angles of the inclination of the seat are observed. The obtained
values of fatigue and discomfort, as well as body parts discomfort, were different for all kinds of subjects. In this way,
car designers can know exactly how their car corresponds to anthropometric characteristic of any population, before
the actual process of making cars.
KEYWORDS: anthropometry, vehicle, driver, discomfort, Ramsis
INTRODUCTION
During on design of the car, special attention is paid to the comfort of the driver. The comfort of the driver is influenced
by the seat construction and the angle of seating. The driver's seat position is one of the essential factors for the car
production. When designing, it is necessary to take care of driver’s health and safety in driving condition. During the
drive and the full pedal activation, driver fatigue occurs. A long drive causes a back pain or other musculoskeletal
change. Modern vehicle testing is done through virtual testing using the virtual models [1]. In this way, the errors in
the initial phase are significantly reduced. The process of vehicle design requires knowledge of all anthropometric
characteristics of a large number of populations. Driving fatigue depends on the anthropometry of a driver [2, 3].
Different populations have a different influence on comfort [4]. Numerous studies have studied the impact of different
seat angles on the driver's comfort [5-7]. In this paper, analyzes were made for 5th percentile woman ("small woman")
and 95th percentile man ("tall man"). The position of the human model was investigated in the virtual environment of
the middle class vehicle (Figure 1). The comfort of the human model was examined during the driving. For this
purpose, the Catia v5 R18 software package (Part Design and Assembly modules) and Ramsis software packages
were used [8].
1
Slavica Mačužić, PhD student, University of Kragujevac, Faculty of Engineering, Kragujevac, Serbia, Sestre Janjic
6, slavicamacuzic89@gmail.com
2
Jovanka Lukić, Prof. Ph D, University of Kragujevac, Faculty of Engineering, Kragujevac, Serbia, Sestre Janjic 6,
lukicj@kg.ac.rs
473
Figure 1 Dimensions of the vehicle
METHODS
Digital human models have been used for testing in the automotive industry for many years. They help to define the
design of the vehicle and the position of the various components in the vehicle. The dimensions of the driver's body
represent the starting points in the dimensioning of the vehicle. In this paper, the first task was to create an interior
of a vehicle model. For this purpose, the Catia V5 R18 software package is used. The next task was to set the model
on the seat and set boundary conditions, using the Ramsis program package (Figure 2).
Figure 2 Mannequin in passenger space after setting boundary conditions
This software has the ability to work with two types of human models - kinematics and geometry. In order to achieve
the comfort of driving a seating position must always be in a comfortable condition [9]. This is specialized software
for ergonomics, which helped analyze seating comfort for different angles of sitting. A geometric model of the
mannequin with five fingers and foot legs with shoes was used. A virtual human model consists of two parts, muscular
tissue and skeleton. Human skin was not considered in the analysis because it is geometrically very complicated and
has little impact on the pressure distribution analysis [10]. This paper analyzes the comfort of eleven different male
and female populations during the driving process, for different seatbacuret angle. The anthropometric characteristics
are presented in Table 1 [11].
Table 1 The anthropometric data of 95th percentile man and 5th percentile woman populations
Height (mm) Sitting height (mm) Foot length (mm) Shoulder breadth
(mm)
Population 5% 95% 5% 95% 5% 95% 5% 95%
West Africa 1440 1785 741 881 212 278 355 450
North India 1450 1765 771 919 205 270 315 415
Eastern Europe 1540 1845 831 959 225 285 371 489
North Europe 1585 1910 846 999 227 280 367 498
Australia 1565 1885 831 979 220 280 345 489
South East Europe 1525 1825 811 949 220 285 372 485
Central Europe 1560 1805 826 989 220 288 376 495
South East Africa 1480 1775 771 909 210 280 362 465
Middle East 1529 1800 801 944 222 280 360 469
South India 1395 1715 746 874 200 265 330 435
North Asia 1504 1820 801 961 207 275 355 465
Ramsis software provides highly accurate analysis and comfort assessment using different human models.
Discomfort and fatigue were defined by numerical values based on the experiments performed. Comfort is presented
with a value of 2.5. Estimated value above 5.5 represents the occurrence of discomfort.
474
RESULTS
Ramsis software is based on a statistical analysis of the results obtained experimentally [12]. Using this software an
analysis of the interaction between driver and seat was conducted. Based on this, fatigue and discomfort values have
been obtained. An example of the obtained results is shown in Figure 3.
Figure 3 Example of the obtained results - the results of discomfort assessment in the case of West Africa 5th
percentile woman
The red column represents the reference value, that is, the values when the driver occupies during the driving
process, while the yellow column shows the resting state. Difference column is the difference between these two
columns and is an indicator of how much the driving position has a comfortable position or the opposite. This estimate
of discomfort gives the whole body fatigue in a given position, as well as the individual discomfort of each part of the
body separately.
Results of fatigue analysis for 95th percentile man for different angles of sitting are shown in Figure 4.
Figure 4 Fatigue analysis of 95th percentile man for four angles of sitting
From Figure 4 it can be noticed that the 90° angle of sitting is most comfortable to the West Africa and North India
populations, while the seating angle of 95° corresponds to the West Africa, North India and Eastern Europe
populations. The North Europe population corresponds most to the angle of the seat of 100°, while the Middle East
population corresponds to an angle of 105°. An analysis of fatigue results for female 5% population is shown in Figure
5.
475
Figure 5 Fatigue analysis of 5th percentile woman for four angles of sitting
Figure 5 shows that the 90° angle corresponds to the West Africa population, and the angle of 95° corresponds to
the Middle East and South India populations. The North Asia, South India and Middle East populations correspond
to the seating angles of 100 and 105°.
An analysis of discomfort for 95th percentile man for different angles of sitting is shown in Figure 6.
Figure 6 Discomfort analysis of 95th percentile man for four angles of sitting
From Figure 6 it can be seen that the greatest discomfort for all seating angles suffers the populations: Australia,
South East Europe, Central Europe and South East Africa. Figure 7 shows a discomfort analysis for female 5%
populations for different angles of sitting.
Figure 7 Discomfort analysis of 5th percentile woman populations for four angles of sitting
Figure 7 show that the smallest discomfort for seating angles of 95, 100 and 105° has the populations of Middle East,
South India and North Asia. The smallest discomfort for the 90° angle of seating has the populations from West Africa
and North India. On the basis of the results shown, it can be concluded that the lowest comfort is given by a seat
with a 105° backrest. Table 2 gives descriptive characteristics of discomfort and fatigue for both populations for the
angle of sitting of 105 degrees.
476
Table 2 Descriptive characteristics of discomfort and fatigue for 95th percentile man populations and 5th percentile
woman populations
Fatigue Discomfort
Man Woman Man Woman
Mean 4.34 3.77 7.02 4.61
Standard Error 0.17 0.04 0.39 0.04
Median 4.6 3.8 7.7 4.7
Mode 4.8 3.8 7.9 4.7
Standard Deviation 0.57 0.14 1.3 0.15
Sample Variance 0.33 0.02 1.69 0.02
Kurtosis -2.03 4.90 -0.77 0.32
Skewness -0.25 -1.97 -1.16 -1.03
Range 1.4 0.5 3 0.5
Minimum 3.6 3.4 4.9 4.3
Maximum 5 3.9 7.9 4.8
Sum 47.7 41.5 77.2 50.7
Count 11 11 11 11
Largest(1) 5 3.9 7.9 4.8
Smallest(1) 3.6 3.4 4.9 4.3
Confidence Level(95.0%) 0.38 0.09 0.87 0.10
CONCLUSIONS
The aim of this paper was to determine the fatigue and discomfort of drivers during driving for different angles of
sitting. Digital human models were used to assess discomfort and fatigue. The analyzes were carried out using the
Ramsis software package. Eleven populations of 95th percentile man populations and 5th percentile woman
populations were examined. The results obtained for all populations have a permissible fatigue value below 5.5,
which represents a limit value. Fatigue analysis of male 95th percentile man showed that the 90° angle of sitting is
most suited to the West Africa and North India populations, while the seating angle of 95° corresponds to the West
Africa, North India and Eastern Europe populations. The North Europe population corresponds to the seat angle of
100°, while the Middle East population corresponds to an angle of 105°. In the case of fatigue analysis of 5th
percentile woman population showed that the 90° of seat angle corresponds to the West Africa population and the
angle of 95° corresponds to the Middle East and South India populations. The North Asia, South India and Middle
East populations correspond to the seating angles of 100 and 105°.
Analyzing the driver's fatigue, it is concluded that there is less fatigue in the case of women (5% populations) than in
the case of men (95% populations), for the same angles of sitting. Also, the same case is with discomfort analysis. It
can be concluded that the height of the driver plays a major role in this analysis. Women are lower and consequently
there is less fatigue for the same seat angles compared with man.
Vehicle manufactures have the opportunity, using the Ramsis software package, to graphically display how their
digital human model interacts with its environment inside the vehicle. The benefits of such research during the phases
of vehicle design can save a lot of money and predict where manufacturers can sell their products.
ACKNOWLEDGMENTS
This research supported by the Ministry of Education, Science and Technological Development of Republic of Serbia
through Grant TR 35041.
REFERENCES
[1] Duffy, V. G.: "Handbook of Digital Human Modeling: Research for Applied Ergonomics and Human Factors
Engineering", CRC Press Inc. Boca Raton FL USA, 2008,
[2] Helander, M.G., Zhang, L.: "Field studies of comfort and discomfort in sitting", Ergonomics, Vol. 40, 1997, pp
895–915,
[3] Vink, P.: "Comfort and design: principles and good practice", Boca Raton: CRC Press, 2005,
[4] Bridger, R. S.: "Introduction to Ergonomics", Boca Raton: Taylor & Francis Group, LLC, 2009,
[5] Graf, M.C.: "Effects of seat angle on comfort and lower back pain at work Swiss federal institute of technology",
Zurich: ETH, 2004, pp 182,
477
[6] Lapa, V.V., Neverov, G.I., Poliakov, V.V., Solov'ev, N.N.: "Effect of inclination angles of supporting surfaces of
aircraft seat on pilot's working capacity in longduration flight", Aviakosm Ekolog Med, Vol. 34, No. 3, 2000, pp
17-21,
[7] Schmidt, S., Amereller, M., Franz, M., Kaiser, R., Schwirtz, A.: "A literature review on optimum and preferred
joint angles in automotive sitting posture", Applied Ergonomics, Vol. 45, Issue 2, Part B, March 2014, pp 247-
260,
[8] RAMSIS 3.8 in CATIA V5 User Guide - HUMAN SOLUTIONS GmbH, 29-November-2012,
[9] Bernaer, J. D.: "Ergonomic for beginners", Second Edition, Taylor and Francis, 2001,
[10] Ebe, K ., Griffin, J. M.: "Factors influencing static seat cushion comfort", Taylor and Francis, 2001,
[11] DINED / Anthropometry in design, Available from: http://dined.io.tudelft.nl/en, Accessed 01.05.2017.,
[12] Heiner Bubb, H., Bengler, K., Grünen, R., and Vollrath,: "Automobilergonomie", Springer Vieweg, 2015.
478
Kragujevac, Serbia
MVM2018-057
Slavica Mačužić1
Igor Saveljić2
THREE DIMESIONAL ANALYSES OF SEAT BELT AND

DRIVER IN CASE OF SUDDEN BRAKING
ABSTRACT: A seat belt is a vehicle safety device designed to secure the occupant of a vehicle against harmful
movement that may result during a collision or a sudden stop. In recent years, serious and fatal accidents have
decreased in number. The reason is that more and more people begin to understand the importance of using
safety belts. When driver makes a sudden stop, the speed from the car causes driver to hit the seat belt with
significant force. That force saves life, but can leave passenger with bruises, scrapes and fractures. In this paper
the influence of different force values on the abdomen of the driver by the seat belt is analysed. As a result, the
different speed values of the car, due to sudden braking, give different values of force which the seat belt acts on
the driver.
KEYWORDS: driver, injury, vehicle, finite element method
INTRODUCTION
Traffic accidents are a major source of fatalities and serious injuries [1]. The major cause of car occupant fatalities
in car accidents is head injuries which occurs when vehicle passengers are thrown out through the windshield or
when they collide with the vehicle seat or dashboard. Based on research the World Health Organization [2] injury in
road traffic crashes is expected to become the third most important cause of the global burden of injuries/diseases
by 2020. More than 1.25 million people die each year as a result of road traffic accidents. Globally, road traffic
crashes are a leading cause of death among young people, and the main cause of death among those aged 15–29
years (Figure 1).
Based on all the research, a lot of effort has been invested to change the awareness of drivers and all passengers
in vehicles. The most important form of passenger protection in the car is the use of the seat belt. Seat-belts limit
the movement of vehicle occupants in the event of a crash in order to reduce the force that currently operates on
the passenger in the vehicle. Wearing a seat-belt reduces the risk of a fatality among drivers and front-seat
occupants by 45–50%, and the risk of minor and serious injuries by 20% and 45% respectively. In the case of
passengers in the rear seat, the use of seat belts reduces fatal and serious injuries by 25% and minor injuries by
up to 75% [3]. Over time, there is progress and development of awareness about the application of good laws in
the binding of seat belts, but still only half of the country's world has adopted a good seat-belt laws. About 4.8
billion people, from 105 countries have seat belts in the back seat.
1
Slavica Mačužić, Phd student, University of Kragujevac, Faculty of Engineering, Sestre Janjić 6, 34000
Kragujevac, Serbia, slavicamacuzic89@gmail.com
2
Igor Saveljić, Ph.D., University of Kragujevac, Faculty of Engineering, Sestre Janjić 6, 34000 Kragujevac, Serbia,
isaveljic@kg.ac.rs
479
Figure 1 The most common causes of death of young people under 30 years [2]
Figure 2 Seat-belt laws in the world
Experience from developed countries indicates that the use of seat belts is one of the most effective ways to
reduce road accident fatalities [4, 5]. In this paper, using numerical simulations, it has been shown that the force of
the sudden braking extends through the torso of an adult male person, previously tied to the safety belt.
MATERIALS AND METHODS

The finite element model of a human torso was developed using the CT scans. The high resolution CT data were
read into Mimics 14.0 ((Materialise Inc., Leuven, Belgium) visualization software, where the images were
segmented by thresholding to obtain 3D model (Figure 3).
480
a) b)
Figure 3 The process for the 3D FE model: a) CT scan and 3D reconstruction of the CT scans, and b) meshing of
the 3D FE model
The whole FE torso model was divided into two parts, including thoracic and abdominal soft tissues and thoracic
cage (Figure 4). In this paper, the aim was to examine the values of the load on the chest, and the pressure forces,
so we did not take into account the human internal organs.
Figure 4 Finite element model of the human torso with tissue and thoracic cage
The material properties assignment is detailed in Table 1.
Table 1 Material properties of used materials

Parts Material type Parameters Reference
ρ = 2.0 g/cm3, E = 11.5
Rib (cortical) Elastic [6]
GPa, µ = 0.3
ρ = 1.0 g/cm3, E = 0.04
Rib (trabecular) Elastic [6]
GPa, µ = 0.45
ρ = 1100 kg/mm3, E =
Tissue Elastic [7]
10 MPa, µ = 0.3
In this paper the model of an adult man weighing 85 kg, height 180 cm was used. Three cases of speed of car
movement and sudden shock of an irreversible obstacle were analysed. The speed of the car was 40, 50 and 60
km/h. Also, the driver used the seat belt. According to [9], the following force values applied to the driver were
used: 24.7 kN, 38.5 kN and 55.5 kN, for 40, 50 and 60 km/h, respectively.
One of the basic principles of continuum mechanics is the principle of virtual work. Starting from the equilibrium
equations [8] by applying the boundary conditions can be equal to the virtual work of internal and external forces
Wint  Wext (1)
Virtual work of the previous equation in matrix form can be written as:
481
 Wint    eT σdV  Wext    uT FV dV    uT F S dV   uT Fi  (2)
V V S i
Applying the principle of virtual work and the constitutive relations for linear elastic material in matrix form
σ  Ce (3)
and by applying the concept isoparametric interpolation [9] in the finite element, on the basis of which the
coordinates and displacements at any point within the element is
x  NX u  NU (4)
we can write the equation of equilibrium finite elements:
KU  Fext (5)
where K is element stiffness matrix, B - strain relation matrix - displacements at the nodes, which contains excerpts
interpolation function, C - elastic constitutive matrix, e = BU - matrix deformation, U - displacements at the nodes, X
- coordinates of nodes, N - matrix of interpolation functions, Fext - external forces in the element nodes.
In the linear analysis of solids a basic assumption is that the moving solids are infinitesimally small and that the
material is linearly elastic. Also, the assumption is that the nature of the boundary conditions remains unchanged
under the action of external loads. Under these assumptions, the equations of equilibrium are derived for finite
element structural analysis. Equation (5) is related to the linear analysis of solids because the moving U is linear
function of external forces Fext. In the case when displacement is not linearly dependent of the load, nonlinear
analysis is applied.
In the linear analysis, the assumption that the displacement must be small is applied in calculating the stiffness
matrix and force vector because all the volume integrals are applied to the original volume of the finite elements.
Also, the matrix B, which relates strain and nodal displacement is constant for each element and is independent of
the displacement element nodes. For a linear elastic material is assumed that the constitutive matrix is constant.
Fixed boundary conditions are also the default in the linear analysis.
RESULTS
Numerical simulation was performed using the Ansys software package 14.5. Three cases of speed of car
movement and sudden shock were analysed. The speed of the vehicle was 40, 50 and 60 km/h. Following force
values applied to the driver were used: 24.7 kN, 38.5 kN and 55.5 kN, for 40, 50 and 60 km/h, respectively. Figure
5 represent the results of the first simulation.
Figure 5 The numerical result of the von Misses stress when a force of 24.7 kN was applied on the chest part
The result of this simulation shows that the highest stress occurs on the chest part, due to the influence of the belt,
where the maximum value was 2.15 MPa. Figure 4 and 5 shows the stress distribution for the remaining cases.
482
From the previous figures, Figure 6 and Figure 7, it can be concluded that the maximum value of the stress is
always in the chest. The highest stress values of the last two measurements are 3.35 and 4.84 MPa, respectively.
It can be concluded that the dependence of force and stress is rather linear (Figure 8).
Figure 8 A diagram of the dependence of force and stress generated on the body of the model
CONCLUSIONS
Previous section shows the research results of the numerical simulation driver interactions and seat belts during a
sudden impact on a fixed obstacle. Numerical simulations were done using Ansys 14.5 was used. Finite element
model of the human torso with tissue and thoracic cage was used. Three cases of speed of car movement and
sudden braking of an irreversible obstacle were analysed. The speed of the car was 40, 50 and 60 km/h. Also, the
driver used the seat belt. The following force values applied to the driver were used: 24.7 kN, 38.5 kN and 55.5 kN,
for 40, 50 and 60 km/h, respectively. The results of this study show where the greatest pressure on the human
body was during sudden braking. The largest load zone is the upper part of the chest, where the values of the von
Misses stress reach the highest values of 4.84 MPa, for the speed of a car of 60 km/h.
The direction of future research will be to investigate how internal organs suffer from sudden pressure, as well as
the way the shifts affect them.
483
ACKNOWLEDGMENTS
This research supported by the Ministry of Education, Science and Technological Development of Republic of
Serbia through Grant TR 35041.
REFERENCES
[1] World Health Organization: "United Nation road safety collaboration: a handbook of partner profiles", Geneva:
WHO, 2005, pp 1-5,
[2] World Health Organization: "World report on road injury prevention", Geneva: WHO, 2005, pp 37-39.
[3] Elvik, R.: "The handbook of road safety measures", 2nd ed. Bingley, UK: Emerald Group Publishing Limited,
2009,
[4] Canada: crash modes and kinematics of injury, Accident Analysis and Prevention Journal 26, 1994, pp 207-
214,
[5] Evans, L.: "Safety-belt effectiveness: The influence of crash severity and selective recruitment", Accident
Analysis and Prevention Journal, Vol. 28, No. 4, 1996, pp 423-433,
[6] Li, Z., Kindig, M.W., Kerrigan, J.R., Untaroiu, C.D., Subit, D.: “Rib fractures under anterior–posterior dynamic
loads: Experimental and finite- element study", Journal of Biomechanics, Vol. 43, No. 2, 2010, pp 228–234,
[7] Lizee, E., Robin, S., Song, E., Bertholan, N., Lecoz, J.Y., Besnault, B., Lavaste, F.: "Development of 3D Finite
Element Model of the Human Body", SAE Transactions, SAE Paper number 983152, 1998, pp 2760–2782,
[8] Kojic, M., Filipovic, N., Stojanovic, B., Kojic, N.: "Computer Modeling in Bioengineering - Theoretical
Background, Examples and Software" John Wiley and Sons, 2008, Chichester, England,
[9] Bathe, K.J.: "Finite Element Procedures", Prentice-Hall, Inc., Englewood Cliffs, 1996, N.J.
484
SECTION
Road Traffic Safety
 Safety
 Ecology & Energy
 Legislation
 Socio-economic Issues
 Personal/Public/Freight
transportation
 Maintenance & Aftermarket
485
486
Kragujevac, Serbia
MVM2018-006
Slobodan Mišanović1
Predrag Živanović2
Dušan Savković3
Pavle Krstić4
Slobodan Ivanov5
Slobodan Stević6
TWO-YEAR SUCCESSFUL EXPLOITATION

OF THE ELECTRIC BUSES IN BELGRADE
ABSTRACT: Since, September 1, 2016, new line EKO 1 is established in Belgrade, where are five E-buses in
operation. What are the experiences in exploitation in terms of realized transport, maintenance, electric consumption,
environmental benefits as well as the satisfaction of passengers, who use the line EKO 1, are the subjects of this
paper. In paper will also present future plans for further expansion of the network line in Belgrade, where the electric
buses will operate.
KEYWORDS: e-bus, exploitation, development
INTRODUCTION
City Transport Company "Belgrade'' (JKP GSP ''Beograd''), is the bearer of public transport in Belgrade and one of
the largest public transport companies in South East Europe. Every day in the operation is: 609 buses, 150 trams
and 94 trolleybuses and 5 buses electrically powered. From 1st September 2016, introduced a new line of EKO 1
where operate buses exclusively on electric power. In this way, Belgrade is included in the map of cities in Europe
and the world that have begun using electric buses as a long-term strategy for using this concept of drives, which will
be the main alternative to diesel-powered buses.
The choice of the electric bus concept came after several years of activity in monitoring the development and
application of these buses in many cities in Europe and the world, as well as successful cooperation with many bus
manufacturers (Higer, BYD, Solaris, Siemens-Rampini). Also, City Transport Company "Belgrade'' has been active
in many EU projects related to electric buses (ZEEUS, Hybrid user forum, Civitas, UITP bus committee ...). Specificity
of using electrically powered buses can be seen in terms of line on which they work, vehicle concept, charging
systems, maintenance, exploitation indicators and environmental suitability.
1
Slobodan Mišanović, Project manager, City Public Transport Company “Belgrade“, Kneginje Ljubice 29, 11000
Belgrade, PhD candidate, University of Kragujevac, Faculty of Engineering, slobodan.misanovic@gsp.co.rs
2
Predrag Živanović, PhD, prof.assist. Faculty of Transport and Traffic Engineering, University of Belgarde,
p.zivanovic@sf.bg.ac.rs
3
Dušan Savković , Msc.M.E, Coordinator for tehnical jobs, City Public Transport Company “Belgrade“, Kneginje
Ljubice 29, 11000 Belgrade, dusan.savkovic@gsp.co.rs
4
Pavle Krstić, Msc.E.E, City Public Transport Company “Belgrade“, Kneginje Ljubice 29, 11000 Belgrade,
pavle.krstic@gsp.co.rs
5
Slobodan Ivanov, Msc.E.E, City Public Transport Company “Belgrade“, Kneginje Ljubice 29, 11000 Belgrade,
slobodan.ivanov@gsp.co.rs
6
Slobodan Stević, Msc.E.E, City Public Transport Company “Belgrade“, Kneginje Ljubice 29, 11000 Belgrade,
s.stevic@gsp.co.rs
487
CITY LINE EKO 1 (VUK'S MONUMENT – BELVIL)
The choice of a new line (Figure 1) on which electric buses operate, comes after a detailed analysis of the
fulfilment of the following criteria [1]:
 A central city line, so that the environmental impact of the "0" emission is the biggest.
 The high attractiveness of line from the aspect of passenger requirements.
 Suitability of the line or terminal from the aspect of providing energy requirements for chargers
 Line length, such that at the end of the journey there is a minimum of 20% of the available power in the
supercapacitors.
Figure 1 Route of line EKO 1 Figure 2 E-bus HIGER KLQ6 at the “Vuk's
monument’’ charging station
The mean length of the EKO 1 line is 8 km. The line is with a flat configuration with a slight climb. On the line work 4
buses ''Higer KQL6'' City E-bus bus, with an average interval of 20 minutes.
''Higer KQL6'' (Figure 2), City E-bus bus has the capacity of minimum 80 passengers. E-bus is equipped with two
traction asynchronous motors “Siemens IPV5135’’, 2x67 kW nominal power, 2x90 kW peak power. System for storing
electricity realised by supercapacitor “Aowei“, 20 kWh capacity [2]. Chargers for fast charging have the power of 150
kW. The charging station on “Vuk’s Monument” terminus is connected to the DC voltage (DC) from the tram contact
grid and the charging station on “Belvil’’, to the three-phase AC voltage from the public power grid. The charging time
at the terminals is 5-10 minutes.
EXPLOITATION INDICATORS ON THE LINE EKO 1

Analyzing the period of operation of buses on electric power from 01/09/2016 to 1/9/2018, E-buses have had the
following results [3]:
 Working hours per vehicle per day: 1618 h;
 Average daily mileage per vehicle: 190215 km;
 Exploitation speed: 14.8 km·h¯1;
 Daily number of passengers transported per vehicle: 9001200 passengers;
 Reliability of work on the line: 97.5%;
 Depending on operating mode, number of passengers, traffic conditions, driving style, impact of the system
for heating and air conditioning of the vehicle), electric consumption may vary:
 In the spring/autumn period, direction "A" 0.821.15 kWh·km¯1 , direction "B" 1.21.45 kWh·km¯1
 In the summer period average consumption is higher by 23.3%, than the transition period;
 In the winter period average consumption is higher by 45.4%, compared with transition period;
 Loss of electricity in the charging phase (network, charger, pantograph, super capacitor): about 5%;
 E-bus realized recovery of electricity in the braking phase of about 25-30% compared to the energy
consumed to drive.
An example of the exploitation indicators of the operation of the E-bus (garage number 2103) on the line EKO 1, for
the day 18 April 2018, is presented in Tables 1 and 2. The outside temperature was min 12 ˚C, max 18 ˚C [4].
488
Table 1 Direction’'A'' (Vuk's-Belvil)-Expoloitation indicators, 18 April 2018
Electric Loss in Electric
Timetable energy- the charging *SOC SOC consumption Electric Driving Exploatation
[hh:mm:ss charging phase [Vuk] [Belvil] SOC (SOC+loss) consumption time speed
start end [kWh [kWh [% [%] [%] [kWh] [kWh·km-1] [hh:mm:ss [km·h-1]
5:14:00 5:45:00 5.00 0.25 98.00 62.00 36.00 7.38 1.02 0:31:00 14.46
7:13:00 7:45:00 10.87 0.54 99.00 62.20 36.80 7.54 1.08 0:32:00 14.01
8:46:00 9:21:00 10.17 0.51 99.60 55.20 44.40 9.10 1.29 0:35:00 12.81
10:20:00 10:49:00 9.86 0.49 99.00 69.00 30.00 6.15 0.89 0:29:00 15.46
11:55:00 12:27:00 10.70 0.54 100.00 66.00 34.00 6.97 1.00 0:32:00 14.01
13:31:00 14:05:00 9.75 0.49 94.75 55.00 39.75 8.15 1.16 0:34:00 13.18
15:04:00 15:40:00 11.30 0.56 99.00 62.60 36.40 7.46 1.07 0:36:00 12.45
16:39:00 17:14:00 10.21 0.51 88.00 49.70 38.30 7.85 1.12 0:35:00 12.81
18:12:00 18:42:00 11.34 0.57 94.00 55.00 39.00 8.00 1.15 0:30:00 14.94
19:46:00 20:18:00 11.69 0.58 98.00 70.00 28.00 5.74 0.85 0:32:00 14.01
21:19:00 21:46:00 10.25 0.51 98.00 65.00 33.00 6.77 0.97 0:27:00 16.60
*SOC (State of charge supercapacitor)
Table 2 Direction’'B'' (Belvil-Vuk's)- Expoloitation indicators,18 April 2018

Electric Loss in Electric
Timetable energy- the charging *SOC SOC consumption Electric Driving Exploatation
[hh:mm:ss charging phase [Belvil] [Vuk] SOC (SOC+loss) consumption time speed
start end [kWh [kWh [%] [%] [%] [kWh] [kWh·km-1] [hh:mm:ss [km·h-1]
6:22:00 6:53:00 8.18 0.41 101.90 48.00 53.90 11.05 1.35 0:31:00 16.45
8:01:00 8:37:00 7.79 0.39 100.00 46.00 54.00 11.07 1.35 0:36:00 14.17
9:34:00 10:06:00 7.54 0.38 99.00 50.00 49.00 10.05 1.23 0:32:00 15.94
11:09:00 11:42:00 9.39 0.47 101.00 50.90 50.10 10.27 1.26 0:33:00 15.45
12:44:00 13:20:00 6.36 0.32 100.00 47.80 52.20 10.70 1.30 0:36:00 14.17
14:18:00 14:52:00 7.18 0.36 101.00 47.20 53.80 11.03 1.34 0:34:00 15.00
15:52:00 16:28:00 9.43 0.47 101.00 43.90 57.10 11.71 1.43 0:36:00 14.17
17:26:00 18:05:00 7.67 0.38 100.00 38.20 61.80 12.67 1.54 0:39:00 13.08
18:59:00 19:37:00 10.31 0.52 100.00 38.70 61.30 12.57 1.54 0:38:00 13.42
20:33:00 21:07:00 9.23 0.46 100.00 41.00 59.00 12.10 1.48 0:34:00 15.00
22:08:00 22:37:00 6.15 0.31 100.00 48.00 52.00 10.66 1.29 0:29:00 17.59
*SOC (State of charge supercapacitor)
Realized mileage per vehicle a given period is presented in table 3.
Table 3 Mileage per vehicle

E-bus Period [1/9/2016 - 1/9/2018*
Mileage [km]
2101 104.400
2102 106.320
2103 109.050
2104 100.350
2105 103.100
* In the period 14/8-6/10/2017, E-buses not in operation because of the works in Roosvelt’s street.
MAINTENANCE OF E-BUSES
Buses on an electric drive as a relatively new concept of buses used in public transport are characterized by certain
specifics when maintenance is concerned [3]. Some of these specificities compared to diesel buses are:
 Simpler maintenance compared to the diesel bus;
 Vital parts of the electrical components (inverters, converters, air compressors, steering pump,
supercapacitor) are a modular type (Figure 3,4);
 Lower maintenance costs (compared to a diesel bus about 3 times, E-bus 3000 Euros per year, Diesel 9000
Euros per year);
 Diagnosis of defects are identified on the instrument panel;
 Short replacement time.
489
Figure 3 Traction inverters Figure 4 Compressor, steering pump
In Table 4, shows the components and systems on the buses with a diesel-powered and electric-powered bus.
Table 4 Components and systems on the buses with a diesel-powered and electric-powered bus
The electrically powered bus has a significantly smaller number of components and systems, which makes it more
efficient and cheaper in terms of regular and corrective maintenance. In E-buses, regular servicing is performed
according to the defined checklist at every 20,000 km.
ECOLOGICAL EFFECTS ON THE LINE EKO 1

One of the main reasons for introducing E-buses on the line EKO 1 is the environmental effects compared to diesel
buses [3]. This relates primarily to:
 The smaller the level of noise, compared to a diesel bus lower by 13 dB(A) [3].
 "0" emission of harmful gases. Comparison of the emissions of harmful gases of one E-bus and diesel buses
on the line ECO 1 for annual mileage 60,000 km, consumption 44 L/100 km, present in Table 5.
Analysing the period of operation of buses on electric power from 01/09/2016 to 15/5/2018, E-buses have had the
following results [3]:
490
Table 5 Comparison of the emissions of harmful gases from one E-bus and diesel buses on the line ECO 1
Emissions of diesel Emissions of diesel Emissions of diesel Emissions of diesel
E-bus
Pollutant buses [Euro 3 buses [Euro 4] buses [Euro 5-EEV] buses [Euro 6]
[kg/year] [kg/year] [kg/year] [kg/year] [kg/year]
CO 0 206 147 147 147
CxHy 0 64.7 45.2 24.3 12.75
NOx 0 490.4 342.1 195.7 39.2
PM 10 0 9.8 1.96 1.95 0.95
E-BUS LINE- USERS SATISFACTON

Three months after the EKO1 line started its operation; a survey of the route’s passengers was conducted in order
to get acquainted with their opinions and first impressions. The survey was conducted in November 2016, during
vehicles operation and lasted for three days [5]. The largest group of the EKO1 line passengers composed of
employed users and the largest percentage of the passengers have been using the line daily.
Questions from the survey related to the satisfaction of passengers from the EKO1 line included the following:
 line route,
 arrival interval of vehicles,
 features of e-buses,
 comfort and e-bus equipment
(interior and exterior design, noise and safety and other information in a vehicle, wi-fi ...).
Around 80% of the passengers are satisfied with the line route, and 2% of the passengers are completely dissatisfied
with it.
The vehicles interior and exterior design, air conditioning and heating system and the noise level in them got a
similarly high score. Even 99% of the passengers feel completely safe in the vehicles with purely electric drive and
over 90% of the passengers are satisfied with the information system.
The EKO1 line passengers give it the highest grades. Even 73% of the passengers give it to grade 5, and 25% of
them give it a high grade 4. Thus, the average grade for the EKO1 line in the opinion of its users is 4.7.
Figure 5 The total grade score for the EKO1line
FUTURE PLANS
After two years of successful exploitation of E-buses in Belgrade and satisfactory results, JKP GSP ‘’Beograd’’ and
the City of Belgrade are planning to develop the concept of the city electric bus drive on city lines [6].
The plan is to purchase of 80 E-buses in the next 3-4 years, which will replace trolleybuses that work on lines:
19,21,21,29. These lines pass through the city centre and the idea is to dismantle the contact network, due to the
high costs of maintenance and frequent failures that initiate large congestion and crowds in the centre of the city. For
491
the realization of this project, it is necessary to complete the construction of the new central city terminal in Dunavska
Street (work in progress), ensuring sufficient capacity of the electric network for the operation of chargers than 360
kW. The spatial position of the lines: 19, 21, 22, 29 on which electric buses will operate as well as the required
number of chargers at the terminals is presented in Figure 6.
Eko 19, 21, 22, 29
EKO 21
EKO 22
EKO19
EKO 29
Figure 6 New E-bus lines in Belgrade
It is planned that E-buses will replace diesel buses on line 77 (Zvezdara-Bezanijska Kosa) that passes through the
city centre in order to reduce the emissions of harmful gases in the city centre, as well as the introduction of new E-
bus line will be connected the airport ‘’Nikola Tesla’’ with the central city zone.
CONCLUSIONS
The introduction of buses with electric drive in the regular operation is a significant step in the development of public
transport system in Belgrade. The concept of E-buses with the pantograph charging system on the termini completely
meets the conditions of exploitation in terms of electricity supply, the daily autonomy, and the passengers’ transport
demands.
The new EKO 1 bus line with E-buses will enable the full effect of the use of this concept in terms of environmental
requirements. Tracking the results of the use of electric buses on the EKO 1 line will serve as the best argument for
defining future strategies of public transport in Belgrade, regarding this concept of buses and its mass application.
The introduction of electric buses on lines: 19, 21, 22, 29, 77 is compliance with the EU transport policy, which
determines that participation of "clean" buses in the cities of Europe will be 50% by 2025 (amendment of the directive
EC/33/2009) or 30% for cities from EU countries with less economic power.
REFERENCES
[1] Mišanović S., Tica S., Milković Ž., Krstić P., Milovanović B.: "Ecology and energy aspects of exploitation fully
electrical buses on the new line in public transportation Belgrade", International Congress Motor Vehicles &
Motors 2016, Kragujevac, 6-7 October 2016, pp 33-40, ISBN 978-86-6335-037-3,
[2] Higer Electric bus – Tehnička dokumentacija, 2016,
[3] Mišanović, S., Milković, Ž., Savković, D., Krstić, P., Stević, S.: "Specificity of exploitation and maintenance of
electric bus in Belgrade", 3rd Maintenance Forum 2018- Maintenance and Industry 4.0, Belgrade 24-26th May
2018, Proceedings, ISBN 978-86-84231-43-9, COBISS.SR-ID 264166668, pp 119-127,
[4] ЈКП ГСП ''Београд'', Baza podataka potrošnje električne enegije E-busa, interni dokument, 2017- 2018,
[5] ЈКП ГСП ''Београд'', Истраживање задовољства путника на линији ЕКО 1 у Београду, интерни документ,
2016,
[6] Fabula Konsultant, Beoconex Beograd, Studija opravdanosti nabavke 80 elektroautobusa u Beogradu, 2018.
492
Kragujevac, Serbia
MVM2018-007
Milja Simeunović1
Zoran Papić2
Milan Simeunović3
Pavle Pitka4
THE ROLE OF APPLYING ELECTRIC VEHICLES IN

