Guideline For Seismic Vulnerability Reduction in The Urban Environment

EDITOR: A.
PLUMIER
Guidelines for Seismic
Vulnerability Reduction
in the Urban Environment
IUSS Press
Istituto Universitario di Studi Superiori di Pavia
L
E
S
S
L
O
S
S

R
e
p
o
r
t
-

2
0
0
7
/
0
4
Risk Mitigation for Earthquakes and Landslides
LESSLOSS Report - 2007/04
G
u
i
d
e
l
i
n
e
s

f
o
r

S
e
i
s
m
i
c

V
u
l
n
e
r
a
b
i
l
i
t
y

R
e
d
u
c
t
i
o
n

i
n

t
h
e

U
r
b
a
n

E
n
v
i
r
o
n
m
e
n
t
This Report presents a series of developments which aim at the
reduction of the seismic vulnerability of buildings and
infrastructures, focusing on practical applications through the
introduction of case studies. The first case study concerns a rapid
screening technique to evaluate buildings on an urban scale in order
to identify the need for retrofitting, followed by a description of
conventional retrofitting methods; the remaining case studies
concentrate on new retrofitting techniques, which is the main
subject of the present report. The application of Fibre Reinforced
Polymers (FRP) for the retrofit of existing structures is presented
by introducing a user friendly tool for analysis and design that
considers the contribution of steel rebars and FRP to resistance,
supported by the presentation of experimental data on durability
and fatigue; the effectiveness of FRP solutions is demonstrated
through the retrofit of masonry infills against in- and out-of-plane
deformations. The use of steel dissipative devices to reduce the
vulnerability of precast concrete frames and steel frames with
concentric bracings is then presented, followed by the displacement
based design of a base isolation solution for the seismic upgrading
of historical buildings. Guidance for the mitigation of hammering
between buildings, including a design methodology for various
situations is also presented. The report closes by describing a
displacement based methodology for the analysis of underground
structures in soft soils.
cyan_magenta_giallo_nero

LESSLOSS RISK MITIGATION FOR EARTHQUAKES AND LANDSLIDES

LESSLOSS Report No. 2007/04

Guidelines for Seismic Vulnerability Reduction in the
Urban Environment

Editor
Andr Plumier

Reviewer
Gian Michele Calvi

July, 2007

Nessuna parte di questa pubblicazione pu essere riprodotta o trasmessa in qualsiasi forma o con qualsiasi mezzo
elettronico, meccanico o altro senza lautorizzazione scritta dei proprietari dei diritti e delleditore.

No parts of this publication may be copied or transmitted in any shape or form, and by any type of electronic,
mechanical or different means, without the prior written permission of the copyright holder and the publisher.

prodotto e distribuito da:
produced and distributed by:

IUSS Press
Via Ferrata 1 - 27100 Pavia, Italy
Tel.: (+39) 0382.516911 - fax: (+39) 0382.529131
Email: info@iusspress.it - web: www.iusspress.it

ISBN: 978-88-6198-008-2

FOREWORD
Earthquake and landslide risk is a public safety issue that requires appropriate mitigation
measures and means to protect citizens, property, infrastructure and the built cultural
heritage. Mitigating this risk requires integrated and coordinated action that embraces a
wide range of organisations and disciplines. For this reason, the LESSLOSS Integrated
Project, funded by the European Commission under the auspices of its Sixth Framework
Programme, is formulated by a large number of European Centres of excellence in
earthquake and geotechnical engineering integrating in the traditional fields of engineers
and earth scientists some expertise of social scientists, economists, urban planners and
information technologists.
The LESSLOSS project addresses natural disasters, risk and impact assessment, natural
hazard monitoring, mapping and management strategies, improved disaster preparedness
and mitigation, development of advanced methods for risk assessment, methods of
appraising environmental quality and relevant pre-normative research.
A major objective of the project is to describe current best practice and advance
knowledge in each area investigated. Thus, LESSLOSS has produced, under the
coordination of the Joint Research Centre, a series of Technical reports addressed to
technical and scientific communities, national, regional and local public administrations,
design offices, and civil protection agencies with the following titles:
Lessloss-2007/01: Landslides: Mapping, Monitoring, Modelling and Stabilization
Lessloss-2007/02: European Manual for in-situ Assessment of Important Existing
Structures
Lessloss-2007/03: Innovative Anti-Seismic Systems Users Manual
Lessloss-2007/04: Guidelines for Seismic Vulnerability Reduction in the Urban
Environment
Lessloss-2007/05: Guidelines for Displacement-based Design of Buildings and Bridges
Lessloss-2007/06: Probabilistic Methods to Seismic Assessment of Existing Structures
Lessloss-2007/07: Earthquake Disaster Scenario Predictions and Loss Modelling for
Urban Areas
Lessloss-2007/08: Prediction of Ground Motion and Loss Scenarios for Selected
Infrastructure Systems in European Urban Environments

LIST OF LESSLOSS-2007/04 CONTRIBUTORS
Section Name Institution
1. Seismic safety screening
method
M. Hasan Boduroglu, Pinar
Ozdemir, Alper lki, Ergun
Binbir
Istanbul Technical University
(ITU), Faculty of Civil
Engineering, Istanbul, Turkey
2. Seismic upgrading of
structures using
conventional methods
M. Hasan Boduroglu, Engin
Orakdogen, Konuralp Girgin,
Berna Buyuksisli, Ergun Binbir
3.1. Guidelines for the
application of FRP
retrofitting
Juan Manuel Mieres, Ignacio
Calvo, Javier Bonilla
ACCIONA, Spain
3.2. Integration of
knowledge on FRP
retrofitted structures
Xavier Martinez, Sergio Oller,
Pablo Mata, Alex Barbat
International Center for
Numerical Methods in
Engineering (CIMNE)
3.3. Experimental data on
durability and fatigue
resistance
Juan Manuel Mieres, Ignacio
Calvo, Javier Bonilla
ACCIONA, Spain
3.4. Computation of the
resistance of structural
elements considering steel
& FRP
Alper Ilki, Cem Demir, Nahit
Kumbasar, Onder Peker, Emre
Karamuk, Dogan Akgun
3.5. Urban rehabilitation
using FRP
Polat Glkan, Ahmet Yakut,
Bar Binici, Gney zcebe,
Haluk Sucuolu
Middle East Technical University
(METU), Turkey
3.6. Design of FRP
reinforcement of masonry
infill walls against
transverse move
Colin Taylor, Luiza Dihoru University of Bristol (UBRIS),
Bristol, UK
4.1. RC structures Alex H. Barbat, Sergio Oller M.,
Pablo Mata A., Xavier Martinez.
International Center for
Numerical Methods in
Engineering (CIMNE)
4.2. Precast concrete portal
frames
Nicolas Hausoul, Andr
Plumier
University of Liege (ULIEGE),
Departement ArGEnCo, Liege,
Belgium
LESSLOSS Risk Mitigation for Earthquakes and Landslides

iv
4.3. Steel frames with
concentric bracings
Andr Plumier, Hugo Tedoldi,
Catherine Doneux
University of Liege (ULIEGE),
Belgium
5.1. Displacement based
design models for base
isolated historical buildings
Lus Guerreiro Instituto Superior Tcnico (IST),
Dept. of Civil Engineering,
Lisbon, Portugal
5.2. Non linear method for
control of auto adaptative
semi active base isolator
Unal Aldemir, Melih Ozdilim,
Selcuk Ozbas
6. Mitigation of
hammering between
buildings
Viviane Warnotte University of Liege (ULIEGE),
Belgium
7. Methodology of analysis
for underground structures
in soft soils
Mrio Lopes, Antnio Brito Instituto Superior Tcnico (IST),
Dept. of Civil Engineering,
Lisbon, Portugal

ABSTRACT
This report deals with the reduction of the seismic vulnerability of buildings and
infrastructures. This can correspond to very different interventions, as there are many
types of structures, many materials and many ways to reduce vulnerability. This explains
that a variety of topics is treated.
The first chapter deals with the screening of buildings on an urban scale to identify which
need retrofitting. In the second chapter, conventional methods for retrofitting are
described.
In all the following chapters, new techniques for retrofitting are presented.
The application of Fibre Reinforced Polymers (FRP) on existing structures is a technique
which has developed a lot recently. The content of Chapter three is related to the design
of FRP solutions: a user friendly design tool, experimental data on durability and fatigue
and a design method considering the contribution of steel rebars and FRP to resistance.
An effective numerical model for composite is presented. Chapter three also describes
experimental studies on masonry infill which FRP can effectively reinforce against
transverse move and for their in-plane strength. Rehabilitation using that technique can
be applied at an urban scale.
The use of dissipative devices to reduce the vulnerability of structures is the subject of
Chapter four. The technique is applied to precast concrete portal frames and to steel
frames with concentric bracings. The use of base isolation for seismic upgrading of
historical buildings is developed in Chapter five, in which a displacement - based method
is applied to a light house. The mitigation of hammering between buildings, with a
methodology to face various situations, is the subject of Chapter six. A displacement
based methodology of analysis for underground structures in soft soils is presented at
Chapter seven.
This Report focuses on practical applications rather than on theory. Detailed information
on the research topics can be found in the specific deliverables of the Lessloss project
available on the Lessloss website (www.lessloss.org).

TABLE OF CONTENTS
FOREWORD................................................................................................................................................. i
LIST OF LESSLOSS-2007/04 CONTRIBUTORS............................................................................... iii
ABSTRACT................................................................................................................................................... v
TABLE OF CONTENTS......................................................................................................................... vii
1. SEISMIC SAFETY SCREENING METHOD................................................................................... 1
1.1 INTRODUCTION.................................................................................................................................. 1
1.2 SEISMIC SAFETY SCREENING METHOD (SSSM) ............................................................................ 1
1.2.1 Concept of SSSM................................................................................................................... 2
1.3 SEISMIC INDEX, IS.............................................................................................................................. 3
1.3.1 Seismic Capacity Index, P..................................................................................................... 3
1.3.1.1 Strength Index, C.................................................................................................... 4
1.3.2 Structural Irregularity Index, D............................................................................................ 7
1.3.3 Time Dependent Deterioration Index, K........................................................................... 8
1.4 REQUIRED SEISMIC CAPACITY INDEX, ID..................................................................................... 8
1.5 PILOT REGION STUDY ...................................................................................................................... 9
1.5.1 Application of SSSM to a Sample Building......................................................................13
1.6 CONCLUSIONS ..................................................................................................................................18
2. SEISMIC UPGRADING OF STRUCTURES USING CONVENTIONAL METHODS ......19

viii
2.1 EVALUATION OF SEISMIC SAFETY OF EXISTING BUILDINGS AND RETROFITTING
TECHNIQUES........................................................................................................................................... 19
2.1.1 The Procedure for Determination of the Seismic Safety of Existing Buildings......... 20
2.2 OBSERVATIONS ON IMPERFECTIONS CAUSING THE COLLAPSE OR DAMAGE IN
RESIDENTIAL BUILDINGS AND COMMON STRENGTHENING TECHNIQUES IN TURKEY............ 20
2.2.1 Strengthening the Existing Buildings by Additional Shearwalls ................................... 22
2.2.2 Construction Rules for Additional Shearwalls ................................................................ 23
2.3 NUMERICAL EXAMPLE ................................................................................................................... 24
2.3.1 Performance Evaluations of Strengthened Building According to Non-linear Push-
over Analysis Methods of TEC-2006, FEMA-440 and ATC-40.............................................. 27
2.4 CONCLUSIONS.................................................................................................................................. 32
3. SEISMIC UPGRADING OF STRUCTURES USING FIBRE REINFORCED POLYMERS33
3.1 GUIDELINES FOR THE APPLICATION OF FRP RETROFITTING............................ 33
3.1.1 Introduction......................................................................................................................... 33
3.1.2 Materials characterisation................................................................................................... 33
3.1.2.1 Fibres...................................................................................................................... 34
3.1.2.2 Resins ..................................................................................................................... 35
3.1.2.3 Lamina and Laminates......................................................................................... 35
3.1.3 Design of the FRP reinforcement..................................................................................... 36
3.1.4 Adhesiveness........................................................................................................................ 38
3.1.4.1 Structural adhesives.............................................................................................. 38
3.1.4.2 Surface preparation .............................................................................................. 39
3.1.4.3 Adhesion mechanism........................................................................................... 40
3.1.5 Placing on site...................................................................................................................... 40
3.1.5.1 Wet lay-up/Hand lay-up...................................................................................... 41
Guidelines for Seismic Vulnerability Reduction in the Urban Environment