SUSTAINABLE URBAN MOBILITY
ABSTRACT: Urban agglomerations are increasingly burdened by motorized vehicles, not only in terms of over-
occupation of the city space, but also by the noise and pollution they produce. On the one hand, it is necessary for
all inhabitants to provide transportation options that enable access to key destinations and services, and on the other
hand, they raise issues of sustainability and environmental protection. One of the key environmental influences is
traffic. For decades, alternative modes of transport have been sought, which would have the effect of reducing the
environmental impact of motor vehicles. Use of electric vehicles is one of the acceptable solutions that has been in
expansion for the last few years. Electric vehicles are still not sufficiently present in traffic due to aggravating
circumstances, of which the most important ones are the prices of these vehicles in relation to motor vehicles and
underdeveloped charging grids. Some countries have timely recognized benefits of this mode of transport, thus
introducing a series of extenuating measures for buyers of electric vehicles. In this paper, the role of electric vehicles
in the existing traffic system will be presented, as well as the future forecasts expected in this field.
KEYWORDS: electric vehicles, charge network, emission, environmental protection
INTRODUCTION
Today, we are in a situation that after a little more than a century we are returning again with electric vehicles. The
development of electric vehicles began simultaneously with the development of combustion engines vehicles.
However, at the time, combustion engines vehicles had an advantage over electric vehicles. The main failing of
electric vehicles were heavy batteries and the possibility of passing significantly shorter distances compared to
combustion engines vehicles that could be supplemented anywhere. Electric vehicles are quickly suppressed from
use and practically forgotten because of that.
There is also an interesting story about it [1]: “In 1888 Bertha Benz, the wife of the inventor of the first car designed
to use an internal combustion engine, took the mark 3 model on a trip to visit her mother some 60 miles away. The
journey was a success and proved that automobiles were ready to be taken seriously for long distance travel. The
secret lay in the fuel supply: when Bertha ran out of fuel, she just stopped at the nearest chemist and bought some
more petrol, filled the car up and continued.”
1
Milja Simeunović, Ph.D. assistant professor, Faculty of technical sciences, Trg Dositeja Obradovića 6, Novi Sad,
mlekovic@uns.ac.rs
2
Zoran Papić, Ph.D. associate professor, Faculty of technical sciences, Trg Dositeja Obradovića 6, Novi Sad,
njele@uns.ac.rs
3
Milan Simeunović, Ph.D. associate professor, Faculty of technical sciences, Trg Dositeja Obradovića 6, Novi Sad,
milansim@uns.ac.rs
4
Pavle Pitka, Ph.D. assistant professor, Faculty of technical sciences, Trg Dositeja Obradovića 6, Novi Sad,
pitka@uns.ac.rs
493
It is re-thinking about the use of electric vehicles, nowadays, in which traffic is one of the important air pollutants and
factors affecting climate change. Production and use of electric vehicles becomes a strategy for the future.
Automotive industry, public authorities and many other stakeholders are trying to integrate their interests into common
goals and increase the use of electric vehicles by common forces.
Within this paper will be presented importance and role of the application of electric vehicles, the advantages that
can be realized, as well as the deficiencies that slow down their massive use.
CONTRIBUTION OF ELECTRIC VEHICLES TO SUSTAINABLE MOBILITY

"Sustainable System", "Sustainable Development" and the like are concepts that nowadays have a central place in
considering the long-term perspective of survival and progress of humanity. The developed strategies expanded to
all aspects of our lives. An exception to this is not the transport system, which is tending to fall into certain sustainable
frameworks. Currently, traffic accounts for 30% of total energy consumption in the EU. Half of the total fuel used in
road traffic burns in cities [2]. Consequently, the concept of sustainable mobility has been singled out as a separate
concept of sustainable development. This concept should serve as a guidepost for activities and infrastructural
development that will ensure the desired level of mobility and reduce greenhouse gas emissions from the transport
sector. Sustainable mobility is a trend that is very actual at the world level and which primarily involves the use of
alternative modes of transport, including the use of electric vehicles. Promotion of the use of electric vehicles is an
important aspect that can significantly contribute to solving the issue of urban pollution by traffic.
The seriousness of the problem of sustainable mobility is also indicated by the fact that there are decisions and
directives at the European level related to reduction in the dependence of transport on oil. Directive 2009/28/EC of
the European Parliament and of the Council sets a market share target of 10% of renewables in transport fuels [3].
In addition, European Parliament 94/2014 requires the development of new technologies and innovations, with the
support of the development of electric vehicles. Within the DIRECTIVE 2014/94/EU of the European Parliament and
of the Council it is pointed out that electricity is the source of energy necessary for the development of electric vehicles
that can contribute to the improvement of air quality and noise reduction in urban/suburban agglomerations and other
densely populated areas. Public authorities should take measures to assist users of such vehicles by ensuring the
installation of adequate infrastructure with sufficient charging points for electric vehicles [3].
Since electric vehicles have the ability to move without emission, considering from this aspect, the use of these
vehicles is undoubtedly a big step in the sustainable mobility. However, although it is undeniable that electric vehicles
without emission significantly contribute to environmental protection, the questions arises: what is the energy needed
to produce one such vehicle? What is the energy used to fill the vehicle?
Various studies have been carried out to compare electric and conventional vehicles. Many consider that there is no
justification for replacing conventional vehicles with electric vehicles, if they look at overall lifetime that includes the
production and the use of the vehicles. Some of the done researches showed that electric vehicles generate more
carbon emissions during their production than current conventional vehicles.
Study of electric car production in Malaysia estimated a compact electric car production release 5,791 kg CO 2 per
unit against conventional vehicles 4,166 kg CO2. Further, electric car components contain many lightweight materials
that require a lot of energy to produce and process, such as aluminium and carbon-fiber-reinforced polymers. Based
on research published in the Journal of Industrial Ecology named "Comparative environmental life cycle assessment
of conventional and electric vehicles", it has been concluded that the global warming potential of the process used
to make electric cars is twice that of conventional cars [4]. On the other hand, an EU study shows that an electric car
using electricity generated solely by an oil-fired power station would use only two-thirds of the energy of a petrol car
travelling the same distance [5].
Lithium-ion batteries are also a problem. Their lifetime in electric cars is currently 7 to 10 years old, so it is necessary
to make a strategy for their storage.
ELECTRIC VEHICLES CHARGING

A problem existing at the beginning of the production of electric vehicles, just over a century ago, is still not fully
solved. Inability to cross larger distances with existing batteries continues to be cited as one of the biggest failing in
nowadays. Existing electric vehicles can cross the distance up to 250 km. The basic precondition for unlimited
movement of electric vehicles is a well-developed charging network. There are three main EV charging
speeds. Rapid charging units (43, 50, or 120 kW) can provide an 80% charge in around 30 minutes; Fast charging
points (7-22 kW) can fully recharge some models in 3-4 hours; Slow charging points (up to 3 kW) are used for longer
494
charging times, around 6-8 hours [6]. Slow charging is mainly provided for home use. It is the most convenient and
at the same time the cheapest way to charge an electric vehicle. Fast charging units are used for public charging in
places such as parking areas, shopping centers, etc. It is necessary to install fast charging units, on more significant,
especially underground roads because it takes up to several hours for charging with slower chargers, which
significantly extends travel time. Public authorities and manufacturers of electric vehicles are intensively done on
development of the charging network, for enabling undisturbed movement of electric vehicles.
Teams from Bristol University and Surrey University developed a next-generation material for supercapacitors, which
store electric charge and can be replenished faster than normal batteries. This could allow cars to recharge in 10
minutes and the technology has sufficient energy density to comfortably surpass the 320 to 560 km ranges of leading
battery-powered cars such as Teslas [7].
The European Parliament recommended one of the solutions and it is ability to replace batteries on electric vehicles.
Public electric vehicle charge points would already have charged batteries that would change in a few minutes, which
would significantly shorten the waiting time for charging electric vehicles.
Some countries invest a large part of the funds in development of road infrastructure adapting to the movement of
electric vehicles. Different tests are carried out to determine the ability to charge electric vehicles while they are
moving. One of these projects has been implemented in the UK. A select number of cars is fitted with the requisite
wireless charging technology, and a test road is built to show how smaller sub-stations, charge converters, and power
transfer loops can provide inductive charging built into the road itself (Figure 1) [8]. A similar idea is used in South
Korea, where specially modified electric buses use shaped magnetic field in resonance technology built into the road
surface to receive a charge as they move along [8].
Figure 1 Electric re-charging lane [8]
It is essential that manufacturers of electric vehicles adapt to new technologies so that these ideas can be applied in
practice. It takes time to change the overall concept of the production of electric vehicles. Car makers will need to
design the next generation of electric vehicles to incorporate induction charging pads into their technical make-up.
Road infrastructure companies, meanwhile, will need to install multiple numbers of pads under road surfaces to make
the idea a reality [9].
EXPERIENCES OF INDIVIDUAL COUNTRIES

Developed countries have timely started creating necessary conditions for the use of electric vehicles. Numerous
European cities, promote, subsidies and emphasize their use. The Oslo in Norway is the best examples, which has
a number of benefits for owners of these vehicles: no purchase/import taxes, exemption from 25% VAT on purchase,
low annual road tax, no charges on toll roads or ferries, free municipal parking, access to bus lanes, 50% reduced
company car tax, etc. Recent updates on Norwegian electric vehicles policy: access in bus lanes in Oslo require
carpooling with at least one passenger during rush hours (2015), free municipal parking up to cities to decide (2017),
zero annual road tax (2018), 40% reduced company car tax (2018), 50 % price reduction on ferries (2018), zero re-
registration tax for used zero emission cars (2018) [10] Norway is a country that generally has the largest number of
electric vehicles per habitant.
An analysis of the latest global sales of electric vehicles found that nearly half the vehicles registered in Norway in
the first three months of 2018 were electric (48%), compared to just over a third (35%) during the same period in
2017 [11]. Numbers include passenger cars and light commercial vehicles. The following Figure shows the number of
495
registered electric vehicles in the territory of Norway until March 31, 2018. The number shown includes passenger cars and
light commercial vehicles.
Figure 2 Electric vehicle fleet in Norway [10]

(Note: BEV – battery electric vehicles; PHEV – plug-in hybrids)
In 2018, purely battery-powered vehicles make up less than 1 percent of the 46 million cars on German roads. It is
estimated that by 2035, one in three cars in Germany may be electric [12]. However, there are plenty of those
who believe that such increase in the number of electric vehicles would lead to an overload of energy
networks, so that many parts of the country would often remain without electricit y. For this reason, it is
necessary first to create conditions for use of electric vehicles. The 2017 is a record year for the number of
registered electric vehicles in Germany. According to available data [13], 54,492 electric vehicles were
registered in 2017, up 217% the previous year. The German market is also be gathering momentum, with the
share of electric vehicles rising from 1.2% in 2017 to 2%. At the same time, new registrations of diesel vehicles are
in freefall. According to the latest data, diesel’s market share is 31.3% – a decline of 25.4% on the year before [11].
According to data from 2017, Sweden at that time had about 3.4% of electric vehicles. The Swedish government has
set up subsidies of around 4,000 euros for purchase of an electric car. In addition, electric vehicles are allowed to
enter the narrowest city zones, parking is free, traffic lanes for public transport are permitted, charging of electric
vehicles on public charging points is free, etc. However, although there are incentives for the purchase of electric
vehicles, people still rarely decide to buy them because they are still very expensive and have no scope to compete
with cars for petrol or diesel [14].
Figure 3 Cumulative sales until December 2017 by country/region [13]
496
Data from January 2018 show that there are about 130000 electric vehicles in the UK. It is predicted that the number
of electric vehicles will grow rapidly in the future. This is supported by the fact that the electric vehicle charging
network in the UK is expanding rapidly. 50 locations were identified on the road network where it is necessary
installing chargers so that electric vehicles can be moved all over the country. Graeme Cooper, the project director
of electric vehicles, said that fifty strategic locations means to an extent that 96% of UK drivers would be able to drive
in any direction from any location in the UK and be in 50 miles of an ultra rapid small charger [15]. It is expected the
number of public electric vehicles charging locations will reach 7,900 by August 2020 [16]. The UK government
recently launched a £20m fund for research into such vehicle-to-grid technology, which has previously been confined
to private pilots, but will now be open to consumers [17].
According to one of the available sources [13], the distribution of electric vehicles by regions and countries ending in
2017. is shown in the Figure 3.
CONCLUSIONS
The production and use of electric vehicles has expanded in the last few years. A trend using of electric vehicles is
becoming increasingly popular on a global scale. This is in line with the strategy to reduce as much as possible air
pollution, noise, greenhouse gas emissions and energy consumption. There are different discussions about the
energy consumption for production and use of electric vehicles. Opinions are divided and there is not enough
research in this field, in order to make some general conclusions. It is crucial to create a strategy for the development
of renewable energy sources, such as solar energy or wind energy. In the near future electric vehicles are expected
to have lower operational costs than conventional cars, if renewable energy sources are used. In addition, it is
necessary for public authorities to promote electromobility with monetary and regulatory incentives (subsidies for the
purchase of electric vehicles, free registration of electric vehicles, use of bus lanes, etc.). Practically, these economic
incentives can be result of cost savings by pollution environment reducing. It is also necessary to work on expansion
of charging network. For longer distance trips, a well-organized charging network must be in place. A broad charging
network is the key to long-term plans and strategies for applying electric vehicles.
It is quite certain that electric vehicles have great potential and their use can significantly contribute to the future of
sustainable mobility.
REFERENCES
[1] heinakroon.com, Available from: https://heinakroon.com/2011/01/14/why-electric-cars-havent-made-it/,
Accessed 22.03.2018.,
[2] Sustainable urban transportation, 2016 (on-line), Available from: https://stateofgreen.com/files/download/9642,
Accessed 25.06.2017.,
[3] EUR-Lex, Available from: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32014L0094,
Accessed 24.06.2018.,
[4] Wikipedia, Available from: https://en.wikipedia.org/wiki/Environmental_aspects_of_the_electric_car, Accessed
06.06.2018.,
[5] The Guardian, Available from: https://www.theguardian.com/football/ng-interactive/2017/dec/25/how-green-are-
electric-cars, Accessed 24.06.2018.,
[6] Zap-Map, Available from: https://www.zap-map.com/charge-points/connectors-speeds/, Accessed 24.06.2018.,
[7] The Guardian, Available from: https://www.theguardian.com/environment/2018/feb/26/energy-storage-
supercapacitors-electric-car-batteries-charging-times, Accessed 06.06.2018.,
[8] Science alert, Available from: https://www.sciencealert.com/the-uk-is-trialling-a-new-road-surface-that-charges-
your-electric-car-as-you-drive, Accessed 06.06.2018.,
[9] Autocar, Available from: https://www.autocar.co.uk/car-news/new-cars/road-charges-your-electric-car,
Accessed 06.04.2018.,
[10] Norsk elbilforening, Available from: http://elbil.no/english/, Accessed 06.06.2018.,
[11] The Guardian; Available from: https://www.theguardian.com/environment/2018/apr/19/e-vehicles-electric-cars-
uk-sales-rise-norway-china, Accessed 12.06.2018.,
[12] Handelsblatt Global, Available from: https://global.handelsblatt.com/companies/electric-cars-blackouts-german-
cities-878678, Accessed 25.06.2018.,
[13] Wikipedia, Available from: https://en.wikipedia.org/wiki/Plug-in_electric_vehicles_in_Germany, Accessed
12.06.2018.,
[14] Energy portal, Available from: https://www.energetskiportal.rs/wp-content/uploads/2017/07/08-EP-bilten-EKO-
MOBILNOST-1.pdf, Accessed 06.04.2018.,
[15] The Guardian, Available from: https://www.theguardian.com/environment/2018/jan/30/bp-charging-points-
electric-cars-uk-petrol-stations, Accessed 12.06.2018.,
497
[16] The Guardian, Available from: https://www.theguardian.com/environment/2016/aug/04/electric-vehicle-charge-
points-to-outnumber-petrol-stations-by-2020-say-nissan, Accessed 12.06.2018.,
[17] The Guardian, Available from: https://www.theguardian.com/business/2017/oct/02/electric-car-battery-savings-
nissan-leaf-ovo, Accessed 12.06.2018.
498
Kragujevac, Serbia
MVM2018-008
Goran Štetin1
Zoran Papić2
Nenad Saulić3
Andrijana Jović4
ANALYSIS OF VEHICLE CRASH DEFORMATIONS USING

FE MODELS
ABSTRACT: Methods used for estimation of vehicle collision speed with non-deformable, non-movable obstacles
are related to measuring damaged area and estimation of the energy needed for such damages, considering specific
stiffness of the vehicle. These stiffnesses are different for every vehicle model, and data for each model must be
obtained by crash tests. The second method is comparing the damaged area of the vehicle with a similar damage of
another vehicle from the EES catalog. When frontal collisions are analyzed, estimation of the EES value can lead to
significant errors if it is based only on external damage rating, and vehicle construction is not taken into consideration.
Errors can also be made due to non-uniformity of the vehicle stiffness and various construction solutions of vehicles.
At higher collision speed, the engine block is suppressed toward the firewall. In this case, surface damages can not
indicate to energy that was spent on deformation and suppression of internal parts and was taken over by crumple
zone elements. This kind of analysis is very hard to carry out in practice but can be easily accomplished in computer
simulations using FE models. Therefore, the conducted research represents the analysis of the vehicle front parts
deformation mechanism via the FE models.
KEYWORDS: FE models, vehicle collision, deformation, vehicle stiffness, crumple zone
INTRODUCTION
Each year road accidents take many victims, so a very important characteristic of every vehicle is the passengers
protection ability during the road accidents, for the users and manufacturers. The ability of passengers protection
can be checked by crash tests in specially equipped labs/polygons and/or computer simulations with specially
designed models of vehicles using finite element method (FEM). This kind of simulations allows a detailed review of
each part of the vehicle, acting forces as well as collision deformations intensity.
By the observing collision simulations with FE models of the vehicles with non-deformable obstacles and way of the
vehicle’s front components deformation, can be concluded that maximum depth of that deformations, in the case of
constant test speed, depends negligible on the width of the obstacles which contacted with the vehicle, so the
maximum deformation’s depth mostly depends on the vehicle’s construction. That happens in the cases when energy
spent on deformations is enough to cause engine block suppressing and his contact with the firewall.
1
M.Sc. Goran Štetin, Ph.D. student, University of Novi Sad, Faculty of Technical Sciences, Department of Traffic
Engineering, Trg Dositeja Obradovića 6, 21000 Novi Sad, Serbia, stetinns@gmail.com
2
Ph.D. Zoran Papić, professor, University of Novi Sad, Faculty of Technical Sciences, Department of Traffic
Engineering, Trg Dositeja Obradovića 6, 21000 Novi Sad, Serbia, njele@uns.ac.rs
3
M.Sc. Nenad Saulić, assistant, University of Novi Sad, Faculty of Technical Sciences, Department of Traffic
Engineering, Trg Dositeja Obradovića 6, 21000 Novi Sad, Serbia, nenadsaulic@gmail.com
4
M.Sc. Andrijana Jović, junior researcher, University of Novi Sad, Faculty of Technical Sciences, Department of
Traffic Engineering, Trg Dositeja Obradovića 6, 21000 Novi Sad, Serbia, andrijanajovic992@gmail.com
499
This claim was verified by simulation using the FE models of the vehicles Ford Taurus (1996) and Geo Metro (1997).
Results are presented in terms of dependence between damage’s depth and folding width, when speed is constant
and its value is 60 km/h. During the choosing of collision speed, was taken into account that speed needs to be
sufficient to provide deformations of the front parts of the vehicle and engine block suppressing toward the firewall.
VEHICLE CONSTRUCTION AND DEFORMABILITY

Vehicle construction and its design should fulfill high standards in terms of safety, economy and aesthetics. The
structural circuits of the vehicles are the most important elements of passive passenger safety and they are designed
to provide adequate protection against external impact forces. Generally speaking, frontal collisions have greater
intensity and therefore the front parts of the vehicle are designed to absorb most of the energy. Energy-absorbing
parts of the vehicles are connected to make so-called crumple zone to provide as small as possible speed change
of the cabin where passengers are located. Materials and construction geometry plays a key role in determining the
ways for load movement [4].
Vehicle consists of two zones and they have various requirements in this regard (Figure 1).
Figure 1 Crumple zone and passenger protection zone [4]
The passenger protection zone is designed as a high-stiff safety cage that prevents deformations of the passenger's
cabin and protects passengers. The crumple zone is located on the front and rear of the vehicle and it has the task
to absorb as much energy as possible during the frontal and rear-end collisions, thus reducing the impact by
prolonging the deceleration time.
Figure 2 Pressure (MPa) and percentage elongation ratio for various types of steel and their use in vehicle
structure [4]
500
From Figure 2 it can be seen that various vehicle segments must fulfill different requirements in terms of stiffness.
Red color marked parts represent circuits that are critical to preserving the passenger protection zone and are made
of steel with high resistance to deformation. In recent times, steels with boron are used for these purposes because
it provides good firmness. During the collision, steels deform and take over part of the energy. Those steels are
marked with blue color in Figure 2. These circuits are made from steels which must provide a good balance between
firmness, workability, energy absorption and durability. The body of the vehicle is covered with a steel of low firmness
and excellent workability and it can fulfill users aesthetic requirements. Low firmness steel must provides a light
deformability because it is important for the human body protection, during the collision with the vehicle. These parts
are represented by the green color in Figure 2 [4].
The engine block and firewall plays a very important role in the collision process, although they are not components
of energy absorption. A firewall is a barrier that separates the space of the drive unit from the passenger
compartment. Its main purpose is to prevent the heat, noise, and gases from entering the passenger compartment.
In the cases of collision, firewall represents a powerful barrier that prevents the front parts of the vehicle from
penetrating into the passenger compartment.
An interesting test was done using the FE model [1], where the resulting depth of damage of that vehicle model was
analyzed for the different stiffness of the front parts of that vehicle. While the deformation of less stiff front parts of
the vehicle was more severe, the higher stiffness of the vehicle resulted in a greater penetration of the front circuits
into the interior of the passenger compartment, i.e. the circuits pushed towards the firewall made greater pressure
on the firewall, thereby increasing the risk of injuries to the driver and front seat passenger.
Figure 3 Deformation of the firewall in frontal collisions at the same speeds, and for different stiffness values of the
same vehicle [1]
During a collision, under the influence of numerous elements located in the front of the vehicle, the supporting
structure, and the firewall, the engine block will not move in the force moving direction. Engine block’s movement will
have a more complex trajectory with a rotation around the axis of suspension on the body of the vehicle.
When estimating the EES value, it often happens that the same vehicles are not compared, that the data of different
vehicle stiffness are not taken into account or data about vehicle stiffness are not available, so it considers that their
stiffness is equal. Deformation at a certain speed for one vehicle does not have to be the same for another vehicle,
at the same speed. These values can be very different in practice. Since the vehicles construction and materials
have been changing for years, the stiffness of the vehicle undoubtedly largely depends on the year of production of
the vehicle. By comparing the parameters of mass, wheelbase and year of production, and analyzing the data of the
stiffness of these vehicles by crash tests, it has been concluded that the highest degree of correlation exists between
vehicle stiffness and year of production. In this case, the vehicle stiffness is in a slight increase [1].
According to statistical data which are presented with graph, the percentage of vehicles that have been suffered
major damages (VDI6=3, VDI6=4 and VDI6=5+) is decreasing in recent models. Older models of vehicles suffer more
severe deformations for the same EES values than newer models. This means that older vehicles suffer more severe
deformations than new ones, when absorbing the same amount of deformation energy, which can lead to a significant
error in determining the collision speed over the EES value, if vehicles from different age groups are used during the
assessment [1].
501
Figure 4 Deformations intensity expressed through VDI6 index for vehicles from different age groups [1]
Also, the size of the collision frontal zone (the distance between front bumper and firewall of the vehicle) plays a very
important role for the specific stiffness of the vehicle. Researches have shown that the stiffness is rising with the
shortening of this part of the vehicles [1].
The depth and shape of the damages play a very important role in passenger safety and the estimation of collision
speed over the EES and/or EBS values, so this problem required a deeper analysis. By observing the virtual crash
tests using the FE model of the vehicle, it was noticed that the front parts deformation depth of the vehicle very slightly
depends on the obstacle’s width, but in tests with a constant obstacle width, the correlation between the speed and
the deformation depth showed great significance. This can be explained by the good distribution of impact forces at
the front parts of the vehicle.
THE FINITE ELEMENTS METHOD (FEM) AND CREATION OF FE MODELS

The finite element method belongs to modern numerical analysis methods. Its application first began in the field of
engineering construction calculations. FEM is based on a so-called physical discretization of the continuum. This
basic idea is very old, as much as the human efforts to replace complex problems with simpler, with the purpose of
the easier solution finding.
The basic FEM principle consists of dividing the considered area into a finite number of smaller elements, so a whole
area can be analyzed by analyzing individual elements, assuming interconnectedness between the individual
elements. This approach in the analysis goes from a separate to the general, from the individual to the universal.
Conclusions about the whole area can be made by analysis of the parts, and it is a known inductive approach, which
is applied in many areas of science. This approach also found application in analyzes related to vehicle collisions. It
is important to note that each part of the FE model of a vehicle must be modeled specifically, and a complete model
is obtained by fitting and connecting those parts. Dividing a wider area to a finite number of smaller elements, or
generating a finite element network, is done automatically by specialized software.
Figure 5 Parts of vehicles made of finite elements incorporated as a whole vehicle area
Therefore, the essence of the FEM is in the following:

 The considered area is divided into many smaller areas which have finite dimensions, using imaginary lines
or surfaces.
502
 Some smaller areas are called finite elements, and their set for the entire area is called a set or a network
of finite elements.
 It is assumed that the finite elements are interconnected in the final number of points, which are adopted on
the contour of the element. These points are called nodal points or nodes.
 The condition in each final element (e.g. the field of displacement, deformation, etc.) is described by means
of interpolation functions and the finite number of parameters in nodes, and they represent the basic
unknown variables in the FEM.
Depending on the type of problem that needs to be solved, the corresponding variables are obtained as solutions.
Thus, in the case of calculating the voltage-deformation condition of the structure, as in the case of the deformations
of complex structures analysis such as vehicles, the results are the displacements of each node in the structure and
the voltages within each element. Movements and voltages are the results of the action of an external load. These
loads in the collision analysis are impact forces that affect the vehicle, pedestrians or passengers in the vehicle,
depending on what is the subject of observation.
In order to arrive at concrete values and to properly define them in space, as well as to determine all the constraints
and loads, a software package based on a visual approach need to be used. Today there are numerous software
packages dealing with FEM, which can be divided into those designed for modeling (preparation phase), analyzing
the given problem and applying animation in post-processing (visual representation of the results obtained for the
model). Softwares that fully integrate all phases (model development, analysis and post-processing) are becoming
increasingly popular on the market. Some of the well-known software solutions that can be found, which deal with
this issue are NASTRAN, ANSYS, LS-DYNA, HYPER WORKS etc. [6].
APPLICATION IN AUTO - INDUSTRY

Vehicle safety during the collision is always checked via the crash tests. Those tests can be carried out physically,
on vehicle prototypes, or in computer simulations.
Crash tests represent deliberate, carefully and precisely prepared collisions, and they were, until now, the only
reliable way for determining the exact dynamics of the collision and thus improve road safety. The problem connected
with this method is that the equipment is expensive, the vehicle can be tested only once, and the time for preparing
and analyzing the data, provided by the expert teams, is long lasting and therefore there is no possibility of testing
many vehicles. Virtual crash tests have brought a revolution in the automotive industry in recent years. Collision
calculations conducted by especially computerized vehicles has proven to be a very reliable and practical method.
The advantage of this method is that it can be repeated an unlimited number of times, under the same or different
conditions. Also, this kind of models allows the examination and study of each individual element or the construction
of the vehicle. In cases when it is detected a weak point or disadvantage during the analysis, the model can be
redesigned and the test can be repeated under the same conditions. There is a question why tests with the real
vehicles still perform? The answer to this question is that the accuracy of virtual crash tests goes to about 90% [6].
The precision of virtual crash tests depends on the quality of the model, on the number of finite elements in the model,
as well as the competence and precision of the designer. Precision is greater when the model is composed of a
larger number of elements, but increasing in the number of elements also leads to increasing of the calculation time.
The calculation time is inversely proportional to the processing power of the computer and the number of elements,
so for more complex models (composed of several million elements), must be used supercomputer that contains
several thousands of processors. These processors work simultaneously. Therefore, it can be concluded that
precision requires large initial investments, which exceed the capabilities of the average user [5].
FE MODELS
Complex models of vehicles used by car industry are still not practical for use because of the fact that they are not
publicly available. The second reason is a weaker processing power of personal computers. Currently, we need to
satisfy with simplified, publicly available models. Models must be validated by experimental results, and all the
limitations of the FE method analysis should be considered.
In the PC Crash software database there are vehicle models, which were downloaded from the US NCAC5, created
in collaboration with the research teams of the Federal Highway Administration, NHTSA6 and George Washington
University. Currently, there are few vehicle's models: 1994 Chevrolet C2500, 1996 Ford Taurus, 1997 Dodge
5
National Crash Analysis Center – U.S. Department of Transportation
6
National Highway Traffic Safety Administration
503
Caravan, 1997 Geo Metro i 1998 Ford Econoline. Parts of the vehicle are represented by various colors (Figure 6)
[7].
Figure 6 FE vehicles models which are available in PC Crash database
SIMULATION SETTING
The first phase is the preparation for the simulation. Selection and assignment of the model’s movement parameters,
and ultimately the calculation of the collision is done through the Gmsh software, that is located within the PC Crash
software package. Gmsh is designed to integrate all stages of model development, analysis and postprocessing and
it has an excellent visual overview of all elements. After placing the FE vehicle model in front of a non-deformable,
non-movable obstacle, it is necessary to select the desired collision speed and determine the duration of the
simulation, which should not exceed 200 ms. The calculation duration for the average computer is about 1h/100ms.
Since the simulations duration in the particular case is limited to 200 ms, each calculation is solved for a period of 2
to 2.5 h.
The maximum depth of damage was measured as follows:
Figure 7 Measuring the damage’s depth on the FE model of the 1996 Ford Taurus
Figure 8 Measuring the damage’s depth on the FE model of the 1997 Geo Metro
From Figures 7 and 8 it can be seen that the measurement was done by simply comparing the undamaged vehicle’s
bumper height and the deformed vehicle's deepest part of the damage.
504
Analysis of the values obtained by simulations
Frontal collision analysis is carried out over the FE models of 1996 Ford Taurus and 1997 Geo Metro for a given
speed of 60 km/h and different obstacle's widths. The total analysis time is 200 ms, while the total collision time
during which vehicle contacts with an obstacle, is about 80 ms. The deformation at the end of the simulation can be
considered as final. The obtained results confirm the existence of assumed dependence between the maximum
deformation and the obstacle's width, at the same speed.
Figure 9 The graph which represents the dependence between damage’s depth and the obstacle’s width for the
1997 Geo Metro
By graph observing, it can be noted that the maximum depth of the resulting deformation gradually decreases when
obstacle’s width increases. Although the trend lines undoubtedly indicate to a gradual decrease of the maximum
damage’s depth, it must be taken into account that these reductions are minimal and difficult to notice without
measurement. The maximum damage’s depth (0.63 m) can be seen in a case when deformation occurs as a result
of the collision with 0.1 and 0.25 m obstacle’s width, at 60 km/h. This also represents the maximum deformation for
the observed speed. When obstacles overlapping the entire width of the vehicle, there is value for the minimum depth
of the maximum deformation.
A similar result was obtained during the collision analysis using FE model of the vehicle 1996 Ford Taurus.
Figure 10 The graph which represents the dependence between damage’s depth and the obstacle’s width for the
1996 Ford Taurus
f x   0.06681x  0.7595 (1)
f x   0.0945x  0.6498 (2)
The curve (1) represents the dependence between maximum damage’s depth and the obstacle’s width, at a constant
speed of 60 km/h, for FE model of the vehicle 1996 Ford Taurus. After comparison with curve (2), which represents
the same dependence for the FE model of the vehicle 1997 Geo Metro, it can be noted that curve for 1996 Ford
Taurus mildly decreases.
505
Figure 11 The graph which represents dependence between damage’s depth and the obstacle’s width for both
vehicles
This means that changing of the obstacle’s width has a smaller impact on maximum damage’s depth change for FE
model 1996 Ford Taurus. This can be explained by the existence of a larger crumple zone of the 1996 Ford Taurus,
by the greater stiffness, but also by the different constructive solutions of both vehicles.
During the simulation, vehicle 1997 Geo Metro had significant deformations located not only on front parts, but also
on the entire vehicle. Therefore, the absorption of kinetic energy was significantly higher on vehicle 1997 Geo Metro,
which supports the claim that this vehicle has less stiffness in comparison with the 1996 Ford Taurus (Figure 12).
Figure 12 Different stiffness vehicle’s deformations during the collision with a non-movable obstacle, at the same
speed
In this case, it's easier to determine the EES value by observing the vehicle's 1996 Ford Taurus damages. It can be
noticed that the vehicle's 1997 Geo Metro damages are complex and they are located over the entire construction of
the vehicle. It is very difficult to determine speed equivalent for this kind of damages, and a significant error in
estimation can occur if analysis includes only the front parts of the vehicle.
CONCLUSIONS
Expansion of numerical simulations leads to wider application of the finite element method and it is possible to create
complex systems such as a vehicle. With the development of various specialized softwares and research areas and
their application to the problem of collision processes, high reliable vehicle models have been developed. This
vehicle's models can be used in the collision simulations. That is very important for researchers in the field of traffic
forensic engineering science because expensive and complicated crash tests became accessible to them, for the
first time.
In this paper one of the many possibilities of FE model application is shown, by analyzing the dependence between
vehicle deformations depth and the obstacles width, at the choosen, equal collision speed. Based on the conducted
analysis, it can be concluded that the various vehicles deformations depth differs depending on the crumple zone
length, the stiffness, as well as the constructive solutions of the vehicle. During the collision between the vehicles
front parts and non-movable obstacle, intensive deformations occurs, in cases when the vehicle with less stiffness
506
participate in a frontal collision. In addition, due to the greater absorption of kinetic energy with less stiffness
participating vehicles in a collision, deformations of the complete structure of the vehicle may occur. During the speed
equivalence determination, vehicle's stiffness must be taken into consideration, with the purpose of avoiding speed
equivalence estimation errors.
REFERENCES
[1] Ablaßmeier, W., Boning, S., Ebner, H., Hartlieb, M., Kamm, M., O'Brien, S, Schwarz, T., Slaba, T., Wallner, S.,
Wrobel, B., Zobel, R.: "Opportunities for a World Wide Compatibility Evaluation - German Car Manufacturer’s
Position Paper on Crash Compatibility", 20th International Technical Conference on the Enhanced Safety of
Vehicles (ESV), 07-0323,
[2] Burg, H., Moser, A.: "Handbuch Verkehrsunfallrekonstruktion", Friedr. Vieweg & Sohn Verlag | GWV
Fachverlage GmbH, Wiesbaden 2007, Printed in Germany, ISBN 978-3-8348-0172-2,
[3] Gaffney, T. L., Winter, B., Elston, A., Sandvik, A., Pok, T., Richardson, S., Josevski, N.: "Method for Estimating
Vehicle-Specific Frontal Stiffness Values in the Absence of an Applicable Crash Test Using Methodically-Distilled
Data from the NHTSA Crash Database (Phase 1)", SAE Technical Paper 2015-01-0027, 2015,
doi:10.4271/2015-01-0027,
[4] Keeler, S., Kimchi, M.: "Advanced High-Strength Steels Application Guidelines", Version 5.0, May 2014,
WorldAutoSteel,
[5] Marzougui, D., Brown, D., Park, H. K., Kan, C. D., Opiela, K. S.: "Development & Validation of a Finite Element
Model for a Mid-Sized Passenger Sedan", 13th International LS-DYNA Users Conference, June 2014.,
Dearborn, Michigan, USA, pp 8-10,
[6] Richardson, S., Moser, A., Orton, T., Zou, R.: "Simulation of Vehicle Lateral Side Impacts with Poles to Estimate
Crush and Impact Speed Characteristics", SAE Technical Paper 2015-01-1428, 2015, doi:10.4271/2015-01-
1428,
[7] Wach, W., Han-Wiercińska, E.: "Simulation of Vehicle Accidents using PC-Crash", Institute of Forensic Research
Publishers, 2011, Krakow, Poland, ISBN 83-87425-68-0.
507
508
Kragujevac, Serbia
MVM2018-013
Vase Jordanoska1
Igor Gjurkov2
Darko Danev3
COMPARATIVE ANALYSIS OF CAR FOLLOWING MODELS

BASED ON DRIVING STRATEGIES USING SIMULATION
APPROACH
ABSTRACT: Transportation and traffic affect all the aspects of everyday life. To better understand traffic dynamics
traffic models are developed. On microscopic level, car-following models are developed and improved during long
period of time. They are used in traffic simulation tools or are the basis for operation in some advanced vehicle
systems. Car-following models describe traffic dynamics through movement of individual vehicle-driver units. This
paper compares Gipps model and Intelligent Driver Model (IDM) as car-following models based on driving strategies.
These models are derived based on assumptions such as keeping safe distance from the leading vehicle, driving at
a desired speed and producing accelerations within a comfortable range. The models are implemented and simulated
in MATLAB environment and the results are discussed in terms of the ability to reproduce real driving behaviour in
car following scenarios.
KEYWORDS: traffic dynamics, traffic simulation, car-following model, driving behaviour
INTRODUCTION
One of the key elements for good functioning and progress of modern societies is to have an efficient transport
system. Limited road capacity and hence frequent traffic jams is the problem many of them face. Expanding road
infrastructure is expensive and it cannot be done in dense urban areas. The solution should be sought in more
effective usage of existing road infrastructure through application of new technologies. It is how the research area of
intelligent transportation system emerged.
To better understand traffic situations and states, mathematical models are developed. These models also enable
analysis and definition of traffic and transport problems, and they give possibility of predicting future conditions, as
well as development of proposals for solutions. There are a lot of available traffic related data such as acceleration
of individual drivers and vehicles, macroscopic data obtained from static detectors complemented by data from GPS,
wireless LAN and mobile phone applications which can be used for modelling as well [9]. Experimental
measurements also serve as a basis for mathematical modelling.
Traffic dynamics which can be mathematically interpreted, describes the interaction between number of vehicles and
drivers. The interaction of complex so-called “driver-vehicle” units, leads to new joint traffic effects which do not
1
Vase Jordanoska, M.Sc., Teaching Assistant , University Ss. Cyril and Methodius, Faculty of Mechanical
Engineering - Skopje, ul. Rugjer Boshkovikj 18, 1000 Skopje, Republic of Macedonia, vase.jordanoska@mf.edu.mk
2
Igor Gjurkov, Ph.D., Professor, University Ss. Cyril and Methodius, Faculty of Mechanical Engineering - Skopje,
ul. Rugjer Boshkovikj 18, 1000 Skopje, Republic of Macedonia, igor.gjurkov@mf.edu.mk
3
Darko Danev, Ph.D., Professor, University Ss. Cyril and Methodius, Faculty of Mechanical Engineering - Skopje,
ul. Rugjer Boshkovikj 18, 1000 Skopje, Republic of Macedonia, darko.danev@mf.edu.mk
509
depend on the details of the single units. Example can be stop-and-go waves but also more complex spatial – time
patterns of congested traffic.
TYPES OF TRAFFIC FLOW MODELS

The types of traffic flow models can be divided on different basis, by the level of aggregation (the way in which the
reality is presented), by the mathematical structure or by the conceptual aspects. When it comes to traffic models
classification it is usually thought of the level of aggregation. According to it there are three ways to mathematically
model real traffic events: macroscopic, microscopic and mesoscopic [1, 6, 9]. By the mathematical structure traffic
flow models can be represented as partial differential equations, coupled ordinary differential equations, coupled
iterated maps, cellular automata, discrete state variables - continuous time or static models. Also, classification can
be made by conception foundation, identical versus heterogeneous drivers and vehicles, constant versus variable
driving behaviour or single-lane versus multi-lane models.
Macroscopic models describe the traffic flow as the flow of fluids. Dynamic variables are locally aggregated quantities,
such as the traffic density ρ(x,t), flow Q(x,t), mean velocity V(x,t) or the change in velocity σV2(x,t). Macroscopic
models describe collective phenomena such as the evolution of congested regions or the propagation velocity of
traffic waves.
Microscopic models include car-following models and cellular automata. These models describe individual "driver-
vehicle" units as particles α, which form the traffic flow. Microscopic models describe the reaction of each driver
(acceleration, braking or lane change) depending on the surrounding traffic. Dynamic variables are position xα(t),
velocity vα(t) and acceleration 𝑣̇ α(t) of the vehicle.
Mesoscopic models are hybrids of the microscopic and macroscopic approach. In local field models, parameters of
the microscopic model may depend on macroscopic quantities such as traffic density or local velocity and velocity
changes.
Microscopic models
Microscopic traffic flow modelling is based on description of the motion of each individual vehicle which is a part of
the traffic stream. It implies modelling the actions i.e. accelerations, decelerations, and lane changes of each driver-
vehicle unit in relation to the surrounding traffic.
In cellular automata models variables are discrete. The space is divided into fixed cells and the time is updated at
fixed intervals. The status of each cell is 0 ("no vehicle") or 1 ("vehicle" or "part of a vehicle"). Cell occupancy is
determined at each time step and depends on the occupation in the previous step.
In this paper the focus of interest are the car-following models which describe traffic dynamics from perspective of
individual "driver-vehicle" units. Literally, car-following models describe the behaviour of the driver (vehicle) only in
case of interaction with other vehicles, while the free flow is described by a separate model. However, a car-following
model is considered complete if it can describe all situations, including acceleration and free flow, following other
vehicles in stationary and non-stationary situations, approaching slow or stopped vehicles and a red traffic light. The
first car-following models were proposed as early as the 1950's by Reuschel (1950) and Pipes (1953). These two
models contained one of the basic elements of modern microscopic modelling which is the minimum distance, from
bumper to bumper to the leading vehicle, known as "safe distance", which should be proportional to the speed. The
elementary car-following models also include optimal velocity model, full velocity difference model and Newell’s car-
following model.
By the logic used car-following models can be classified in three categories [2]:
 Gazis-Herman-Rothery models (GHR) which state that the following vehicle’s acceleration is proportional
to the speed of the follower, the speed difference between follower and leader and the space headway.
 Safety distance models which are based on the assumption that the follower always keeps a safe distance
to the leader vehicle. These are also known as car-following models based on driving strategies. Model
examples that fall into this category are Gipps, IDM, MITSIM model.
 Psycho-physical models use thresholds for, e.g., the minimum speed difference between follower and leader
perceived by the follower. These are also known action-point models. Representatives in this category are
Wiedemann and Fritzsche models.
Car-following model is implemented in every available traffic simulation tool like VISSIM, Paramics, Aimsun,
MITSIMLab, SUMO, etc. [1].
510
CAR FOLLOWING MODELS BASED ON DRIVING STRATEGIES
Car-following models based on driving strategies are derived on assumptions from real driving behaviour such as
keeping a safe distance to the leader vehicle, driving with the desired velocity or accelerating in real and comfortable
range [9]. The relation between the braking distance and the velocity is also considered.
Driver-vehicle unit α is described with state variables vehicle’s position xα(t) and velocity vα(t) in function of time t and
vehicle’s length lα (figure 1). From the positions and lengths of vehicles, the distance between them is obtained:
sα = xα−1 − lα−1 − xα = xl − ll − xα (1)
Figure 1 State variables in car-following models
This distance sα along with vehicles’ velocities represents the main input in microscopic models. Of course, depending
on the model, additional variables are needed. In time-continuous models driver reaction is given in relation to the
acceleration function amic(s,v,vl), or when instead of leader velocity vl, velocity difference (∆vα = vα − vl ) is given, then
it is 𝑎̃𝑚𝑖𝑐 (s,v,Δv). In discrete-time models time is not modelled as a continuous variable but is discretized in finite and
constant time steps. The driver reaction is given in relation to the velocity function vmic(s,v,vl).
Gipps model
Gipps model is named after Peter G. Gipps who developed it in the late 1970's and published it in 1981. The model
proposed by Gipps is one of the most extensively used. Studies show that it produces unrealistic acceleration profile
because there is no difference between comfortable and maximum braking [3, 4, 9]. Still it is the simplest and
complete model without accidents which is accomplished by introducing a safe velocity vsafe(s,vl).
The original formulation of Gipps model states:
v(t + ∆t) = min[vacc (t + ∆t), vdec (t + ∆t)] (2)
where
v(t) v(t)
vacc (t + ∆t) = v(t) + 2.5 ∙ a ∙ ∆t ∙ (1 − ) ∙ √0.025 + (3)
v0 v0
presents the limitation of the acceleration process, and
vl 2
vdec (t + ∆t) = b ∙ ∆t + √b 2 ∙ ∆t 2 − 2 ∙ b ∙ (s − s0 ) − v(t) ∙ ∆t − (4)
b̂
presents the limitation of the braking process. This model is used in traffic simulation package Aimsun and a modified
version is used in DRACULA [1]. In 1998 Krauss did a modification of the model which now is used in SUMO traffic
simulator [2, 7].
The simplified Gipps model is defined as a discrete-time model with vsafe as the main component:
v(t + ∆t) = min[v + a ∙ ∆t, v0 , vsafe (s, vl )] (5)
where v0 is the desired velocity.
Braking manoeuvres are performed with constant deceleration b, which means there is no difference between
comfortable and maximum deceleration. Braking distance, which the vehicle leader should pass to full stop, is given
by ∆xl = vl2 ⁄2b.
511
Constant reaction time Δt exists. So for complete stop of the current vehicle to occur it is not necessary for the vehicle
just to pass the braking distance v2/2·b, but also the additional distance v·Δt that is passed during the reaction time:
v2
∆x = v ∙ ∆t + .
2b
Even in situations where the vehicle leader suddenly slows down and brakes to a full stop, the distance from the
vehicle to the vehicle leader should not be less than the minimum defined s0:
v 2 vl 2
s ≥ s0 + v ∙ ∆t + − (6)
2b 2b
The safe velocity is:
vsafe (s, vl ) = −b ∙ ∆t + √b 2 ∙ ∆t 2 + vl 2 + 2 ∙ b ∙ (s − s0 ) (7)
Intelligent Driver Model (IDM)

The intelligent driver model (IDM) is also simple and complete, accident-free model but produces more realistic
acceleration profiles. The model is developed by Treiber, Hennecke and Helbing who published it in 2000 [8].
IDM is time-continuous model which has the following characteristics [5, 6]:
 The equilibrium distance to vehicle leader cannot be less than the safe distance s0+v·T, where s0 is minimum
distance and T is time gap to the leading vehicle.
 It has braking strategy i.e. intelligent control for approaching slower vehicles. Under normal conditions the
braking is smooth; the deceleration gradually increases to value b and gradually decreases to 0 before
reaching the situation of a steady-state car-following or complete stop. In critical situations deceleration
exceeds the comfort value until the danger is avoided.
 Transition between driving modes is smooth which means that the time derivative of the acceleration
function which is the jerk (m/s3) is finite at any time. Equivalently, the acceleration function amic(s,v,vl) or
ãmic(s,v,∆v) is differentiable by all three variables.
IDM acceleration function has the form ãmic(s, v, ∆v), which is a continuous function of the velocity v, the gap s, and
the velocity difference ∆v to the leading vehicle:
2
v 2 s ∗ (v, ∆v)
v̇ = a [1 − ( ) − ( ) ] − IDM (8)
v0 s
and it is consisted of two parts. The first part is comparing the current velocity v with the desired v0 and the second
is comparing the current distance s with the desired s*.
v ∙ ∆v
s ∗ (v, ∆v) = s0 + max (0, v ∙ T + ) (9)
2 ∙ √a ∙ b
When there is a situation of an approaching to a stopped vehicle or red traffic light ∆v=v, the equilibrium part s0+v·T
of the dynamical desired distance s* (Eq. 9) can be neglected and deceleration function gets the form:
2
s∗ 2 a ∙ v 2 ∙ (∆v)2 v2 1 bkin 2
v̇ = −a ( ) = − = − ( ) ∙ = − (10)
s 4 ∙ a ∙ b ∙ s2 2∙s b b
v2
which defines kinematic deceleration bkin = . Kinematic deceleration is the minimum deceleration to avoid collision.
2∙s
A critical situation is considered if bkin is being greater than the comfortable deceleration b (bkin>b). In regular situation
the actual deceleration is less than the kinematic deceleration (bkin< b), which means bkin increases over time and
approaches the comfortable deceleration.
Intelligent driver model development was mainly been in direction of using it in ACC (Adaptive Cruise Control)
systems in vehicles [5, 6, 9].
512
SIMULATION
In order to compare the properties of Gipps and IDM model, both models were implemented in MATLAB script and
simulations were performed within the same scenario. The scenario consists of 1km of road in urban conditions that
come to intersection from where vehicles can go left or right (T-shaped intersection, figure 2). So, every vehicle which
comes to the intersection point needs to stop before it makes the turn manoeuvre. Five vehicles are included in
simulations which last for 120 seconds.
Figure 2 T-shaped intersection
In both car-following models the desired velocity is 54km/h and the minimum distance gap is 2m. The time and the
velocity of appearance of a new vehicle in traffic are same in both simulations. The comparative results are given in
figure 3. The figure includes results about the gap between vehicles (top graphs), velocities (middle graphs) and
accelerations of the vehicles (bottom graphs). Simulation parameters regarding desired velocity, maximum
acceleration / deceleration, comfortable deceleration and minimum gap are according to the recommendation of [9].
Figure 3 Gipps (left column) and IMD (right column) models' simulation results
In both models, acceleration, car-following and deceleration regimes are clearly visible. In view of simplicity Gipps
model produces good results, but the acceleration profile is unrealistic. There are just three acceleration values: a, b
and 0. Figure 3 graphs presenting the velocity and the acceleration show resemblance of a robotic driving. Gipps
model does not have comfortable and critical deceleration. So if b value is set for critical deceleration, every braking
will be done with full braking intensity which is uncomfortable in real driving behaviour. Also, transitions between
acceleration / deceleration or no acceleration regimes are hasty and unrealistic.
513
Intelligent driver model produces more realistic acceleration / deceleration profile but also during the process of
acceleration / deceleration the value does not increase gradually. It starts with a maximum (a or b) or near maximum
value and then the transition continues smoothly.
As it can be seen from the figure 3 it is a situation with loose traffic. The time interval of appearance of a new vehicle
is quite long, up to 8 seconds and the vehicles’ gap is large too. If the time of appearance of a new vehicle is reduced
the unrealistic acceleration profile of Gipps model is more noticeable (figure 4).
Figure 4 Gipps (left column) and IMD (right column) models' simulation results with a shorter time of appearance
A lot of fluctuations in vehicles’ accelerations can be noticed in Gipps model which is not the case in real driving
behavior. It is due to the model's formulation which does not include velocity difference between vehicles. The gap
between vehicles in Gipps model changes all the time and in IMD it is clearly seen when a vehicle gets into a car-
following regime and the desired gap is kept.
CONCLUSIONS
Car-following models should reproduce the general behaviour of drivers and vehicles in the most realistic way
possible. Gipps and IDM are simple and produce accident free results but their disadvantages need to be improved.
Because of the unrealistic acceleration Gipps model produces, AIMSUN adopted a modified version which has
different strategies for the selection of 𝑏̂ (the most severe braking that the leading vehicle wishes to undertake), it
allows more realistic distances to be kept among vehicles and introduced parameter that can be calibrated [4, 7].
The model used within SUMO is the Gipps model modified by Krauss in 1998, which defines new way of computing
the safe velocity. Results of the simulation of Gipps model in MATLAB (given in figure 3 and 4) clearly show that the
model gives an unrealistic acceleration profile, which is more apparent when the traffic is dense. When the number
of vehicles is higher and the time gap lower, the unrealistic character of the acceleration is more prominent. Regime
changes from acceleration mode to car-following mode and vice versa, are harsh. That is not the case in reality, nor
in the design of autonomous driving. The gap between vehicles in more frequent traffic is constantly changing, and
that does not correspond to the objective which is car-following regime in a case of dense traffic.
IDM model gives much better results in terms of acceleration profile, but still improved acceleration function is done
in the Improved Intelligent Driver Model (IIDM). The modification is in direction to improve the behaviour near the
desired speed. Further modifications of the model are done in terms of adapting for ACC systems [5, 6]. Simulation
results show that the disadvantage of the intelligent driver model is the initial large acceleration, which would be more
realistic if it gradually increases. The same happens in the braking process, with the exception of the first vehicle that
moves in free flow regime. In the car-following regime the model produces good results and that can be seen from
514
the gap graph (first row of figure 3 and 4). Comparing the simulation results of both models, it can be concluded that
the IDM model produces better results and is more stable than the Gipps model.
In summary, due to its simplicity Gipps model is one of the most used car-following models in traffic modelling and
simulations. IDM is a relatively new model, but it is also simple and gives realistic results in terms of the longitudinal
parameters of vehicles’ dynamics.
REFERENCES
[1] Barceló, J.: "Fundamentals of Traffic Simulation", Springer, 2010, New York,
[2] Brackstone, M., McDonald, M.: "Car-following: a historical review", Transportation Research Part F: Traffic
Psychology and Behaviour, Vol. 2, Issue 4, 1999, pp 181-196,
[3] Gipps, Peter G.: "A behavioural car-following model for computer simulation", Transportation Research Part B:
Methodological, Vol. 15, Issue 2, 1981, pp 105-111,
[4] Ciuffo, B., Punzo, V., Montanino, M.: "Thirty Years of Gipps' Car-Following Model: Applications, Developments,
and New Features", Transportation Research Record: Journal of the Transportation Research Board, Vol. 2315,
2012, pp 89-99,
[5] Kesting, A., Treiber, M., Helbing, D.: "Enhanced intelligent driver model to access the impact of driving strategies
on traffic capacity", Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and
Engineering Sciences, Vol. 368, 2010, pp 4585-4605,
[6] Kesting, A.: "Microscopic modeling of human and automated driving: Towards traffic-adaptive cruise control",
Doctoral Thesis, Technische Universität Dresden, Germany, 2008,
[7] Krajzewicz, D., Christian, R.: "SUMO (Simulation of Urban MObility)-an open-source traffic simulation",
Proceedings of the 4th Middle East Symposium on Simulation and Modelling (MESM20002), 2002,
[8] Treiber, M., Hennecke, A., Helbing, D.: "Congested traffic states in empirical observations and microscopic
simulations", Physical review E, Vol. 62, 2000, pp 1805-1824,
[9] Treiber, M., Kesting, A.: "Traffic flow dynamics: data, models and simulation", Springer Science & Business
Media, 2012.
515
516
Kragujevac, Serbia
MVM2018-020
Jelena Eric Obucina1
Nikola Kostic2
Zarko Djordjevic3
Sasa Babic4
Branimir Milosavljevic5
SAFETY ASPECTS OF THE HYDRAULIC SERVO STEERING

GEAR IN THE VEHICLE SYSTEM
ABSTRACT: The safety requirements of the vehicle in traffic are largely based on the functionality, efficiency and
reliability of the braking system and control systems. The basic function of the power steering is to achieve the
necessary moments to move the steering wheels. The use of hydraulic power steering at first started with heavy
trucks and buses, then on medium heavy, with a tendency to be applied and on light commercial vehicles. This is
particularly the case with vehicles that are driving in urban traffic. Practically, it can be said that the widest application
in cars has hydraulic power steering. This paper describes a hydraulic power steering from the PPT Trstenik
production program, but with a reference to the safety role in the transport vehicle.
KEYWORDS: safety, hydraulic servo steering gear
INTRODUCTION
Cancellations or changes in geometry in the control system can have a direct impact on the occurrence of the
accident. Therefore, the functionality and reliability of the control system are very important for the active safety
parameters of the vehicle. In terms of safety, the structural characteristics of the steering system, and in particular
the transmission ratio, the size and position of the steering wheel, as well as the required energy (torque and stroke)
for its launch, also have a significant role to play. In the servo system, it is important that the driver is given the sense
of resistance by turning the steering wheels, which should be administratively proportional to the corner of the turn.
A special requirement is that the steering wheels and steering wheel, when moving the vehicle, after turning and
releasing from the driver's action, themselves return to the neutral position. [1, 8, 4, 9]
The characteristics of the tire, the control system and the suspension system are the characteristics of the stability
or behavior of the vehicles on the road, and represent a group of active parameters of the vehicle safety.
Traffic safety requirements are largely based on the functionality, efficiency and reliability of the braking system and
steering systems [8].
1
Jelena Eric Obucina, College of Tehnical and Mehanical Engineering Trstenik, R. Krstica 19 Trstenik,
jelena.obucina@vtmsts.edu.rs
2
Nikola Kostic, College of Tehnical and Mehanical Engineering Trstenik, R. Krstica 19 Trstenik,
kosticn83@gmail.com
3
Zarko Djordjevic Faculty of Mechanical Engineering in Kosovska Mitrovica, Serbia, zaredj@gmail.com
4
Sasa Babic, College of Tehnical and Mehanical Engineering Trstenik, R. Krstica 19 Trstenik,
babicsf@nadlanu.com
5
Branimir Milosavljevic, College of Tehnical and Mehanical Engineering Trstenik, R. Krstica 19 Trstenik,
lenka1809@gmail.com
517
POWER STEERING SYSTEM
In case the wheel turning resistance is so large that excessive force is required at the steering wheel - the steering
wheel of the vehicle, sometimes special drivers or servos are installed in the control system beyond the driver's
ability. The servers create opportunities for very easy operation, which is why they are being used not only on freight
and duty vehicles, but also on heavy-duty passenger vehicles, that is, working volumes. In this latter case, especially
at higher speeds, greater safety of control is ensured, so also medium-speed speeds can be increased. It is also very
important to emphasize that with the servomotor you can easily keep the vehicle on the road, and when suddenly
there is suddenly a sudden discharge of air from one tire at a higher speed, which in other cases is very difficult. In
this sense, the servomotor control systems are widely used today - the so-called power steering systems.
The introduction of a servomotor represents a significant increase in the degree of mechanization of the control
system. The main task of the servomotor is to amplify the signal by which the driver acts on the wheel of the steering
wheel during its transmission to the driven wheels. In Figure 1 the hydraulic power steering type 5033 100 Prva
Petoletka from Trstenik is shown.
Figure 1 Hydraulic power steering 5033 100 [1]
Hydraulic power steering 5033 100 is intended for vehicles where the permissible load on the steering axle does not
exceed 30 kN. This range of vehicles includes a large number of light and medium commercial and delivery vehicles.
Hydraulic power steering 5045 188 applies to vehicles with axle load up to 75 kN. Such a control system design was
developed primarily for heavy commercial vehicles.
Hydraulic power steering 5033 100 and 5045 188, since they make up one family of hydraulic power steering, but for
different axle loads, they had to satisfy a large number of requirements with their designs and technical
characteristics:
• Ensuring high maneuverability of the vehicle (steering wheel shift of 90 ° and 96 °);
• Easy to operate in and out of the vehicle (corresponding hydraulic torque);
• Rolling wheels without side slip;
• Minimal transmission of steering wheel impacts;
• Kinematical approval of wheel turning;
• Compliance with the return of wheels in a neutral position;
• Work reliability, long service life and convenience of maintenance;
• Good keeping of the direction with allowable oscillation of the control wheels (minimum gaps in the system).
Additional and aggravating factors in the design of the power steering were:
• Design constraints (installation measures, connecting points, teeth, etc.);
• Approval of work with other systems and aggregates (wing propulsion);
• Maximum unit unification.
The hydraulic power steering 5033 100 is constructed as a classic version of the power steering with balls in short
performance. It was created by combining more reliable and tested solutions according to the predetermined
parameters.
518
TESTING THE MANAGEMENT SYSTEM
It is known from theory that testing of motor vehicles, their assemblies and parts is carried out according to specific
methodologies. In principle, all these methodologies can be classified into four groups [1]:
 testing the performance or functional characteristics of the test object,
 testing workloads in real exploitation conditions,
 reliability testing, i.e. the ability to preserve output-functional characteristics over a certain period of time,
under certain workloads and working conditions, and
 еxamination of vehicle safety, ie characteristics that determine the safety of vehicles in traffic.
In each of these groups, more or less, their place also has a management system.
The main goal of workload testing is to obtain knowledge about the size and character of the voltage in the average
and extreme conditions of exploitation. This has created the basis for programming adequate laboratory testing of
reliability, ie testing of critical loads. On this basis, it can follow the implementation of appropriate calculations of the
reliability of the elements and assemblies or directly determine the performance of the derived construction.
The development of scientific achievements contributes to the creation of a large number of methodologies.
Depending on the specific conditions, needs and possibilities, a specific methodology is designed for each situation.
Mechanical testing
The friction torque and friction tensile test is performed with the fully mounted steering wheel (without oil) placed in a
horizontal position and measurement is performed using the torque measuring device.
Table 2 Results of friction torque testing and adjustment of the attack point [1, 9]
Measurement Total frictional moment Adjustment of the attack point

no. outside the attack point ( Ncm ) ( Ncm )
on left on right on left on right
1 95 80 130 120
2 80 75 115 115
3 80 75 115 115
4 80 70 120 115
5 80 70 120 115
6 89 70 120 115
7 80 70 120 115
8 80 70 120 115
9 80 70 120 110
10 80 80 115 110
Figure 2 The results of the friction moment test and the setting of the attack point [1]
The measured value of the total friction torque of the fully mounted steering wheel outside the attack point is within
the allowed limits (70-150 Ncm). The measured value of the friction moment in the attack point is within the allowed
519
limits of (110 - 210 Ncm). When the steering column spindles evenly, the torque of the torque outside the attack point
is max. 40 Ncm.
The conclusion of this test is that the power steering meets the set conditions.
Hydraulic testing
Continuous pressure test (p) between the thrust and return terminal of the controller is performed by examining the
pressures that occur in the pressure and return water, as follows:
 p1 - pressure in the pressure line,
 p2 - pressure in return line.
Continuous pressure between the thrust and return terminal of the controller is calculated as Δp = p1-p2.
Table 3 Results of the hydraulic test [1, 9]
Measurement
p1 (bar) p2 (bar) Δp (bar)
no.
1 5 3 2
2 5 3 2
3 5 3.5 1.5
4 5 3.5 1.5
5 5 3.5 1.5
6 5 3.5 1.5
7 5 3.5 1.5
8 5 3.5 1.5
9 5 3.5 1.5
10 5 3.5 1.5
Figure 3 Continuous pressure test results (p) between the push-pull and
return control of the steering wheel [1]
The maximum measured value is within the prescribed limits (Δpmax=7 bar). The conclusion of this test is that the
power steering meets the set conditions.
System approach to risk management in the case of power steering with balls
Vehicles with PPT hydraulic power steering must be delivered to an authorized service for the control of the power
steering and high pressure pump, after passing the following kilometers [1].
520
Table 4 Controls that influence the risk reduction
Driving on the field and

Long-line traffic Mixed short-line traffic
construction site
80000 km
I control review 100000 km 100000 km 2500 h
150000 km
II control review 200000 km 175000 km 4000 h
200000 km
III control review 300000 km 250000 km 6000 h
For vehicles that do not have a built-in timer or hour counter, a consumed amount of fuel, which corresponds to the
specified intervals, is taken as the orientation value. At I and II inspections, no steering and vehicle pumps are
required unless the overhaul of the steering wheel or other defects is detected during the inspection. For the sake of
greater traffic safety, during the III check check, it is recommended to build and disassemble the steering and pump,
check the parts - that they did not wear out or have drawings (for heavily loaded parts). On this occasion, all sealing
parts should be replaced. In order to avoid waiting for a repair, it is recommended to purchase a spare aggregate. In
order to have a clearer picture of the condition of the vehicle and power steering, and for the comparison of the power
steering behavior before and after the inspection, a test drive is recommended.
CONCLUSIONS
Safety is one, perhaps the most important management system item, and through this work, as the priority is given
to power steering, it explains why power steering has earned such a priority over some other control devices. Of
course, the essential characteristic that is related to each product, as well as the control system of the freight vehicles,
is the risk of using that product. Then, if in this respect the product deserves to be used, then without accepting it
from the market, without proper economic affinity, it is still not fully accepted. The positive solution highlighted in this
paper, the hydraulic servo system in trucks, is justified from the aspect of risk and, of course, the markets. To make
all this profitable, to see how much the vehicle's steering system is carrying, and to answer why the vehicles have a
power steering, the whole system must be viewed from multiple angles. Multidisciplinary research will provide an
answer to why hydraulic power steering devices are used and justify the use of a servo unit in terms of safety.
REFERENCES
[1] Eric, Ј., Masters work, Faculty of Mechanical Engeneering in Kraljevo, 2007,
[2] Eric, Ј., Final work, Faculty of Mechanical Engeneering in Kraljevo, 2004,
[3] Rajamani, R.: "Vehicle Dynamics and Control", University of Minnesota, 2005, USA,
[4] Minic, M.: "Sistemi za upravlјanje teretnih vozila", ABC Glas, 1992, Beograd,
[5] Fitech, J. W.: "Motor Truck Engineering Handbook", SAE, Warendalle, 1993, USA,
[6] Jankovic, D., Todorovic, J.: "Teorija kretanja motornih vozila", Masinski fakultet Beograd, 1990,
[7] Simic, D.: "Motorna vozila", Naucna knjiga, 1988, Beograd,
[8] Jankovic, A., Simic, D.: "Bezbednost automobila", DSP-mecatronic, 1996, Kragujevac,
[9] Technical documentation IHP PPT Trstenik, Serbia.
521
522
Kragujevac, Serbia
MVM2018-025
Saša Vasiljević1
Nadica Stojanović2
Ivan Grujić3
Bojana Bošković4
Milan Mladenović5
MEASUREMENT OF VEHICLE BRAKE COEFFICIENT AS

VEHICLE MAINTENANCE PARAMETER ON TECHNICAL
INSPECTION LINES
ABSTRACT: Technical inspection of the vehicle represents activities for the purpose of determining and assessing
the technical condition of the vehicle, as well as compliance with the minimum technical regulations that are legally
prescribed. One of the most important systems on the vehicle is the braking system. Without this system, it is not
possible to stop the vehicle in a timely manner and therefore there is a danger of accident occurrence. In order not
to have a previously dangerous situation, technical inspection of the vehicle is performed by measuring the brake
coefficient of the vehicle at the technical inspection line. The brake coefficient can be used as a parameter of
maintenance of the brake system on vehicles. The minimum value of the braking coefficient is prescribed by the
legislation in the Republic of Serbia, that is, the minimum values of the braking coefficient are prescribed, depending
on the category of vehicle, which the vehicle needs to satisfy. In this paper, an overview of the procedure for
measuring the brake coefficient on the technical inspection line is given, with a special focus on specific cases and
legal regulations.
KEYWORDS: brake, coefficient, inspection, technical inspection line
INTRODUCTION
The vehicle is a set of components and therefore a set of different systems, which the driver can operate in the most
secure way in motor vehicles. If the vehicle is viewed as one circuit of the systems, it is logical that, with failure of
one of the systems, the vehicle comes into situation, which may result in its unsafe use and the occurrence of an
accident, in this case, a traffic accident. Based on the above, all in-vehicle systems must be properly maintained and
preventive maintenance measures should be taken to increase the vehicle safety. According to the previously
described problem, the parameters have been developed that can describe the technical condition of the system.
The braking system is one of the systems for which a parameter is specified by which to determine the correctness
and technical condition. Checking the braking system on vehicles is done on the technical inspection line, where,
based on the measured braking forces, the braking coefficient is determined. Compared to the legally prescribed
1
Saša Vasiljević, M. Sc., High technical school of professional studies, Kragujevac, vasiljevic.sasa036@gmail.com
2
Nadica Stojanović, M.Sc., assistant, University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić Str., 34000
Kragujevac, Serbia, nadica.stojanovic@kg.ac.rs
3
Ivan Grujić, M.Sc., assistant, University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić Str., 34000
4
Bojana Bošković, B.Sc., College of Technical and Mechanical Engineering, Trstenik, bokaev92@gmail.com
5
Milan Mladenović, B.Sc., University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić Str., 34000 Kragujevac,
m.mladenovic141@gmail.com
523
standard values for this coefficient for a certain category of vehicles, technical condition the braking system on the
vehicle can be determined. By measuring the braking force, it can also be determined whether corrective or
preventive maintenance measures are required, with only one wheel or a part of the system, if the difference in the
braking force relative to the other wheels is noticed.
In this paper, a method for measurement of braking forces and the calculation of braking coefficient is presented.
Also, special attention is paid to the legislation related to the manner of carrying out the technical inspection and the
listed minimum of technical requirements. Also, in addition to the above, it is important to note the problems that can
arise in the measurement of the braking coefficient (in terms of malfunctioning of the braking system), which in
practice must be noted and certain maintenance measures must be taken.
STATISTICAL DATA ON THE VALIDITY OF THE VEHICLE AS A FACTOR OF