ix
3.1.5.2 Vacuum bagging....................................................................................................41
3.1.5.3 Filament winding...................................................................................................41
3.1.5.4 Prepregs..................................................................................................................42
3.1.5.5 Resin Film Infusion (RFI)....................................................................................42
3.1.6 Quality control .....................................................................................................................43
3.1.6.1 Check-points during installation .........................................................................43
3.1.6.2 Site tests..................................................................................................................45
3.1.6.3 Inspection recommendations ..............................................................................45
3.2 INTEGRATION OF KNOWLEDGE ON FRP RETROFITTED STRUCTURES....................................47
3.2.1 Introduction..........................................................................................................................47
3.2.2 Formulation used to simulate RC structures reinforced and/or retrofitted with CFRP48
3.2.2.1 Simulation of Composite Materials ....................................................................48
3.2.2.2 Other Formulations Developed to Simulate CFRP Reinforcements ............52
3.2.2.3 Efficiency Improvement of the Developed Code............................................55
3.2.2.4 Making the Code More User Friendly................................................................58
3.2.3 Numerical examples of the formulation proposed. CFRP reinforcements of RC
structures...........................................................................................................................................61
3.2.3.1 Code Validation: Bending Reinforcement of a RC Beam...............................61
3.2.3.2 Code Validation: CFRP Retrofitting of a Beam................................................63
3.2.3.3 Reinforcement of a Framed Structure using CFRP .........................................64
3.2.4 CONCLUSIONS ................................................................................................................71
3.3 EXPERIMENTAL DATA ON DURABILITY AND FATIGUE RESISTANCE..............73
3.3.1 Durability ..............................................................................................................................73
3.3.1.1 Fibres environmental degradation .....................................................................73

x
3.3.1.2 Accelerated Ageing Models................................................................................. 74
3.3.2 Fatigue of composites......................................................................................................... 78
3.3.2.1 Factors affecting the fatigue life. Damage mechanisms .................................. 79
3.3.2.2 Comparison of the fatigue behaviour in the different types of composites. 81
3.3.2.3 Conclusions........................................................................................................... 84
3.4 COMPUTATION OF THE RESISTANCE OF STRUCTURAL ELEMENTS CONSIDERING STEEL &
FRP 84
3.4.1 Introduction......................................................................................................................... 84
3.4.2 Testing program.................................................................................................................. 86
3.4.2.1 Outline of The Characteristics of The Tested Specimens .............................. 86
3.4.2.2 Loading and Data Acquisition Setup................................................................. 90
3.4.2.3 Test Results and Discussions.............................................................................. 90
3.4.3 Computation of strength and deformability.................................................................... 95
3.4.4 Verification of proposed computation methodology..................................................... 98
3.4.5 Conclusions.......................................................................................................................... 99
3.5 URBAN REHABILITATION USING FRP ........................................................................................ 100
3.5.1 Introduction....................................................................................................................... 100
3.5.2 Seismic Performance Assessment Procedures .............................................................. 101
3.5.2.1 The Walkdown Evaluation Procedure............................................................. 101
3.5.2.2 Preliminary Assessment..................................................................................... 102
3.5.2.3 Detailed Assessment Procedure ....................................................................... 104
3.5.3 Application to Zeytinburnu............................................................................................. 109
3.5.3.1 Walkdown Survey............................................................................................... 109
3.5.3.2 Preliminary Assessment..................................................................................... 110

xi
3.5.3.3 Detailed Assessment...........................................................................................111
3.5.4 Analysis and Design of FRPs for Seismic Retrofit........................................................112
3.5.4.1 Strengthening with FRP.....................................................................................113
3.5.4.2 Proposed Analytical Model................................................................................115
3.6 DESIGN OF FRP REINFORCEMENT OF MASONRY INFILL WALLS AGAINST TRANSVERSE
MOVE....................................................................................................................................................116
3.6.1 Scope of research...............................................................................................................116
3.6.1.1 Objectives and strategy of the present studies................................................116
3.6.2 Experimental programme.................................................................................................120
3.6.2.1 Materials and panel configuration.....................................................................120
3.6.2.2 Instrumentation...................................................................................................124
3.6.2.3 Seismic tests input motions ...............................................................................124
3.6.2.4 Quasi-static tests experimental observations ...............................................125
3.6.2.5 Seismic tests .........................................................................................................128
3.6.2.6 Seismic tests experimental observations and numerical simulations ........130
3.6.2.7 Conclusions from the experimental programme ............................................134
3.6.3 Analytical studies................................................................................................................135
3.6.4 Conclusions ........................................................................................................................137
4. SEISMIC DESIGN AND RETROFIT OF STRUCTURES USING DISSIPATIVE
DEVICES ..................................................................................................................................................139
4.1 RC STRUCTURES ..............................................................................................................................139
4.1.1 Introduction........................................................................................................................139
4.1.2 Nonlinear analysis of beam structures ............................................................................143
4.1.2.1 Finite deformation initially curved beams .......................................................144

xii
4.1.3 Nonlinear constitutive models ........................................................................................ 147
4.1.3.1 Degrading materials: damage model ................................................................ 148
4.1.3.2 Plastic materials................................................................................................... 150
4.1.3.3 Mixing theory for composites........................................................................... 152
4.1.3.4 Energy Dissipating Devices .............................................................................. 153
4.1.4 Numerical implementation .............................................................................................. 154
4.1.4.1 Tangential stiffness tensors............................................................................... 154
4.1.4.2 Cross sectional analysis...................................................................................... 156
4.1.5 Numerical examples.......................................................................................................... 158
4.1.5.1 Nonlinear Seismic Response of Planar Frame ............................................... 158
4.1.5.2 3D Precast concrete building............................................................................ 159
4.1.6 Conclusions........................................................................................................................ 161
4.2 PRECAST CONCRETE PORTAL FRAMES....................................................................................... 162
4.2.1 Post-earthquake surveys................................................................................................... 162
4.2.2 INERD pin connection.................................................................................................... 164
4.2.3 Bracings using INERD pin connections........................................................................ 164
4.2.4 Design Model for the systems......................................................................................... 166
4.2.4.1 Definition of the model ..................................................................................... 166
4.2.4.2 Plastic hinges at column bases.......................................................................... 168
4.2.4.3 Design of INERD pin connection .................................................................. 169
4.2.4.4 Static non linear analysis (Pushover analysis) ................................................. 173
4.2.4.5 Dynamic non linear analysis (Time history analysis) ..................................... 177
4.2.4.6 Analysis of the results ........................................................................................ 179

xiii
4.3 STEEL FRAMES WITH CONCENTRIC BRACINGS. CONNECTIONS...............................................182
4.3.1 Reasons for using dissipative connections in frames with concentric bracings and
purpose of the research activity in LESSLOSS. ........................................................................182
4.3.2 The INERD pin connection geometry and properties ................................................183
4.3.3 Code rules for braced frames with pin INERD-connections......................................184
4.3.4 Practical design procedure................................................................................................185
4.3.5 Application of dissipative connections to a tall office building with X bracings ......187
4.3.5.1 Design stage.........................................................................................................187
4.3.5.2 Pushover analysis ................................................................................................189
4.3.5.3 Dynamic Non linear Time History Analyses ..................................................192
4.3.5.4 Conclusions of the application of dissipative connections to a tall office
building with X bracings...................................................................................................195
4.3.6 Application of dissipative connections to a tall industrial building with V bracings 196
4.3.6.1 Reference structure and design conditions. .....................................................196
4.3.6.2 Comparison of mass of designed structures....................................................198
4.3.6.3 Influence of superstructure design on the dimensions of foundations. ......200
4.3.6.4 Results of the analysis.........................................................................................201
4.3.6.5 Conclusions of the design stage. .......................................................................202
4.3.6.6 Push-over analysis...............................................................................................202
4.3.6.7 Conclusions from the application of dissipative connections to a tall
industrial building with V bracings..................................................................................204
industrial building with V bracings..................................................................................205
4.3.7 General conclusions on the use of dissipative connections in frames with bracings205
5. SEISMIC UPGRADING OF STRUCTURES USING BASE ISOLATION............................207

xiv
5.1 DISPLACEMENT BASED DESIGN MODELS FOR BASE ISOLATED HISTORICAL BUILDINGS... 207
5.1.1 Introduction....................................................................................................................... 207
5.1.2 The proposed methodology............................................................................................. 207
5.1.3 The Capelinhos lighthouse............................................................................................... 211
5.1.3.1 Introduction ........................................................................................................ 211
5.1.3.2 The Capelinhos Lighthouse non linear model................................................ 214
5.1.3.3 Capacity curve definition................................................................................... 216
5.1.4 Conclusions........................................................................................................................ 220
5.2 NONLINEAR METHOD FOR CONTROL OF AUTO-ADAPTIVE SEMI ACTIVE BASE ISOLATOR. 220
5.2.1 Structural System............................................................................................................... 220
5.2.2 System Dynamcs .............................................................................................................. 221
5.2.3 Optimal and Sub-optimal Control .................................................................................. 222
5.2.4 Passive Viscous Damping Control And on-off Cases ................................................. 222
5.2.5 Causal semactve control ................................................................................................. 223
5.2.6 Numerical application....................................................................................................... 223
5.2.7 Derivation of the Linear Control Law ........................................................................... 225
5.2.8 Conclusions........................................................................................................................ 230
6. MITIGATION OF HAMMERING BETWEEN BUILDINGS................................................ 233
6.1 INTRODUCTION............................................................................................................................. 233
6.2 ANALYSIS OF POUNDING BETWEEN BUILDINGS AND MITIGATION BY LINKING
INTRODUCTION TO THE NUMERICAL STUDY ................................................................................... 234
6.2.1 Assumptions and limitations ........................................................................................... 234
6.2.2 Design of the structures considered in the analysis...................................................... 234
6.2.3 On the use of linear or nonlinear analysis to evaluate pounding effects ................... 236

xv
6.3 OVERVIEW OF POUNDING BETWEEN ADJACENT BUILDINGS..................................................237
6.3.1 Pounding modelling ..........................................................................................................237
6.3.2 Pounding effects ................................................................................................................239
6.3.2.1 Case A: Adjacent buildings with equal height and with aligned floor levels239
6.3.2.2 Case B: Adjacent buildings of unequal height and with aligned floor levels241
6.3.2.3 Case C: Adjacent buildings of similar or different height and with not
aligned floor levels .............................................................................................................241
6.3.2.4 Conclusions..........................................................................................................242
6.4 OVERVIEW OF POUNDING MITIGATION BETWEEN ADJACENT BUILDINGS...........................242
6.4.1 The seismic gap..................................................................................................................242
6.4.2 Increasing the stiffness of one or both buildings ..........................................................243
6.4.3 Supplemental energy dissipation......................................................................................243
6.4.4 Strengthening......................................................................................................................244
6.4.5 Alternative load paths........................................................................................................244
6.4.6 Strong shear wall ................................................................................................................244
6.4.7 Primary structure away from property limits .................................................................244
6.4.8 Reconnection......................................................................................................................245
6.5 RECOMMENDATIONS FOR THE MITIGATION OF POUNDING PROBLEMS BETWEEN ADJACENT
BUILDINGS..............................................................................................................................................245
6.5.1 Guidance to mitigate pounding with a PRDs ................................................................245
6.5.1.1 Adjacent buildings of equal height, with aligned floor levels and similar
structural types, in particular their stiffness....................................................................246
6.5.1.2 Adjacent buildings of equal height, with aligned floor levels and different
structural types ...................................................................................................................247
6.5.1.3 Adjacent buildings of unequal height, with aligned floor levels and same
structural types ...................................................................................................................248