ACCIDENT
The simplest way of determining the cause of accidents is by monitoring the traffic safety situation and by identifying
the causes that induced these traffic accidents. The Road traffic safety Agency (hereinafter: the Agency) is one of
the organizations for monitoring the traffic safety in the Republic of Serbia. The Agency is a body established by the
Republic of Serbia, whose work is regulated by the Law on Traffic Safety. The Agency monitors the state of traffic
safety at the state level and monitors the number of traffic accidents, but also the causes for the occurrence of a
traffic accident.
According to the data of the Agency, 579 persons died in traffic accidents in 2017, which is 4.6% less than in 2016
in the Republic of Serbia. However, based on the total number of 21856 persons who died in traffic accidents in 2017,
it can be seen that it is 2.6% higher than in 2016. If the incidence of traffic accidents with fatal injuries are analysed,
the most common reason is "taking ruthless actions by the driver", with a total of 354 persons. In traffic accidents
that are due to the inadequacy of the vehicle, the number of persons killed amounts to 32, while 336 persons were
injured in these traffic accidents, and the number of accidents in which material damage was caused is 321 [1].
Bearing in mind that the failure of the vehicle is a very broad concept, it is possible to look at the reasons for the
failure that caused the accident. As said, the factor of vehicle failure is the cause of the death of 32 people in traffic
accidents. If statistics are considered of which failure is the leading cause of the death, it is the failure of the vehicle
lightning device or direction indicator (12 people died). The failure of the braking system is the cause of the death for
7 persons in traffic accidents. Inadequate labelling of the vehicle caused the death of 2 people, while the same
number of people were killed in traffic accidents with the failure of the steering system. According to the Agency, the
other injuries caused by the vehicle as a factor of the traffic accident are also cited in addition to previously mentioned
ones, as causing the death of 11 people [1].
Based on their activities and knowledge of the importance of the technical proper function of the vehicle, many
organizations allow users to check the proper function of the vehicle. According to the data of the Automobile and
Motorcycle Association of Serbia (AMSS) and based on many years of work, it has been determined that one third
of the vehicles, when checked for proper function, have had a problem with defective brakes, one-tenth with a steering
system and one-third with vehicle lightning system. Less than one-tenth of failures apply to the driving system,
exhaust system and the like. All of the listed failures can lead to serious consequences and endanger not only the
drivers of the vehicles, but also the other road users [2].
The data available from the vehicle testing companies from our region are very similar to those in the Republic of
Serbia. According to one Croatian company, which has inspected the technical proper function of 2,065,038 vehicles,
439,999 defective vehicles have been identified. A total number of 1,579,100 defects were identified in the defective
vehicles, with the largest share of failures in the braking system. The share of failures in the braking system in the
total number of failures is 29.87%, while 22.34% of failures belong to the vehicle lightning systems and turn signals.
Based on the number of vehicles inspected by the Automobile and Motorcycle Association of Serbia, it was also
found that the greatest number of failures is related to the braking system not fulfilling the minimum required value
of the brake force [3].
Despite statistical data related to the vehicle failures and the other factors leading to traffic accidents, there are no
precise data on whether a traffic accident could have been avoided if a system had had a larger reliability. On the
basis of everything presented, it can be concluded that the most important part of the failure statistics belong to the
systems that are part of the active vehicle safety. The failure in these systems directly increases the possibility of the
occurrence of a traffic accident. For these reasons, the vehicle should be regularly maintained, but also there must
be some parameters to determine whether it is necessary to take any preventive maintenance measures for the
vehicle or the on-board diagnostics system [4].
524
TECHNICAL INSPECTION OF THE VEHICLE AND LAW REGULATIONS
The Rulebook on technical inspection of vehicles defines the technical inspection of a vehicle as a set of work
operations in which the technical proper function of the vehicle devices and equipment is determined, using
appropriate measurements and comparisons of the measured values with the prescribed values and using visual
inspection, with or without appropriate tools and without significant disassembling. The technical proper function of
the vehicle as a whole is determined and whether the vehicle meets the prescribed conditions and technical norms
for participation in road traffic. The regulation prescribes all actions that must be taken during the technical inspection
of the vehicle, who can perform the technical inspection and what one facility must fulfil. The Law on Traffic Safety
states that it is mandatory to carry out technical inspection of the vehicle once a year. A six months inspection period
is introduced for certain types of vehicles, including: motor vehicles and trailers for public transportation, for
transportation of dangerous goods, for training the candidates for drivers, with built-in devices for provision of special
light and sound signals and with maximum permissible weight greater than 3,500 kg, buses, rent-a-car vehicles and
motor vehicles more than 15 years old [7-9].
The control of the proper function of the vehicle braking system is one of the activities defined in the Rulebook on
technical inspection of vehicles. The braking system is tested using dynamometer rollers for measurement of the
brake forces. The legal regulations in the field of technical inspections require the use of such equipment that can
measure brake forces. Determination of the value of the braking coefficient can be done by measuring the
deceleration or by measuring the braking force on the vehicle wheel and expressing it in percentages. In the case of
braking force measurement, the value of the braking coefficient is calculated as the ratio of the braking force
measured on the wheel circumference to the total weight of the vehicle representing the product of the total mass of
the vehicle and the acceleration. Exceptionally, when calculating the braking coefficient of the braking system of
semi-trailers, instead of the total weight, the weight that represents the sum of axle loads of the semi-trailer multiplied
by the gravity acceleration is used. The axle load data should be provided from the manufacturer's plate, from
technical documentation of the vehicle manufacturer, from document on the determination of technical data issued
by the competent organization or from the axle load measurement. Using the equation (1), a braking coefficient can
be calculated. The braking coefficient is the ratio of the sum of the measured braking forces and the total weight of
the vehicle ready for driving.
𝐹𝑝𝑙 + 𝐹𝑧𝑙 + 𝐹𝑝𝑑 + 𝐹𝑧𝑑

𝐾𝑘 = ∙ 100 (1)
𝐺𝑠𝑧𝑣
Members of the equation (1) representing the measured brake forces at all four wheels are denoted by Fpl, Fzl, Fpd
and Fzd and the weight of the vehicle ready to drive is denoted by Gszv. The Rulebook on classification of motor
vehicles and trailers and the technical conditions for vehicles in the road traffic prescribes the minimum values of the
braking coefficient that the vehicle must satisfy during the test. The minimum values of brake coefficients that the
vehicle must satisfy during the test are prescribed for different vehicle categories. However, different brake
coefficients for vehicles that were first registered before January 1st, 2017 and for vehicles that were first registered
after January 1st, 2017, were also prescribed. Table 1 gives the minimum values of braking coefficients prescribed
by the legislation for driving for auxiliary and service brakes. The measured values of the braking coefficients must
be greater than or equal to the legally listed values when examining the vehicle [9].
In addition to the legally prescribed values of the brake coefficient for the service and auxiliary brake, the values of
the brake coefficients for a parking brake are also prescribed and amount to 15%. The variation of braking forces on
the wheel can not be higher than 20% during one turn. Of course, legal regulations also stipulate differences in
braking forces between two wheels of the same axle. The maximum allowed braking force difference for service
brakes on the wheels of the same shaft, at any time from the moment when the first braking force reaches one third
of its maximum value, up to the moment when the maximum braking force is achieved at both wheels, is 30%. The
difference in the braking force can be calculated using the expression (2), in case the braking force at the left wheel
(Fl) is greater than the braking force on the right wheel (Fd).
𝐹𝐿 − 𝐹𝐷
𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑖𝑛 𝑏𝑟𝑎𝑘𝑖𝑛𝑔 𝑓𝑜𝑟𝑐𝑒 = ∙ 100 (2)
𝐹𝐿
525
Table 1 Minimum prescribed values of braking coefficients that the vehicle must meet at the technical inspection [9]
Minimum vehicle brake coefficient values for Minimum values for brake coefficient for
vehicles registered before January 1st, 2017 vehicles registered after January 1st, 2017
Vehicle category
[%] [%]
Service brake Auxiliary brake Service brake Auxiliary brake
L1 40 20 42 20
L2, L6 40 20 40 20
L3 40 20 50 20
L4 40 20 46 20
L5, L7 40 20 44 20
M1 50 20 58 20
M2, M3 50 20 50 20
N1, N2, N3 45 20 50 20
45 (semi- trailers)
O 40 - -
50 (trailers)
T, C, K5a 25 (50) - 35 -
R, S, K5b 25 (50) - 35 -
The process of measuring the braking coefficients on the technical

inspection line
As noted above, a number of actions related to vehicle testing were identified during the technical inspection of the
vehicle. Legislation regulates the process of vehicle testing and these actions are classified in three test queries. In
the second vehicle test, the braking forces are examined and the brake coefficient is determined. The Rulebook on
the division of motor vehicles and trailers and technical conditions for vehicles in road traffic defines what needs to
be met in terms of equipment and simulation of wheel turning, and it defines that the vehicle must move on a
horizontal surface that is dry and has a modern pavement. It also defines the speed at which the vehicle must move
at the start of the braking as 50 km/h for passenger cars or 40 km/h for other motor vehicles, and for vehicles that
can not achieve this speed it is at least 80% of their highest designed speed.
Measurement of the brake forces is usually done by using dynamometer rollers. This device is situated on the floor
where vehicles are moving and it is popularly called “rollers”, because that's exactly what it is - because they rotate
and then brake. Figure 1 shows dynamometer rollers for measurement of the braking forces of passenger cars,
buses, terrain vehicles (a) and motorcycles (b).
Figure 1 Devices for measuring the braking force on vehicle wheels [10]
The principle of operation is very simple. The vehicle is placed on the device with dynamometer rollers. Then the
mass of the vehicle with another 75 kg added is typed into the computer. The signal is given by the remote control
from the vehicle that the vehicle is ready for measurement. The computer displays a sign or writes on the monitor
when the worker or the person in the vehicle is to be engaged. The braking force of the auxiliary brake is measured
first and then the braking force on the front axle and then on the rear axle. The layout of the computer and the
software for entering the vehicle mass is shown in Figure 2.
Finally, when measuring the braking forces on the wheels, the measured results, which are shown in the form of
diagrams, are checked showing the results for the axle, for the wheels on the one axle, for the service and for the
parking brake and the differences in forces between axles. After that, the results are printed and they are one of the
documents for the purpose of proof of the proper function or failure of the in-car braking system. The document itself
shows the measured forces on the brake force-time diagram. Of course, the document also records the values of the
measured brake forces and the brake coefficient and shows whether the measured values are within the legal norms.
Figure 3 shows the cut-out of the measured braking force document. The calculation of the braking coefficients can
be made based on the measured forces and according to the equation (1).
526
Figure 2 Computer and software for measuring the braking forces on the vehicle wheels
Figure 3 Part of the document that shows the measured brake forces on the vehicle wheels
PREVENTIVE MAINTENANCE BASED ON THE RESULTS OF BRAKING

FORCE MEASUREMENTS
Based on the results of the measurement of braking forces, it is possible to calculate the coefficient based on which
it can be determined whether the braking system is correct or, on the other hand, whether it is unacceptable and it is
necessary to take some preventive measures in order to increase the safety of the vehicle. Based on the braking
coefficient, it is definitely possible to determine whether it is necessary to perform some of the maintenance
measures. The same is the case with monitoring the braking forces. With the reduction of the brake coefficient, the
braking distance of the vehicle increases, due to the reduction of the vehicle friction. On the other hand, it is necessary
to take measures to increase the brake coefficient by maintaining the vehicle braking system.
However, in practice, in addition to the brake coefficient, to check the proper function of the braking system, it is
necessary to follow the braking diagrams (braking force-time diagrams). From these diagrams, it is possible to notice
the faults in the operation of the braking system, and as such, notice the possible danger of failure of the same
system. The reason for this is that when braking on some wheels, the maximum force can be reached, or at the end
of the calculation a coefficient can be obtained that is within the legal norms and that the braking system, however,
is not functioning properly. Such phenomena can easily be seen in practice with the previously mentioned diagram.
Of course, such phenomena should also be removed, because such system decreases its reliability and
consequently leads to a traffic accident.
Based on the above, the brake force-time diagrams are shown in Figure 4. Diagram showing the failure of the braking
system can be determined, although all the values of the brake forces are within the limits of the prescribed norms
and can give a satisfactory brake coefficient to their values. Figure 4a shows the case when both left and right wheel
reaches the maximum braking force, however, the right wheel reaches the maximum force later. Figure 4b shows a
diagram showing that the pressure drop in the brake system is noticeable, and in addition to achieving maximum
brake forces, the system is not functioning properly. Figure 4c displays a case when the braking system is not
functioning properly, in addition to achieving a massive force because the frictional plates are unevenly worn or there
527
is the oval flap or right-wheel drive discs (the test piece is visible in the figure). Figure 4d also shows a faulty braking
system. The fault is that, in addition to achieving maximum braking force, the blocking of the left-hand wheel causes
the blocking of the left-hand wheel. In each of these specific cases, it is necessary to notice any kind of failure and
remove it from the vehicle, because some specific road conditions may result in failure or even more damage to the
vehicle can occur.
Figure 4 Specific brake force-time diagrams which indicate faulty braking system
CONCLUSIONS
Maintenance of a vehicle as one complex system is imposed as one obligation for safe traffic. In order to achieve
increased traffic safety, different norms have been set for in-vehicle systems. Inspection of the legally minimum
standards for technical requirements for vehicles is done on the lines of technical inspections, whose work is
prescribed by the laws and the bylaws.
In addition to all the systems that are inspected on the vehicle, a braking system is also checked as a part of an
active safety system whose failure can have traffic safety implications in terms of traffic accidents. Analysing the
traffic accidents, as well as statistical data, companies that test the vehicle's proper function, have found that the
most common failure of the vehicle is related to the braking system of the vehicle.
By applying the established methods of inspecting the proper function of the braking system on the technical
inspection line, it is possible to directly determine the malfunction or a certain defect in order to prevent the failure of
the braking system on the vehicle. The developed methods primarily apply the parameters that are prescribed by the
law, while, on the other hand, they use the diagrams that can be formed when testing vehicles on the technical
inspection line. Of course, all systems must be properly maintained in order to prevent the occurrence of traffic
accidents. Legislative regulation should definitely be tougher, but in the meantime, is necessary for the car companies
to understand the serious issues they deal with and adhere to the ethical codes of their business.
REFERENCES
[1] Road traffic safety Agency of the Republic of Serbia, “Statistical report on the state of traffic safety in the Republic
of Serbia in 2017” (in Serbian), 2018, Belgrade
[2] The Automobile and Motorcycle Association of Serbia (AMSS), “Technically correct vehicle - safe vehicle” (in
Serbian), Preventive safety action, 2018, Available from: http://www.amss.org.rs/index.php?option=com
_content&view=article&id=213&Itemid=161, Accessed 11.09.2018,
[3] Centre for Vehicles of Croatia (CVH), “Review of the number of detected malfunctions at regular technical
inspections by individual assemblies” (in Croatian), 2018, Available from: https://www.cvh.hr/media/
2770/s03__pregled_broja_neispravnosti_po_sklopovima_2017.pdf, Accessed 12.09.2018,
[4] Auto-moto Association of the Republic of Serbian (AMS RS), “Do you drive a technically correct vehicle – check
2017” (in Serbian), 2017, Available from: www.ams-rs.com/preventiva/istrazivanja/tehnicka-ispravnost/
item/download/ 143_fea96f4755a3776a07c4a456dcb3fd74, Accessed 12.09.3018,
[5] Krstić V., Krstić I., Krstić B., Vukić L.: “Influence of age structure of vehicles and their technical sheet as at road
safety”, Proceedings of the 9th scientific-technical meeting with international participation ”QUALITY 2015”, 2015,
Neum, Bosnia and Herzegovina, pp 517-523,
528
[6] Pešić Z.: "Technology of motor vehicle maintenance” (in Serbian), Vojnoizdavački zavod, 2009, Belgrade,
[7] Law on Road Traffic Safety, Official Gazette of the Republic of Serbia, 41/2018, 2018, Belgrade,
[8] Rulebook on technical inspection of vehicles, Official Gazette of the Republic of Serbia, 31/2018, 2018, Belgrade,
[9] Rulebook on classification of motor vehicles and trailers and the technical conditions for vehicles in the road
traffic, Official Gazette of the Republic of Serbia, 45/2018, 2018, Belgrade,
[10] Beobalans, “Measurement of braking force of a motorcycle”, Available from: http://beobalans.rs/tehnicki-
pregled/mb-30-pc, Accessed 14.09.2018.
529
530
Kragujevac, Serbia
MVM2018-026
Bojana Bošković1
Milan Mladenović2
Saša Vasiljević3
Marina Teofilović4
Nadica Stojanović 5
Ivan Grujić6
PERSONS WITH DISABILITIES AS PARTICIPANTS IN THE

TRAFFIC
ABSTRACT: Being different is very expensive today. When the problems of people with disabilities are viewed
from a broader perspective, fulfilling their needs remains a challenge. People with disabilities meet daily with various
types of rejection and difficulties with whom they have to fight on their own. In Serbia, on average, about 200 persons
with disabilities take a driving test every year. It is necessary to provide adequate conditions for persons with
disabilities to become or continue to be equal traffic participants, to take driving tests at affordable prices and to have
the benefit of modifying and purchasing new vehicles on the reputation of some foreign countries. In addition, it is
necessary for participants in traffic, who are not persons with disabilities, to approach problems with which persons
with disabilities as participants in traffic meet on a daily basis in order to raise their awareness. It is also necessary
to provide adequate education programs through which drivers who are not persons with disabilities who share a car
that is modified for the needs of the person with disabilities, they could use the car in the least possible way.
KEYWORDS: people with disabilities, problems, education programs, modified vehicles
INTRODUCTION
The problems of persons with disabilities when are observed from the broader perspective, meeting their needs is
still a challenge. Although some countries have achieved a lot in the field of respecting the rights and equal
opportunities of persons with disabilities, Serbia, unfortunately, is still far behind. An enormous problem is the legal
regulations, which are and when they are adopted, or inadequately or not used at all.
People with disabilities meet daily with various kinds of rejection and difficulty, with which they have to make their
own choices. Such is the situation with those who want to take the driving test, where the only criterion is the market,
and needs of people with disabilities are seen exclusively as an additional unnecessary expense. The only way to
change such situation is the real reaction of the state, which, as we have explored, is largely absent.
1
Bojana Bošković, B.Sc., College of Tehnical and Mehanical Engineering Trstenik, bokaev92@gmail.com
2
Milan Mladenović, B.Sc., Faculty of Engineering University of Kragujevac, Kragujevac,
m.mladenovic141@gmail.com
3
Saša Vasiljević, M.Sc., High technical school of professional studies, Kragujevac, vasiljevic.sasa036@gmail.com
4
Marina Teofilović, B.Sc., Faculty of Engineering University of Kragujevac, Kragujevac, Kragujevac,
marinatea92@gmail.com
5
Nadica Stojanovic, M.Sc., assistant, University of Kragujevac, Faculty of Engineering, 6 Sestre Janjić STR.,
6
531
The law of traffic safety does not oblige the driving-schools to provide conditions for practical training for persons
with disabilities, while for owners of auto-schools this is not profitable, so that wheelchairs users still have to train
passes and take the exam in their vehicles [1] which is not acceptable from the point of safety. Driving can also
contribute to improvement of quality of life and health [2]. However, many people with mobility impairments cannot
use a conventional car unless it is adapted.
Among persons with disabilities the share of women (58.2%) is higher than that of men (41.8%), show in Figure 1.
The average age of women with disabilities is about 69 years, and for men is about 64 years. Observed by type of
problems, the biggest share of persons with disabilities is due to a problem with walking: for 5.8% of the total number
of women and for 3.6% of the total number of men in the Republic of Serbia [3].
Figure 1 Total population and persons with disability according to age and sex, Republic of Serbia,
2011 census [4]
Studies in research for disabled people was heavier than that for unimpaired people, and it was also discovered that
there are some difficult situations which exist for disabled people when they try to drive in the same way as unimpaired
people. Driving environments are designed basically for use by unimpaired people, and building highways with small
and sharp curves is inevitable. In order to support safe driving by disabled people, the following are required [5]:
 To investigate the driving load for disabled people based on their driving and traveling situations and to
reduce their driving load as much as possible;
 To predict danger and to develop traffic-warning systems equipped with functions for predicting dangerous
situations.
Тhe installation of driving-aid devices to vehicles built basically for the movements of unimpaired people and the
alternation of driving methods are not enough to support driving by disabled people. When disabled people are driving
vehicles, one of their hands is mainly used for steering the wheel and another hand is used for controlling acceleration
and braking. They cannot control the steering wheel with both hands; thus their posture is unstable when they are
driving. Тhe following are required [5]:
 Driver's seats which can stabilize the driver's posture even on curves;
 Development of a human interface system for disabled people; and
 A support system for safety, such as access to information on road situations and warning systems when
disabled people drive into the starting point of curves.
Entering and leaving the vehicle for people with disabilities should be made in the possible easiest way, Figure 2. It
is exist a wide range of vehicles that can be modified for the needs of persons with disabilities [6]. It is possible to
modify a classic vehicle to the vehicle which can be controlled by hand controls. If a classic vehicle is modified to a
vehicle for the needs of people with disabilities, the manufacturer's consent and guidelines are required. After that,
the owner of the vehicle can make a rebuilding of the vehicle. It then receives a confirmation that the installation is
done according to the manufacturer's guidelines and can then attest to its vehicle [6]. Of course, there are vehicles
532
that are designed just for them. How today attention is mostly concentrated on the of environmental pollution
reduction in the city, there are electric vehicles for people with disabilities, Figure 3. In addition, it is also necessary
to enable them to operate the vehicle in an easier way. In addition, it is very important to adjust the vehicle for people
with disabilities so they can easy and simple to control the vehicle, Figure 4.a. In the event that the upper extremities
of such persons are weak, it is possible for the vehicle to operate through the joystick, Fig. 4.b. When designing one
such vehicle, it is necessary to take care of ergonomics. Dols et al. in their research [7]:
 Produce information about the disabled population;
 Interpret data useful for designing a vehicle adaptation;
 Simulate in a suitable space the driving position, the driver, and the interaction between them and so create
a repeatable process which allows the optimization of a driving adapted position.
Which will be further tested for the range, the distance movement of the controls and the necessary activation force.
In addition, traffic testing will be carried out, analysing the physical and mental abilities in order to obtain a driver's
license.
Figure 2 Toyota Porte [8]
Figure 3 Mini-cars and steering system [9]
a) Driving with steering wheel b) Joystick for driving

Figure 4 Hand control drive system [10, 11]
533
It is very important that vehicles for persons with disabilities are equipped with active safety systems. For example,
if the vehicle is equipped with the Adaptive Cruise Control (ACC) system, people with disabilities would be much less
tired and their driving would be more comfortable [12].
Disability is a human rights issue because [13]:

 People with disabilities experience inequalities – for example, when they are denied equal access to health
care, employment, education, or political participation because of their disability.
 People with disabilities are subject to violations of dignity – for example, when they are subjected to violence,
abuse, prejudice, or disrespect because of their disability.
 Some people with disability are denied autonomy – for example, when they are subjected to involuntary
sterilization, or when they are confined in institutions against their will, or when they are regarded as legally
incompetent because of their disability.
Research shows that discrimination of people with disabilities in Serbia is a direct and indirect, widespread
phenomenon in all areas of social life, especially in the fields of education, work and employment, community life,
equality before the law, access to services and information. Also, people with disabilities and their families often
suffer from stigmatization and segregation due to deeply rooted and widespread prejudices and lack of information
[14].
The main purpose of the paper is to show the position of people with disabilities in traffic and what other traffic
participants think about it.
PARKING SPACES FOR VEHICLES OF PERSONS WITH DISABILITIES

In the world in which we live today, there is a growing motorization and travel on a global basis, so the vehicle is an
inevitable part of people with disabilities. Parking permits can help overcome barriers to their use by providing
reserved and privileged parking spaces in the immediate vicinity of the destination.
Parking spaces should have such position to be in the front of the building, entrance or if possible in the yard of the
building. In public garages and parking lots, there are marked parking spaces for vehicles of peple with disabilities,
on carts and sidewalks (Figure 5), as well as on outside car parks [15]. If these parking spaces are side by side, a
minimum distance of 1.3 m should be left (Figure 6). In no case should the number of accessible parking spaces be
less than 4 for the first 100 spaces (i.e. 1:25 parking spaces) plus 2 spaces for each additional 100 parking spaces
(i.e. 1:50), Table 1. No less than 1 accessible parking space should be provided where the number of parking spaces
provided is less than 25 [16].
Figure 5 Parking place for people with disabilities on road [16]
534
Figure 6 Marked out shared space between two standard bays [17]
Table 1 Requirements for parking places for people with disabilities [18]
Total number of parking space Min number of accessible spaces
1 to 25 1
26 to 50 2
51 to 75 3
76 to 100 4
101 to 150 5
151 to 200 6
201 to 300 7
301 to 400 8
401 to 500 9
501 to 1000 2% of total
1001 and over 20, plus 1 for each 100, or fraction thereof, over 1000
RESULTS OF THE POLL "PERSONS WITH DISABILITIES AS PARTICIPANTS

IN TRAFFIC"
Based on a survey conducted on 50 respondents, some of whom are and some of them not drivers, in order to arrive
at a conclusion about the situation of persons with disabilities and how other participants in the traffic are seen by
them.
The survey was conducted by random selection, which can be concluded based on the question "Are you a driver?".
The essence of this research is to conclude how other participants in the traffic see persons with disabilities, whether
they are drivers or not. Of the fifty respondents, 72% are drivers.
In addition, it can be said that their small check of knowledge was made, because the questionnaire contained the
question: "What is the sign shown in the picture (Figure 7)". As the question did not contain the answers offered, the
respondents were supposed to say what the given sign represents. The answers were different, but everyone knew
they were people with disabilities.
535
Figure 7 The International Symbol of Access (ISA), also known as the (International) Wheelchair Symbol
When asked if there is a person with disabilities in your environment, 54% of respondents answered “Yes” it can be
seen on Figure 8, which means that this phenomenon is not foreign to them, even more so that they are accustomed
to their presence, both in the neighbourhood and in the traffic.
Figure 8 Graphic representation of the answer to the question: Is there a person with disabilities in your
environment?
As the majority response was that in the environment of respondents are persons with disabilities, a different
awareness was developed in these people. So when parking, they don’t park their vehicle on place marked for
persons with disabilities, regardless of the circumstances and crowds in the parking lots, will not take the place
reserved for people with disabilities. However, there are a few who are still parked in places that are designed for
people with disabilities (Figure 9).
Figure 9 Graphic representation of the answer to the question: If you are a driver, have you ever parked in a place
marked as a parking place for people with disabilities?
It can be seen from the figure that even 14% of the drivers parked their vehicle at a parking spot for people with
disabilities, although they know it is forbidden by law if they do not have a sticker on their vehicle (Figure 10), as well
as a special parking card which persons with disabilities receive.
536
Figure 19 Sticker and card for persons with disabilities
The answer to questions about whether people with disabilities threaten traffic and whether they can be equal
participants in traffic, 39 out of 50 respondents answered that people with disabilities can be equal participants in
traffic, Figure 11.
Figure 11 Graphic representation of the answer to the question: Do you consider that persons with disabilities can
be equal participants in the traffic?
From Figure 12 can be seen that even 12% of the 50 respondents said that people with disabilities, as drivers, were
jeopardizing traffic, and even seven of them were asked if they had ever parked in a place marked as a parking spot
for people with disabilities, has answered affirmatively. In addition, we are not competent to give an opinion whether
someone is able to operate the vehicle or not. Such an opinion can only be given by a doctor. The reason for such a
way of thinking that is present in people is their social consciousness.
Figure 12 Graphic representation of the answer to the question: Do you consider that people with disabilities, as
drivers, are endangering traffic safety?
537
At the end of the survey, with the aim of raising the awareness of respondents about people with disabilities, the
following Figure 13 was shown, with the possibility of writing the answer, 39 of them responding with "I do not want"
and the rest was left blank.
Figure 13 Message that should raise awareness of people
CONCLUSIONS
People with disabilities meet daily with various kinds of rejection and difficulty, with which they have to struggle. In
order to facilitate their departure to the desired place, car manufacturers give them a wide choice of vehicles for their
needs. Also, the local community allows them to have a reserved place for them. This is because they need a place
for the vehicle, as well as an additional area for going out and entering the vehicle with wheelchairs. The poll found
that in our country there are those who think that they are not equal participants in traffic and that their situation can
endanger others. The reason for this is the social consciousness that has developed in a few people. This was also
established on the basis of previous research. However, there are those who do not think so. Further, people with
disabilities should be promoted as drivers, thereby demonstrating that they are equal participants in traffic and that
they cannot in any way cause the occurrence of a traffic accident.
ACKNOWLEDGMENTS
This paper was realized within the researching project “The research of vehicle safety as part of a cybernetic system:
Driver-Vehicle-Environment” ref. no. TR35041, funded by Ministry of Education, Science and Technological
Development of the Republic of Serbia.
REFERENCES
[1] Law on Road Traffic Safety, „Official Gazette of the RS“, No. 24/18,
[2] Fulland J., Peters B.: “Regulations and routines for approval of passenger cars adapted to drivers with disabilities
– including an international survey”, VTI, Linköping, 1999,
[3] Women and Men in the Republic of Serbia, First Edition, Statistical Office of the Republic of Serbia, Belgrade,
2014, ISBN 978-86-6161-133-9,
[4] Statistical yearbook of the Republic of Serbia, Statistical Office of the Republic of Serbia, Belgrade, 2015, ISSN
0351-4064,
[5] Ikeda, H., Ikeda, H., Mihoshi, A., Hisara, Y.: “Physical Load Related to Highway-Driving Among Disabled
People”, IATSS Research, Vol. 31, 2007, pp 100-109,
[6] Road Traffic Safety Agency, Available from: http://www.abs.gov.rs/sr/vozila/ispitivanje-vozila/ispitivanje-
prepravljenih-vozila/podnosenje-zahteva-za-ispitivanje-prepravljenih-vozila, Accessed 04.08.2018.,
[7] Dols, J.F., García, M., Sotos, J.J.: „Procedure for improving the ergonomic design of driving positions adapted
for handicapped people“, Boletín Factores Humanos, Vol. 12,1996, pp 7-17,
[8] Al Palmer Repairs, Available from: https://alpalmer.com.au/imports.html, Accessed 03.08.2018.,
[9] Der-Querschnitt, Available from: https://www.der-querschnitt.de/archive/13694, Accessed 03.08.2018.,
[10] Research Institute for Disabled Consumers, Available from: https://www.ridc.org.uk/content/custom-built-car-
controls, Accessed 03.08.2018.,
538
[11] Superior Van & Mobility, Available from: http://www.superiorvan.com/uploads/hand-controls.jpg, Accessed
03.08.2018.,
[12] Peters, B.: “ADAPTATION EVALUATION: An Adaptive Cruise Control (ACC) System Used by Drivers with
Lower Limb Disabilities”, IATSS Research, Vol.2 25, 2001, pp 51-60,
[13] Quinn G, Degener T. A survey of international, comparative and regional disability law reform. In: Breslin ML,
Yee S, eds. Disability rights law and policy - international and national perspectives. Ardsley, Transnational,
2002,
[14] Special Report on the Discrimination of Persons with Disabilities in Serbia, 2013, Belgrade,
[15] Kostić, S., Davidović, B.: “Parkiranje i javne garaže”, Fakultet tehničkih nauka u Novom Sadu, 2012, Novi Sad,
[16] Toronto, Available from: https://www.toronto.ca/wp-content/uploads/2017/08/8fcf-accessibility_design_
guidelines.pdf, Accessed 08.08.2018.,
[17] City of Melton, Available from: https://www.melton.vic.gov.au/files/assets/public/services/building-planning-amp-
transport/engineering/infrastructure-planning/car_parking_guidelines_-_2015.pdf, Accessed 08.08.2018.,
[18] Illinois, Available from: https://www2.illinois.gov/sites/iced/training/Documents/ICED_Presentation_Accessibility
Requirements_State_FederalLaw_3-9-16.pdf, Accessed 08.08.2018.
539
540
Kragujevac, Serbia
MVM2018-032
Clio G. Vossou1
Dimitrios V. Koulocheris2
A COMPUTATIONAL MODEL FOR THE RECONSTRUCTION

OF VEHICLE COLLISIONS
ABSTRACT: The reconstruction of a vehicle collision consists of the processes of investigation, analysis and
conclusions about the causes and the events during the traffic accident. In this perspective, a vehicle collision is
considered in three distinct phases, the pre-collision, the collision and the post-collision phase. For the analysis of
the collision phase two main approaches exist in the literature, the energy based and the momentum based one. The
latter has been described in details by Brach et al. and can find a solution to a given set of parameters to reconstruct
a collision. These parameters can be known or assumed using monitoring systems of the vehicle or physical
evidence. In the present paper a computational model implementing the Planar Impact Mechanics (PIM) collision
model has been set up in Matlab ® and its coupling with the least squares method has been investigated. As test
cases, the RICSAC database, which consists of twelve staged collisions, have been used. Special attention has been
given in the number of parameters which have been considered known (or assumed). The results indicate the
importance of each parameter.
KEYWORDS: traffic accident reconstruction, Planar Impact Mechanics, RICSAC, least squares method,
computational model
INTRODUCTION
Traffic accident reconstruction involves the qualitative and quantitative estimation of the way such an accident
occurred. This process includes the use of engineering, scientific and mathematical laws and it is based on data and
physical evidence collected from the accident scene.
A typical traffic accident involves the collision of two vehicles and it can be considered in three distinct time phases,
the pre-collision, the instantaneous collision and the post-collision phase. In the literature, two broad approaches are
used for the simulation of the collision phase, the one based on the conservation of linear and/or angular momentum
and the one based on the conservation of energy. The momentum based models are these defined by Brach (Brach,
1987) and Ishikawa (Ishikawa, 1993) while the model considering the conservation of energy is this defined by
McHenry (McHenry, 1981).
The momentum based collision simulation model presented by Brach is based on the second law of Newton and the
principle of impulse – momentum. This collision model is referenced in the literature as Planar Impact Mechanics
model and it consists of six algebraic equations. Since the movement of a vehicle is quantified with three velocity
components, the normal, the tangential and the rotational velocity, twelve velocity components, six initial (three per
vehicle) and six final ones are involved in this model. Furthermore, this model incorporates three impact coefficients
quantifying physical constraints of the collision, namely, the restitution coefficient, the equivalent coefficient of friction
and the momentum coefficient of restitution. All fifteen parameters, velocity components and impact coefficients,
need to be known, assumed or calculated in order for a traffic accident reconstruction to be accomplished.
1
Clio G. Vossou, Researcher, Vehicles Laboratory, School of Mechanical Engineering, National Technical
University of Athens, Greece, Zografou Campus, Iroon Polytexneiou 9, 157 80, klvossou@mail.ntua.gr
2
Dimitrios V. Koulocheris, Assistant Professor, Vehicles Laboratory, School of Mechanical Engineering, National
Technical University of Athens, Greece, Zografou Campus, Iroon Polytexneiou 9, 157 80, dbkoulva@mail.ntua.gr
541
In the present paper the Planar Impact Mechanics collision model, which is presented in details in the following
section, has been set up in the programming environment of Matlab ®. Planar Impact Mechanics collision model has
been coupled with the least squares method in order to calculate the unknown velocity components and/or the impact
coefficients. A minimization problem has been set up, using the least squares method equation as a cost function.
Using the deterministic optimization method of SQP, implemented in Matlab through the fmincon function, the values
of the impact coefficients where determined.
As test cases, the RICSAC database, which consists of twelve staged collisions, have been used. Special attention
has been given in the number of velocity components which have been considered known (or assumed) exploring
the behaviour of three different cost functions. For the evaluation of the performance of the optimization procedure
and the efficiency of each cost function the measured velocity values for each RICSAC test have been used. The
results consider the performance of the optimization set up, the performance of each cost function and the importance
of the impact parameters.
MATERIALS & METHODS

In this section the equations of the Planar Impact Mechanics collision model are going to be introduced. Moreover,
the test collisions, belonging in the RICSAC database, are going to be presented. Finally, the least squares method
along with the optimization procedure are going to be described providing information on the design variables, the
evaluated cost functions and the applicable constraints.
Planar Impact Mechanics collision model

In 1977 Brach (Brach, 1977) presented a set of six linear algebraic equations which simulate the planar collision of
two vehicles. These equations are based on the Newton’s second law and the impulse - momentum principles and
achieve the calculation of the velocity changes of two particles due to impact (Brach, 1983). This set of equations
can be applicable to the simulation of a collision of two vehicles provided that the following assumptions are satisfied
(Brach, 1984):
1. The resultant intervehicular impulse is much larger than the impulses of other forces such as friction with the
ground, drive train drag and aerodynamic drag which are neglected.
2. The resultant impulse vector of the intervehicular force acts at a single point, assumed to be know, called
the centre of impact.
3. Changes in the position of the centre of mass and in angular orientation in every vehicle are small over the
time interval of contact.
4. A hypothetical, fixed contact surface is presumed in such a way that motion normal to this surface is due to
deformation while motion parallel to this surface has the nature of relative motion corresponding to frictional
sliding.
5. The time duration of contact is small.
Vehicular collisions typically have contact times less than 0.2 s. Time intervals of this magnitude, coupled with the
assumption of large forces cause large accelerations, finite velocity changes and small displacements. All of these
considered together, usually, cause the above assumptions to be satisfied for the study of vehicle collisions.
Figure 1 Free body diagram of the collision of two vehicles (Brach, 1983)
542
In Figure 1, the free body diagram of two vehicles (Vehicle 1 and Vehicle 2) in the collision phase is presented. These
vehicles are in contact along the crush surface (line forming angle Γ with y-axis of the global Cartesian Coordinate
System (CCS)). Each vehicle has an a priori known mass and inertia, 𝑚𝑖 , 𝐼𝑖 , respectively.
Since the momentum is conserved in each axis of the global CCS Eq. 1 and Eq. 2 are constructed.
𝑚1 ∙ (𝑉1𝑓𝑥 − 𝑉1𝑖𝑥 ) + 𝑚2 ∙ (𝑉2𝑓𝑥 − 𝑉2𝑖𝑥 ) = 0 (1)
𝑚1 ∙ (𝑉1𝑓𝑦 − 𝑉1𝑖𝑦 ) + 𝑚2 ∙ (𝑉2𝑓𝑦 − 𝑉2𝑖𝑦 ) = 0 (2)
The indices 1 and 2 represent each vehicle while the indices i and f stand for the initial (i.e. start of collision phase)
and the final (end of collision phase) values, respectively. These indices are used in all following equations.
In Figure 1 the distances 𝑑1 and 𝑑2 correspond to the distance between the centre of mass of each vehicle and the
centre of impact. Furthermore, 𝜑1 and 𝜑2 represent the angle between 𝑑1 and 𝑑2 and the longitudinal axis of each
vehicle while with 𝜃1 and 𝜃2 the angle between the the longitudinal axis of each vehicle and the x-axis of the global
CCS is denoted. Using these geometrical quantities and the principle of conservation of angular momentum Eq. 3 is
derived.
𝐼1 ∙ (𝛺1𝑓 − 𝛺1𝑖 ) + 𝐼2 ∙ (𝛺2𝑓 − 𝛺2𝑖 ) + 𝑚1 ∙ (𝑑𝑏 + 𝑑𝑑 ) ∙ (𝑉1𝑓𝑦 − 𝑉1𝑖𝑦 ) + 𝑚2 ∙ (𝑑𝑎 + 𝑑𝑐 ) ∙ (𝑉2𝑓𝑥 − 𝑉2𝑖𝑥 ) = 0 (3)
In Eq. 3 𝑑𝑎 , 𝑑𝑏, 𝑑𝑐 , 𝑑𝑑 are correlated to 𝑑1 and 𝑑2 with the following trigonometrical functions.
𝑑𝑎 = 𝑑2 ∙ 𝑠𝑖𝑛(𝜃2 + 𝜑2 ) (3a)
𝑑𝑏 = 𝑑2 ∙ 𝑐𝑜𝑠(𝜃2 + 𝜑2 ) (3b)
𝑑𝑐 = 𝑑1 ∙ 𝑠𝑖𝑛(𝜃1 + 𝜑1 ) (3c)
𝑑𝑑 = 𝑑1 ∙ 𝑠𝑖𝑛(𝜃1 + 𝜑1 ) (3d)
The following three equations (Eq.4 – 6) are provided considering the impact coefficients. The coefficient of
restitution, 𝑒, is used to model energy loss due to material deformation in a mode normal or perpendicular to the
crush surface.
(𝑉1𝑓𝑦 − 𝑑𝑑 ∙ 𝛺1𝑓 − 𝑉2𝑓𝑦 − 𝑑𝑏 ∙ 𝛺2𝑓 ) ∙ 𝑠𝑖𝑛𝛤 + (𝑉1𝑓𝑥 − 𝑑𝑐 ∙ 𝛺1𝑓 − 𝑉2𝑓𝑥 − 𝑑𝑎 ∙ 𝛺2𝑓 ) ∙ 𝑐𝑜𝑠𝛤
=𝑒
∙ [((𝑉1𝑖𝑦 − 𝑑𝑑 ∙ 𝛺1𝑖 − 𝑉2𝑖𝑦 − 𝑑𝑏 ∙ 𝛺2𝑖 ) ∙ 𝑠𝑖𝑛𝛤) + (𝑉1𝑖𝑥 − 𝑑𝑐 ∙ 𝛺1𝑖 − 𝑉2𝑖𝑥 − 𝑑𝑎 ∙ 𝛺2𝑖 ) (4)
∙ 𝑐𝑜𝑠𝛤]
The equivalent friction coefficient, 𝜇, corresponds to the ratio of the tangential to normal impulse components which
develop between the vehicles. The tangential impulse is typically attributed to and referred to as friction, though shear
deformation is probably equally significant (Brach 1987).
𝑚1 ∙ (𝑉1𝑓𝑦 − 𝑉1𝑖𝑦 ) ∙ (𝑐𝑜𝑠𝛤 − 𝜇 ∙ 𝑠𝑖𝑛𝛤) + 𝑚2 ∙ (𝑉2𝑓𝑥 − 𝑉2𝑖𝑥 ) ∙ (𝑠𝑖𝑛𝛤 + 𝜇 ∙ 𝑐𝑜𝑠𝛤) = 0 (5)
Finally, the third coefficient, the moment coefficient of restitution, 𝑒𝑚 , governs the rotational effects. A value of this
coefficient of unity (1) implies that no moment impulse is developed between the vehicles during collision and that
the centre of impact is known. Otherwise any value in the range of [-1, 0] implies that the centre of impact is not
known (Barch 1987).
(𝛺1𝑓 − 𝛺2𝑓 ) ∙ (1 − 𝑒𝑚 )
= −𝑒𝑚
(𝑉1𝑓𝑥 − 𝑉1𝑖𝑥 ) (𝑉1𝑓𝑦 − 𝑉1𝑖𝑦 )
∙ [((𝛺1𝑓 − 𝛺1𝑖 ) − 𝑚1 ∙ 𝑑𝑐 ∙ + 𝑚1 ∙ 𝑑𝑑 ∙ − (𝛺2𝑓 − 𝛺2𝑖 ) − 𝑚2
𝐼1 𝐼1 (6)
(𝑉2𝑓𝑥 − 𝑉2𝑖𝑥 ) (𝑉2𝑓𝑦 − 𝑉2𝑖𝑦 )

∙ 𝑑𝑎 ∙ + 𝑚2 ∙ 𝑑 𝑏 ∙ )]
𝐼2 𝐼2
543
The Planar Impact Mechanics collision model through the use of these coefficients models the energy loss which is
present in all real collisions, since there is loss of kinetic energy, mostly through deformation, friction and vibrational
energy. According to the literature (Brach, 1987), typical values of energy loss due to collision range from 25% to
95%.
It is noteworthy that Eq. 1 – 6 correlate (a) six initial velocity components (three for each vehicle – 𝑉𝑖𝑥 , 𝑉𝑖𝑦 , 𝛺𝑖 ), (b) six
final velocity components (three for each vehicle – 𝑉𝑓𝑥 , 𝑉𝑓𝑦 , 𝛺𝑓 ), (c) vehicle inertial properties (𝐼, 𝑚) and (d) the
collision geometry (𝑑𝑎 , 𝑑𝑏 , 𝑑𝑐 , 𝑑𝑑 , 𝛤). The aforementioned equations written in a matrix form (Eq. 7) constitute the
mathematical collision model (Brach, 1983).
𝐴 ∙ 𝑉𝑓 − 𝐶 ∙ 𝑉𝑖 = 0 (7)
RICSAC database
In the 1970’s, while the first computational methods for the simulation of traffic accidents appeared in the literature,
the need for a database with fully defined vehicle collisions arose. Within this context a research project named the
“Research Input for Computer Simulation of Automobile Collisions” (RICSAC) and funded by the National Highway
Traffic Safety Administration (NHTSA), provided the researches with a test matrix of 12 full – scale crash tests. Within
this project cameras and accelerometers were attached to the vehicles and a set of staged collision scenarios has
been performed resulting into a test matrix of 12 crashes.
For each RICSAC test, at least 13 accelerometers were mounted on each vehicle in order to monitor the acceleration
components. At three locations, triaxial (XYZ) packages were installed to provide coverage between the front and
the rear of the vehicle. The front steer angles were measured on each vehicle by a linear stroke potentiometer
attached to the vehicle steering linkage. The time history of the change in vehicle yaw, pitch and roll angles and yaw
rate were recorded by two degrees of freedom, free gyroscopes and rate gyro (McHenry, 1987). The final test reports
include, also, objective information on the impact speeds, vehicle weights, vehicle dimensions, weight distributions,
spin-out trajectories and positions of rest.
RICSAC test database contains vehicle collisions engaging six different vehicles included in four categories of vehicle
sizes. The different vehicles used in RICSAC tests are namely (V1) Chevrolet Chevelle, (V2) Ford Pinto, (V3) Ford
Torino, (V4) Honda Civic, (V5) VW Rabbit and (V6) Chevrolet Vega. The tests can be classified into four impact
configurations (IC1 – IC4) (Figure 2) according to the relative orientation of the vehicles at the time of collision. In the
IC1 belong the Tests no. 1, 2, 6 and 7, in the IC2 belong the Tests no. 3, 4 and 5, in IC3 belong the Tests no. 8, 9
and 10 and in IC4 belong the Tests no. 11 and 12. Each crash test involved vehicles of different size categories,
except for Test no. 8, which involved two intermediate vehicles. In the front-to-rear collisions (Figure 2b) the car
struck in the rear was stopped while in all other tests, both cars were moving.
(a)
(b) (c) (d)
Figure 2 (a) IC1 - Front corner to corner at 60 o, (b) IC2 - Rear offset oblique at 10 o (c) IC3 - Side
perpendicular offset and (d) IC 4 - Frontal offset oblique at 10o (Struble, 2013)
In Table 1, that follows, the measured velocity components of both vehicles for each RICSAC test are presented.
This and the following tables are organized per IC and the data of RICSAC test no. 2 are not included due to loss of
experimental measurements.
544
Table 1 Linear and angular velocity for each vehicle for all RICSAC tests
Impact Configuration (IC)
1 2 3 4
RISAC Units 1 6 7 8 9 10 3 4 5 11 12
Vi1x m/s -8.95 -9.61 -13.01 -9.3 -9.48 -14.89 -9.48 -17.30 -17.75 -9.12 -14.08
Vi1y m/s 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 -4.28
Ωi1 rad/s 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Vi2x m/s 4.43 4.66 6.50 0.00 0.00 0.00 0.00 0.00 0.00 8.98 13.87
Vi2y m/s 7.67 8.32 11.27 9.3 9.48 14.89 0.00 0.00 0.00 -1.58 2.44
Ωi2 rad/s 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Vf1x m/s -3.76 -5.69 -7.74 -3.12 -0.86 -1.55 -5.34 -8.94 -10.46 1.77 4.28
Vf1y m/s 2.41 1.26 1.48 3.27 4.52 8.59 -0.32 -0.44 0.17 0.62 -0.49
Ωf1 rad/s -1.57 -0.52 -0.52 -1.99 -3.14 -5.24 -0.42 -0.65 -0.21 0.52 1.57
Vf2x m/s -2.07 -1.28 -2.22 -3.66 -3.02 -4.44 -6.73 -9.92 -11.32 1.96 1.93
Vf2y m/s 5.17 5.49 8.64 6.01 7.38 11.14 1.14 0.42 0.84 -1.26 -2.94
Ωf2 rad/s 0.00 -3.14 -3.35 -0.31 0.79 1.26 -0.42 -0.52 -1.22 0.00 1.05
In Table 2 the geometrical properties of each collision are presented, i.e. the crush angle (Γ) for each IC along with
the initial heading angle (θi) and the angle to center of collision (φi). Both angles θi and φi are presented for each
vehicle while angle φi is also presented for each RICSAC.
Table 2 Crush angle per RICSAC, initial heading angle and angle to center of collision per vehicle and RICSAC
1 2 3 4
RISAC Units 1 6 7 8 9 10 3 4 5 11 12
Γ -30.0 0.0 -10.0 0.0
θ1 0.0 0.0 0.0 -10.0
θ2 deg 60.0 90.0 170.0 0.0
φ1 -19.8 -17.9 17.9 0.0 6.0 0.0 -17.0 -18.2 -20.7 9.4 9.6
φ2 -38.7 -90.0 -90.0 -68.8 -29.7 -29.2 171.4 171.7 -168.0 11.3 10.3
Least squares method and minimization problem

As mentioned above, the least squares method is utilized as a means of retrieving a combination of unknown
parameters in a way that the equations of the Planar Impact Mechanics model are satisfied and the specified velocity
components are closely matched to the estimated ones. The assumed values of the velocity components may result
from monitoring devices mounted on the vehicle and/or physical evidence. Such a monitoring device installed in,
more vehicles as time progresses, is the Event Data Recorder (EDR). Thus, in the least squares method the velocity
components have been included, since an estimate of their value might be available.
In order to utilize the least squares method the vector of all the velocity components included in Eq. 7 is renamed to
the vector U in the following way:
{𝑈1 𝑈2 𝑈3 𝑈4 𝑈5 𝑈6 𝑈7 𝑈8 𝑈9 𝑈10 𝑈11 𝑈12 }𝑇

𝑉𝑓2𝑦 𝛺𝑓2 }𝑇 (8)
= { 𝑖1𝑥 𝑉𝑖1𝑦 𝛺𝑖1 𝑉𝑖2𝑥 𝑉𝑖2𝑦
𝑉 𝛺𝑖2 𝑉𝑓1𝑥 𝑉𝑓1𝑦 𝛺𝑓1 𝑉𝑓2𝑥
In order to evaluate the effect of the number of the known and/or estimated velocity components, Eq. 9 – 11 have
been set up.
6
𝑄1 = ∑ 𝑤𝑘 ∙ (𝑈𝑘 − 𝑈𝑘𝑒𝑠𝑡 )2
𝑘=1 (9)
545
12
𝑄2 = ∑ 𝑤𝑘 ∙ (𝑈𝑘 − 𝑈𝑘𝑒𝑠𝑡 )2 (10)
𝑘=7
12
𝑄3 = ∑ 𝑤𝑘 ∙ (𝑈𝑘 − 𝑈𝑘𝑒𝑠𝑡 )2 (11)
𝑘=1
In all three equations 𝑤𝑘 , are weighting factors in the range of (0,1] allowing for the definition of different confidence
levels for each estimate. In the present paper all 𝑤𝑘 have been considered equal to unity stating that all the estimates
have the same level of confidence. The experimental estimates 𝑈𝑘𝑒𝑠𝑡 are considered having the corresponding values
available in Table 1, i.e. the measured values available in the RICSAC database.
In Q1 (Eq. 9) all the final velocity components have been considered exactly know and their values, equal to those
appearing in Table 1, are used in Eq.7. As far as the initial velocity components are concerned, estimates have been
considered available, and they were treated as unknowns in Eq.7. On the contrary, in Q2 (Eq. 10), all the initial velocity
components have been considered exactly know (Table 1), while experimental estimates have been considered
available for the final velocity components. Finally, in Q3 (Eq.11), no velocity component has been considered a priori
known, but estimates are considered to exist for all of them. For all three equations, the geometrical properties of the
collisions are considered exactly known and having the values appearing in Table 2 while for the impact coefficients
(e, μ and em) no experimental values have been considered available.
In order to retrieve the impact coefficients (e, μ and em) these three equations (Eq. 9 – 11) have been coupled with
an SQP based optimization subroutine (fmincon) available in Matlab ® in order to achieve the minimization of each
Q, forming an equal number of optimization problems.
𝐹𝐶𝑖 = min(𝑄𝑖 ) , 𝑖 = 1 − 3 (12)
Each optimization procedure (Eq. 12) provides the values of the design variables, being the three impact coefficients
(e, μ and em), which minimizes Qi with 𝑈𝑘 values calculated via the least squares method. The Planar Impact
Mechanics collision model are included in the optimization procedure in the form of non-linear constraints. Other
active constraints of the optimization procedure are the boundary values of the velocity components and these of the
design variables. In more details, the linear velocities have been considered in the range of [-20, 20] m/s and the
angular velocities have been considered in the range of [-5, 5] rad/s. Respectively, the ranges of the design variables
have been considered as: 𝑒 ∈ [0,0.2], 𝑒𝑚 ∈ [−1,0] and 𝜇 ∈ [0,1.2]. The constraints are imported in the optimization
procedure through the augmented Lagrange function (𝐿).
RESULTS
In the results section the outcome of all three optimization procedures is presented in terms of minimum values of
cost function (FCi).
In order to investigate the feasibility of the solution of each minimization procedure along with the value of the cost
function, also the values of the derivative of the augmented Lagrange function (∇𝐿) and the vector of the linear
constraints (C) are monitored. Moreover, the optimized values of the impact coefficients are introduced and the results
regarding the velocity component values are presented in terms of absolute error with respect to the measured,
corresponding values. All results are organized per IC and cost function.
In Figure 3 the results of the optimization procedure, for all RICSAC tests, is presented. Each bar represents a
different cost function 𝐹𝐶𝑖 . It is obvious that 𝐹𝐶2 , which considers the initial velocity components as known quantities
provides considerably higher values of cost function in tests 1, 6, 9, 10 and 11. For the rest of RICSAC tests the
minimum value of all 𝐹𝐶𝑖 is almost the same regardless cost function.
In Table 3, for tests 1, 6, 7, 3, 5 and 11, can be observed that the design variable vector which minimizes FC2 leads
to the violation of the constraints. This violation is quantified through the values of the derivative of the augmented
Lagrange function (∇𝐿) which has a value significantly greater than 0 in the tests 1, 6, and 11. The value of C is
significantly greater than 0 in tests 1, 6, 7, 3 and 5.
546
Figure 3 Value of cost function for all cases at the end of the optimization procedure
Table 3 Constraint satisfaction for all optimization problems per RICSAC test
IC1 IC2
RICSAC1 RICSAC6 RICSAC7 RICSAC8 RICSAC9 RICSAC10
∇𝐿 C ∇𝐿 C ∇𝐿 C ∇𝐿 C ∇𝐿 C ∇𝐿 C
F1 0.001 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.002 0.002
F2 102.570 0.732 0.122 2.620 0.000 2.830 0.000 0.000 0.000 0.000 0.000 0.003
F3 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.000 0.001 0.000 0.002 0.000
IC3 IC4
RICSAC3 RICSAC4 RICSAC5 RICSAC11 RICSAC12
∇𝐿 C ∇𝐿 C ∇𝐿 C ∇𝐿 C ∇𝐿 C
F1 0.000 0.000 0.001 0.000 0.000 0.000 0.001 0.000 0.000 0.000
F2 0.005 0.260 0.003 0.000 0.000 1.010 23.060 0.003 0.004 0.001
F3 0.001 0.000 0.000 0.000 0.001 0.000 0.000 0.000 0.001 0.000
In the following Tables 4, 5 and 6 the absolute error values for each velocity component and every RICSAC test for
all three cost functions (FC1 – FC3) are presented.
Table 4 Absolute error for all velocity components in m/s for each vehicle for all RICSAC tests for the cost function
FC1
1 2 3 4
RISAC 1 6 7 8 9 10 3 4 5 11 12
Vf1x 0.25 -0.61 -1.18 1.07 1.01 1.50 0.08 1.63 1.13 0.10 -0.18
Absolute error
Vf1y -1.23 -0.47 -0.85 -0.22 0.76 0.33 0.40 0.52 0.92 -0.07 0.99
Ωf1 0.35 1.52 2.17 0.13 -1.44 -1.13 0.36 0.86 1.01 0.17 -0.97
Vf2x 0.94 1.48 0.38 1.20 0.49 1.35 -0.29 0.57 -0.12 0.13 -0.68
Vf2y 0.71 1.53 1.75 0.39 -0.33 -0.61 0.00 1.07 1.13 0.67 -2.44
Ωf2 0.84 -1.10 -0.66 0.90 -0.41 -1.09 0.36 0.99 0.00 0.17 0.21
547
FC2
1 2 3 4
RISAC 1 6 7 8 9 10 3 4 5 11 12
Vi1x -4.16 -5.56 0.34 -1.00 2.43 4.71 0.10 -1.68 -1.07 -7.25 -0.16
Vi1y 2.07 1.79 0.48 0.03 2.25 1.86 0.84 0.68 1.21 0.62 -0.99
Absolute error
Ωi1 2.67 1.68 1.79 0.42 3.15 2.78 0.79 0.30 2.06 1.11 1.98
Vi2x -3.92 -0.68 -0.81 1.27 2.07 4.38 0.08 0.49 0.01 -7.09 0.69
Vi2y -2.00 -3.69 -1.21 -0.15 1.01 1.36 0.08 1.32 1.06 -0.32 -1.44
Ωi2 3.24 3.61 0.65 1.73 0.46 0.49 0.79 1.49 0.61 0.06 2.41
FC3
1 2 3 4
RISAC 1 6 7 8 9 10 3 4 5 11 12
Vi1x -0.34 -0.33 0.13 -0.53 -0.25 -0.36 0.03 -0.81 -0.56 0.03 -0.09
Vi1y 0.47 0.14 0.14 0.09 0.45 0.24 0.41 0.25 0.63 0.03 -0.49
Ωi1 0.44 0.39 1.07 0.15 0.46 0.35 0.45 0.60 1.30 0.08 0.89
Vi2x 0.23 0.09 -0.29 0.61 0.39 0.95 0.11 0.29 0.04 -0.08 0.34
Absolute error
Vi2y -0.20 -0.37 -0.57 -0.18 0.20 0.34 0.02 0.53 0.57 -0.33 -1.71
Ωi2 0.47 0.19 0.20 0.39 0.22 0.01 0.23 0.30 0.38 0.08 1.32
Vf1x -0.02 -0.86 -1.00 0.53 0.38 0.69 0.00 0.81 0.56 0.06 -0.09
Vf1y -0.68 -0.54 -0.64 -0.09 0.45 0.24 0.00 0.25 0.29 -0.03 0.49
Ωf1 -0.05 1.23 1.39 -0.08 -1.10 -0.92 0.00 0.38 -0.08 0.09 -1.11
Vf2x 0.60 1.27 0.58 0.59 0.27 0.62 0.00 0.29 -0.04 0.06 -0.34
Vf2y 0.41 1.06 1.08 0.18 -0.20 -1.06 0.00 0.53 0.57 0.33 -1.71
Ωf2 0.45 -1.39 -1.06 0.61 -0.61 -1.07 0.00 0.51 -1.09 0.08 -1.00
The maximum absolute error for the use of cost function FC1 is -2.44 m/s and it appears in RICSAC test 12 in the
velocity component Vf2y. If FC2 is used then the maximum absolute error is -7.25 m/s and it appears in RICSAC 11
in the Vi1x. Finally, the maximum absolute error using FC3 is -1.71 m/s and it appears again in RICSAC 11 in the
velocity component Vi2y.
In Figure 4 the values of the impact coefficients are presented in the form of bar diagrams. For all impact coefficients,
the vertical axis represent the applicable range of values while the horizontal axis is organized per RICSAC test. The
different color in bars indicates use of different cost function, 𝐹𝐶𝑖 . Lack of a bar indicates that using the corresponding
cost function 𝐹𝐶𝑖 the optimization procedure converges for the minimum value of this impact coefficient.
In RICSAC tests 9, 10 and 3 the coefficient of restitution (Figure 4a) reached its maximum allowable value regardless
cost function. In the rest of the tests every cost function provides different value for the coefficient of restitution. In all
RICSAC tests, except RICSAC 5 FC2 provides the lowest value of coefficient of restitution. Furthermore, FC3 provides
higher value for the coefficient of restitution compared to FC1 with the exception of RICSAC 5.
The value of the moment coefficient of restitution (Figure 4b) is more uniform. In all RICSAC tests except 1, 7, 4 and
11 all cost functions provide the same value for this coefficient which is either its lower (6, 3, 5) or its upper bound
(8, 9, 10, 12).
As far as the values of the equivalent coefficient of friction (Figure 4c) is concerned they show significant non-
uniformity. Nevertheless in RICSAC tests belonging to IC1, FC2 seems to provide the highest values, followed by
FC3. In RICSAC tests 9 and 10 the opposite seems to happen, while in tests 3, 5 and 12 all cost functions provide
the same value of friction coefficient.
548
(a) (b)
(c)
Figure 4 Optimized value of the (a) restitution coefficient e, (b) moment coefficient of restitution em and (c) coefficient
of friction μ for all RICSAC and all cost functions
DISCUSSION
In Figure 3 is obvious that FC3 provides the overall minimum value for all RICSAC tests, while FC2 provides the
highest value in most RICSAC tests. The fact that FC3 provides results of better quality is also obvious in Table 6
where the absolute error for all velocity components are presented. In Table 3 where the values of ∇𝐿 and C are
presented, is obvious that FC2 leads to violation of the constraints in tests 1, 6, 7, 3, 5 and 11, meaning that Eq. 7 is
not satisfied, i.e. these solution vectors cannot be taken under consideration.
In order to quantify, in more details, the quality of each solution the error of the change of velocity in every vehicle is
also considered. The components of the vehicle’s change of velocity (ΔVi) were computed by subtracting the initial
velocity at impact from the velocity at the time of separation. The procedure was performed for the X and Y velocity
component separately. The measured change in velocity ΔV is provided in the literature (Brach, 1982) as an overall
measure of the severity of a collision. In Table 7 the measured velocity changes are presented per IC and RICASC
test.
In Figure 5 the relative computed velocity changes are presented for Vehicle 1 (Figure 5a) and Vehicle 2 (Figure 5b)
for the cost functions FC1 and FC3. The relative computed velocity change was calculated as the difference between
the measured and the computed velocity change divided by the measured one. FC2 was omitted for Figure 5 since,
as it was discussed earlier, it leads to solution vectors that may violate both the linear and the non-linear constraints.
The accuracy of both cost functions as far as the velocity change is concerned is fairly good. The relative computed
velocity change ranges from 0 to 25% for FC1 and from 0 to 36% for FC3. The maximum value of relative computed
velocity change appears for both cost functions in RICSAC test 10 and Vehicle 2. In general, the values of the relative
computed velocity change are larger in Vehicle 2 where 36% is met, than in Vehicle 1. The highest value Vehicle 1
is 19%.
549
Table 7 Absolute measured ΔV for each vehicle and RICSAC test
RICSAC 1 6 7
IC1 Vehicle 1 2 1 2 1 2
Measured ΔV (m/s) 5.72 6.96 4.12 6.58 5.47 9.11
RICSAC 8 9 10
IC2 Vehicle 1 2 1 2 1 2
Measured ΔV (m/s) 6.99 4.92 9.73 3.68 15.87 5.81
RICSAC 3 4 5
IC3 Vehicle 1 2 1 2 1 2
Measured ΔV (m/s) 4.15 6.83 8.37 9.93 7.29 11.35
RICSAC 11 12
IC4 Vehicle 1 2 1 2
Measured ΔV (m/s) 10.91 7.03 18.75 13.10
(a) (b)
Figure 5 Relative velocity change for cost functions 1 and 3 for vehicle (a) 1 and (b) 2
As far as the impact coefficients are concerned, in Figure 4, is obvious that only equivalent friction coefficient depends
on the cost function used. Both the restitution coefficient, and the moment coefficient of restitution demonstrate a
dependence on the test conditions. Furthermore, the moment coefficient of restitution in most tests obtains one of its
boundary values.
CONCLUSIONS
In the present paper the Planar Impact Mechanics collision model has been implemented in Matlab® and coupled
with the least squares method. Three different cost functions based on least squares method have been minimized
with the deterministic optimization method of SQP in order to calculate the velocity components of two vehicles in
collision and provide the impact coefficients which verify the PIM collision model. The presented methodology is a
way to utilize EDR data (Brach, 2011). Comparing three scenarios through different cost functions it was indicated
that the most reliable results where produced when estimates for all velocity components were available (FC3).
The values of absolute error of the velocity components are comparable for all RICSAC tests and all cost functions,
but are quite high. It is worth mentioning that a concern has risen in the past, with respect to the accuracy of the
velocity components documented in RICSAC reports. In 1997 a re-evaluation of the provided data has been
proposed, since it was realized that the measuring devices were not placed on the centre of mass of each vehicle
(McHenry and McHenry, 1997). Furthermore, in the reports of the RICSAC tests (Jones, 1978) it is acknowledged
that the value of the separation velocity in all tests was contaminated by the effects of rotation of the vehicles between
impact and separation. The abovementioned facts influence the performed analysis within this study, in terms of
relative values, since the re-evaluated values have not been taken under consideration.
Finally, as far as the impact coefficients are concerned, they were successfully computed regardless cost function,
with slight differences in most RICSAC tests for both the moment coefficient of restitution and the coefficient of
550
restitution. On the other hand each cost function leads to a different value for the friction coefficient, leading to the
conclusion that this is the more decisive impact coefficient.
REFERENCES
[1] Ishikawa, H.: "Impact model for accident reconstruction-normal and tangential restitution coefficients", (No.
930654) ", SAE Technical Paper, 1993,
[2] Brach, R., Brach, R.: "A review of impact models for vehicle collision", SAE Technical Paper, No. 870048, 1987,
[3] McHenry, R,: “CRASH3 Users Guide and Technical Manual” NHTSA, DOT Report HS, 805, 732, 1981,
[4] Brach, R.: “Least Squares Collision Reconstruction”, SAE Technical Paper, No. 870429, 1987,
[5] Brach, R.: "An impact moment coefficient for vehicle collision analysis", SAE Transactions, 1977, pp 30-37,
[6] Brach, R.: "Identification of vehicle and collision impact parameters from crash tests", Journal of vibration,
acoustics, stress, and reliability in design, Vol. 106, No. 2, 1984, pp 263-269,
[7] Brach, R.: "Momentum and Energy Analysis of Automobile Collisions", Trans. Structural Impact and
Crashworthiness, Elsevier Appl. Sci. Publ, 1984,
[8] Brach, R.: "Nonlinear Parameter Estimation of a Vehicle Collision Model", 13th Annual conference on Modeling
and Simulation, April 1982, University of Pittsburgh,
[9] McHenry, B., McHenry, R.: "RICSAC-97 A Re-evaluation of the Reference Set of Full Scale Crash Tests”, SAE
Technical Paper, No. 970961, 1997,
[10] Jones, I., Baum, A.: "Research Input for Computer Simulation of Automobile Collisions, Volume IV: Staged
Collision Reconstructions", DOT HS-805, 40, 1978,
[11] Struble, D.: "Automotive accident reconstruction: practices and principles", CRC Press, 2013.
551
552
Kragujevac, Serbia
MVM2018-039
Žarko Đorđević1
Nikola Kostić2
Jelena Erić Obućina3
Saša Babić5
CONTRIBUTION TO THE RESEARCH OF PARAMETERS,

EFFECTS AND METHODS OF ECO DRIVING
MEASUREMENTS
ABSTRACT: Eco-driving is a driving style that directly provides economical and environmentally friendly vehicle
exploitation, while indirect benefits could be seen through the slowdown in the exhaustion of fossil fuel reserves and
the reduction in the number of traffic accidents. Therefore, the effects of eco- driving are at the center of attention of
all participants in the transport system. However, the effects of eco-driving can be quantified by different
measurement methods, and are achieved by educating and motivating drivers to apply a certain eco-driving
style/mode. This paper presents the effects of eco-driving in real terms, as well as a comparison of the obtained
results and the compatibility of the applied methods.
KEY WORDS: eco-driving, parameters, effects and methods of eco driving measurement
INTRODUCTION
In road traffic, the following terms are increasingly being used: defensive driving, economical driving, ecological
driving, eco driving and environmental protection. The content and meaning of these terms are unknown to most
drivers, so we will briefly explain their content and meaning.
Defensive driving is a manner of managing a vehicle to avoid risky situations. The goal is to prevent entry to risk
situations and thus increase traffic safety. This is achieved by driving at moderate speeds and adapting to speed
conditions in traffic, avoiding sudden braking and acceleration. The result of this way of driving is reducing the number
of traffic accidents, fuel consumption and emissions of exhaust gases.
An economical driving is managing the vehicle with the least cost. The objective of economical driving is
rationalization-reducing fuel consumption, thereby reducing the cost of transport and emissions of harmful gases.
This is achieved by properly managing the fleet, driving with as high a gear as possible with lower revolutions and
avoiding sudden braking and acceleration.
Ecological driving is handling the vehicle with the least harmful effects on the environment. The goal of ecological
driving is the protection of the environment. This goal is achieved by reducing the consumption of fossil fuels and/or
1
2
Nikola Kostić, College of Tehnical and Mehanical Engineering Trstenik, kosticn83@gmail.com
3
JelenaErićObućina,College of Tehnical and Mehanical Engineering Trstenik, jericobucina@gmail.com
4
Branimir Milosavljević, Ph. D., Prof., College of Tehnical and Mehanical Engineering Trstenik,
lenka1809@gmail.com
5
Saša Babić, Ph. D., Prof., College of Tehnical and Mehanical Engineering Trstenik, babicsf@nadlanu.com
553
by using cleaner fuels of alternative propulsion (LPG, methane, hydrogen, bio fuel, electro-propulsion, etc.). In this
way, the emissions of harmful gases, and above all carbon dioxide (CO2), which is the greatest air pollutant, is
significantly reduced.
Eco-driving is a term used by some authors and cited in some sources of literature as a synonym for ecological
driving.
Environmental protection is an effort to make the road traffic (harmful to the environment), as much as possible less
polluting to the environment, such as: water, land and watercourses.
Taking into account the objectives and results of all these driving concepts, we conclude that their partial goals are
realized in different scientific areas (traffic safety, traffic economics, and ecology) and different methods, and that
their common/resultant goal is to reduce: fuel consumption, traffic accidents and emissions of harmful gases.
According to the foregoing, the term "eco-driving" is in the narrow sense synonymous with economical and ecological
driving, and in the broader sense both for defensive driving and environmental protection, because their goals and
results are the same. However, eco-driving is not only a reduction in fuel consumption, exhaust emissions and the
number of traffic accidents, but it is also the attitude, that is, the respect for the vital interests of the society as a
whole. Therefore, in this paper this term is used in its entire meaning.
The attitude of drivers towards eco-driving and their knowledge about this are extremely important, so driver training
in eco driving is imperative to the effects of eco driving. Drivers who know how to apply eco-driving generate
significant savings. The driver plays a key role in rational fuel consumption, as an educated and stimulated driver
maintains the optimal balance between: speed, driving time and fuel consumption. It is the transporter’s responsibility
to recognize the possibility of increasing profits, reducing costs, which can be achieved by eco-driving. Based on
past experience, it is possible to achieve savings of 5-15%, if the drivers have undergone training, and the transporter
establishes an efficient system of stimulation and control of fuel consumption [1]. Therefore, the United Nations
Commission for Europe helps promote the inclusion of the eco-driving school model into training programs for
professional drivers and their subsequent certification. Eco-driving significantly reduces CO2 emissions, making it an
important eco-driving segment to raise awareness about the environmental aspects of transport. The basic principle
of eco-driving is anticipation, that is, the attention and forecasting of the traffic situation, so that drivers avoid risky
situations (one of the principles of defensive driving).
ORGANIZATION AND REALIZATION OF ECO DRIVING TRAINING