xvi
6.5.1.4 Adjacent buildings of unequal height, with aligned floor levels and different
structural types................................................................................................................... 250
6.5.1.5 Adjacent buildings of similar or different height, with not aligned floor levels
and similar or different structural types ......................................................................... 251
6.5.1.6 Buildings with a small seating length (unseating problems) ......................... 252
6.5.2 Some practical indications on the design of Pounding Reduction Devices
(PRD's)Choice of the PRD.......................................................................................................... 253
6.5.2.1 Introduction ........................................................................................................ 253
6.5.2.2 Requirements in selecting PRD's ..................................................................... 253
6.5.2.3 Number and location of the devices................................................................ 254
6.5.3 Models and programs for impact zone .......................................................................... 255
6.6 CONCLUSIONS................................................................................................................................ 256
7. METHODOLOGY OF ANALYSIS FOR UNDERGROUND STRUCTURES IN SOFT
SOILS......................................................................................................................................................... 259
7.1 HISTORICAL BACKGROUND......................................................................................................... 259
7.2 ANALYSIS OF THE SEISMIC VULNERABILITY.............................................................................. 260
7.3 SEISMIC BEHAVIOUR ..................................................................................................................... 261
7.4 CONCEPTION................................................................................................................................. 262
7.5 DESIGN METHODOLOGY ............................................................................................................. 263
7.6 PRACTICAL APPLICATION EXAMPLE ........................................................................................... 265
7.6.1 Structure designed according to current code concepts .............................................. 266
7.6.2 Structure designed according to the proposed methodology...................................... 269
7.6.2.1 Choice of deformation mechanism.................................................................. 269
7.6.2.2 Design of reinforcement ................................................................................... 271
7.6.2.3 Results.................................................................................................................. 275

xvii
7.7 SUMMARY AND CONCLUSIONS .....................................................................................................276
REFERENCES.........................................................................................................................................277
LIST OF SYMBOLS................................................................................................................................295

1.SEISMIC SAFETY SCREENING METHOD
1.1 INTRODUCTION
Turkey is located at the boundary area where the Arabian Plate and African Plate are
moving north towards the Eurasian Plate. Many strong earthquakes have occurred along
the fault which is called North Anatolian Fault (NAF). Marmara and Duzce Earthquakes
in 1999 were two of them which caused catastrophic disaster in Marmara and adjacent
regions. Seismologists are pointing out the possibility of another big earthquake hitting
Istanbul at the western edge of NAF. When the huge existing stock of buildings in
Istanbul is considered, it is clear that it is practically impossible to carry out detailed
structural analysis for all of the buildings. In order to cope with the seismic safety
evaluation of large number of existing buildings, it is necessary to use simplified
techniques, which can predict the seismic safety of the existing buildings in relatively
shorter time. The expected outcome from such an analysis is to determine which
buildings are at relatively higher risk of collapse.
Seismic Safety Screening Method (SSSM), is an adaptation of the Japanese Seismic Index
Method (JSIM) considering the building damages in recent earthquakes and Turkish
Earthquake Resistant Design Code 1998 (TERDC-98), Specification for Structures to be
Built in Disaster Areas, Earthquake Disaster Prevention). The original method have
been applied to a number of buildings damaged during 1992 Erzincan, 1998 Adana-
Ceyhan and 1999 Marmara and Duzce Earthquakes. These results have been used for
adaptation of the original method to Turkey. This rapid seismic safety evaluation method
can be applied for structures having a storey number 6 or less with reinforced concrete
frame, shear wall or dual frame-shear wall structural systems. The calibration of several
coefficients proposed in this method will further be done considering the studies carried
out in various pilot areas like Zeytinburnu.
1.2 SEISMIC SAFETY SCREENING METHOD (SSSM)
The structural characteristics of the building are necessary for calculating the seismic
capacity index. A large number of the buildings in Istanbul have not received proper civil
engineering service and they do not have any civil engineering drawings of the structural
system. The inspection of a large number of buildings after recent earthquakes indicated
that average concrete compressive strength is as low as 10 MPa and the amount of

2
longitudinal reinforcement bar is 1% of gross cross sectional area of columns with the
tensile strength of 220 MPa. Both longitudinal and transverse bars are generally plain bars
with large spacing between stirrups like 20~30 cm. With these facts some assumptions
were made to simplify the quick assessment procedure. For obtaining reliable results, the
method was applied in a certain geographical pilot region, Zeytinburnu, and the results
were then compared with the results of more comprehensive and detailed methods to
calibrate several parameters that are used in SSSM. Equation Chapter 1 Section 1
1.2.1 Concept of SSSM
Basic concept of SSSM and the original JSIM is the equal energy concept. In case of
structures with relatively short natural period, dissipation energy is almost equal to
maximum potential energy of the elastic system as shown in Figure 1-1.

Figure 1-1. Bi-linear model of elasto-plastic response
Based on this assumption, the below equation can be written:
1
2 1
Q
Y
Q
E
(1.1)
where Q
Y
= lateral yield force; Q
E
= lateral elastic force; and = ductility ratio.
When we consider the relation between responses of elastic and elasto-plastic structural
systems:
+
= =
1 0.05 1
0.75
2 1
Q
Y
Q R
E
(1.2)
where R = ductility index.

3
When we write lateral forces with the term of weight of structure W:
= = Q CxW ; Q C xW
Y E E
(1.3)
where C = strength index.
Substitution of Equation (1.3) in Equation (1.2), seismic capacity index P is obtained.
= = = =
1
Q
C
Y
; C CxR P
E
Q C R
E E
(1.4)
It can be assumed that the structures with same seismic capacity index, P value, have
similar seismic performance.
1.3 SEISMIC INDEX, IS
Seismic index of a building, I
s
value, is calculated by using Equation (1.5):
=
s
I PxDxK (1.5)
where P=seismic capacity index; D=structural irregularity index; and K=time dependent
deterioration index.
Seismic index I
s
should be calculated separately for both earthquake directions for all
critical storeys of a building and the most unfavourable result should be compared with
the required seismic index I
D
. Based on this comparison the seismic safety of the building
can be estimated. If for any of the comparison cases I
s
< I
D
then it is concluded that the
behaviour of the structure is at higher risk of collapse and further detailed analysis is
necessary.
1.3.1 Seismic Capacity Index, P
During calculation of seismic capacity index, P, the vertical structural members are
classified into three groups as columns, short columns and shear walls. The definitions of
these groups are given in Table 1-1.


4
Table 1-1. Classification of vertical members
Elements Definition Ductility Index, R
Columns h
0
/D > 2.0 1.0
Short columns h
0
/D < 2.0 0.8
Walls reinforced concrete wall 1.0
h
0
: clear height of column; D: depth of column
Most of the existing medium/low rise reinforced concrete buildings have practically no
shear walls. The calculation of P index differs according to the existence or absence of
short columns in the structure. If there is no short column in the structure then P index is
calculated by Equation (1.6) or (1.7), otherwise by Equation (1.8).
+
=
+
1
( ) ( )
c c
n
P C R
n i
(1.6)
+
= +
+
1
1
( ) ( )
w c w
n
P C a C R
n i
(1.7)
+
= + +
+
2 3
1
( ) ( )
sc w c sc
n
P C a C a C R
n i
(1.8)
where n = number of storeys of a building; i = number of the storey where the
calculation is done ( for a ground floor i=1); C
c
= strength index of a column; C
w
=
strength index of a shear wall; C
sc
= strength index of a short column; R
c
= ductility index
of a column; R
w
= ductility index of a shear wall; R
sc
= ductility index of a short column;
a
1,2,3
= strength reduction factors due to displacement compatibility as shown Figure 1-2,
and a
1
=0.7 ; a
2
=0.7 ; a
3
=0.5.
C
sc
a
2
C
w
a
3
C
c
C
w
a
1
C
c
A B
C
sc
a
2
C
w
a
3
C
c
C
w
a
1
C
c
A B

Figure 1-2. Displacement compatibility of vertical members
1.3.1.1 Strength Index, C
The strength index, C is computed with the lateral strength of vertical members based on
the assumption that beams are strong enough and the expected structural damage on
beams cannot be as important as vertical elements damages. The lateral strength of each

5
vertical member is the minimum of the flexural strength and shear strength, Equation
(1.9).
( )
=
min
;
k mu k su
k
Q Q
C
W
(1.9)
where
k
C = k
th
vertical members strength index;
k
Q
mu
= flexural strength of k
th
vertical
element;
k
Q
su
= shear strength of k
th
vertical element; and W = total weight above the
storey concerned. The calculation should be done for both directions.
For calculation of strength index, C;
Number of storeys of a building,
Drawings of the selected critical storeys under consideration,
Size of vertical elements and their orientations,
Tributary area of columns which will be used to define the column axial forces,
Type of slabs, height of the selected critical storeys, and beam heights, should be
known.
In this method, following assumptions will be made if these are already not available.
During the investigation if the material tests are not carried out, concrete
compressive strength can be taken as 10 MPa.
The amount of longitudinal tensile reinforcement is 0.4% of gross cross sectional
area of columns and tensile strength of reinforcing bars can be assumed as 220 MPa.
Stirrups spacing in columns is not less than 20 cm and the stirrups diameter is not
larger than 8 mm.
Total weight of the structure can be estimated by examining the type of slabs and
beams, usually 10-12 kN/m2 for the unit floor area (including live load for the
seismic design) is acceptable.
The effect of cracks or deterioration of concrete or light corrosion of reinforcement
may be neglected in the calculation of strength index, since it might be considered by
the estimation of time dependent deterioration index, K value. But if extreme
deterioration is observed, special consideration should be given for material
properties.

6
Moment capacity of a member should be known for the calculation of flexural strength
for column. The moment capacity of columns can be calculated by using the approximate
formulas given in Equation (1.10), (1.11) and (1.12) according to the level of axial load.
These equations are the same equations given in the JSIM.
> >
= +
= +
max
2
max max
max
0.4 ;
(0.8 0.12 )( )/( 0.4 )
c
C u t y c c
c g y
N N bDf
M a D bD f N N N bDf
N bDf a
(1.10)
> >
= +
0.4 0;
0.8 0.5 (1 / )
c
C u t y c
bDf N
M a D ND N bDf
(1.11)
> >
= +
=
min
min
0 ;
0.8 0.4
C u t y
g y
N N
M a D ND
N a
(1.12)
where N
max
= axial compressive strength of a column [N]; N
min
= axial tensile strength of a
column [N]; N = axial load of a column [N]; a
t
= area of tension reinforcement [mm
2
]; a
g
=
total area of longitudinal reinforcement [mm
2
]; b = width of compressive side of a column
[mm]; D = cross-sectional depth of a column [mm];
y
= yield strength of longitudinal
reinforcement [MPa]; and f
c
= compressive strength of concrete [MPa].
Ultimate shear strength for column can be calculated by using Equation (1.13) proposed
by Arakawa.

+
= + +

+

=
= = =
0.23
0
0
0.053 (18 )
0.85 0.1 .
/( . ) 0.12
/ /2 /( . ) 3
100 7 /8
. .
t c
C su w s wy
t w
t w
x
f
Q b j
M Q d
M Q h ; 1 M Q d
a a
x ; ; j d
b d b s
(1.13)
where
t
= tensile reinforcement ratio(%);
w
= shear reinforcement ratio; a
w
= total cross
sectional area of a set of stirrups [mm
2
]; a
t
= area of tension reinforcement [mm
2
]; b =
width of compressive side of a column [mm]; d = effective depth of column [mm]; s =
spacing of stirrups [mm];
s
wy
= yield strength of stirrups [MPa];
0
= axial stress of
columns, 0 <
0
< 0.5 f
c
[MPa]; and f
c
Ultimate flexure strength for shear wall can be calculated by the following equation.