Eco-driving positively influences the environment, traffic safety, financial and social aspects of the population, thus
stimulating a sustainable transport system.
The Ministry of Energy, Development and Environmental Protection of the Republic of Serbia and the City of
Belgrade, within the framework of the project "Support to sustainable transport in the City of Belgrade", implemented
the eco-driving training. The aim of the program was to create an initial group of eco-driving instructors, who in the
future would train drivers with basic and advanced eco-driving techniques for the needs of city institutions and public
services. This goal is determined by a multitude of challenges and attributes of modern traffic, on which sustainable
transport depends, as well as methods of training and measuring the effects of eco-driving. The success of training
and the quantification of eco-driving effects depend on the applied methods and the technical means used. The
training was realized with a number of drivers of UTC (Urban Traffic Transport) Belgrade and public utility companies
of Belgrade.
The training was realized in two ways. With the first group of drivers, using the verified IRU Academy method from
Geneva and electronic/instrumental measurement of eco-driving effects and with the other group using conservative
method and manual measurement of eco-driving effects.
Implementation of eco-driving training using IRU Academy verified method

The aim of the training is to train a group of at least 25 eco-driving instructors and 15 drivers, who will, at the end of
training with quantitative parameters, demonstrate that they have mastered eco-driving style.
According to the verified eco-driving program, the licensed eco-driving instructors of the RICo TC accredited training
institute IRU (International Road Transport Union) of the Academy from Geneva rented equipment from EKOmobilis
from Zagreb (Samobor), were trained.
554
This program combines modern technologies, advanced training techniques, monitoring tools and road safety rules.
Qualitative implementation of the program requires expertise, appropriate knowledge and measuring tools. The
instructors measured and evaluated the driving skills and driving style of each driver individually. Adequate measuring
tools track the progress of drivers in eco-driving training, which enabled quantitative measurement of fuel
consumption indicators to evaluate acquired knowledge in eco-driving.+
Reduction in fuel consumption after training represents a savings, which is a direct indicator of the effect of eco-
driving training. Savings, in the long run, can be achieved after a comprehensive investment in eco-driving, which
involves driver training in an eco-driving program and an adequate adaptation of eco-transport vehicles. According
to [2], all investments in eco-driving are returning in three years, with a fuel saving of only 5% per year.
The driver training program according to the verified eco-driving program begins with the assessment of driver's skills
prior to the training (first driving), their training in the eco-driving program and the establishment of the system of
tracking and control after training (second driving). It is especially important to transparently demonstrate to the
drivers the achieved eco-driving effects before driving, point them to the possible improvements and motivate them
to apply the learned (eco-style) on a daily basis, as it is most effective to prevent the return to the previous driving
style.
The eco-driving training program is based on a one-hour training session of 7 hours, which includes two driving in
public traffic, in real terms, on the same route, on a vehicle that the driver most often manages and the necessary
theoretical training in the classroom after the first drive.
In the first driving, the driver drives, without any preparation, the way he normally controls the vehicle on the selected
section of the road, where the start and end of the measurement is determined. The effects of the driver's style,
driving style and skills are measured and recorded automatically with dedicated equipment and software applications.
After the first driving, during the course of theoretical training, the obtained driving results are analyzed with each
driver and suggestions, guidance and instructions are given for the second driving, according to the principles of eco
driving. The second driving is realized on the same route, immediately after theoretical training, under the supervision
of the instructor, who directly monitors the work of the driver and, if necessary, gives specific instructions in order to
achieve the effects of eco driving.
At the end of training with drivers, comparative results of the first and second driving are analyzed and with their
participation a detailed analysis and assessment of the effects of eco-driving training is carried out and the guidelines
for further actions are given to them.
The training included 40 drivers, of which 25 should take on the role of an eco-driving instructor. Bus drivers operated
the MAN SG313 bus on a 15 km long route. The truck drivers operated the DAF LF 55.220 and the Opel Astra 1.7
CDTI passenger vehicle on a 25 km long track (Table 1). Both drives were carried out on the same vehicle and the
same route, but at different times (the first one in the morning hours, and the second in the afternoon hours). On
selected routes there were climbs, downhills, a large number of traffic lights and streets with a high intensity of traffic.
Table 1 Fuel consumption and driving times before and after electronic measurement training
Average consumption Average driving
Number Difference
Ord. L/100km time (min:sec) Difference
Vehicle type of
No. First Second Liter First Second (sec)
drivers %
driving driving (L) driving driving
1. Solaris bus 27 55.07 50.40 -4,67 -8.48 27:40 27:43 +3
Freight vehicle
2. 5 24.28 22.58 -1,70 -7.00 29:37 28:38 -59
DАF LF 55 220
A passenger
3. 8 7.86 6.84 -1,03 -12.98 28:58 28:00 -58
vehicle
Measuring instruments and measurement methods

The application that was used to measure the training result with the first group of drivers was developed by the
company "Key Driving Competencies"6. It provides an objective assessment of driving skills based on eco-driving
principles in real time - while driving.
This application is connected to the computer with the CAN interface, which enables the display of data measured
during driving and prepared for analysis, in order to properly assess the driver's driving skills.
3
http://www.eco-drivingxl.be/
555
The electronic measurement equipment consists of: laptop, software application and diagnostics and connection
cables. During driving on the laptop screen, the current parameters are shown in real time (Figure 1).
Figure 1 Eco-driving parameters of a particular driver in real-time obtained using software application
On the left side of the screen, 4 active diagrams are shown from top to bottom, showing the parameters of the quality
of vehicle management in the last 30 sec. The abscissa of the diagram shows the time in seconds (30 sec.), and on
the ordinate the parameters of the quality of the vehicle's management:
1. Diagram of engine revolution: shows the number of engine revolutions in the last 30 sec;
2. Vehicle speed diagram: shows the speed of the vehicle in km/h in the last 30 sec;
3. Diagram position of the pedal of power (gas pedal in %) and clutch;
4. Fuel consumption diagram: shows fuel consumption in milliliters in the last 30 sec.
At the center of the screen there is a turntable, under which the name and surname of the driver is entered, the
parameters of which are controlled by the vehicle, during driving. In the lower center of the screen (from left to right):
1. Driving: total driving time (upper cursor), total travel time (lower cursor);
2. Driving without gas: total time without using gas (upper cursor), total length the route passed without the use
of gas (lower cursor);
3. Braking: total driving time using the brake pedal (upper cursor), total length of the route crossed in braking
(lower cursor).
On the right side of the screen, from the top to the bottom are shown:
1. Fuel consumption: consumption in L/h (standstill) or L/100 km (in motion);
2. Gear degree: Current gear;
3. The average driving speed: in km/h;
4. Fuel consumption: total fuel consumption of one driver during one driving in L;
5. The average fuel consumption of one driver during one driving: L/100 km;
6. Stopping: number of vehicle stops;
7. Braking: the number of braking (using the brake pedal) of one driver while driving;
8. Time in motion: displayed in minutes and seconds.
A report for each driver is shown in the table (Figure 2). Two types of data are displayed in the working part of the
table. In the upper part, data for the 9 measured eco-driving parameters are given, and the bottom line shows the
analysis of the results through 13 eco-driving indicators. The table shows the comparative data for the first and
second ride, the nominal difference of measured values of the first and second driving and the difference in percent.
In addition to the tabular display, the software application displays 4 histograms on which the measured first and
second driving parameters are compared: total consumption, average consumption, distance travelled and driving
time, in function of the gear (Figure 3).
556
KEY DRIVING TRAINING SYSTEM
Eco proactive behavior of drivers: ,,What you can not measure you can not manage“
Driver Date
07/28/2013
RESULTS
1. Driving 2. Driving Difference %
Total time H:mm:ss 00:24:17 00:25:49 01:32 6.29
Total distance km 14.40 14.40 0.00 0.00
Average speed km/h 35.57 33.47 -2.10 -5,90
Average speed in motion km/h 37.84 35.10 -2.74 -7.24
Fuel consumption in standstill l 0.17 0.21 0.04 19.05
Fuel consumption in motion l 8.91 7.43 -1.48 -16.61
Total fuel consumption l 9.08 7.63 -1.45 -15.97
Average consumption l/100 km 63.1 53 -10.1 -16.01
Average CO2 emission kg/100 km 167.9 141.0 -26.9 -16.02
RESUTS ANALYSIS
Average position of gas pedal % 36 29 -7 -19,44
Maximum position of gas pedal % 100 100 0 0
Driving time without gas in motion mm:ss 06:45 06:32 0:13 3.38
Time using brakes mm:ss 02:50 01:24 01:26 50.77
Total distance without gas km 4.52 3.68 -0.84 -18.58
Total distance using brake km 0.48 0.02 -0.46 -95.83
Breaking number * 49 7 -42 -85.71
Number of stoppings * 6 3 -3 -50.00
Time in standstill mm:ss 01:27 01:12 00:15 17.59
Number of gear changes * 101 86 -15 -14.85
Number of gear changes to higher * 57 44 -13 -22.81
Total engine revolution o/min 27094 27600 506 1,87
Average engine revolution o/min 1110 1064 -46 -4,14
Figure 2 Comparative overview of parameters and analysis of results in first and second driving
Total consumption / Average consumption / Gear
Distance travelled/Gear Driving time/ Gear
Figure 3 Graphic display of the parameters of the first and second driving in function of the gear degree
Realization of eco-driving training with a conservative method

In the absence of measuring equipment, with the other group of 46 drivers, eco-driving training was conducted, using
a conservative method, of which 34 with freight vehicles, and 12 with buses (Table 2). The training was carried out
557
according to the same principle as with the previous group of drivers, while the measurement of fuel consumption
was done manually - according to the principle of full tank.
Table 2 The parameters of fuel consumption and driving times before and after training obtained by manual
measurement
Average consumption Average driving time
Difference
Ord. Number L/100km (min:sec) Differen.
Vehicle type
No. of drivers First Second Litre First Second (sec)
%
driving driving (L) driving driving
1. Solaris bus 12 43.89 43.49 -0.4 -0.92 26:35 27:55 +80

Freight vehicle
2. DАF LF 55 220 19 23.95 19.83 -4.12 - 17.19 33:32 33:16 -16
Freight vehicle
3. DАF LF 55 180 15 16.01 13.23 -2.78 -17.40 33:52 33:44 -8
Each driver after the first and second driving, under the same conditions (the same pump), replenished the tank on
the pump. The filled amount of fuel, according to this method of measurement, is a quantitative indicator of fuel
consumption in given conditions, and hence the effect of eco driving.
Bearing in mind the vehicles with which training was conducted using one method and the other (the Solaris bus and
the truck DAF LF 55.220), it is possible to compare the results of the training of 27 bus drivers from the first group
(Table 1) with 12 drivers from the second group (Table 2) and 5 truck drivers from the first group (Table 1) with 19
drivers from the second group (Table 2).
ANALYSIS OF TRAINING RESULTS

Measured values of average fuel consumption among all drivers in both measurement methods indicate that in the
second driving, after training, fuel savings were achieved (Tables 1 and 2). This is the proof that the effective effects
of eco-driving can be achieved quickly, after training in real-life conditions in a relatively short period of time. The
highest savings after training were recorded by drivers of passenger vehicles (12.63%), according to the method of
electronic measurement (Table 1), and by the manual method (17.40%) the drivers of the truck DAF LF 55 180 (Table
2).
By means of a manual measurement method (Table 2), before training bus drivers spent 43.89 L/100 km on average,
and in the second driving, after training, 43.49 L/100 km. The savings are at least 0.92%, but we must have in mind
the imprecision of the measurement by this method, since the drivers after each ride have filled in the consumed
amount of fuel, and this is taken as the proper consumption. When filling the vehicle with fuels for various reasons
(inclination, vehicle position, fuel spillage in the reservoir, foam phenomenon, temperature), the tank cannot always
be filled identically, showing the imprecision of this method. The results obtained by this method in principle show
the possibility of fuel savings, so they can be accepted as a qualitative indicator, and for the quantification of savings,
more precise measurement is needed.
By means of manual method of measurement, before training the drivers of freight vehicles spent 23.95 L/100 km on
average before training; i.e. 16.01 L/100 km, and in the second driving, after training 19.83 L/100 km; or 13.23 L/100
km, which means that they achieved savings of 17.19%; or 17.40% of fuel consumption (Table 2). The savings are
significant, but one must certainly bear in mind the inaccuracy of measuring the amount of fuel consumed by filling
up to the full tank.
Using the nonparametric U test on the significance of the difference in the fuel consumption of the bus driver in the
first and second driving, based on the obtained results, using manual measurement can be concluded that there is
no significant difference in fuel consumption after eco driving because p> 0.05. This may mean that the measurement
by this method was imprecise, or the bus drivers did not master eco-driving, since it has been proven in practice that
eco driving has a significant impact on fuel consumption, and this test would have demonstrate it.
The difference in DAF vehicle consumption was tested using the same test. Based on the results of the test with a
probability of 95% it can be confirmed that there is a significant difference in fuel consumption in the first and second
driving because p <0.05, which would mean that fuel consumption depends on the applied eco-driving style, that is,
that the drivers of freight vehicles mastered eco-driving.
In the second driving, according to manual measurements, bus drivers spent 80 sec more for the same route (Table
2), which means that the speed after eco driving training was slightly lower.
558
Bus drivers were on average slower in the second driving for 80 sec. Of the 12 tested drivers, 5 of them in the second
driving, were driving more slowly for 264 sec (total 22 min), while three drivers in the second driving crossed the
same section faster for 120 sec (total of 6 min). Four drivers were driving at the same speed the vehicles both in the
first and the second driving.
In the second driving, freight drivers spent slightly less time for the same route, for 16 sec and 8 sec, so it can be
concluded that the speed in the second runwas slightly higher (Table 2).
The drivers of the DAF LF 55 220 truck were on average 16 sec faster in the second driving. Of the 19 tested drivers,
7 were driving more slowly in the second driving for 111 sec (total 13 min), while the 8 drivers in the second driving
were driving the vehicles faster for 135 sec (total 18 min). Four drivers were driving at the same speed both in the
first and second driving on the same section. The maximum driving time on the observed section was 40 min in the
first run, and 36 min in the second. The minimum driving time on the same section is 25 min in the first driving and
26 min in the second.
According to electronic measurements, the difference in average fuel consumption of the bus driver before and after
training, (Table 1), was tested with a T-test (two small dependent samples). The value of the obtained probability is
p = 0.0005 and is significant p <0.05, so the hypothesis is accepted with great probability that training in eco driving
has significantly influenced the reduction in the average fuel consumption of the bus drivers. The interquartile
distance in the first driving was 16.3 L, and in the second 18.8 L, the difference between the two driving had an
interquartile distance of 19.8%. Using the same test the differences for passenger car drivers were also tested, p =
0.0031. At a risk of 5% (p <0.05), it can be concluded that even the drivers of the passenger vehicles achieved
savings in the second driving, after eco-driving training. The number of truck drivers is small, so a reliable
measurement should be done with a larger number of drivers. All five drivers achieved savings in the second driving
on average of 1.70 L or 6.92%, but the number of samples does not give entitle to conclude that the differences are
not random (p = 0.0676) with a lower risk of 5%.
The first driving was realized from 09:00 to 12:00 am and the second from 15:00 to 18:00 pm. In the second driving,
the traffic was significantly of higher intensity, with a greater number of more complex traffic situations compared to
the first driving, which had a negative impact on average fuel consumption and possible savings, thereby masking
the contribution of training to the effects of eco driving.
According to electronically measured results, there are no statistically significant differences in driving time, for both
groups of drivers, for the first and second driving. The drivers of freight vehicles and the drivers of passenger vehicles
in the second driving were on average faster in less than 1 min.
On average, bus drivers were slightly slower in the second driving (on average in 3 sec).
Of the 40 tested drivers, 1/2 was driving the same route section in the second run faster in about 2 min on average
(121 sec), while the rest were slower on average for about 102 sec compared to the first driving. Only 1 truck driver
was slower in the second driving (63 sec), and the other four were faster on average for 88 sec. The five drivers of
the passenger vehicles were faster in the second driving on average for 116 sec, and the three were slower on
average for 38 sec. A larger number of bus drivers was slower in the second driving, 16 of them on average for about
116 sec, and those who were faster were on average faster over 2 min (136 sec). According to the above stated,
temporal analysis shows that eco-driving is not a "slow" driving, as a significant number of drivers think.
Two months after the training, control measurements were carried out with 10 bus drivers and 2 truck drivers, which
included a control group, which in terms of average consumption after training made significant savings of 7.86%
(Figure 4). Previous experiences have shown that most drivers, after eco-driving training, have a gradual return to
previously adopted habits and old driving patterns, and in this case, the trend is reversed. The application of the
learned has not been recalled but has been improved.
Figure 4 Reduction of average fuel consumption after training was measured electronically
559
The results of the control group give the right to claim that the tested drivers applied novel driving skills to the eco-
driving style and improved them, as they achieved savings of 7.86% compared to the after-drive training. That this
difference is not accidental confirms the T-test, in which the probability p = 0.027 (p <0.05) was obtained. During the
testing of the control group, on the section of the road where the measurement was carried out, a traffic accident
occurred, which significantly slowed down the traffic and extended the driving time for one driver, and thus increased
the average consumption by 8.3%, which had a negative effect on the overall results of the control group, but not on
the conclusion on the significance of the difference in fuel consumption in the first and second driving.
CONCLUSION
Drivers dominantly influence the reduction of fuel consumption, that is, eco-driving, and other transport system
operators can improve or reduce eco-driving in road traffic. The achieved fuel savings for all vehicles, after training
are in empirical limits of the results, achieved in Europe [1] and amount to about 9% (8.92), (Figure 4).
The full effects of eco-driving can be achieved after training using the verified method of by the IRU Academy in
Geneva using the appropriate teaching equipment and measuring equipment. Eco-driving, in addition to lower fuel
consumption, reduces CO2 emissions as a more comfortable driving with less stress and aggressiveness is achieved,
less risky situations, with strict compliance with regulations and saving money.
All trained drivers achieved remarkable savings in the second driving, with differences in driving time not significant.
Driving times, when drivers were driving more fuel efficiently, in more complex traffic situations, show that economical
driving is not a "slow" ride, as a significant number of drivers think. Contrary to expectations and previous experience,
30% of trained drivers have shown at the control measurements that they not only retained but also improved the
eco-style of driving.
Quality measuring equipment and software applications enabled objective measurement and complete analysis of
the measured parameters. The advantage of this analysis/method is the possibility of objectively measuring the
analyzed eco-driving parameters. Comparability of data is possible because both evaluations were carried out on the
same vehicle, in the same way on the same itinerary. Precisely derived data gives us a true picture of the current
state and provide the basis for possible system improvements and rational fuel consumption, subject to monitoring
and evaluation of eco driving.
This research has shown that, in the event of a lack of measuring equipment, driver training in eco-driving can be
done with a conservative method, with more or less success. With this method, special attention should be paid to
the way of fueling up to the full tank, after the first and second driving, because one can be significantly wrong in
estimating fuel consumption. Since fuel consumption (total and average) is the most important parameter of the
vehicle's exploitation, in the event of a gross error, the conclusion about the effects of eco driving cannot be reliably
derived. Namely, in this case, after the second drive, the fuel saving principle cannot be reliably determined.
The disadvantages of this analysis stem from the fact that it was created only on the basis of a short course with a
very small sample of drivers compared to the general population. Objective measurement and quick success rate
greatly influence driver motivation for further work on improving knowledge and eco driving skills.
REFERENCES
[1] Мустапић, Н., Шимунић, Н. Шимунић, Д.: "Еко-вожња као важна мјера реализације енeргетске
учинковитости и смањења емисије стакленичких плинова у сектору транспорта", Зборник радова,
Копаоник, 2014,
[2] ECOEFFECT.ORG, Available from: www.ECOeffect.org, Accessed 20.08.2014.,
[3] Rico Training, Available: from:www.ricotrainingcentre.co.rs, Accessed 22.08.2014.,
[4] IRU, Available from:www.iru.org/sites/default/files/2016-07/en-eco-driving-train-the-trainer-course-outline.pdf,
Accessed 01.08.2018.,
[5] IRU, Available from:www.iru.org/system/files/IRU%20Academy%20FS%20ECO%20Driving%20EN.pdf,
Accessed 01.08.2018.,
[6] EKOmobilis, Available: from:www.ekomobilis.hr, Accessed21.07.2018.,
[7] Eco-drivingxl, Available: from:www.eco-drivingxl.be/, Accessed04.09.2014.
560
Kragujevac, Serbia
MVM2018-041
Goran Pejić1
Marko Bunjevac2
Matija Pečet3
Zoran Lulić4
IMPACT OF INTRODUCTION OF LOW EMISSION ZONES IN

THE CITY OF ZAGREB
ABSTRACT: Despite significant improvements in vehicle technology in recent years, vehicle emissions remain a
major influencing factor on air quality. Although air quality in the City of Zagreb is still relatively acceptable, official
standards, i.e. limiting values are under threat. More than ever, policymakers are actively exploring options for new
vehicle emission control and local fleet management policies, as well as the new monitoring technologies to
support these activities.
The goal of the research was to assess the influence of the introduction of Low Emission Zone on the total
emission from the road traffic in the City of Zagreb. To make a model of emissions from road traffic, emissions on
selected roads in the City of Zagreb using measured traffic activity data and adequate vehicle fleet characteristics
description are calculated. Traffic activity data were obtained from traffic study of the City of Zagreb, while vehicle
fleet characteristics are extracted from the database maintained by Center for Vehicles of Croatia and the Ministry
of the Interior of the Republic of Croatia. The database contains all the data collected during the homologation
process of the vehicles registered in Croatia as well as the data collected during the vehicle periodical technical
inspection. It is important to note that the database contains exact data on the date, time and site of periodical
technical inspection.
Fleet data for vehicles registered in Zagreb was compared with the fleet data for the entire Republic of Croatia. The
comparison shows the sensitivity regarding the fleet characteristics and that it is possible to improve air quality with
simple regulation, i.e. introduction of the Low Emission Zone.
KEYWORDS: vehicle emissions, air quality, low emission zone, road traffic, vehicle fleet
INTRODUCTION
According to the European Environment Agency (EEA), around 30% of total carbon dioxide emissions comes from
traffic-related activities [1]. Air pollution represents a significant problem for environment and human health and
traffic-related activities are major contributing factor. Harmful emissions (tailpipe gases and particulate matter)
which occur after fuel combustion in vehicles internal combustion engines cause respiratory, cardiovascular and
neurological diseases [2]. Beside the air pollution, studies have shown that traffic-related pollution is the reason for
the increase in global mean surface temperatures. For example, simulations performed in [3] suggest that German
transport emissions induce an increase in the global mean surface temperature of the order of 10 mK during the
21st century.
1
Mr. sc. Goran Pejić, Centre for vehicles of Croatia, Capraška 6, HR-10000, Zagreb, Croatia, goran.pejic@cvh.hr
2
Marko Bunjevac, bacc. ing. mech., student, Faculty of Mechanical Engineering and Naval Architecture, University
of Zagreb, mb196921@stud.fsb.hr
3
Matija Pečet, bacc. ing. mech., student, Faculty of Mechanical Engineering and Naval Architecture, University of
Zagreb, mp195277@stud.fsb.hr
4
prof. dr. sc. Zoran Lulić, professor, Faculty of Mechanical Engineering and Naval Architecture, University of
Zagreb, zoran.lulic@fsb.hr
561
With the goal of reducing air pollution, the European Parliament and Council have adopted Directive 2008/50/EC
on ambient air quality and cleaner air for Europe. The Directive 2008/50/EC states that the public has a right to
demand from local authorities to provide a quality environment for life, meaning that authorities should ensure that
the air pollution is within permissible limits, which are set within the Directive [4].
Main products of combustion within internal combustion engine can be split into two distinctive groups, those which
are harmful to human health and those who are not. Carbon dioxide (CO2) and water vapour (H2O) are not harmful
to human health, but are closely related to the Greenhouse Gasses and global warming, while carbon oxide (CO),
nitrogen oxides (NOx) and particulate matter (PM) are. Today, major attention is given to the reduction of GHG and
minimisation of those which are harmful, especially nitrogen oxides and particular matter.
Emission of nitrogen oxides (NOx) occurs in an engine cylinder by oxidation of atmospheric nitrogen typically at
temperatures higher than 1,800 K [5]. They can further be divided into nitrogen monoxide (NO) and nitrogen
dioxide (NO2). Nitrogen monoxide converts to nitrogen dioxide in the atmosphere which can later form acid rain.
Furthermore, nitrogen oxide can contribute to forming low-level ozone which is also harmful to human health.
Alongside with harmful influence on the environment, nitrogen oxides have a negative influence on human health,
causing eye and lung irritation, fatigue and nausea. With long-term exposure, there is an increased chance of
asthma occurrence and permanent lung and nervous system damage, and may even cause genetic disorders.
Nitrogen oxides, like the carbon oxide (CO), can also prevent the normal flow of oxygen throughout the human
blood system.
Particulate matter (PM) mostly consists of soot. According to its average diameter, it is divided into PM10 and
PM2.5. PM10 represents particles whose medium diameter is smaller than 10 micrometres, and PM2.5 represent
particles with a medium diameter smaller than 2.5 micrometres. Such fine particles are severely hazardous to
human health because they can infiltrate human respiratory and cardiovascular system [6]. It is proven that
particulate matter increases mortality by causing chronic obstructive pulmonary disease (COPD), ischemic heart
disease and lung cancer [7].
Low Emission Zone (LEZ)

In its wider meaning, Low Emission Zones (LEZ) represent areas in which traffic is regulated according to vehicle
category or emission standard. Restriction implies fully banning certain vehicles or charging their entrance into the
zone. LEZ can include from a few main roads up to wide city areas. They are implemented in the areas where air
pollution represents a significant threat to human health. Since the grates threat is from particulate matter and
nitrogen oxides which are common for diesel-powered vehicles, LEZ often has stricter regulations for those types
of vehicles. Studies have shown that heavy duty vehicles (HDV) which commonly run on diesel fuel have up to 30
times greater emission of PM2.5 and up to 26 times greater emission of nitrogen oxides (NOX) than light-duty
vehicles (LDV) [8]. Therefore, those kinds of vehicles are often not allowed to enter LEZ, or their entrance is
regulated in some way.
World Health Organisation (WHO) states that air pollution is responsible for the premature death of nearly 310,000
people in Europe which is more than deaths involved in traffic accidents. Besides air pollution having an adverse
impact on human health, it consequently causes great financial losses. Cost of treatment for people with air
pollution-related diseases is estimated somewhere between 427 and 790 billion EUR annually [9].
As mentioned earlier, a great number of LEZs ban or regulate the entrance of heavy-duty vehicles with great
emissions, while some even ban or regulate the entrance of coaches, passenger vehicles and motorcycles. Most of
LEZs are active throughout the entire year, while other are active only occasionally, e.g. in the rush hour, in winter
or in summer. Currently, there are 258 Low Emission Zone in 23 European countries with over 230 major access
regulations and 14 charging schemes [9].
Air Quality in the City of Zagreb

In Zagreb, capital of the Republic of Croatia, there are six stations for continuous measurement and monitoring of
air quality. Locations of the station can be seen in Figure 2. The measuring station Zagreb 1 is very interesting
because it is located near to the crossing of Vukovarska and Miramarska Street which represents crossing with one
of the densest traffic in the city because it is located near the centre of the city.
By using the developed application, the data from all measurements stations in Zagreb have been collected and
analysed from January 1st 2015 to December 31st 2015. The greatest attention was given to the following
pollutants: nitrogen oxides (NOx) and particulate matter (PM). A limit value for PM2.5 according to Air Quality
Standards is 25 μg/m3, while the PM10 is defined with the limit value of 50 μg/m3, and the permitted exceedances
each year is 35 times.
562
Green dots represent values below the limit, red dots represent values over the limit.
Figure 1 Hourly exceedances (shown as red dots) of PM in measuring station Zagreb 1 (left) and exceedances of
NOX in measuring station Zagreb 3 (right) during 2015 [10]
As graphically represented in Figure 1, analysis has shown that the limit for PM10 is exceeded more than 35 times,
but it is also evident that pollutant concentrations of PM are lower in those months when Zagreb’s inhabitants are
using their vacations [10].
METHOD
It is well known that for calculation of vehicle emissions it is necessary to know vehicle emission factors and vehicle
activity. For the purpose of emission calculation in this study, the emission factor implemented in the COPERT:
Street Level were used, and they can be seen in the Figures 6, 7, 8 and 9. Vehicle activity was recorded on
multiple sites throughout the City of Zagreb as shown in Figure 2.
Figure 2 Location of traffic flow recording sites and stations for air quality measurement in City of Zagreb
563
Traffic Flow
Traffic flow is a simultaneous movement of a number of vehicles on a certain road segment in some ordered way.
To describe traffic flow, it is necessary to define the basic traffic flow parameters. For the determination of these
parameters, traffic was recorded using a GoPro Hero camera which has the ability to capture 50 frames per second
at 2704x1520p (2.7k) resolution. Purpose of this camera was to accurately determine the exact moment when a
vehicle passing across the detection line, as it is shown in Figure 3. Traffic flow recording locations are marked with
red squares in Figure 2.
Figure 3 Example of observed road segment with detection lines (left - scheme, right - real situation) [11]
Basic traffic flow parameters are:

 Traffic flow –number of vehicles passing through the observed cross-section of the road in the unit of time,
qveh [veh/h];
 Traffic flow density – number of vehicles in the observed traffic lane per length unit, g [veh/km];
 Traffic flow speed – median value of speed from all vehicles involved in the traffic flow, vtf [km/h];
 Vehicle passing interval – time between the passing of two consecutive vehicles through the observed
cross-section on the road, th [s].
For the purpose of emissions calculation, it was necessary to determine traffic flow and its speed. This was done by
analysis of recorded videos of traffic. For each vehicle passing through detection line, the exact frame number of
passing through first and second line was marked. By knowing the distance between detection lines (17 m), the
exact frame of crossing and the camera frame rate it was possible to determine vehicle passing time with the
equation 1.
f L 2  f L1
th , x  (1)
50
Where:
Δth,x - vehicle x passing time, s;
fL1 - frame number of front axle passing over the first detection line;
fL2 - frame number of front axle passing over the second detection line
After calculation of vehicle passing time, it was possible to determine vehicle speed according to equation 2.
∆𝑠
𝑣𝑥 = (2)
∆𝑡ℎ,𝑥
Where:
vx - vehicle x speed, m/s;
Δth,x - vehicle x passing time, s;
Δs - distance between detection lines (17 m).
Traffic flows were recorded in February and March 2017 at eight locations on working days. Most of the city roads
have similar peak hour interval as shown in Figure 4, so data was analysed for the period from 16:15 to 17:15.
Figure 4 shows hourly traffic flow during the whole week at the intersection of Ljubljanska Avenue and Puljska
Street and afternoon peak hour can be clearly seen.
564
3500
Number of vehicles per hour, veh/h

3000
2500
2000
1500
1000
500
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hours
Sun Mon Tue Wen Thu Fri Sat
Figure 4 Example of weekly traffic flow
As it is earlier stated, for emission calculation it is important to define the traffic flow speed. After calculating the
speed for all vehicles, the median value was used. The median value represents value separating higher and lower
half of all calculated speeds. The median value is used instead of the arithmetic mean value because it is less
sensitive to extreme value oscillations. Figure 5 shows an example of the median value of speed at Slavonska
Avenue 1.
120
110
100
90
Speed vp, km/h
80
70
60
50
40
1 101 201 301 401 501 601 701 801 901 1001 1101
Vehicle number
Figure 5 Example of the median value of speed at Slavonska Avenue 1
Vehicle fleet
For the purpose of calculating emission of harmful substances, it is necessary to determine vehicle fleet on the
observed roads. One of the ways would be to record every vehicle registration plate and verify vehicle
characteristics through the database of the Ministry of Interior, which is very slow and time-consuming process. For
that type of vehicle fleet recognition, it would be better to use the specialised camera for number plates recognition.
Other way, used for the purpose of this research, is to use predefined vehicle fleet composed of all the vehicles
that had an active periodic technical inspection (PTI) in the City of Zagreb on the date of January 1st 2018. This
way vehicle fleet definition was much easier and quicker, while still precise enough for the calculation of roadside
emission. By using this method, the majority of the vehicles registered in the City of Zagreb that are not used in the
city are left out and vice versa. Fleet is structured from all the vehicles from M1 and N1 category as they represent
more than 90% of all vehicles entering the city. In addition, all the vehicles of M1 and N1 category which use
compressed natural gas (CNG) or liquefied petroleum gas (LPG) are not included in the fleet mostly because they
represent a small fraction, but also because it was not possible to precisely define their emission factor.
565
Table 1 Number of vehicles that had an active periodic technical inspection (PTI) in the City of Zagreb on the date
of January 1st 2018 according to the emission standard
Category Fuel type EURO 0&1 EURO 2 EURO 3 EURO 4 EURO 5 EURO 6
PETROL 9,390 25,856 28,840 44,518 20,730 13,872

CNG (total) 13
M1
DIESEL 6,526 12,293 34010 44,680 43,445 28,316
LPG (total) 11,908
M1 (total) / 27,897 38,149 62850 89,198 64,175 42,188
PETROL 41 57 128 171 239 120
CNG (total) 32
N1
DIESEL 778 1,716 3,895 6,716 8,893 2,939
LPG (total) 224
N1 (total) / 1,115 1,773 4,023 6,887 9,132 3,059
As it is shown in Table 1, there are 324,457 vehicles of M1 category in everyday traffic across the City of Zagreb. It
can be clearly seen that diesel vehicles are most represented type with 52.17%, gasoline vehicles are following
with 44.14%, vehicles powered by LPG are present with 3.67% while CNG vehicle holds a neglectable share of the
M1 fleet.
Following the M1 category, there is N1 category with 25,989 vehicles operating across the City of Zagreb. Again,
diesel vehicles are the most represented type with 95.95% followed by gasoline vehicles with only 2.91%. Vehicles
of N1 category powered by liquefied petroleum gas and compressed natural gas hold 1.14% combined.
Overall, it can be assumed that there are around 350,446 vehicles involved in everyday traffic across the City of
Zagreb. Average vehicle age of these vehicles is over 10 years. Disregarding vehicle powered by LPG and CNG
and looking by emission standard it can be seen that Euro 0, Euro 1 and Euro 2 represent high 16.75% and if we
add up Euro 3 vehicles, number crosses over one-third of entire vehicle fleet with 36.52%, which is extremely high.
Emission Calculation
After traffic flow calculation and vehicle fleet description, the last step in emission calculation is the definition of
emission factors. Emission factors need to be from reliable source, e.g. internal combustion engine manufacturer,
environmental agency etc. Emission factors used for this research are compiled from emission factors from
COPERT: Street Level, a commercial software package used for calculation of emission on street level [12].
Emission factor represents a functional link between vehicle fleet characteristics, vehicle activity and emission.
Emission factors of particulate matter and nitrogen oxides can be seen in Figure 6, 7, 8 and 9.
0,30 2,1 4,5
1,8 4,0
0,25
Emission factor PM, mg/km
Emission factor NOX, g/km
3,5
1,5
0,20 3,0
1,2 2,5
0,15
0,9 2,0
0,10 1,5
0,6
1,0
0,05 0,3
0,5
0,00 0,0 0,0