7
= + +
. 0.5 ( ). 0.5 .
W u t y W Wy Wy W W
M a l a l N l (1.14)
where N = total axial load of a wall [N]; a
t
= area of tensile reinforcement at the walls
end zone [mm
2
]; a
Wy
= total area of vertical reinforcement [mm
2
]; l
W
= distance between
centre of wall and end zones (wall end zones may be within the wall itself or within an
adjoining wall or in an enlarged section at the edge of the wall) [mm];
y
= yield strength
of longitudinal reinforcement [MPa]; f
c
The flexural strength can be calculated according to moment capacity of a member and
the inflection height along the member, Equation (1.15).
( ) = /
mu u o
Q M h (1.15)
where is the coefficient for inflection height, and h
0
is the clear height of column or
wall. If column is not connected to the frame is equal to 0. If it is connected then on
the strong direction of column can be taken as 2 while on the weak direction it is 1.7.
When walls are under consideration, coefficient becomes 1.
Ultimate shear strength for shear wall can be calculated by Equation (1.16),

+
= + +

+

= = = = =
0.23
0
0.053 (18 )
0.85 . 0.1 .
/( . ) 0.12
/( . ) 3
100
. . .
te c
W su se wy e e e
t h
te e se 0e e W
e e e
f
Q b j
M Q l
1 M Q l
A a a N
x ; b ; ; ; j l
b l l b s b l
(1.16)
where
te
= equivalent tensile reinforcement ratio(%); a
t
= area of tensile reinforcement at
the walls end zone [mm
2
]; b
e
= equivalent thickness of wall [mm]; l = wall length [mm];
A = cross sectional area of wall;
se
= equivalent lateral reinforcement ratio; a
h
= cross
sectional area of lateral reinforcement area [mm
2
]; s = spacing of lateral reinforcement
[mm];
wy
= yield strength of lateral reinforcement [MPa];
0e
= axial stress of wall,
0e
<
0.5 f
c
[MPa]; and f
c
1.3.2 Structural Irregularity Index, D
Structural irregularity index is related with the irregularities in plan and irregularities in
elevation of a building. In the Specification for Structures to be Built in Disaster Areas,
Earthquake Disaster Prevention (TERDC-98) seven irregularities are described. Four of
these are in plan and three in elevation. Since the seismic index, I
s
is a production of three

8
sub-indices (eq. 1.5) structural irregularity index is the modification factor for the seismic
capacity index, P as well as time dependent deterioration index, K.
If building is an irregular structure, the seismic index, I
s
could be reduced 40% by the
structural irregularity index, D. The details of structural irregularity index, D will not be
given in this study. This is an ongoing research project.
1.3.3 Time Dependent Deterioration Index, K
Influence of structural defects such as deflection, cracking, and corrosion is taking into
account with time dependent deterioration index, K on the seismic index, I
s
. Site
inspection is necessary for the estimation of the time dependent deterioration index, K.
Since this is an ongoing research project, it is adopted from the JSIM as it is. The seismic
index, I
s
could be reduced 30% by the time dependent deterioration index, K.
1.4 REQUIRED SEISMIC CAPACITY INDEX, ID
The required seismic index, I
D
should be calculated by Equation (1.17).
=
0 D
I P xBxZxU (1.17)
where P
0
= index for basic seismic demand; B

=index for regional seismicity; Z

= index
for local soil conditions; and U = index for usage. Index B should be taken as 1.0 for high
seismicity and never lower than 0.5 in any case. Index Z varies for soil conditions from
0.80(for rock or stiff soil) to 1.00 (soft soil). Usage index, U is 1.0 for residential
buildings, 1.5 for buildings to be utilized immediately after the earthquake.
The basic seismic demand index, P
0
is the most critical parameter for the method. In
TERDC-98, Total Equivalent Seismic Load (base shear), V
t
acting on the entire building
in the earthquake direction considered will be determined by Equation (1.18).
=
0
0
( )
0.10
( )
t
a
WxA xIxS T
V xA xIxW
R T
(1.18)
where W

=total weight of the building (G+nQ); A
0
= effective ground acceleration
coefficient; I = building importance factor; S(T)= spectrum coefficient, R
a
(T)

= seismic
load reduction factor; and T= natural period of the building.
The lateral load carrying capacity of the ground floor, Q
g
can be expressed by Equation
(1.19) in the SSSM, and it should be greater than the base shear, V
t
.

9
=
0
( )
( )
g g g t g
a
A xIxS T
Q P xW ; Q V ; P
R T
(1.19)
where P
g
= seismic capacity index of ground floor; and W

=total weight of the building.
When a non-ductile (R
a
(T)=R=4.0), residential building (I=1.0) is considered, whose
natural period is relatively short and nearly matching with soils natural period (S(T)=2.5)
and located on seismic zone 1 (A
0
=0.40), its seismic capacity index of ground floor (P
g
)
should be equal to or greater than 0.25.

=

0.40 1.00 2.5
0.25
4
g
x x
P (1.20)
If we calculate the seismic index, I
s
for the same case with assuming structural irregularity
index, D and time dependent deterioration index, K equal to 1.0 we can say that I
s
should
be equal to or greater than 0.25. Then; the required seismic index, I
D
value should be
equal to 0.25 at least because the seismic index, I
s
value will be judged against to the
required seismic index, I
D
value. If we repeat the same conditions (B=1; Z=1; U=1) in
Equation (1.17) to calculate I
D
, the result implies that P
0
should be equal to 0.25 as the
first approach.
( ) ( )
= =
0
0.25 1.0 1.0 1.0 .1.0 1.0
s D
s D
I I
I x x I P x x x
(1.21)
The value of 0.25 is very conservative as a basic seismic demand index P
0
in case of
buildings with high ductility. For calibration of this index, a pilot study was carried out in
Zeytinburnu, Istanbul.
1.5 PILOT REGION STUDY
The SSSM is applied to 2401 buildings in Zeytinburnu, Istanbul. The cumulative
frequency distribution of the seismic capacity index values of these buildings are given in
Figure 1-3.

10
Cumulative Frequency Distrubution
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55
Seismic Capacity Index, Pmin
F
r
e
q
u
e
n
c
y
(
P
m
in
)

Figure 1-3. The cumulative frequency distribution of the minimum seismic capacity index, Pmin
values of 2401RC building in Zeytinburnu
For the calibration of the SSSM, the results of nonlinear static analysis carried out for 12
of these buildings are utilized. During the nonlinear analyses, actual (as-built) structural
characteristics of these buildings are used for input. The value of the seismic capacity
index, P has a fundamentally important effect on the decisions about the basic seismic
demand index, P
0
.
During all steps of nonlinear analysis, the displacement profile is taken from the shape of
the first mode in the corresponding direction. This assumption does not cause misleading
results, because 12 buildings among 2401 buildings were chosen such that the effects of
higher modes were minimal. Plastic hinge lengths of columns and beams were assumed as
half of the section depth (h/2). From the results of these analyses, base shear force
capacities corresponding to collapse prevention performance level were determined for
each building. Then the ratios of base shear force capacity to structural weight were
calculated. The ratios for x and y directions (Vx/W, Vy/W) and the structural seismic
capacity indices (Px, Py), which are the ratios of strength indices to the structural weight
determined by SSSM are presented in Table 1-2.
Table 1-2. V/W and P values obtained by pushover analysis and SSSM
Building Code Vx/W Px Vy/W Py
Min
(Vx/W; Vy/W)
Min
(Px,Py)
1597944 0.14 0.21 0.12 0.14 0.12 0.14
1610028 0.10 0.09 0.09 0.11 0.09 0.09
1597778 0.08 0.13 0.12 0.27 0.08 0.13
1597914 0.09 0.13 0.12 0.13 0.09 0.13

11
Building Code Vx/W Px Vy/W Py
Min
(Vx/W; Vy/W)
Min
(Px,Py)
1597818 0.07 0.06 0.07 0.10 0.07 0.06
1609263 0.07 0.07 0.08 0.08 0.07 0.07
1597776 0.14 0.15 0.15 0.24 0.14 0.15
1599432 0.09 0.10 0.12 0.11 0.09 0.10
1597884 0.10 0.11 0.09 0.13 0.09 0.11
1597799 0.10 0.11 0.10 0.12 0.10 0.11
1610164 0.18 0.28 0.22 0.23 0.18 0.23
1610166 0.11 0.17 0.08 0.12 0.08 0.12

The relationships between V/W and P values are shown in the Figure 1-4, below. As it
can be seen from these figures, the results obtained by using SSSM are similar to those
obtained by pushover analysis qualitatively. The minimum V/W values obtained by
pushover analysis are approximately 80% of the corresponding minimum P values
obtained by SSSM.
For the calibration of P
o
value, the structural performance indices for two principal
directions (P
x
, P
y
) are compared with the ratios of base shear capacities determined by
pushover analysis to the structural weight, (V
x
/W, V
y
/W). After the calibration, it is
determined that the main comparison index I
D
should be considered as 0.18.
min (Vx/W, Vy/W) and min (Px,Py)
y = 0.80x
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
0 0.05 0.1 0.15 0.2 0.25
min (Px, Py)
m
i
n

(
V
x
/
W
,

V
y
/
W
)

Figure 1-4. The relation between results obtained by using SSSM and pushover analysis
During pushover analyses, for seismic demand, the NEHRP acceleration spectrum for
50% exceedence probability in 50 years, is considered. S
s
and S
l
characteristic spectral
acceleration coefficients are determined by Bogazici University based on local seismicity

12
and ground conditions for each building. This information was used to estimate the index
for regional seismicity, B and index for soil-structure behavior (local soil conditions), Z
for the calculation of the required seismic index, I
D
. Usage index was taken as 1.0 since all
buildings are employed as residential purposes. Evaluation of the buildings using SSSM is
given in Table 1-3.
Table 1-3. Evaluation of the buildings using SSSM
Building
Code min(P
x
,P
y
)
Irregularity
Index D
Time Det.
Index K Seismic index,I
s
Required
seismic
index, I
D

1597944 0.14 1.00 1.00 0.14 0.15
1610028 0.09 1.00 1.00 0.09 0.15
1597778 0.13 1.00 1.00 0.13 0.14
1597914 0.13 1.00 1.00 0.13 0.15
1597818 0.06 1.00 1.00 0.06 0.15
1609263 0.07 1.00 1.00 0.07 0.15
1597776 0.15 1.00 0.95 0.14 0.15
1599432 0.10 1.00 1.00 0.10 0.14
1597884 0.11 0.95 1.00 0.10 0.14
1597799 0.11 1.00 1.00 0.11 0.14
1610164 0.23 0.90 1.00 0.21 0.15
1610166 0.12 0.90 1.00 0.11 0.18

The estimation of the seismic safety of the building is based on the comparison of the
seismic index I
s
and the required seismic index, I
D
. It should be noted that the decision on
the seismic safety is not based on the seismic capacity index, P and the basic seismic
demand index, P
0
.
According to the evaluation, for only one building out of twelve satisfies I
s
>I
D
condition.
This shows that for the seismicity of this region, only one building out of examined
twelve buildings has sufficient seismic safety. For determining the seismic performance of
the remaining 11 buildings, a decision should be given based on a more detailed analysis.
It is interesting to note that according to the pushover analyses results, only one building,
which was determined as earthquake resistant by the SSSM, had sufficient seismic safety.
For other buildings, which had I
s
values smaller than I
D
, the results obtained by pushover
analysis are also in good agreement with the results obtained by SSSM.

13
1.5.1 Application of SSSM to a Sample Building
Figure 1-5 shows the ground storey plans of sample building with material quality of C9
concrete and S220 steel.

Figure 1-5. Ground floor plan of sample building
A four storey reinforced concrete building structure having equal storey height of 2.70 m
is investigated in details for the purpose of comparing non-linear static analyses results
with the SSSM index values. The plan dimensions of the building are 10.10 m x 9.90 m.
The NEHRP acceleration spectrum exceedance probability being 50% in 50 years is
considered for seismic demand. Characteristic spectral acceleration coefficients, S
s
=
0.692 and S
l
= 0.412 are determined by Bogazici University based on local seismicity and
ground conditions for the building. Fundamental periods of the building, T
x
= 0.766 sec
and T
y
= 0.602 sec. Pushover analyses of the building were realised by ETABS Nonlinear
V.8. (Figure 1-6) then performance level of them according to the NEHRP and TERDC-
2006 are determined (Figure 1-7). The ratios of base shear force capacity to structural
weight are obtained as 0.135 and 0.149 in directions of X and Y, respectively.

14
1597776 X_DIRECTION PUSHOVER CURVE
0.0
100.0
200.0
300.0
400.0
500.0
600.0
700.0
0.000 0.050 0.100 0.150 0.200
TOP DISPLACEMENT(M)
B
A
S
E
S
H
E
A
R
(K
N
)
X_PUSHOVER CURVE BILINEARIZATION
1597776 Y_DIRECTIONPUSHOVERCURVE
0.0
100.0
200.0
300.0
400.0
500.0
600.0
700.0
0.000 0.050 0.100 0.150 0.200 0.250
TOPDISPLACEMENT(M)
B
A
S
E
S
H
E
A
R
(K
N
)
Y_PUSHOVERCURVE BILINEARIZATION

Figure 1-6. Pushover curve of residential building (direction x, y) and bilinear form
1597776XDIRECTIONCAPACITYDIAGRAMDEMANDESTIMATION
0.0000
1.0000
2.0000
3.0000
4.0000
5.0000
6.0000
7.0000
8.0000
0.0000 0.0200 0.0400 0.0600 0.0800 0.1000 0.1200 0.1400 0.1600 0.1800 0.2000
SPECTRALDISPLACEMENT(M)
S
P
E
C
T
R
A
L
A
C
C
E
L
E
R
A
T
IO
N
(M
/S
N
2
)
BILINEARIZEDX_PUSHOVER SPECTRUM CORR.LINEARSYS.
1597776YDIRECTIONCAPACITYDIAGRAMDEMANDESTIMATION
0
1
2
3
4
5
6
7
8
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
SPECTRALDISPLACEMENT(M)
S
P
E
C
T
R
A
L
A
C
C
E
L
E
R
A
T
IO
N
(M
/S
N
2
)
BILINEARIZEDY_PUSHOVER SPECTRUM CORR.LINEARSYS.

Figure 1-7. Capacity diagram, demand estimation of residential building (direction x, y)
SSSM is applied to the building with the help of MS-Excel. Excel sheet for this building is
shown in Figure 1-8. Data input part of the sheet contains number of storeys, height and
number of investigated storey, type of slab, areas of ground and normal floors, columns
and beams dimensions and effective areas for calculation of axial forces acting on
columns. Those data capture from the text file by macros, and then calculation of seismic
capacity index, P is started automatically by considering the assumptions and pre-
formulated cells. For determination of seismic index I
s
, structural irregularity index, D and
time deterioration index, K should be calculated for the building. Structural irregularity
index, D is calculated basically as 1.0 since building has neither irregularity, nor
complexity of structural configuration. Time deterioration index, K was considered as
0.95 because evenly corrosion of longitudinal bars inside some columns was reported.

15
Building ID No 1597776
Investigated Floor Ground
Story Height (m) 2.70
n = number of stories of a building 4 DATA INPUT
Type of slab slab-beam
Area of ground floor (m2) 99.99
Area of Normal Floor (m2) 100.28
i =investigated story's number 1
1 2 3 4 5 6 7 8 9
X DirectionY Direction +X-side -X-side +Y-side -Y-side
S1 25.00 40.00 2.70 0.00 50.00 50.00 0.00 0.69
S2 40.00 25.00 2.70 0.00 50.00 50.00 0.00 1.69
S3 25.00 40.00 2.70 0.00 50.00 50.00 0.00 2.03
S4 25.00 40.00 2.70 0.00 50.00 50.00 0.00 1.03
S5 25.00 40.00 2.70 0.00 50.00 50.00 0.00 3.12
S6 40.00 25.00 2.70 0.00 50.00 50.00 0.00 7.78
S7 40.00 25.00 2.70 0.00 50.00 50.00 0.00 9.40
S8 35.00 40.00 2.70 0.00 50.00 50.00 0.00 7.19
S9 25.00 40.00 2.70 0.00 50.00 50.00 0.00 3.34
S10 25.00 40.00 2.70 0.00 50.00 50.00 0.00 10.38
S11 25.00 75.00 2.70 0.00 50.00 50.00 0.00 9.79
S12 25.00 60.00 2.70 0.00 50.00 50.00 0.00 5.63
S13 25.00 60.00 2.70 0.00 50.00 50.00 0.00 13.10
S14 25.00 60.00 2.70 0.00 50.00 50.00 0.00 6.20
S15 25.00 40.00 2.70 0.00 50.00 50.00 0.00 2.21
S16 40.00 25.00 2.70 0.00 50.00 50.00 0.00 5.56
S17 25.00 40.00 2.70 0.00 50.00 50.00 0.00 6.67
S18 25.00 40.00 2.70 0.00 50.00 50.00 0.00 4.48
Column
No
DIMENSIONS (cm)
Height of
column
(m)
BEAMS' DEPTH (cm)
Tributary
Area (m
2
)

Figure 1-8. MS-Excel worksheet for application of SSSM to sample building (direction x, y)

16
10 11 12 13 14 15 16
Table
1.1
Table
1.1
o
h
ox
/D
x
h
oy
/D
y
Q
mux
Q
muy
Q
sux
Q
suy
0.3809 8.80 5.50 201.63 322.61 581.59 799.69
0.9318 5.50 8.80 409.63 256.02 843.76 625.66
1.1206 8.80 5.50 273.07 436.91 640.76 858.86
0.5686 8.80 5.50 220.94 353.50 596.60 814.70
1.7222 8.80 5.50 322.00 515.21 688.90 906.99
4.2946 5.50 8.80 701.50 438.44 1112.78 894.68
0.0000 5.50 8.80 1173.17 733.23 769.22 551.12
2.8349 6.29 5.50 766.01 875.44 1204.44 1295.93
1.8437 8.80 5.50 330.88 529.41 698.61 916.71
0.0000 8.80 5.50 733.23 1173.17 551.12 769.22
2.8822 8.80 2.93 737.11 2211.32 1284.06 2385.35
2.0718 8.80 3.67 519.99 1247.97 976.44 1639.52
4.8208 8.80 3.67 665.59 1597.41 1306.32 1969.40
2.2816 8.80 3.67 540.18 1296.42 1001.62 1664.69
1.2199 8.80 5.50 281.72 450.75 648.71 866.81
3.0691 5.50 8.80 642.76 401.72 1014.75 796.65
3.6818 8.80 5.50 424.35 678.95 845.67 1063.76
2.4730 8.80 5.50 371.52 594.44 748.96 967.05
w [kN/m
2
]
f
c
[MPa]
y
[MPa]
wy
[MPa]
t

w
d' [cm]
12 10 220 220 0.004 0.002 2.5
ASSUMPTIONS
Hoop diameter (mm) Hoop spacing (cm)
8 20
Eq.(1.10, 1.11,
1.12, 1.15)
Eq. (1.13)

Figure 1-9. (Cont.) MS-Excel Worksheet for application of SSSM to sample building (dir. x, y)


17
17 18 19 20 21 22 23 24 25 26 27 28
Eq.
(1.19)
Fig
1.2
Eq.
(1.19)
Fig
1.2
C
x
a
x
C
y
a
y
C
x
a
x
C
y
a
y
R
x
R
y
C
x
R
x
C
y
R
y

w

w
0.0036 1.0 0.0058 1.0 0.0036 0.0058 1.0 1.0 0.0036 0.0058 0.001 0.002
0.0074 1.0 0.0046 1.0 0.0074 0.0046 1.0 1.0 0.0074 0.0046 0.002 0.001
0.0049 1.0 0.0079 1.0 0.0049 0.0079 1.0 1.0 0.0049 0.0079 0.001 0.002
0.0040 1.0 0.0064 1.0 0.0040 0.0064 1.0 1.0 0.0040 0.0064 0.001 0.002
0.0058 1.0 0.0093 1.0 0.0058 0.0093 1.0 1.0 0.0058 0.0093 0.001 0.002
0.0127 1.0 0.0079 1.0 0.0127 0.0079 1.0 1.0 0.0127 0.0079 0.002 0.001
0.0139 1.0 0.0100 1.0 0.0139 0.0100 1.0 1.0 0.0139 0.0100 0.002 0.001
0.0138 1.0 0.0158 1.0 0.0138 0.0158 1.0 1.0 0.0138 0.0158 0.001 0.001
0.0060 1.0 0.0096 1.0 0.0060 0.0096 1.0 1.0 0.0060 0.0096 0.001 0.002
0.0100 1.0 0.0139 1.0 0.0100 0.0139 1.0 1.0 0.0100 0.0139 0.001 0.002
0.0133 1.0 0.0399 1.0 0.0133 0.0399 1.0 1.0 0.0133 0.0399 0.001 0.002
0.0094 1.0 0.0225 1.0 0.0094 0.0225 1.0 1.0 0.0094 0.0225 0.001 0.002
0.0120 1.0 0.0289 1.0 0.0120 0.0289 1.0 1.0 0.0120 0.0289 0.001 0.002
0.0098 1.0 0.0234 1.0 0.0098 0.0234 1.0 1.0 0.0098 0.0234 0.001 0.002
0.0051 1.0 0.0081 1.0 0.0051 0.0081 1.0 1.0 0.0051 0.0081 0.001 0.002
0.0116 1.0 0.0073 1.0 0.0116 0.0073 1.0 1.0 0.0116 0.0073 0.002 0.001
0.0077 1.0 0.0123 1.0 0.0077 0.0123 1.0 1.0 0.0077 0.0123 0.001 0.002
0.0067 1.0 0.0107 1.0 0.0067 0.0107 1.0 1.0 0.0067 0.0107 0.001 0.002
C
x
a
x
C
y
a
y
(C
x
a
x
)
*
(minR
x
)
0.15 0.24 0.15 0.24
P
x
P
y
Table
1.1
(C
y
a
y
)
*
(minR
y
)
Eq. (1.6, 1.7,
1.8)

Figure 1-10. (Cont.) MS-Excel worksheet for application of SSSM to residential building (dir. x, y)
In this example, seismic capacity index for the X direction is found as the minimum
value with the score of 0.15. From the results of non-linear analyses, the minimum value
of base shear force to weight ratio is obtained as 0.135 for the X direction, too.
According to TERDC-2006, the buildings performance level is lower than collapse
prevention level.

18
1.6 CONCLUSIONS
A general conclusion after applying Seismic Safety Screening Method (SSSM) to the
buildings that are examined by the nonlinear static analysis is that SSSM can be used for
estimating the existing seismic safety levels for medium/low rise RC buildings since the
results obtained by the screening method and the structural analysis are in good harmony.
The basic seismic demand index, P
0
value appeared to be 0.18 in the pilot region. But still
much effort is needed to apply this method to a large selection of RC buildings in
different seismic regions, and verify the results by the structural analyses.
Future development for improvement of this method will focus on approximate formulas
given in Equation (1.10) to (1.16), and adding some experimental studies on sub-
structural elements besides to estimate the structural irregularity index, D and time
dependent deterioration index, K.