0 10 20 30 40 50 60 70 80 90 100 110 120 130 0 10 20 30 40 50 60 70 80 90 100 110 120 130
Speed, km/h Speed, km/h
Euro 2 Euro 3 Euro 4 Euro 5 Euro 6 Euro 0* Euro 1* Euro 0 Euro 1 Euro 2 Euro 3 Euro 4 Euro 5 Euro 6
Figure 6 Emission factors for NOX (left) and PM (right) for M1 vehicle category (petrol) - * secondary axis
566
1,4 0,40
0,35
1,2
Emission factor PM, g/km

0,30
1,0
0,25
0,8
0,20
0,6
0,15
0,4
0,10
0,2
0,05
0,0
0,00
0 10 20 30 40 50 60 70 80 90 100 110 120 130
0 10 20 30 40 50 60 70 80 90 100 110 120 130
Euro 0 Euro 1 Euro 2 Euro 3 Euro 4 Euro 5 Euro 6 Euro 0 Euro 1 Euro 2 Euro 3 Euro 4 Euro 5 Euro 6
Figure 7 Emission factors for NOX (left) and PM (right) for M1 vehicle category (diesel)
0,8 8,0
0,7 Emission factor PM, mg/km 7,0

0,6 6,0
0,5 5,0
0,4 4,0
0,3 3,0
0,2 2,0
0,1 1,0
0,0 0,0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 0 10 20 30 40 50 60 70 80 90 100 110 120 130
Euro 1 Euro 2 Euro 3 Euro 4 Euro 5 Euro 6 Euro 1 Euro 2 Euro 3 Euro 4 Euro 5 Euro 6
Figure 8 Emission factors for NOX (left) and PM (right) for N1 vehicle category (petrol)
2,0 1,0
1,8 0,9
1,6 0,8
Emission factor PM, g/km
1,4 0,7
1,2 0,6
1,0 0,5
0,8 0,4
0,6 0,3
0,4 0,2
0,2 0,1
0,0 0,0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 0 10 20 30 40 50 60 70 80 90 100 110 120 130
Euro 1 Euro 2 Euro 3 Euro 4 Euro 5 Euro 6 Euro 1 Euro 2 Euro 3 Euro 4 Euro 5 Euro 6
Figure 9 Emission factors for NOX (left) and PM (right) for N1 vehicle category (diesel)
Vehicle activity is represented through vehicle fleet characteristics, traffic flow, length of a road segment and
vehicle speed. The basic equation for calculation of roadside emission is:
Emission = Vehicle activity(speed, type, number, ...) x Emission factor (3)
567
By implementing this method, it is possible to calculate roadside emission from traffic in the specific time period on
the observed road segment. In this research, the emission is calculated for eight observed road segments.
For emission calculation, according to this method, it was necessary to define overall vehicle number and divide
them according to fleet characteristics, add travelled distance and median speeds to all vehicles and allocate them
their unique emission factor as shown in Equation 4.
n
 g 
Ep  g      EF (EURO, v)  km   s km
p (4)
FUEL EURO j 1
Where:
Ep - Overall emission of a specific pollutant,
FUEL - Fuel type (petrol or diesel),
EURO - Emission standard,
n - Vehicle number,
EFp - Emission factor of specific pollutant as a function of emission standard and vehicle speed,
p - Specific pollutant (NOX or PM),
v - Vehicle speed,
s - Travelled distance.
Scenarios
Before the development of different scenarios, the base case (Scenario 0) was calculated and analysed for before
mentioned composition of the vehicle fleet. After the base case was established, several other scenarios were
developed. Scenario 1 represents LEZ with the ban of all Euro 2 or lower emission standard vehicles with petrol
engines and all Euro 3 or lower emission standard vehicles with diesel engines and the same goes for Scenario 2.
The difference is that Scenario 1 represents a condition in which total traffic flow vehicle number is lowered for the
share of banned vehicles which in return represents starting state of LEZ implementation. In other hand, Scenario 2
represents a state where the number of all vehicles banned by LEZ implementation is returned to the normal state
by other vehicles which in return shows long-term effect.
Table 2 Vehicles permitted to enter LEZ by emission standard and category

Vehicle category
Scenario Fuel type Euro 0 Euro 1 Euro 2 Euro 3 Euro 4 Euro 5 Euro 6
M1 N1
Petrol         
Scenario
1&2
Diesel         
Petrol         
Scenario 3
Diesel         
Scenario 3 offers an insight into emission reduction if all M1 category vehicles were allowed to enter the LEZ, while
entire N1 category was left out. This type of LEZ would rather be used as a measure against traffic congestions
rather than high emissions, but it is useful to show how it would affect air quality.
RESULTS
The busiest street during the peak hour with 2296 vehicles is Hrvatske Bratske Zajednice Street and is closely
followed by Selska Street and Slavonska and Ljubljanska Avenue. Miramarska Street has least vehicle passages
during the peak hour which is slightly under 1600 vehicles. On the other hand, Zagreb bypass has the fastest
moving traffic with a median speed of 93 km/h, while the slowest traffic can be observed on Slavonska Avenue if a
traffic jam occurs. If there is no traffic congestion, Miramarska Street traffic becomes the slowest moving with the
median speed of 37 km/h and Slavonska Avenue traffic becomes one of the fastest with median speed over 70
km/h. Largest observed road section is Zagreb bypass with well over 4 kilometres, while the length of Ljubljanska
Avenue observed road section is only 380 meters.
568
Table 3 Road measurement data and specific road emissions with possible reductions to the base scenario
Traffic Vehicle Observed Scenario 0 Scenario 1 Scenario 2 Scenario 3
Harmful
Location flow, speed, road section, (Base case),
emission g/km % g/km % g/km %
veh/h km/h km g/km
Slavonska NOX 552 296 46% 313 43% 518 6%

Avenue 2059 72 1.65
(regular traffic) PM 40 24 80% 8 79% 39 3%
Slavonska NOX 1536 527 66% 962 37% 779 49%
Avenue 1616 15 1.65
(traffic jam) PM 64 12 81% 35 46% 24 62%
Hrvatske bratske NOX 740 429 42% 440 41% 622 16%
2296 50 0.55
zajednice Street PM 47 11 77% 11 77% 44 8%
Marin Držić NOX 532 303 43% 312 41% 446 16%
1721 57 1.17
Avenue PM 35 8 78% 8 78% 32 7%
Ljubljanska NOX 666 371 44% 379 43% 550 17%

2166 78 0.38
Avenue PM 47 11 78% 11 78% 42 11%
NOX 750 442 41% 456 39% 633 16%
Selska Street 2123 40 0.52
PM 46 12 75% 12 75% 44 4%
Miramarska NOX 634 397 38% 417 34% 543 14%

1596 37 0.58
Street PM 36 9 76% 9 76% 34 5%
Zagreb NOX 596 335 44% 344 42% 489 18%

1759 93 4.24
bypass PM 43 8 81% 8 80% 38 12%
Emission calculation for the base case (Scenario 0) shows that despite the small number of 1,616 vehicles per
hour, congestion on Slavonska Avenue causes greatest specific emission of NOX (1,536 g/km) and PM (64 g/km).
The reason behind that is a very small median traffic speed of just 15 km/h. According to the emission factor in
Figures 6, 7, 8, 9 and 10 low vehicle speed can cause a significant increase in harmful emission. If only roads with
regular traffic were observed, greatest specific emission of NOx would be on Selska Street with 750 g/km, while for
particulate matter it would be on Ljubljanska Avenue and Hrvatske Bratske Zajednice Street with 47 g/km.
Scenario 1 shows large reduction in specific emission for all observed roads. Emission of NOX can be reduced in
the range of 38% to 66% while particulate matter shows an even greater reduction in the range of 75% to 81%. For
nitrogen oxides, the greatest reduction would be achieved during the congestion on Slavonska Avenue while on the
other roads reduction would be around 40%. Once again, it is necessary to mention that this scenario shows a
reduction that would occur right after Low Emission Zone implementation and mainly as a result of vehicle number
reduction.
After the introduction of LEZ, it takes time for vehicle number to increase to its regular value. Scenario 2 shows
specific emission values after banned vehicles are replaced with newer ones. It can be seen that after the vehicle
fleet stabilisation reduction from 34% to 43% can be achieved for NOx and 46% to 80% for particulate matter. The
minimum reduction is achieved on Miramarska Street because of a small number of vehicles, while the greatest
reduction is on Slavonska and Ljubljanska Avenue. In the case of particulate matter, the majority of streets could
achieve a reduction up to 80%, while congested Slavonska Avenue could achieve minimal but still significant
reduction of 46%.
Scenario 3 in which ban of all N category vehicles is being simulated represents a solution for incensement of
traffic speed. Vehicles of this category are often responsible for slowing down traffic. With this kind of restriction, it
is possible to reduce NOx and particulate matter emission by less than 20%. A significant difference can be seen
during traffic jam where 49% of NOx emission reduction is achieved and 62% particulate matter emission
reduction. This fact shows how vehicles of this category contribute to roadside emission during the traffic jam. If
some type of congestion regulation was implemented, it would have a two-sided effect. Firstly, it would greatly
reduce emission and secondly it would increase traffic speed which in return would have an even further reduction
in emissions.
Despite the highest number of vehicles on Hrvatske bratske zajednice Street does not stand out in specific
emission nor in emission reduction. Reason for this lies in the traffic speed. Emission factor shows that traffic speed
of around 60 km/h has the advantage of low emission of harmful substances. The same thing goes for Marin Držić
Avenue where the lowest specific emission is achieved for both particulate matter (35 g/km) and nitrogen oxides
(532 g/km) because of its median speed of 57 km/h.
569
DISCUSSION
The most of Low Emission Zones already introduced in cities of Europe have less restrictive emission standard
than one proposed in this research. Reason for the proposal of such a high standard lies in the fact that every
major LEZ was introduced 10+ years ago and have set higher goals for the future. Implementation of LEZ in
Zagreb takes time, and by the time it would be introduced, LEZ with lower standards will have the same or stricter
standards than those proposed in this research. For example, authorities in London have set a goal for Euro 4
petrol vehicles and Euro 6 diesel by 2020 which is a lot higher standard than proposed. Similarly, Brussels plans to
introduce Euro 2 petrol and Euro 4 diesel by 2020 and Euro 3 petrol and Euro 6 diesel by 2025 [13].
Traffic was analysed on distinctive road segments across City of Zagreb for the purpose of traffic modelling [11]. As
it is evident, each and every observed road segment is specific and different from others, but the result of
emissions reduction are still very similar and intuitive. This states that overall percentage emission reduction does
not lie within the road segment type or traffic flow but in its speed and vehicle fleet composition.
Vehicle fleet composition and emission factors showed that traffic speed has significant influence in the specific
emission of harmful substances. When talking about traffic speed, there are two relevant facts that need to be
considered. Firstly, because of the large variety of traffic speeds over the observed road segments, it can be
considered that same results could be applied on roads across proposed LEZ according to their own traffic speeds,
meaning that the same results would be achieved over roads with traffic speed similar to those observed.
Secondly, traffic congestion has the largest specific emissions, and the reason for this is low vehicle median speed,
despite small overall vehicle number. To ensure reduction of congestions specific emission, it is necessary to
increase medial vehicle speed towards optimum of approx. 60 km/h. This could be achieved with better traffic
organisation during daily traffic peak hours or by some sort of congestion charge.
The biggest concern lies in particulate matter (PM), especially PM2.5 because the general rule says that the
smaller the PM, the more dangerous it is to human health. Also, it is quite difficult to define minimum value at which
particulate matter is harmless to human health, so this represents the reason why it should be closely monitored
and kept at its minimum level. As it is shown earlier, depending on the type of the street, particulate matter can be
reduced from 46% to 80%. Studies have found that there is a 7% increase in mortality with every 5 micrograms per
cubic metre increase in the particulate matter [14]. Therefore a huge accomplishment could be achieved if some
type of LEZ is implemented in the City of Zagreb.
Although particulate matter represents the biggest concern in cities, by implementing Low Emission Zone nitrogen
oxides would also be affected. As it is earlier shown, by implementing some kind of restriction it is possible to
reduce nitrogen oxides in the range of 34% all the way up to 43%. Nitrogen oxides are a big concern not only to
human health but the environment as well [15].
Many European cities struggle with the balance of congestion, ‘liveability’, air pollution, noise levels, accessibility,
damage to historic buildings and other pressures of urban life. As shown earlier, many cities have levels of pollution
that adversely affect health. Also, congested, polluted, noisy cities are not attractive for businesses, residents or
tourists. Congestion also has a significant impact on the economy, costing nearly €100 billion, or 1% of the EU's
GDP [16]. The different types of traffic regulation zones can reduce traffic and congestion in a city, and ensure that
those that need to travel with a vehicle can travel rather than sit in a traffic jam. In addition, noise contributes to at
least 10,000 cases of premature mortality each year and almost 90% of the health impact caused by noise
exposure is associated with road traffic noise [17].
According to the data gathered by Center for Vehicles of Croatia, average M1 category vehicle age stands at 12.95
years while for N1 category is slightly lower at 11.09 years. Older vehicles increase safety concerns in traffic.
Research done by the Center for Vehicle of Croatia show that vehicle age has a high influence on vehicles
technical validity and roadworthiness. By implementing the Low Emission Zone, it would help to reduce the high
age of the fleet and ensure better traffic safety.
CONCLUSION
Conducted research has shown that the City of Zagreb deals with the problem of high emission of harmful
substances derived from traffic. In order to deal with this problem, it is necessary to implement some sort of a Low
Emission Zone which primary purpose would be to ensure greater air quality and cleaner environment for all of the
city’s residents. By conducting this research on various types of urban roads during peak hours of traffic, it is
shown that it is possible to improve current condition regarding roadside air pollution by the introduction of LEZ.
Measured traffic flow parameters, in terms of vehicle activity, were combined with vehicle fleet composition and
appropriate emission factors. This made it possible to develop emission equations which were used for theoretical
570
determination of roadside emission and possibilities for its reduction. Results showed that it is possible to achieve
great emission reduction by changing vehicle fleet composition. Furthermore, it would also help reduce overall
vehicle number which would increase road safety.
By conducting this method during the traffic peak hour, it is estimated that the introduction of LEZ would help
achieve improvement in air quality in the City of Zagreb. By applying Scenario 1, it is possible to reduce NOx
emissions by significant 66% and PM emissions all the way up to 81%. Once again, it is important to note that
Scenario 1 represents a starting state of LEZ implementation with much lower traffic flow. Long-term achievement
is accomplished with Scenario 2 which represents traffic flow stabilisation. By Scenario 2 it is possible to reduce
around 40% of NOx emission and 80% PM emission. Scenario 3 showed that great reduction could be achieved by
regulating light-duty vehicles during traffic congestions, while not much reduction is achieved during regular traffic.
In order to have a more detailed look into roadside emission and impact of LEZ introduction, further research
should be directed towards including more road segments from inside the LEZ. Automatic counters, speed
measuring tools and plates recognition software should be used in order to minimize field work and reduce
measuring error possibilities. That kind of automated analysis would make it possible to develop software for
gathering real-time emission estimation and combine it with air quality measuring stations in order to see their more
detailed correlation.
REFERENCES
[1] EEA. European Union emission inventory report 1990–2015 under the UNECE Convention on Long-range
Transboundary Air Pollution (LRTAP). No 9/2017, https://www.eea.europa.eu/publications/annual-eu-
emissions-inventory-report,
[2] Costa, L.G., Cole, T.B., Coburn, J., Chang, Y., Dao, K., Roqué, P.J.: "Neurotoxicity of traffic-related air
pollution", Neuro Toxicology, Vol. 59, 2017, pp 133-139, doi. 10.1016/j.neuro.2015.11.008,
[3] Hendricks, J., Righi, M., Dahlmann, K., Gottschaldt, K., Grewe, V., Ponater, M., Sausen R., Heinrichs, D.,
Winkler, C., Wolfermann, A., Kampffmeyer, T., Friedrich, R., Klötzke, M., Kugler, U.: "Quantifying the climate
impact of emissions from land-based transport in Germany", Transportation Research Part D: Transport and
Environment, 2017 July, doi. 10.1016/j.trd.2017.06.003,
[4] Directive 2008/50/EC of the European Parliament and of the Council of 21 May 2008 on ambient air quality
and cleaner air for Europe, http://eur-lex.europa.eu/legal-content/HR/TXT/?uri=CELEX:32008L0050a,
[5] Božić, M., Vučetić, A., Sjerić, M., Kozarac, D., Lulić, Z.: "Experimental Study on Knock Sources in Spark
Ignition Engine with Exhaust Gas Recirculation”, Energy conversion and management, 2018, Vol. 165, pp 35-
44, doi. 10.1016/j.enconman.2018.03.053,
[6] Kurt, O.K., Zhang, J., Pinkerton, K.E.: "Pulmonary health effects of air pollution", Current Opinion in Pulmonary
Medicine, pp 138–143, doi. 10.1097/MCP.0000000000000248,
[7] Poorfakhraei, A., Tayarani, M., Rowangould, G.: “Evaluating health outcomes from vehicle emissions exposure
in the long range regional transportation planning process", Journal of Transport & Health, Vol. 6, 2017, pp
501-505, doi. 10.1016/j.jth.2017.05.177,
[8] Holman, C., Harrison, R., Querol, X.: "Review of the efficacy of low emission zones to improve urban air quality
in European cities”, Atmospheric Environment, Vol. 111, 2015, pp 161-169, doi.
10.1016/j.atmosenv.2015.04.009,
[9] Urban Access Regulations In Europe, http://urbanaccessregulations.eu,
[10] Bonačić Bartolin, P., Bunjevac, M., Lulić, Z.: "Vehicle Emissions and Air Quality in the City of Zagreb", Digital
proceedings of the 8th European Combustion Meeting, Dubrovnik, Croatia, 8-21.4. 2017,
[11] Horvat, R., Kos, G., Ševrović, M.: "Traffic Flow Modelling on the Road Network in the Cities”, Technical
Gazette, Vol. 22, No. 2, 2015, pp 475-486. doi. 10.17559/TV-20150127093334,
[12] Ntziachristos, L., Gkatzoflias, D., Kouridis, C., Samaras, Z.: "COPERT: A European Road Transport Emission
Inventory Model", Information Technologies in Environmental Engineering, 2009, pp 491-504, doi.
10.1007/978-3-540-88351-7_37,
[13] Urban Access Regulations In Europe, Overview of Low Emission Zones, http://urbanaccessregulations.eu/
overview-of-lezs,
[14] Eeftens, M.: "Spatial variation of PM2,5, PM10, PM2,5 absorbance and PM coarse concentrations between
and within 20 European study areas and the relationship with NO2 – Results of the ESCAPE project",
Atmospheric Environment, Vol. 62, 2012, pp 303-317. doi. 10.1016/j.atmosenv.2012.08.038,
[15] Boningari, T., Smirniotis, P.G.: "Impact of nitrogen oxides on the environment and human health: Mn-based
materials for the NOX abatement", Current Opinion in Chemical Engineering, Vol. 13, 2016, pp 133-141. doi.
10.1016/j.coche.2016.09.004,
[16] European Commission, Green Paper - Towards a new culture for urban mobility {SEC(2007) 1209}. https://eur-
lex.europa.eu/legal-content/en/ALL/?uri=CELEX%3A52007DC0551,
[17] EEA (2014), Noise in Europe 2014, EEA Report No 10/2014, https://www.eea.europa.eu/publications/noise-in-
europe-2014.
571
572
Kragujevac, Serbia
MVM2018-058
Kjosevski Stevan1
Kochov Atanas2
Kostikj Aleksandar3
AN INDICATORS BASED APPROACH TOWARDS DECISION

MAKING AND POLICY MAKING REGARDING
INTRODUCING ELECTRIC VEHICLES
ABSTRACT: The dynamic development of the societies and cities is significantly based on the role of transport.
That means a number of issues need to be addressed in order to reduce emission, keep traffic noise on acceptable
level, increase energy efficiency and mitigate long-term threats. One of the solutions is the implementation of electric
and hybrid cars. This problem can be solved thanks to the sustainable mobility concept, in which transport needs are
limited at the stage of planning spatial development from one side or in implementation of solutions based on electric
& hybrid cars. This leads toward research for defining the indicators which would provide information to the policy
makers for creating solutions for sustainable transportation and environmental protection. One of the scientific
methods used in the research conducted in 6 Western Balkan countries is the AHP methodology and multi criteria
decision making process for defining four pillars of indicators, social, economic, cultural, environmental indicators for
possible implementation of different types of cars toward the fulfilment of the sustainability goals for the region of
WBC’s. WBC’s as developing countries are still on the first stage in the process of defining indicators for policy
making based on scientific methodology. This paper presents the scientific approach towards decision-making and
policy creation for introduction of electric vehicles. The first step, identification of indicators, suitable for the electric
vehicles, and developing countries region, has been taken. The paper shows a proposal for indicators on a
state/region level, but also such list for personal/company need. The list of indicators is prepared based on wide and
deep references analysis, and also on view of big number of experts in the region.
KEYWORDS: sustainable development, transport, electric vehicle, indicator, policy
INTRODUCTION
The world is increasingly talking about sustainable development in the broadest sense and makes efforts to reach it.
Transport is one of the largest economic branches globally. This sector has strong impacts on the three main pillars
of sustainable development: economic, social and environmental. Also, in accordance with the current trends, it is
projected to double the world population of motor vehicles over the next twenty years, and as a consequence, an
increase in energy consumption in all regions is expected globally. Knowing that fossil fuels are finite resources and
will be exhausted, as well as the need of reduced emissions due to pollution, solutions have been found in the
implementation of electric and hybrid vehicles. The need of research for defining indicators to help policy makers
create solutions for sustainable transport and environmental protection came to life. The scientific method used in
the research is the AHP method and multi criteria decision-making process for defining the main pillars of indicators:
economic, social and environmental. Taking into account that the WBC’s are developing countries, the identification
of the indicators for implementation of electric and hybrid vehicles has been made. The indicators are made to fill the
1
Kjosevski Stevan, University Mother Theresa, Skopje, R. Macedonia, stevan.kjosevski@gmail.com
2
Kochov Atanas, University “Ss. Cyril and Methodius” in Skopje, Faculty of Mechanical Engineering – Skopje,
Skopje, R. Macedonia, atanas.kochov@mf.du.mk
3
Kostikj Aleksandar, University “Ss. Cyril and Methodius” in Skopje, Faculty of Mechanical Engineering – Skopje,
Skopje, R. Macedonia, aleksandar.kostikj@mf.du.mk
573
needs on state and regional level as well as for personal and company use. For their use, it is necessary to build a
personal and institutional capacity that would guarantee the reduction of risks in making important, costly and far-
reaching decisions.
METHODOLOGY FOR DECISION-MAKING

Making decisions is a complex process that can have far-reaching positive or negative consequences. It is a process
of choosing an alternative, from a group of alternatives, in a systematic and logical way. This is why this process has
been and is the subject of many scientific researches that result in various methodologies.
Process of decision making

Making a decision is a process of choosing an alternative, from a group of alternatives, in a systematic and logical
way. The basic step by step process used in decision making is called the decision-making process.
Figure 1 Basic diagram of decision making
The individual steps, as shown in Figure 1, are:

1. Defining the decision-making problem (goal)
2. Identification of criteria
3. Determining the alternatives, etc.
Depending on the type of problem to be decided, there are two types of decision: decision-making based on a single
criterion and decision-making based on multiple criteria. A decision-making based on multiple criteria is often referred
to as Multi Criteria Decision Making – MCDM, [2].
The AHP method
The AHP method is based on the experience of its inventor T.L. Saaty. It is developed in response to the realization
that there is a great lack of a general method that is easy to understand and apply, which will allow making complex
decisions. Since then, the simplicity and power of AHP has led to its wide application in many domains in every part
of the world. AHP has found applications in businesses, governments, social studies, research and development,
defence and other areas that involve decisions that require choice, priority, and prediction, [3].
Basically, AHP helps to structure the complexity, measurement, and synthesis of the ranking. These features make
it suitable for a wide range of applications. AHP has been tested as theoretically clear, tested on the market and
accepted methodology. Its almost universal scalability as a new paradigm for decision-making, along with its easy
application and understanding make up its success. Moreover, it proved to be capable of yielding results that are
consistent with perceptions and expectations.
ANALYSIS OF THE CURRENT STATE IN THE WESTERN BALKANS REGION

AND DEFINING ITS PROBLEMS
The Western Balkan region covers countries that are not members of the European Union, which at the same time,
in the majority, are candidates for its members. In the direction of efforts to become EU members, the countries of
574
the Western Balkans must accelerate their development in a sustainable way. Therefore, the postulates and the
mechanisms of sustainable development must be mastered and implemented.
Transport, as one of the main components of development, must also be sustainable. The significance of transport
is great, not only in the economic pillar but also in the other two main pillars of sustainable development: social and
environmental. In this process, politics (governance) and planning have a lot of significance in the process, and more
recently the significance of culture which in the region of the Western Balkans brings a series of specifics in relation
to that of the older EU members.
The countries of the region are conditioned to implement this process and accelerate it in a period of intense technical
and technological innovation.
Electric vehicles are one of the most typical examples of a complex challenge, given that they are in the middle of
multiple pillars of the sustainable development of countries. This arises from several key aspects that are particularly
relevant for this region, which by their nature are included in the four groups of SWOT analyses as show in the table
below (Table 1).
Тable 1 SWOT analysis of the state of transport in the Western Balkans region
Strengths (advantages) Weaknesses (constraints)
 Road transport, and especially passenger transport,  The percentage of new vehicles is small compared to
are the dominant segments of the economy and the total number of registered vehicles
mobility  Electric vehicle prices are too high compared to
 Electric vehicles are environmentally superior conventional ones
 The performance of electric vehicles continuously  There are no plans for development of transport in
improves which the potential of electric vehicles is clearly
 The offer of electric vehicles is improving elaborated, as well as strategies for their introduction
 It is not clear the connection between the needs and
the possibilities of the power system to support the
development of a system from fillers
 There are no national strategies for the development of
electric chargers
 There are no national fleet renewal strategies
 There is no research on the level of public awareness
about ecology
 There is no research on the perception of electric
vehicles by potential buyers
Occasions Threats
 Large proportion of the population gravitates towards  In recent years there has been a rapid increase in the
the major urban centers, where most of the economic number of registered vehicles
activities  Most of the re-registered vehicles are used with low
 Large part of the urban transport takes place with emission characteristics
passenger vehicles  A very small proportion of new vehicles are hybrid, an
 In recent years there has been a rapid increase in the even smaller number are plug-in hybrids, and the
number of registered vehicles smallest number are electric
 The condition with the fillers is such that there are only  The condition with the fillers is such that there are only
sporadic efforts to place in a limited number of sporadic efforts to place in a limited number of
separate locations separate locations
 The age structure of the vehicle is extremely  The age structure of the vehicle is extremely
unfavorable unfavorable
 The level of pollution in urban areas is, as a rule, one  The level of pollution in urban areas is, as a rule, one
of the highest in Europe of the highest in Europe
 The post-sales support of electric vehicles has not yet
been built and so on.
Choosing a scientific research method

The previous chapter of this paper provides a description of the theoretical foundations of multi-criteria decision-
making methods. Given that this is a method that will be applied in the field of transport, an additional literature
analysis was made, which led to the conclusion that good results from the use of the AHP method can be expected.
In the part of the calculations according to the AHP method, several approaches are possible. It can be done with
classical calculations, or by using certain software tools, starting from Excel, to special program packages.
This research uses the Expert Choice software package. There are more reasons for this. Among them is its
availability and the already existing local experience. There are also numerous sources with examples that indicate
the compatibility of this software package with the AHP method.
575
Expert Choice, with its adaptability to the user, certainly has a great contribution to the success of the AHP method.
It includes intuitive user interfaces, automatic calculation of priorities and inconsistency, and several ways to conduct
sensitivity analysis.
Defining Indicators and determining alternatives

The chosen method of research relies on recognizing and selecting indicators. Typically, the stage of choosing the
indicators begins with a thorough search of literature. Broadly speaking, the number of papers that in this way
analyses sustainable development is very large. Usually, the minimum number of indicators encountered is eight,
and the maximum eighty-seven. The average is about twenty-eight indicators. These indicators are mainly divided
into segments: economic, social, ecological and institutional, etc [4].
From the analysed literature a selection of nine literary sources was made, which in their completeness and specifics
are in direct relation to the topic of the research. From all of them, by analysis and filtering, are extracted the most
relevant from a large number of indicators and are organized into groups according to their similarity.
In accordance with the AHP method and the way the Expert Choice software package works, it is necessary to
choose alternatives whose influence will be analysed in the light of the selected indicators.
At the current level of the development of automotive technology, the supply of the market and the presence of
vehicles in the traffic, the following alternatives of means of transport were selected:
 Alternative A1: electric vehicles
 Alternative A2: plug-in hybrid vehicles
 Alternative A3: hybrid vehicles
 Alternative A4: vehicles with alternate fuel
 Alternative A5: vehicles with petrol engines
 Alternative A6: vehicles with diesel engines
With the alternatives listed, further research was carried out.
USABILITY OF THE RESULTS
Indicators that can serve to make decisions in the context of the use of
electric vehicles
When compiling the list of indicators and their hierarchical placement in the questionnaire that conducted the
experimental examination, it started from a thorough review of the literature, with the intention not to make special
focus or exclusion of some of them [1].
The consistency and robustness of the results obtained in the ambient of a number of indicators and alternatives,
together with the applied scientific method for multi-criteria decision making, make the results of this research more
useful in many ways. Their use is possible for making policy assessments from the perspective of central or local
government, planning and making corporate, family or personal decisions when purchasing vehicles and the like.
If a list of indicators for a narrower and more specific group of respondents is required, or if the need for a decision
concerns a number of aspects and alternatives, it is also possible to produce a list of indicators for such a purpose.
As a result of the narrowing of the volume of interest, the number of indicators, probably the number of hierarchical
levels, will be reduced, and certain indicators will be shifted along the levels.
In the area of sustainable transport, in the light of the use of electric vehicles, the number of papers is very limited.
Therefore, the identification and selection of research indicators in this context has been accessed in the manner
described below.
From the analysed literature, a selection of nine literary sources was made, which in their completeness and specifics
are in direct relation with the topic of the research. All of them, by analysis and filtering, extracted the most relevant
from a large number of indicators and are organized into groups according to their similarity. Then, "condensation,
i.e. compression" of similar indicators was made, with more than the literally determined ones being replaced by one.
In addition to the indicators that are taken from the literature, indicators are defined and presented, which, according
to the techno-economic, social and other aspects, are specific for the use of electric vehicles. Here, above all, is a
more detailed expression of individual indicators in order to express the specifics of the particular research. Thus, in
the economic indicators are treated the price of vehicles, the costs for their exploitation (range of movement,
576
performance) and maintenance, the infrastructural aspects related to the conditions for charging electric vehicles and
similar. The group of social indicators detail the elements related to the suitability of passenger vehicles for users
with special needs, then the impact of obstacles such as noise and the like. In the ecological group of indicators,
adjustments are made according to the structure of the emissions of individual alternative vehicles, noise etc. For the
first time (within the information derived from the analysed literature), the alternatives in the column of indicators
called culture were made, [1].
Overview of the indicators from the literature and the chosen indicators is shown below:
1. From literature:
 minimal number of indicators is 8,
 maximal number of indicators is 87,
 average number of indicators is 28.
2. Chosen:
 total number of indicators is 90,
 number of hierarchy levels is 5,
 indicators at the first hierarchical level: economic, social, environmental, good governance and planning
and culture.
The chosen indicators can be used for policy-making as well as decision making on national as well as regional level.
Due to overcompensation these indicators are not presented, but are available in the reference number 1.
If a list of indicators for a narrower and more specific group of respondents is required, or if the need for a decision
refers to a smaller number of aspects and alternatives, it is also possible to produce a list of indicators for such a
purpose. As a result of the narrowing of the volume of interest, the number of indicators, probably the number of
hierarchical levels, will be reduced, and certain indicators will be shifted along the levels.
Potential examples of the specific scope of the analysis:
I. Analysis of the needs for conducting economic and environmental policies by central or local government;
II. analyses for the needs of making business decisions in the business segment (corporate decisions);
III. analyses for the needs of making personal or family decisions.
As an example of preparing such a separate list of indicators, Table 2 shows a list that could be used in the variant
III.
Тable 2 List of indicators that can be used for making personal and family decisions when procuring a new vehicle
in the context of sustainable development
Level 1 2 3
Price of a new vehicle
Vehicle prices Price of second-hand
vehicle
Urban
Type of transport
Suburban
Rank / autonomy
Performance
Dynamic characteristics
Density of electric charging
stations / fuel pumps per km
With / without electric
Economic Public
charging stations
Parking Private
charging stations
Work
charging stations
Tolls, parking, etc.
Costs for using a vehicle Environmental tax
Maintenance
Price of fuel / electricity
Costs for capital maintenance
Length and frequency of
Mobility
travel
Need for management of a person
Social
with special needs
Health disorders with exposure to
emission and noise
Technical level
Environmental
Emission properties of the vehicle
577
CONCLUSIONS
The experiences from the conducted research and the results obtained in its separate phases allow for the conclusion
of several conclusions, as well as perceiving the needs and directions for further research.
Sustainable development and, within that framework, sustainable transport are not only concepts, but also messages
for the present and future generations in order to preserve what is inherited from nature and to continuously improve
the quality of life of the present and future generations.
The literature, as well as numerous documents in this area, are a good basis for further efforts to achieve sustainable
transport, and hence sustainable development.
Fossil fuels have limited resources and the result of their use is a major contribution to the road vehicles to the air
pollution and the environment as a whole.
The use of electric vehicles, accompanied by their technical advancement and the development of plug-in hybrids,
gives new hope that they can represent an alternative that is more likely to support the development of sustainable
transport, that is, sustainable development in the broadest sense.
Scientifically based methods for multi-criteria decision-making, developed in recent decades, already have validated
applicability in different conditions. The AHP method used in this test shows a high level of flexibility and adaptability
to the task set.
The analysed scenarios for supporting the use of electric vehicles make it possible to assess the potential effect of
the usual measures that should also be complex and of a different nature.
REFERENCES
[1] Kjosevski, S.: “Decision-making and policy to introduce electric vehicles as a contribution to regional sustainable
development”, PhD thesis, 2018, Skopje, Republic of Macedonia,
[2] Gade, P.K., Osuri, M.: “Evaluation of Multi Criteria Decision Making Methods for Potential Use in Application
Security”, Master Thesis, Electrical Engineering, School of Computing, Blekinge Institute of Technology, 2014,
Karlskrona, Sweden,
[3] Saaty, T.L.: "Decision making with the analytic hierarchy process", International Jurnal Services Sciences, Vol.
1, No. 1, 2008, pp 83-98,
[4] Buzási, A., Csete, M.: "Systainability Indicators in Assessing Urban Transport Systems", Periodica Polytechnica
Transportation Engineering, Vol. 43, No. 3, 2015, pp 138-145.
578
Kragujevac, Serbia
MVM2018-061
Aleksandar Jovanović1
Katarina Kukić2
INTEGRATED CONTROL OF DIVERGING DIAMOND