2.SEISMIC UPGRADING OF STRUCTURES USING
CONVENTIONAL METHODS
Turkey is located at one of the active earthquake zones of the world and many
earthquakes have been occurred in the past resulting the life and property loses. For these
reasons, retrofitting or strengthening of existing buildings against the future earthquakes
is very important for avoiding the potential life and property loses. The problem is
especially important for reinforced concrete buildings which have been built during the
last sixty years according to the existing codes of their times. Most of these early codes
have considered the seismic loads as some percentage of the total live and dead loads of
the building. A linear and approximate method of analysis has been used under this type
of seismic load. Experiences from the past earthquakes in Turkey have shown that large
earthquakes can severely damage the existing buildings designed according to the old
code provisions. The research in this chapter is taken from deliverables 106 which is
available from www.lessloss.org. The purpose of Deliverables 106 was to reach the
reliability and performance of a low disturbance strengthening technique depending on
adding the new shearwalls to the building by using the non-linear analysis which was not
applied in common practice after the past four major earthquakes in Turkey. In the
report moreover, the non-linear behaviour and seismic performances of the strengthened
buildings are also evaluated considering the soil structure interaction using a convenient
analysis model.
2.1 EVALUATION OF SEISMIC SAFETY OF EXISTING BUILDINGS AND
RETROFITTING TECHNIQUES
The analytical and experimental studies on real behaviours of structures under seismic
loads have led to either revisions in code provisions or more realistic earthquake resistant
design of buildings built in seismic zones. On the other hand, It is a reality that there are a
number of existing buildings under seismic risk due to constructional problems such as
poor concrete quality, insufficient dimensions of structural elements, insufficient and
wrong placed reinforcement etc., in Turkey. Thus, first, determination of the seismic
safeties of existing buildings and then developing the economical and practical retrofitting
techniques for the existing buildings which have insufficient seismic safety are very
important.

20
Pushover analysis method is an effective tool used for determination of seismic safety of
existing buildings. In this method, mathematical model of the existing structure is
analyzed under the monotonically increasing gravity and earthquake loads, and plastic
hinge formations are observed until the structure is collapsed. The results of this type of
analysis are very important from the viewpoint of design engineer, since they show the
weak parts or potential damage points of the structure and give an idea for existing
seismic safety.
2.1.1 The Procedure for Determination of the Seismic Safety of Existing
Buildings
The procedure consists of the following steps:
i) Providing and evaluation of the documents of the existing building
ii) Inspections on the existing building.
iii) Preparation of the mathematical model of the existing structure
iv) Pushover analysis of the mathematical model of the existing building.
v) Evaluation of the seismic safety of the building.
vi) Steps of the given procedure is explained in Deliverable-106 in detail.
2.2 OBSERVATIONS ON IMPERFECTIONS CAUSING THE COLLAPSE OR DAMAGE IN
RESIDENTIAL BUILDINGS AND COMMON STRENGTHENING TECHNIQUES IN
TURKEY
Frequently observed imperfections causing the collapse or heavy damages in residential
buildings from the past earthquake experiences such as Erzincan (1992), Dinar (1995),
Ceyhan (1998), Kocaeli and Dzce (1999), are as follows:
inappropriate regulation of lateral load carrying structural system elements,
inadequate reinforcement and poor detailing,
foundation and geotechnical problems,
very poor strength of concrete,
poor workmanship, short columns,
soft storeys, weak column- strong beam connections,

21
lack of construction control organization in Turkey.
As the main part of the building stock in Turkey are suffered from the deficiencies
mentioned above, strengthening the existing buildings by an economic, efficient, rapid
and low disturbance technique against to the future earthquakes are of great importance.
Some common upgrading (retrofitting or strengthening) techniques are summarized as
follows:
There are a lot of strengthening techniques have been considered and applied in the
actual practice of the existing reinforced concrete structures. Some of them are in the
following:
i) reinforced concrete jacketing ,
ii) epoxy bonded steel plates,
iii) steel bracing,
iv) steel frames and trusses,
v) pre-cast reinforced infill walls,
vi) additional cast-in-place concrete concentric or eccentric shear walls and new
foundations,
vii) wrapping the existing structural elements with FRP,
vii) application of the base isolation and damping devices.
In Turkeys experience and practice however, even if it is expensive, upgrading the
existing structures by additional cast-in-place RC shearwalls is found to be more effective
technique comparing the other alternative upgrading techniques for the following
reasons:
i) From the experiences of past earthquakes, the collapsed residential buildings
generally had 3-5 storey and collapses were occurred due to the storey mechanism called
as pan-cake mechanism.
ii) The most of the buildings do not have a design project prepared by an engineer or an
architect.
iii) The concrete strength is very poor due to the non-proportional mixture of water,
aggregate and cement. Moreover, the mixture includes sea sand.
iv) The buildings have bad formwork and reinforcement placement workmanship.
v) Lateral load carrying elements have inadequate stirrups and inadequate concrete
covers.

22
vi) The buildings have many irregularities in plan and in view, most of them have short
columns and soft storeys.
vii) The columns of the basements are faced the intensive corrosion problem due to the
damp, especially in the buildings of stanbul and its vicinity.
viii) It is the lowest disturbance upgrading method among the other alternatives.
2.2.1 Strengthening the Existing Buildings by Additional Shearwalls
In Turkey most of the low-rise residential RC buildings are not well designed and
constructed, thus, lateral rigidities and lateral load carrying capacity of them are not
sufficient. Furthermore, as collateral type of construction is very common, the storey
drifts should be limited. In order to overcome the mentioned deficiencies above, adding
new shear walls to the existing structure has been preferred as the most common
upgrading technique in Turkey.
If new added shear walls are selected as an upgrading technique, for the sake of
completeness the positive contributions of them to the structural behaviour have been
given below:
i) storeydrifts will be reduced, thus, partitioning walls made of brittle material can not
be damaged due to excessive storey drifts,
ii) storeymechanism will be diminished,
iii) short column effects can be eliminated,
iv) efficiency of the moment-resistant frames will be reduced so that beam-column
connections will not be effective in controlling the lateral load behaviour of the structure,
v) by placing the suitable orientation of the additional shear walls, some of the plain
and elevation irregularities of the existing structure can be eliminated as much as possible,
vi) existing lateral load carrying elements of the structure will keep their reserve strengths
since the additional shear walls carry the biggest portion of the lateral load.
The major and common difficulty is the integration of the newly added shear walls to the
existing structure. In general, since the concrete strength of the existing structures in
Turkey is very poor, it becomes very difficult to have shear connectors to join the
additional shearwalls to the existing one. In practice, it is very convenient to design the
new shear walls as they cover the existing beams and columns for full integration. Epoxy
injection may be needed for bonding between the new and existing structural elements.
Construction procedure for additional shearwalls will be given in the next chapter. The
designing and construction of the foundations of additional shear walls are another

23
problem since the axial loads transferred to the additional shear walls are not big enough.
Rotation of the foundations cannot be restricted, unless they are integrated to the other
existing foundations. For this reason, adequate transfer of internal forces must be
satisfied by picking up newly added shear walls in order to limit rotation of foundation
system.
2.2.2 Construction Rules for Additional Shearwalls
Strengthening by adding the new shearwalls to the existing structure may increase either
the seismic safety of the building and also limits the drift ratio. But, as the new shear walls
increase the total stiffness of the building, total earthquake load and its distribution over
the new and existing structural elements will change. Thus, the existing structural
elements should be checked by the new internal force distribution and they should be
also retrofitted if necessary.
The following measures should be considered in strengthening by new shear walls:
i) At least two shear walls should be placed in each direction. The dimensions of the
shear walls should be selected by considering the total plan area of the building and the
number of the storeys. Approximately 80 % of the total seismic base shear should be
resisted by the new added shearwalls. Furthermore, wall dimensions should satisfy the
minimum requirements imposed by the code.
ii) Shearwalls may be placed at the external axes, blind sides and the stairwell sides of
the building. Other internal axes may also be selected as shear wall places if the
architectural provisions permit and as not to interfere the buildings function.
iii) New shearwalls should provide sufficient torsional rigidity to the existing structure.
iv) Shear walls should be placed between the two adjacent existing columns and they
should also surround the columns for transmitting the existing gravity loads of the
structure to the foundation.
v) The connections of the new and existing structural elements such as shear wall-
column, shear wall-beam and shear wall-floor connections should be carefully built and
the anchorages should be designed for the integration between existing and new
structural elements.
vi) The new foundations for the new shear walls should be designed and joined with the
existing ones for full integrity.
vii) New shearwalls should also allow for the economical design of the foundation
system.

24
viii) 3D analysis should be performed for the design of strengthened building and actual
stiffness properties of existing and new structural elements should be used in the analysis.
2.3 NUMERICAL EXAMPLE
In this chapter, performance levels of a sample strengthened building chosen from the
Deliverable-106 are evaluated considering various code provisions including ATC-40,
FEMA-440 and TEC-2006 and conclusions on the obtained results are given. Soil-
structure interaction is included for the performance analyses using non-linear static
analysis by a mat foundation on tensionless Winkler soil model. Non-linear pushover
analyses and performance evaluations are performed by SAP2000 computer package. The
calculations are performed for two major earthquake directions X and Y, but only the
results for the X direction are given for the sake of brevity.
The sample building has a roof storey in 2.5 m, ground storey in 3.15 m height and three
normal storeys in 2.75 m heights each one and plan dimensions of the building are 10.20
m x 6.80 m. Typical storey plans for the existing and strengthened buildings are given in
Figure 2-1 and Figure 2-2. Two shearwalls are added to the building for strengthening in
each direction. 95% of the total base shear is carried by the additional shearwalls and the
remaining is carried by the existing elements. The coefficients for the equivalent lateral
load calculations are taken as effective ground acceleration, (Ao) = 0.40, intensity factor
of design earthquake, E = 1.0, building importance factor, (I) = 1.0, Soil class = Z3,
characteristic spectral periods, T
A
= 0.15 sec, T
B
= 0.60 sec.
For the performance evaluations of the building with and without soil-structure
interaction, two foundation models are considered. For the first model, the foundations
of the building are considered as fixed base. At the other hand, the second one is
considered that it has a mat foundation on tensionless Winkler foundation. The
mathematical model used for the analysis is shown in Figure 2-3. Fundamental period of
the fixed base building at the X direction is T
x
= 0.343 sec. and the one for the mat
foundation model is T
x
= 0.414 sec. The total storey weights, masses and the Equivalent
Earthquake Loads of the strengthening building calculated according to the TEC-2006
are given in Table 2-1. The Equivalent Earthquake Loads of the strengthening buildings
with fixed base and with mat foundation are the same since the spectrum coefficients are
equal to 2.5.


25

Figure 2-1. Normal storey plan of existing building

Figure 2-2. Normal storey plan of strengthened building

26

Figure 2-3. The mathematical model used for the analysis of building with mat foundation
Table 2-1. Equivalent earthquake loads of strengthened building (TEC-2006)
StoreyNo Total storey
Weight
(kN)
StoreyMass
(kNs
2
/m)
Equivalent
Earthquake
Load (kN)
(Direction X)
4 448.20 45.60 194.20
3 878.10 89.50 312.20
2 878.10 89.50 236.80
1 878.10 89.50 161.50
Ground 881.60 89.80 86.20
Total 3964.10 990.90


27
2.3.1 Performance Evaluations of Strengthened Building According to Non-linear
Push-over Analysis Methods of TEC-2006, FEMA-440 and ATC-40
Performances evaluations of the strengthened building by non-linear pushover method
are carried out by the PPS-2006 (explained in D106) excel macro according to the TEC-
2006. The same evaluations according to FEMA-440 and ATC-40 are performed by using
SAP2000 package. Pushover analyses are also carried out by SAP2000. The mathematical
model of the mat on tensionless Winkler foundation is given in Figure 2-3. According to
the model, the mat discretized by thick plate finite elements and tensionless Winkler soil
is represented by equivalent frame members joining the plate nodes and subjected to only
axial forces. Tension limit for the frame members representing the tensionless soil is
equal to zero and compression limit is equal to allowable soil stress under earthquake
effects. The allowable soil stress under earthquake effects are taken as 1.5 times of the
one under gravity loads. The pushover curves and performance point calculation graphics
according to TEC-2006, FEMA-440, and ATC-40, are given Figure 2-4 to Figure 2-11
respectively. The soil modulus is taken as k
s
=30000 kN/m
3
and allowable soil stress is
taken as q=125 kN/m
2
*1.5 convenient with the Z3 class soil. The mat thickness is equal
to 0.7 m and it is divided into 192 shell finite elements with approximately 0.95mx0.95m
dimensions.