INTERCHANGE AND RAMP METERING: FUZZY LOGIC
APPROACH
ABSTRACT: In this paper, we present an approach developed to control one complex system: Diverging Diamond
Interchange (DDI) - Ramp Metering (RM), under oversaturated traffic flows. The DDI is a relatively new concept of
the two intersections that serve left turn movements without conflict. The left turn movements, along with the right
one from the opposite direction, are routed to the ramp to the highway. If the highway flows are closely saturated or
saturated, these additional vehicles can lead to a complete drop in the level of service (traffic congestion). RM is a
device that leaks vehicles from the ramp to the highway, with the aim to avoid breakdown of the main flow. The
purpose of the control strategy is preventing the queue spillback from the metering ramp to DDI. In other words, the
DDI-RM coordination is needed, so that the flows in this system can be sustainable. In this paper, the two separately
way of coordination are presented, based on the fuzzy logic and fixed time control strategy. Fuzzy logic control
strategy implies the real-time approach. Finally, with the example, the efficiency of the proposed algorithms, through
the criteria of vehicle control delays, is shown. For the purposes of comparing the proposed algorithms, a special
simulation method has been developed.
KEYWORDS: diverging diamond interchange, ramp metering, fuzzy logic, oversaturated flows, real-time control
INTRODUCTION
The problem of traffic congestion has been widely researched in many areas and all around the world. The impacts
of this problem are numerous and it influences every day’s life for most people, especially as the negative effects of
traffic congestion in urban areas. Negative consequences such as the ecological and economic influence of the
increased fuel consumption are well known. From the point of the traffic engineer, the main task is to reduce the
vehicles control delays. There are a few ways to solve this problem. The most common is to carry out control of traffic
signals and, when it is possible, to build a new, innovative type of roads. In this paper, we present one a new method,
which combines these two approaches by carrying out the control of the traffic signals on new, innovative type of
road. We combine a control of the traffic signals on diverging diamond interchange (DDI, further in text) and ramp
metering (RM, further in text).
A Diverging Diamond Interchange (DDI) is a relatively new concept of intersection with an alternative geometrical
structure. This type of diamond interchange became very popular, particularly in the USA, because of reducing
vehicles control delays. The DDI also improves traffic safety because of left turn movements haven’t any conflicts.
The basic disadvantages of this concept are the relatively high initial construction costs and the space required for
this design. In our recent paper [14] we developed a fuzzy logic approach to control a oversaturated DDI, and we
obtained significantly better results compared to the classical fixed time approach.
1
Aleksandar Jovanović, PhD Traffic Engineering, University of Kragujevac, Faculty of Engineering, Sestre Janjic 6,
34000 Kragujevac, Serbia, caki1987@gmail.com
2
Katarina Kukić, PhD Mathematics, University of Belgrade, Faculty of Transport and Traffic Engineering, Vojvode
Stepe 305, 11000 Belgrade, Serbia, k.mijailovic@sf.bg.ac.rs
579
In this paper we are dealing with a more complex task: we further generalize our approach developed in [12] and
[14] by adding ramp metering. The idea of ramp metering to determine the metering rate, which defines the number
of vehicles per hour that can be connected to the highway. That defined number of vehicles is controlled by the traffic
light, mainly by allowing one vehicle per green to pass. The green light is often set to 2 s. After that, the red light
controls the metering rate. Ramp Metering can take decisions based on historical data for a certain period of the day
(morning peak hours, afternoon peak hours, etc.) and this type of control is called "fixed time". There is also control
based on the data from the detector, where we have the "real-time" control.
The aim of the integrated system is to control the on-ramp queue, in such a way that it does not exceed the critical value
after which vehicles on the DDI would be blocked. This type of control involves placing detectors in the middle and at
the end of the ramp. Detectors have the task for updating the number of vehicles on the ramp at certain time intervals
(no more than 6 s). Based on this data, the proposed algorithms are making a decision on how to avoid the queue
spillback on DDI.
In the review of hitherto literature, some relevant papers can be found. In [6] authors showed that DDI have better
performance in comparison with a conventional diamond interchange, like control delays and queue lengths, for both
vehicles and pedestrians. In [7] the authors were investigating the case of the separate intersection control mode,
under different geometric and demand conditions. The DDI in Texas, for example, shows the benefits of this
approach. The concept of dynamically reversing lines was proposed in [32]. Results obtained by Binary Mixed Integer
Linear Programming (BMILP) reveals the benefits in capacity expansion of traffic lanes. Authors in [26] discuss five
different signal phasing schemes which could be implemented on DDI. Each of them has some advantages and
disadvantages under specific traffic pattern and geometric layout of DDI. A methodology for offsets optimization on
a corridor including a DDI was proposed in [5] for intersections including DDI, it has been shown that the proposed
methodology provides a saving in travel time. Some aspects of traffic safety at DDI was discussed by the authors [4].
Crash prediction models were developed for DDI in Missouri. The study found that the DDI is safer than the
conventional diamond signalized intersection.
A various number of ramp metering control models has been developed by numerous researchers. Real-time control
can be found in a number of works, whose models are based on a strategy of flow-and-capacity relationship [15, 30,
13, 9, 21, 22]. ALINEA is one of the most implemented strategies in Europe, proposed by [17]. This algorithm adjusts
the metering rate so as not to crash the capacity of the road, which is measured at a certain distance downstream
from the ramp zone.
The approximate reasoning approach to ramp metering control can be found in the following works [3, 16, 23, 29,
31, 28, 24]. A wider and systematic overview of the previous ramp metering algorithms can be found in the following
works [2, 18, 20].
In [27] is proposed a model for integrated control of traffic flows for the highway and the city street network. In the
case of oversaturated flows, they showed the advantages of their model, as well as any disadvantages if such a
control does not exist. In [8] is developed an algorithm which controls the queue spillback from the ramp, to prevent
congestion on the street network. Depending on the traffic conditions on the network, the metering rate takes values
between the desired metering rate and a higher metering rate. The response of the system is set to 10 s. In [25] is
developed an integrated system of classical Diamond Interchange (DI) and ramp metering with the aim to prevent
queue spillback from the ramp to DI. The authors used genetic algorithms and showed all the benefits their approach
have on the numerical example.
A lack of methodology for controlling this type of intersections leaves space for further research. In this paper, we
consider the problem of controlling a oversaturated system of DDI and RM. A mathematical model for controlling an
oversaturated DDI, based on the fuzzy logic system ”Sugeno” type is developed. The criterion functions represent
the control delay of vehicles in a certain period of analysis and average queue lengths on DDI approaches. The
proposed model has been compared with fixed time control.
We considered the case when the intersection is middle oversaturated in all flows (”full oversaturated intersection”)
with a high percentage of left movements. The highway is congested only in one direction. Such a case is possible
in peak periods of traffic load.
This work is organized in the following way: after the introductory review and a brief review of some of the important
papers in this field, the second chapter is devoted to setting up the problems and goals of this paper. The third chapter
provides a methodology for solving a problem based on the fuzzy logic system. In the fourth chapter, the proposed
approach was tested on a hypothetical system of DDI and RM. A simulation was used to compare the results
obtained by the classic approach and the methodology proposed in this paper. Chapter 5 is dedicated to concluding
observations and directions for future research.
580
THE FORMULATION OF THE PROBLEM
In this paper, the problem of integrated control of DDI and RM is considered. The queue spillback from the ramp can
endanger the normal operation on DDI. The aim of this integrated control is to avoid blockage of vehicles on DDI
caused by queue from the ramp. Figure 1 provides a layout of the problem.
Figure 1 A layout of the problem
Most signal plans for DDI works in two phases, but in the case of total oversaturation we didn’t find it as a good
choice, because vehicles from directions C and D would be prevented from entering the intersection by oversaturated
directions A and B. We suggest three phases as in Figure 1. The skipping of the phases is not allowed, neither the
choice between the several phases for a green time assignment.
During the third phase, a certain amount of left movement vehicles are separating from the straight flow D and enters
on the ramp. These vehicles (denoted by ql) "feeds" the ramp, so this phase is called "feeding" phase. During the
other phases, vehicles don’t have access to the ramp. The number of vehicles from flow B to the ramp can be neglect.
Classical DDI design supposes that ql vehicles aren’t controlled by a traffic light, so it can be concluded that ql vehicles
randomly turns. Also, all vehicles arriving on DDI have a random nature. It is well known that vehicles arriving obey
to the Poisson distribution.
Movement marked with r in Figure 1 represents the right turning vehicles that "feeds" the ramp. Usually, these
vehicles are turning without control, but for the purposes of this coordination system, it will be stopped. A separated
traffic light controls these vehicles.
Let us suppose that capacity of the highway is known and denoted by c (veh/h). The capacity of the highway is
measured downstream from the ramp zone (about 150 m from the ramp). Let us denote with q (veh/h) traffic flow on
the highway in the congested direction. Fixed time control strategy for the ramp metering is simple and the metering
rate (MR) is calculated as:
MR = c - q (veh/h) (1)
For example, if the green light is set to 2 s and the calculated metering rate is 360 veh/h, which means that the red
time on the traffic light is set to 8 s.
Fixed time control for DDI, based on the historical value of vehicle flows, optimizes cycles and splits (solutions). Such
predefined solutions are implemented over a certain period of the day. The criterion function for optimization, most
often, is the vehicle control delays. More details about fixed time optimization can be found in [11]. When the
intersection is overloaded in all traffic lines, fixed time control is recommended as it provides the same solutions as
classical real-time control with detectors [19].
581
In this paper, we consider a case when only one ramp is controlled. When recurring congestion occurs, it is usually
only in one direction of the highway. It is a rare case, although it is possible, that we have high traffic demands in
both directions of the highway.
This paper has, basically, two tasks: 1) Develop a solution for integrated control of DDI and RM in real-time mode 2)
Develop a simulation model for the comparison of fixed time and real-time solutions. The comparison criteria are the
total vehicles control delays during the simulation and the average queues of the vehicle on the DDI approach at the
end of the simulation. In the next chapter will be proposed algorithms for solving the placed problems.
METHODOLOGY
We developed a mathematical model for controlling an integrated system of DDI and RM based on fuzzy logic system
“Sugeno” type. The fuzzy logic controller that we developed is based on three input and one output fuzzy set. Input
fuzzy sets are Vap – the number of vehicles in a row that is served during the green phase (vehicles approaching)
and Vq – the number of vehicles in a row waiting for the green time of the next phase (vehicles queuing) and Vr - the
number of vehicles in the ramp during the ”feeding” phase.
When getting a decision whether to break or extend the "feeding" phase the fuzzy logic system includes the third
fuzzy set - Vr,. In the other cases, when “feeding” phase is off, the fuzzy logic system operates only with the first and
the second fuzzy sets - Vap and Vq. The input variables are shown in Figure 2.
The input fuzzy sets consist of three membership functions represented by Gaussian curves: the small, middle, and
the large queue of vehicles. For controlling DDI we set the limit values for the number of vehicles in a row on 80. That
number is obtained for specific flow depending on the distance between the detector and stop line and the number
of traffic lanes. The maximum number of vehicles on the ramp is set to 40.
Figure 2 Input variables in the fuzzy logic system
According to that, the proposed methodology assumes the two fuzzy logic system. The output value of the first is
denoted by EXT1 and, when "feeding” phase control is included, the output value of the second is denoted by EXT2.
The outputs EXT1 and EXT2 are binary variables and have a value of 0 if the current phase is interrupted or 1 if it is
extended by another few seconds.
The fuzzy logic system is formed basis of the "If-Then" rules. "If" part of the rule is a premise, while "Then" represents
a consequence part. In this case, "Then" part of the rule is a binary decision about whether to extend the phase or to
interrupt it. The rule base for both fuzzy logic systems is formed and shown in Table 1.
Table 1 Fuzzy rules base

The fuzzy logic system 1 The fuzzy logic system 2
then
Serial and Vq then Weight Serial if Vap and Vq and Vr Weight
if Vap is EXT2
number is EXT1 is factor number is is is factor
is
1 small Small 1 0.8 1 small medium medium 0 1
2 small medium 0 0.8 2 small big medium 0 1
3 small Big 0 1 3 small medium small 1 0.8
4 medium Small 1 1 4 medium small medium 1 0.8
… …
9 big Big 0 0.6 14 medium big medium 0 1
The values of the EXT and Weight factor are defined by the subjective opinions of the authors.
582
The algorithm 1 control the extension of each of three phase on DDI, while algorithm 2 control the queue spillback
on the ramp. Both algorithms are given in Figure 3.
Figure 3 The algorithms for control of the system of DDI and RM
At the beginning of algorithm 1, each phase has its minimal green time (gmin) and maximal green time (gmax). Maximal
green time is obtained from optimization in the “fixed time” control.
In the next step, if the queue of the vehicles in the current phase is less than 5 or the green time has reached the
maximum value, then the phase is stopped and algorithm switches to the next phase. Depending on the state of
feeding phase (is it on or off) we take in consideration EXT value. For values greater than 0.5, the phase is extended
by an additional 2 or 6 seconds (this choice will be explained later in the text), and vice versa.
Algorithm 2 operates with two detectors, middle and the end one. The first level of coordination between DDI and
RM is to stop ql vehicles during the “feeding” phase if the middle detector is on. The detector (middle or end) is
considered to be on when a part of the ramp between the detector and the highway is filled by vehicles.
The second level of coordination is to set all DDI traffic lights to a minimum green time when both detectors are on.
This measure allows additional all red time in the cycle, which vehicles use to leave the ramp and reduce the queue
spillback.
The simulation model

For calculation of the delay, we used Akçelik’s discrete minimal delay model [1], as in Figure 4 where D is obtained as
a surface of the shaded figure, for the case when there is a oversaturation in the two-phase cycle. In our model, we
have a three phases cycle, and we obtained delay by modifying Akcelik’s minimal delay model, as in Figure 5.
Figure 4 Situation of the queue during oversaturation
We use following nomenclature:

 l(k) - queue length at the beginning of cycle state k
 q(k) - arrival flow rate in the cycle k
 c - cycle length
 s - departure (saturation) flow rate
 g(k) - green time in cycle k
583
 D(k) - total delay in cycle state k
 h(k) - the moment when queue has got empty starting from the beginning of the observed phase
 l1b (k) - queue length at the beginning of cycle state k
 l1e (k) - denotes the queue length at the end of cycle state k
 qi(k) - arrival flow rate for phase i, i =1,2,3 in the cycle k
 gi(k) - is green time for phase i in cycle k
𝐹𝑗
 𝑛𝑖 (𝑘) - the number of approaching vehicles during the all red time after the phase 𝐹𝑗
From Akçelik’s discrete minimal delay model total delay for two phases is obtained from formulae
1
𝐷 = (2𝑙(𝑘)𝑐 + 𝑞(𝑘)𝑐 2 − 𝑠𝑔2 (𝑘)) (2)
2
We slightly modified formulae for the three phases’ case and added possibility when queues become empty in some
moment during the observed phase; see formulas (3) – (10).
The modified Akçelik’s discrete minimal delay model adopted to three phases regime is presented in Figure 5. Since
in simulation we have both cases – when the queue of the vehicles gets empty during the observed phase and the
situation when the queue does not get empty, we show one of the possible cases on Figure 5. In Figure 5 the queue
𝑙1 is empty before the end of phase one (denoted by F1 in Figure 5), and the moment when the queue becomes
empty is ℎ1 and here ℎ1 < 𝑔1 . The queue 𝑙2 on Figure 5 is empty at the end of phase two (F2) and on Figure 5 we
presented the case when the moment the queue gets empty is at the end of the phase, so this would be suited the
case ℎ2 = 𝑔2 . Finally, the queue 𝑙3 in Figure 5 is not empty during the phase three (F3), nor at the end of phase
three, so here also ℎ3 = 𝑔3 .
Figure 5 The discrete minimal delay for three phases cycle
584
In our example, we have all kind of combinations of the situations when some of the queues get empty during the
observed phase and the situation when none of the queues gets empty. Based on outputs we obtained, both for fixed
time control and for fuzzy logic approach control, we calculated total delays for our 15 minutes simulation.
Beside previously introduced nomenclature, with superscript Fi denote the phase for which we calculate the
corresponding variable. All the calculations have been done in Matlab. We present an algorithm for the first phase,
which is repeated for all three phases and for all cycles. The following is algorithm we used for calculations, where k
is the index of the cycle:
1. Determine a value of 𝑔1 (𝑘) from obtained data
2. Determine a value of
𝐹1 (𝑘) 𝐹1 (𝑘)
𝑙1𝑒 = 𝑙1𝑏 + 𝑔1 (𝑘) ⋅ (𝑞1𝐹1 (𝑘) − 4 ∗ 𝑠1 (𝑘)) (3)
1
where: 𝑠1 (𝑘) = .
3
𝐹1 (𝑘)
2.1. If 𝑙1𝑒 ≤ 0 then determine ℎ1 (𝑘) - the moment when the queue has empty
else ℎ1 (𝑘) = 𝑔1 (𝑘)
3. Determine a value:
𝐹2 (𝑘) 𝐹1 (𝑘)
𝑙1𝑏 = 𝑙1𝑒 + 𝑛1𝐹1 (𝑘) (4)
4. Determine values of 𝑔2 (𝑘) and

𝐹2 (𝑘) 𝐹2 (𝑘)
𝑙1𝑒 = 𝑙1𝑏 + 𝑞1𝐹2 (𝑘) ⋅ 𝑔2 (𝑘) (5)
𝐹3 (𝑘) 𝐹2 (𝑘)
𝑙1𝑏 = 𝑙1𝑒 + 𝑛1𝐹2 (𝑘) (6)
5. Determine values of 𝑔3 (𝑘) and

𝐹3 (𝑘) 𝐹3 (𝑘)
𝑙1𝑒 = 𝑙1𝑏 + 𝑞1𝐹3 (𝑘) ⋅ 𝑔3 (𝑘) (7)
6. Determine a value:
𝐹1 (𝑘 𝐹3 (𝑘)
𝑙1𝑏 + 1) = 𝑙1𝑒 + 𝑛1𝐹3 (𝑘) (8)
7. Determine values of delays for each phase and for each all red time from formulas:
𝐹1 (𝑘)
𝐷1𝐹1 (𝑘) = (𝑙1𝑏 𝐹1 (𝑘))
+ 𝑙1𝑒 ⋅ ℎ1 (𝑘)/2
𝐹2 (𝑘) 𝐹2 (𝑘) 𝐹2 (𝑘))
𝐷1 = (𝑙1𝑏 + 𝑙1𝑒 ⋅ 𝑔2 (𝑘)/2
𝐹3 (𝑘)
𝐷1𝐹3 (𝑘) = (𝑙1𝑏 𝐹3 (𝑘))
+ 𝑙1𝑒 ⋅ 𝑔3 (𝑘)/2
𝐹1 (𝑘) 𝐹1 (𝑘) 𝐹2 (𝑘))
𝐷1𝑛 1
= (𝑙1𝑒 + 𝑙1𝑏 ⋅ ∆𝑡1 /2 (9)
𝐹2 (𝑘) 𝐹2 (𝑘) 𝐹3 (𝑘))
𝐷1𝑛2 = (𝑙1𝑒 + 𝑙1𝑏 ⋅ ∆𝑡2 /2
𝐹3 (𝑘) 𝐹3 (𝑘) 𝐹1 (𝑘
𝐷1𝑛 3
= (𝑙1𝑒 + 𝑙1𝑏 + 1)) ⋅ ∆𝑡3 /2
where: ∆𝑡1 = 6𝑠, ∆𝑡2 = 12𝑠, ∆𝑡3 = 6𝑠 are all red times.
8. Finally, we calculate:
𝐷1 (𝑘) = 𝐷1𝐹1 (𝑘) + 𝐷1𝐹2 (𝑘) + 𝐷1𝐹3 (𝑘) + 𝐷1𝑛

𝐹1 (𝑘)
1
𝐹2 (𝑘)
+ 𝐷1𝑛 2
𝐹3 (𝑘)
+ 𝐷1𝑛 3
(10)
According to [10] saturation flow rate (s) can be expressed as 3600/h, where h is average headway. During the "fixed
time" optimization process, the value of saturation flows for straight movements (phases 2 and 3) and turn
movements (phase 1) was adopted by 1800 veh/h and 1200 veh/h, respectively.
Because of that, in the proposed simulation, we consider the case when the saturation flow rate s is different in phase
F1 compared with s in F2 and F3. In phases two and three we considered s constant and equal to 1/2 for each vehicle
approach. In phase one we considered s equal to 1/3.
585
TEST EXAMPLE
As a test example, we take a hypothetical DDI and RM with the layout as in Figure 1. We located detectors in the
approaches A and B on 100 meters from the stop line and in the approaches C and D on 200 meters from the stop
line. In the simulation, we provided that vehicles are arriving according to Poisson distribution. Our simulation process
last 15 minutes.
All red time during the cycle is 24 seconds. The red times (protection times) between phases one and two, and
between phases three and one is 6 s, while the red time between phases two and three is 12 s. During all red time,
the detector sends the information about new arrivals on the stop line. Such protection times are needed to avoid
conflicts between vehicle queues when switching from one phase to another.
Let us suppose that the following values were adopted: Traffic flow for the A, B, C and D is set to 640, 600, 730 and
670 veh/h, respectively; ql is equal 50 % form traffic flow D; qr is equal to 200 veh/h; The middle detector is considered
to be on with 25 vehicles and the end detector is considered to be on with more than 40 vehicles.
Based on the optimization from [11] "fixed time" control is obtained. At the same time, the green timing of the phases
obtained in this way represents the maximum green time in "real time" control. Up to 690 seconds, the solution gives
a cycle value of 120 s and a green time value of 42 s, 28 s and 26 s, respectively in phases. After 690 seconds all
green times on DDI are set to minimum values. The minimum green time for phase 1 was adopted at 12 s, while for
phases 2 and 3 at 10 s.
The fuzzy logic controller uses imbalances in the arriving of the vehicle during the cycle. The solutions offered by the
model developed in this paper are presented in the following form: cycle length C; g1phase, g2phase, g3phase. The
following solutions were obtained: 1) 82; 24, 16, 18. 2) 56; 12, 10, 10. 3) 76; 24, 18, 10. 4) 56; 12, 10, 10. 5) 68; 24, 10,
10. 6) 56; 12, 10, 10. 7) 62; 18, 10, 10. 8) 60; 12, 14, 10. 9) 62; 18, 10, 10. 10) 62; 18, 10, 10. 11) 56; 12, 10, 10. 12)
56; 12, 10, 10. 13) 62; 18, 10, 10, and after 810 seconds all green times on DDI are set to minimum values.
Based on this outputs we calculated total delays both for the “fixed time” and for the fuzzy logic controller as a sum of
all delays for all cycles in the “fixed time” and for all cycles, we obtained for the fuzzy logic control. The delays for fixed-
time control and for fuzzy-logic control we calculated with formulas (10) are presented in Figure 6.
Figure 6 The delay for FTC and FLTC after 15 minutes simulation
After 15 minutes of simulation, the control delays of vehicles in the case of FTC control was 25993 s, while in the
case of FLTC control were 23328 s. Fuzzy logic gave better results for 2665 s or 10.25%. Beside comparison on
total delays by all the vehicles during the simulation, we also compared the sum of the length of all the queues at the
end of the phases. We obtained the following results:
586
Figure 7 The sum of the length of all queues after phases 1,2 and 3 for all cycles for fuzzy-logic time control
As one can see from Figures 7 and 8, this criterium also give better results for fuzzy-logic time control. Average
lengths of all queues in FLTC is less than in FTC, and we obtained that average length of the queue after phase 1 in
FLTC is 21.4 vehicles, and the same variable for FTC is 27.3 vehicles. Similarly, the average length of the queue
after phase 2 in FLTC is 23.4 vehicles, and the same variable for FTC is 30.4 vehicles; and after phase 3 for FLTC
27.5 vehicles, and for FTC 32.9 vehicles.
Figure 8 The sum of the length of all queues after phases 1,2 and 3 for all cycles for fixed-time control
Through numerical example, it is shown that the proposed algorithm for controlling the integrated system of RI and
RM has given improvements comparing to the classical approach.
CONCLUSION
In this paper, we obtained real time controlling of the fully saturated system of DDI and RM. We implemented a similar
methodology as in papers [12, 14], expecting to obtain better results for fuzzy logic controlling than in the case of
“fixed time” control. The results we obtained are even better than we expected and prove that it is possible to make
a real time control of a complex system such as DDI and RM. For the further research, it is possible to test this
algorithm in some cases when traffic flows are very high. Additionally, we can test our algorithm in real time ramp
metering environment.
REFERENCES
[1] Akçelik, R.: "Time-dependent expressions for delay, stop rate and queue length at traffic signals", In:
International Report AIR 367-1. Vermont South, Australia: Australian Road Research Board, 1980,
[2] Bogenberger, K., May, A.: "Advanced coordinated traffic responsive ramp metering strategies", Institute of
transportation studies, University of California, Berkeley, California PATH Working Paper, UCB-ITS-PWP-99-
19, 1999,
587
[3] Chen, L.: "Freeway ramp control using fuzzy set theory for inexact reasoning", Transportation Research A, Vol.
24, No. 1, 1990, pp 15-25,
[4] Claros, B., Edara, P., Sun, C.: "When driving on the left side is safe: Safety of the diverging diamond interchange
ramp terminals", Accident Analysis & Prevention, Vol. 100, 2017, pp 133-142,
[5] Day, C., Lavrenz, S., Stevens, A.: "Extending link pivot offset optimization to arterials with single controller
diverging diamond interchange", Transportation Research Record, 2016,
[6] Edara, P., Bared, J., Jagannathan, R.: "Diverging Diamond Interchange and Double Crossover Intersection-
Vehicle and Pedestrian Performance", In 3rd International Symposium on Highway Geometric Design, 2005,
Chicago, IL,
[7] Engelbrecht, R., Barnes, K.: "Advanced traffic signal control for diamond interchanges", Journal of the
Transportation Research Board, Vol. 1856, 2003, pp 231-238,
[8] Gordon, R.: "Algorithm for controlling spillback from ramp meters", Journal of the Transportation Research
Board, Vol. 1554, 1996, pp 162-171,
[9] Hegyi. A., De Schutter, B., Hellendoorn, J.: "MPC-based optimal coordination of variable speed limits to
suppress shock waves in freeway trafﬁc", In: Proceedings of the 2003 American Control Conference, 2003,
Denver, Colorado, USA,
[10] Highway Capacity Manual (HCM 2010). Transportation Research Board, National Research Council, 2010,
Washington, DC,
[11] Jovanović, A., Teodorović, D.: "Pre-timed control for an undersaturated and over-saturated isolated intersection:
A Bee Colony Optimization approach", Transportation Planning and Technology, Vol. 40, No. 5, 2017, pp 556-
576,
[12] Jovanović, A., Kukić, K.: "Controlling an isolated oversaturated intersection in real time", Military Technical
Courier, Vol. 65, No. 4, 2017, pp 866-881,
[13] Kotsialos, A., Papageorgiou, M., Mangeas, M., Haj-Salem, H.: “Coordinated and integrated control of motorway
networks via nonlinear optimal control", Transportation Research C, Vol. 10, No. 1, 2002, pp 65–84,
[14] Kukić, K., Jovanović, A.: "Fuzzy logic approach on traffic control of an diverging diamond interchange in real
time", 1st International Conference on Computational Methods and Applications in Engineering, 23 – 26 May
2018, UPT, 2018, Timisoara, Romania,
[15] Masher, D., Ross, D.W., Wong, P.J., Tuan, P.L., Zeidler, H.M., Petracek, S.: "Guidelines for design and
operating of ramp control systems", Stanford Research Institute, SRI, 1975, Mendi Park, CA,
[16] Meldrum, D., Taylor, C.: "Freeway traffic data prediction using artificial neural networks and development of a
fuzzy logic ramp metering algorithm", Final technical report. National Technical Information Service, 5301
Shawnee Road Alexandria, VA 22312 USA, 1995,
[17] Papageorgiou, M., Haj-Salem, H., Blosseville, J.: "ALINEA: a local feedback control law for on-ramp metering",
Transportation Research Record, Vol. 1320, 1991, pp 58–64,
[18] Papageorgiou, M., Papamichail, I.: "Overview of Traffic Signal Operation Policies for Ramp Metering",
Transportation Research Board, Vol. 2047, 2008, pp 28-36,
[19] Roess, R., Prassas, E., McShane, W.: "Traffic engineering", 4thed. Upper Saddle River, N.J.: Prentice Hall, 2011,
[20] Shaaban, K., Khan, M., Hamila, R.: "Literature review of advancements in adaptive ramp metering", Procedia
Computer Science, vol. 83, 2016, pp 203-211,
[21] Smaragdis, E., Papageorgiou, M.: "A series of new local ramp metering strategies", Transportation Research
Record, Vol. 1856, 2003, pp 74–86.
[22] Smaragdis, E., Papageorgiou, M., Kosmatopoulos, E.: "A ﬂow-maximizing adaptive local ramp metering
strategy", Transportation Research B, Vol. 38, No. 3, 2004, pp 251–270,
[23] Taale, H., Slager, J., Rosloot, J.: "The Assessment of Ramp Metering Based on Fuzzy Logic", 3rd ITS World
Congress in Orlando, 1996,
[24] Taylor, C., Meldrum, D., Jacobson, L.: “Fuzzy Ramp Metering: Design Overview and Simulation Results”,
Transportation Research Record, Vol. 1634, 2007, pp 10-18,
[25] Tian, Z., Messer, C., Balke, K., Urbanik, T.: "Integration of diamond interchange and ramp metering
operations", Transportation research record, Vol. 1925, No. 1, 2005, pp 100-111,
[26] Tian, Z., Xu, H., de Camp, G., Kyte, M., Wang, Y.: "Readily implementable signal phasing schemes for diverging
diamond interchanges", In 94th Annual Meeting of the Transportation Research Board, 2015, Washington, DC,
[27] Van Aerde, M., Yagar, S.: "Dynamic integrated freeway/traffic signal networks: Problems and proposed
solutions”, Transportation Research A, Vol. 22, No. 6, 1988, pp 435-443,
[28] Vukanović, S., Ernhofer, O.: "Field evaluation of the fuzzy logic based ramp metering algorithm ACCEZZ", In:
Preprints 11th IFAC Symposium on Control in Transportation Systems, 2006, Delft, Netherlands,
[29] Wei, C., Wu, K.: "Applying an Artificial Neural Network Model to Freeway Ramp Metering Control",
Transportation Planning Journal, Vol. 25, No. 3, 1996, pp 335-355,
[30] Zhang, H., Ritchie, S., Recker, W.: "Some General Results on the Optimal Ramp Control Problem",
Transportation Research C, Vol. 4, No. 2, 1996, pp 51-69,
[31] Zhang, H., Ritchie, S.: "Freeway Ramp Metering Using Artificial Neural Networks", Transportation Research C,
Vol. 5, No. 5, 1997, pp 273-286,
[32] Zhao, J., Liu, Y., Yang, X.: "Operation of signalized diamond interchanges with frontage roads using dynamic
reversible lane control", Transportation research C, Vol. 51, 2015, pp 196-209.
588

Proceedings MVM2018 PDF

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Proceedings MVM2018 PDF

Transféré par

Droits d'auteur :

Formats disponibles

th

Motor Vehicles & Motors

October 4th - 5th, 2018

October 4th - 5th, 2018

For Publisher: Prof. dr Dobrica Milosavljević - Dean

Editor: Dr Radivoje Pešić, prof.

Technical preparation: Dr Dragan Taranović, prof.

Picture on the cover: Nemanja Lazarević

Print CD: Faculty of Engineering, University of Kragujevac, Kragujevac

Year of publication: 2018.

Number of copies printed: 200

CIP - Каталогизација у публикацији

INTERNATIONAL Congress Motor Vehicles and Motors (7 ; 2018 ; Kragujevac)

Sistemski zahtevi: Nisu navedeni. - Nasl. sa naslovne strane dokumenta. -

Copyright © 2018 Faculty of Engineering, University of Kragujevac

Publishing of this book is supported by:

Ministry of Education, Science and Technological Development of the Republic of Serbia

President: Prof. Radivoje Pešić, Ph.D., UniKg, FE, Serbia

Secretaries: Assist. Nadica Stojanović, M.Sc. UniKg, FE, Serbia

President: Prof. Jasna Glišović, Ph.D., UniKg, FE, Serbia

Secretaries: Assist. Nadica Stojanović, M.Sc. UniKg, FE, Serbia

 Faculty of Engineering, University of Kragujevac

 Government of the Republic of Serbia, Ministry of Education, Science and Technological

Moto Kongresa „MVM 2018“ je „Ekologija – Vozilo i bezbednost saobraćaja - Efikasnost“.

Jasno je da je automobil, kao multi-disciplinarni objekat razvoja i istraživanja sa jedne strane, i

U Kragujevcu, 4. - 5. oktobar, 2018. godine Naučni i Organizacioni odbor

In Kragujevac, October 4th - 5th, 2018 Scientific and Organizational Board of

WILL ELECTRIC MOTOR SUBSTITUTE INTERNAL

THE ROLE OF INDUSTRIAL AND SERVICE ROBOTS

WILL ELECTRIC MOTOR SUBSTITUTE INTERNAL

SHORT HISTORY OF ELECTRIC VEHICLES

Figure 1 Loner-Porsche electric vehicle at the World exhibition in Paris in 1900

Figure 2 The first hybrid vehicle

STATE OF THE ART IN DEVELOPMENT OF VEHICLE POWER UNITS

Figure 3 Energy needed for 500 km range of driving

Figure 7 Electric bus in Belgrade traffic

Electric drives in water traffic

Electric drives in air traffic

A VIEW INTO THE FUTURE

Figure 9 Modern electric cars (source: ADAC)

Figure 11 Toyota Mirai with fuel cell

ELECTRIC BUS VERSUS BUS WITH COMBUSTION ENGINE

Description of the System Approach and preliminary Remarks

Tabelle1 Antriebstechnologien in der deutschen Linienbusflotte 2017 (Stadt- und Überlandbusse)

Abbildung 1 Potenzielle Marktdurchdringung von Elektrobussen: Entwicklung der Neuzugänge

Zu den aktuell diskutierten, aussichtsreichen Konzepten der Elektromobilität zählen:

Abbildung 3 Fahrgastkapazität von Solo-Linienbussen nach Antriebssystem

Abbildung 9 Relativer Kostenvergleich der Antriebstechnologien für Solobusse (Lebenszykluskosten ohne

Abbildung 10 Relativer Kostenvergleich der Antriebstechnologien für Solobusse (Lebenszykluskosten ohne

Abbildung 11 Relativer Kostenvergleich der Antriebstechnologien für Solobusse (Lebenszykluskosten ohne

Abbildung 12 Relativer Kostenvergleich der Antriebstechnologien für Solobusse (Lebenszykluskosten ohne

Abbildung 13 Relativer Kostenvergleich der Antriebstechnologien für Solobusse (Lebenszykluskosten ohne

FUEL CELLS AND ELECTRIC VEHICLES

KEYWORDS: fuel cells, electric vehicles, hydrogen

Figure 1 Schematic of fuel cell

Figure 2 Schematic of the fuel cell process

Thermo and Electrochemistry of a Fuel Cell

TdS = dU + pdV + δA = dH + δA, J/mol (1)

If we introduce the Gibbs free energy (available work) function ΔG term

H2 + 1/2 O2 = H2O (6)

The process starts with ionization of a hydrogen molecule

H2 = 2H+ + 2e- (7)

E = - ΔG/it = - ΔG/it = - ΔG/q (8)