Figure 2-4. Pushover curve of strengthened building with fixed base

Figure 2-5. Pushover curve of strengthened building with mat foundation

28
TEC 2006 - X DIRECTION
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Spectral Displacement, Sd [cm]
S
p
e
c
t
r
a
l A
c
c
e
le
r
a
t
io
n
, S
a
/g
[
c
m
]
Capacity Spectrum Bilinear Curve
Design Spectrum Tangent of The Bilinear Curve
Performance Point

Figure 2-6. Performance graphic of strengthened building with fixed base (TEC-2006)
TEC 2006 - X DIRECTION
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Spectral Displacement, Sd [cm]
S
p
e
c
t
r
a
l A
c
c
e
le
r
a
t
io
n
, S
a
/g
[
c
m
]
Capacity Spectrum Bilinear Curve
Design Spectrum Tangent of The Bilinear Curve
Performance Point

Figure 2-7. Performance graphic of strengthened building with mat foundation (TEC-2006)

Figure 2-8. Performance graphic of strengthened building with fixed base (FEMA-440)

29

Figure 2-9. Performance graphic of strengthened building with mat foundation (FEMA-440)

Figure 2-10. Performance graphic of strengthened building with fixed base (ATC-40)

30

Figure 2-11. Performance graphic of strengthened building with mat foundation (ATC-40)
The soil stresses at the performance points and design earthquake load levels according to
FEMA-440 and ATC-40 are also given in Figure 2-12 to Figure 2-14. Furthermore,
comparison of performance points calculated according to three different codes are given
in Table 2-2, considering with and without soil-structure interaction.

Figure 2-12. The soil stresses at the performance point and design earthquake load level

Figure 2-13. The soil stresses at the performance point and according to FEMA-440

31

Figure 2-14. The soil stresses at the performance point and according to ATC-40
From the evaluation of Figure 2-12, Figure 2-13 and Figure 2-14 that, allowable stresses
are not exceeded and there is not any tension stress on the winkler soil at the design
earthquake level. However, at the performance point level, the soil stresses in
compression zone increase and tension stresses begin to occur at the other edge of the
mat foundation.
Table 2-2. Comparison of the performance points of the strengthened building with fixed base and
at foundation
Displacement at the
performance point [cm]
Base shear at the performance
point [kN]

Fixed Base Mat
Foundation
Fixed Base
Mat
Foundation
TEC-2006 8.14 11.31 1630.00 1701.04
FEMA-440 8.00 12.60 1663.00 1714.60
ATC-40 17.90 18.30 1755.00 1749.90

It can be shown from the Table 2-2, displacement demands of the building with mat
foundation at the performance points obtained from TEC-2006 and FEMA-440 increase
about 30%, however, the base shears at the same points increase only 4%, comparing to
the building with fixed base. This situation may be explained as the deformations of mat
foundation are small and linear, i.e. the behaviour of the mat is approximately rigid.
Therefore, there is not significant change between the lateral load carrying capacities of
the buildings with and without mat foundation. However, displacement demand of the
building with mat foundation is larger than that of fixed base, due to the rotations at the
foundation. Furthermore, according to the performance evaluation by ATC-40, the
performance points of the buildings with and without mat foundation are similar, since
the effect of the mat foundation to damping of the building is rather small. The
parametric studies about other strengthened buildings explained in Deliverable 106 are
still going on. More realistic results would be obtained and more accurate evaluations
would be given at the end of the studies.

32
2.4 CONCLUSIONS
In this section, seismic performance of a sample building strengthened by added
shearwalls which is the common strengthening technique in Turkey, experienced from
the past earthquakes, is evaluated. The building strengthened by two added shearwalls in
each major earthquake direction demonstrates a linear elastic behaviour at the design
earthquake level. In this case however, it resists the earthquake loads with minimal
damage. Performance evaluations have been comparatively made according to the three
current codes TEC-2006, FEMA-440 and ATC-40. The building strengthened by added
shearwalls has been in the life safety range given in the codes for design earthquakes
considering the drift ratios. From the comparison of the three codes, ATC-40 is found to
be more conservative than FEMA-440 and TEC-2006. However, TEC-2006 and FEMA-
440 is more realistic as it has been regulated to remedy the criticisms on equivalent
linearization method in ATC-40. Soil-structure interaction does not affect the
performance point coordinates obtained from ATC-40 when the mat foundation is
considered. The reason is that the mat foundation is relatively rigid comparing to soil and
there are not significant changes in energy dissipations between the buildings with and
without mat foundation due to the linear elastic behaviours of mat and Winkler soil.
General design and construction rules for strengthening by added shearwalls are also
given in this section. Finally, it is concluded that, the strengthening technique by
additional shearwalls is a reasonable and low disturbance technique for the common
existing buildings in earthquake risk areas of Turkey.

3.SEISMIC UPGRADING OF STRUCTURES USING FIBRE
REINFORCED POLYMERS
3.1 GUIDELINES FOR THE APPLICATION OF FRP RETROFITTING
3.1.1 Introduction
Composites materials are defined as the combining of two or more materials physically
different and mechanically separable, where the blend is made in such a way that the
dispersion of one material in the other may be carried out in a controlled fashion to
achieve optimum properties, obtaining as a result a new material with higher and possibly
unique properties in some specific aspect, in comparison with the original, separate
components.
The fibres and matrices of advanced composite materials may be combined by using a
variety of manufacturing processes, according to their destination. Thus, the orientation
of the fibres, the number and size of filaments and complexity of the fibres are chosen
amongst other things.
The best use possibly known is in the aerospace industry, although they spread their field
of application to the automotive, biomedical and sports articles manufacturing industries.
Moreover, their mechanical, chemical and physical characteristics have led to their being
used in the construction industry as means for rehabilitating, repairing and strengthening
structures.
The properties of composite materials depend on the type of materials of which they are
made. Reference will be made hereafter to the characteristics of the fibres and resins used
and, then to FRPs. Everything described hereafter will focus on the three types of fibre
usually used in civil engineering; carbon fibre, glass fibre and aramid fibre.
3.1.2 Materials characterisation
Before defining some general guidelines on the application of FRP retrofitting, it is
essential to briefly define the basic characterisation of FRP materials, in order to provide
the lecturer with their basic properties. Thus, we will be following distinguishing between
types of fibres, resins and the laminates.

34
3.1.2.1 Fibres
There are several grades of carbon fibre (or of glass fibres or aramid fibres). Table 3-1
shows several types of fibre with their typical properties.
There are several remarks to be made on Table 3-1. Carbon fibres have the highest
modulus of elasticity, but both glass fibres and aramid fibres have a greater tensile
strength. The greater tensile strength of glass fibres and aramid fibres (compared with
carbon fibres) does not turn into greater strength of the composite material, since glass
fibres are highly sensitive to small defects, which may heavily reduce their strength. Both
stiffness and compressive strength are very difficult to estimate in fibres, since they are
very difficult to experimentally test.
It is impossible to draw a conclusion from Table 3-1 of which of the three types of fibre
is structurally most efficient, due to this depending on the surrounding conditions; in
addition, the weight of the composite material will depend on the fibre concentration in
the matrix. Every individual case, either retrofitting or new structure design, must be
individually analysed to choose the optimum and most efficient fibre type.
Table 3-1. Typical values of fibre properties [FIB-CEB, 2004]
Material
Modulus
of
elasticity
(GPa)
Tensile
strength
(MPa)
Ultimate
tensile
strength (%)
Density
(g/cm3)
Heat transfer
coefficient
(x10-6 C-1)
Carbon 1.5-1.6 Long: -1-0
High strength 215-235 3500-4800 1.4-2.0 Trans: 22-50
Ultra high
strength.
215-235 3500-6000 1.5-2.3

High modulus 350-500 2500-3100 0.5-0.9
Ultra high
modulus
500-700 2100-2400 0.2-0.4

Glass 1.2-2.1 Long: 6-10
E 70 1900-3000 3.0-4.5 Trans: 19-23
S 85-90 3500-4800 4.5-5.5
Aramid 1.2-1.5 Long: -6-(-2)
Low modulus 70-80 3500-4100 4.3-5.0 Trans: 60-80
High modulus 115-130 3500-4000 2.5-3.5

35
3.1.2.2 Resins
The properties of resins depend on their internal structure. Table 3-2 hereafter shows the
properties of some typical resins.
As can be seen by comparing Table 3-1 and Table 3-2, most resins contribute very little
to the load capacity of composite materials. In any event, resins may be highly important
in the toughness of composite materials; the more ductile the resin is, the more laminates
toughness will increase. A resins function is to transfer a fibres load to an adjacent one
so that, when a fibre breaks, a ductile resin distributes the force better to several fibres
and not only to the adjacent fibre. This means to say that the laminate will resist more
broken fibres by increasing the toughness of the composite material
Table 3-2. Typical values of resin properties with different materials [FIB-CEB, 2004]
Property (at 20C) Cured epoxy adhesive Concrete Steel

Density (g/cm3) 1100-1700 2350 7800
Modulus of elasticity (GPa) 0.5-20 20-50 205
Shear modulus (GPa) 0.2-8 8-21 80
Poisson ratio 0.3-0.4 0.2 0.3
Tension strength (MPa) 9-30 1-4 200-600
Shear strength (MPa) 10-30 2-5 200-600
Compressive strength (MPa) 55-110 25-150 200-600
Breaking strain (%) 0.5-5 0.015 25
Breaking energy (Jm
-2
) 200-1000 100 10
5
-10
6

Coefficient of expansion (x10
-6
C
-1
) 25-100 11-13 10-15
Vitreous transition temperature (C) 0.1-3 5 0
Water adsorption: 7 days 25C (% w/w) 45-80 - -

3.1.2.3 Lamina and Laminates
A lamina, or ply, is a plane (or curved) layer of unidirectional fibres in a matrix. A
laminate is made up of two or more unidirectional laminae or plies stacked together at
various orientations. The laminae can be of various thicknesses and consist of different
materials.
Most laminates are not single direction but have a variety of orientations in fibres. To
illustrate how it affects composite materials, table 5 shows the strengths calculated for a
Glass/Polyester laminate with several orientations. The notation used for the type of

36
fibre is [
n
/
n
/]s , where is the orientation of the fibres, n is the number of layers
with that orientation and the letter s outside the brackets shows that the layers are
repeated symmetrically.
Table 3-3. Effect of fibre orientation on the tensile strength of a laminate depending on the
material [ACI 440.2R-02].
Direction 0 Direction 90
Type of FRP and
orientation
Modulus
of elasticity
(GPa)
Breaking
tension
(MPa)
Modulus
of elasticity
(GPa)
Breaking
tension
(MPa)
Breaking
strain at
direction 0
High strength carbon /epoxy
[0
4
] 100-140 1020-2080 2-7 35-70 1.0-1.5
[0
1
90
1
]
s
55-76 700-1020 55-75 700-1020 1.0-1.5
[45
1
45
1
]
s
14-28 180-280 14-28 180-280 1.5-2.5
Glass E/epoxy
[0
4
] 20-40 520-1400 2-7 35-70 1.5-3.0
[0
1
90
1
]
s
14-34 520-1020 14-35 520-1020 2.0-3.0
[45
1
45
1
]
s
14-21 180-280 14-20 180-280 2.5-3.5
High strength aramid/epoxy
[0
4
] 48-68 700-1720 2-7 35-70 2.0-3.0
[0
1
90
1
]
s
28-34 280-550 28-35 280-550 2.0-3.0
[45
1
45
1
]
s
7-14 140-210 7-14 140-210 2.0-3.0

3.1.3 Design of the FRP reinforcement
The calculation process to design FRP reinforcements follows a number of steps which
are commented here below, and which pretends to provide the designer with a number of
applicable rules.
i) Verification of the unreinforced section: checks are carried out on the current status
of the unreinforced cross section and whether it is necessary to reinforce it or not, as per
the cross section global safety coefficient.
ii) Sizing of the reinforcement cross section: the required reinforcement cross section is
defined in order that the section achieves a proper global safety coefficient.
iii) Sizing of the reinforcement length: we calculate the required laminate length in order
to transfer the grade load between the concrete cross section and the polymer composites
cross section.

37
Verification of the unreinforced section
To check the unreinforced section of the beam we shall use the ULS theory on the lower
part of the unreinforced section. To see if the section is suitable for strengthening, we
shall calculate the global safety coefficient (C.G.S.) of the section by means of calculating
the depth of the neutral fibre with relation to the upper one and the moment that the
section can withstand, as a standard reinforced concrete calculation.
Thus, making use of the safety coefficients of the standard and the resisting moment
(M
r
), this must be compared with the design moment (M
d
):
If M
d
>M
r
: The structure is safe and no strengthening shall be required (C.G.S.>1)
If M
d
<M
r
: It is necessary to strengthen the structure (C.G.S.<1).
Sizing of the reinforcement cross section
To calculate the size of the strengthening strip we have to insert in the beam lower
section we shall use the Ultimate Strength theory of reinforced concrete. We shall assume
for all intents and purposes that the strengthening strip placed on the lower part behaves
like the reinforcement embedded in the concrete, establishing two possible failure modes:
Concrete failure
Strengthening strip failure
The calculation procedure shall be as follows: In both cases, we shall initially calculate the
depth of the neutral fibre in the cross section and then the strains in the upper and lower
fibres.
If we have assumed that the section undergoes a concrete failure, the strain in the upper
fibre shall amount to 3.5 per thousand and the strain in the lower fibre shall be smaller
than the strengthening ultimate stress; if we have assumed that a strengthening failure
occurs in the section, the strain in the lower fibre shall be the same as the ultimate one in
the strengthening and the strain in the upper fibre shall be smaller than the concrete
ultimate strain (3.5 per thousand).
Whatever the section failure case, we shall use the following formulation to calculate the
depth of the neutral fibre, the fibre strains and finally the section resisting moment.
Consequently, once we have described the general procedure, the sizing procedure is
based on the performance of simple iterations, consisting in increasing the thickness of

38
the strengthening in proportion to the laminate thickness until the calculation moment is
greater than the resisting moment, at which time we shall stop the iterations and obtain
the required strengthening thickness
3.1.4 Adhesiveness
This section will provide an overview on what a structural adhesive is, the preparation of
the surface of the element in needs of reinforcement and the adhesion mechanisms that
rule the adhesion process.
3.1.4.1 Structural adhesives
Structural adhesives are used in conjunction with structural materials and operate in such
a way as to prevent the joint, or bonding line, from failing when the material is subject to
loads of up to its yield strength, while allowing the structure to accommodate its
mechanical and physical requirements. In general, these adhesives consist of two
polymers, one thermoset (modified or otherwise) and another thermoplastic, including
some elastomers.
The border line between structural and non-structural adhesives is not precise, but it is
commonly accepted the definition of the structural adhesive being that adhesive which
exceeds a tensile load of 10 MPa, at the ambient temperature
The structural systems of the adhesives are variable, each presenting different properties,
what makes the choice of the proper adhesive a relevant issue. The following table (Table
3-4) provides a comparison among the most commonly used material-based adhesives.
Table 3-4. Comparative properties of adhesives
STRUCTURAL SYSTEMS COMPARATIVE ASSESSMENT
SYSTEM ADVANTAGES DISADVANTAGES
Epoxy High final bond
Durability
Any thickness
2% shrinkage
Surface preparation
Mix life
Curing cycles
Dosing and mixing of components
Polyurethanes Flexibility
Better for plastics
Lesser final bond than epoxy
Worse for glass and metals
Acrylates Good final bond
Ease of application
Quick curing
Surface preparation
Layer thickness
7% shrinkage
Durability

39
3.1.4.2 Surface preparation
The quality of the bonded union depends both on the quality of the adhesive and on the
preparation of the substrata surfaces. The more thorough the cleaning of the surfaces, the
greater the saturation of the adhesive, as no grease or dirt is present. When structural
bonding is pursued, this cleaning operation is crucial.
However, some metallic materials (due to the presence of rust on the surface) or plastics
(due to their apolarity) must be subjected, in addition to the cleaning operation, to a
surface treatment.
Generally speaking, these preparations read as follows:
Mechanical treatment: sanding, steel shot blasting, sandblasting, etc.
Surface chemical treatment (chemical attack).
Miscellaneous treatments: electrical (corona discharge), flaming, priming.
However, the level of cleanliness required before proceeding with a bonding operation
depends both on the substrata and on the adhesive used. Thus, some adhesives, such as
hotly-applied epoxy-based ones, play a certain dissolving role of the grease that might
exist on the substrata surfaces. The same holds true for acrylates cyano-acrylates, which
do not contain dissolvents, but their chemical composition is of a dissolvent nature,
provided the polluting substance of the substrata does not inhibit curing. Some plastisoles
used in the construction industry are directly applied to greased materials, such as steel.
The recommended surface treatment depends upon the surface conditions, which, in
turn, is a function of the type of material and of its previous history. As a general rule, the
substrata composition is always different from that of the base material.
The simplest of the treatments are based on the cleaning action of dissolvents or
abrasives to remove pollution which, otherwise, would prevent the wetting of the base
material by the adhesive.
Other treatments have been developed further and tend to increase adhesion by
encouraging the process whereby adhesion takes place. That is the case of chemical
treatments, which modify the internal structure, from a chemical and physical point of
view, to increase the specific adhesion properties. The formation of surface roughness
serves to increase the adhesive mechanical adhesion.

40
Often, the prepared surfaces can be preserved for subsequent use, by covering them with
a removable protection layer. Furthermore, it is recommended that the treated
substratum is covered with adhesive as soon as possible.
3.1.4.3 Adhesion mechanism
In order to achieve a good bonded union, a free energy reduction must take place when
adhesive and substratum come into contact. It shall therefore be necessary for the
adhesive to perfectly wet the surface of the material. In this way, we arrive at the
interdiffusion of the molecules between the faces of the contact.
Generally speaking, the adhesion can take place in accordance with one or more of the
following items:
Mechanical adhesion: the adhesive enters the pores, inside which it solidifies. Better
results are obtained if, on top, surfaces are also rough.
Electrochemistry: based upon electrostatic phenomena between materials having this
type of affinity with the adhesive.
The ideal scenario for a union would be that in which the two types of adhesions take
place. This is not always possible with the emergence of a combination of both
mechanisms.
It should also be taken into account that, in the case of thin layer structural adhesives, the
adhesion forces between adhesive and substrata are greater than the adhesive internal
cohesion forces. For this reason, attempts should be made, whenever possible, to look
for thin adhesive thicknesses (smaller than 0.2 mm) so that the impact of the former over
the latter prevails. In this way, we shall accomplish the transmission of stresses, without
cohesive rupture of the adhesive.
There is, however, the case of certain types of adhesives, such as polyurethane-based
adhesives which require a given layer thickness (between 3 and 5 mm) to perform its
adhesion role. These types of unions lose stiffness and also resistance against the strictly
structural ones. However, they are very useful where vibrations, irregularities and noise
should be absorbed between faces.
3.1.5 Placing on site
On this section we will summarize the most commonly construction processes for FRP
components manufacturing.

41
3.1.5.1 Wet lay-up/Hand lay-up
Resins are impregnated by hand into fibres, which are in the form of woven, knitted,
stitched or bonded fabrics. This is usually accomplished by rollers or brushes, with an
increasing use of nip-roller type impregnators for forcing resin into the fabrics by means
of rotating rollers and a bath of resin. Laminates are left to cure under standard
atmospheric conditions.
The main advantage of this method is the facility in use, its wide application along the
years and it can be used with most of the fibres.
Its main disadvantages are that the resin mixing, laminate resin contents, and laminate
quality are very dependent on the skills of laminators, and it is therefore to count with a
team of skilled professionals.
3.1.5.2 Vacuum bagging
Vacuum bagging procedure is an extension of the wet lay-up process described above
where pressure is applied to the laminate once laid-up in order to improve its
consolidation. This is achieved by sealing a plastic film over the wet laid-up laminate and
onto the tool. The air under the bag is extracted by a vacuum pump and thus up to one
atmosphere of pressure can be applied to the laminate to consolidate it.
It presents an additional advantage to the above methodology, and it is that higher fibre
content laminates can usually be achieved than with standard wet lay-up techniques.
But labour worker must be higher skilled than in the previous process, and it is produced
an extra-cost due to the extra-material and equipment required.
The vacuum bagging process is one of the most commonly used processes in the
composites industry.
3.1.5.3 Filament winding
This process is primarily used for hollow, generally circular or oval sectioned
components, such as pipes and tanks. Fibre tows are passed through a resin bath before
being wound onto a mandrel in a variety of orientations, controlled by the fibre feeding
mechanism, and rate of rotation of the mandrel.
The main advantages of this method are clear: resin content can be controlled; structural
properties of laminates can be very good since straight fibres can be laid in a complex
pattern to match the applied loads.

42
However, it also presents disadvantages, such as the fact that this process is applicable
only in convex shaped components, fibre cannot easily be laid exactly along the length of
a component and the mandrel cost for large components can be quite high.
3.1.5.4 Prepregs
Fabrics and fibres are pre-impregnated by the materials manufacturer, under heat and
pressure or with solvent, with a pre-catalysed resin. The catalyst is largely latent at
ambient temperatures giving the materials several weeks, or sometimes months, of useful
life when defrosted. However to prolong storage life the materials are stored frozen. The
resin is usually a near-solid at ambient temperatures, and so the pre-impregnated materials
(prepregs) have a light sticky feel to them, such as that of adhesive tape. Unidirectional
materials take fibre direct from a creel, and are held together by the resin alone. The
prepregs are laid up by hand or machine onto a mould surface, vacuum bagged and then
heated to typically 120-180C. This allows the resin to initially reflow and eventually to
cure. Additional pressure for the moulding is usually provided by an autoclave (effectively
a pressurised oven) which can apply up to 5 atmospheres to the laminate.
Main advantages of prepregs are that materials have excellent health and safety
characteristics and are clean to work with and fibre cost is minimised in unidirectional
tapes since there is no secondary process to convert fibre into fabric prior to use.
Main disadvantages are the cost of the material, which it is usually higher, and that usually
autoclaves are required.
3.1.5.5 Resin Film Infusion (RFI)
Dry fabrics are laid up interleaved with layers of semi-solid resin film supplied on a
release paper. The lay-up is vacuum bagged to remove air through the dry fabrics, and
then heated to allow the resin to first melt and flow into the air-free fabrics, and then
after a certain time, to cure.
Main advantages are that high fibre volumes can be accurately achieved with low void
contents and good health and safety and a clean lay-up, like prepreg.
Main disadvantages are the fact that this is method is not widely proven outside the
aerospace industry and that tooling needs to be able to withstand the process
temperatures of the resin film (which if using similar resin to those in low-temperature
curing prepregs, is typically 60-100C).

43
3.1.6 Quality control
Last point on this section deals with the quality control, which not only becomes an
important issue, but it can be essential to assure the structural integrity of the designed
component or reinforcement.
During the fabrication of the different FRP applications on the projects, a number of
control points, to ensure the quality of the methods used, shall have to be considered.
These control points are divided into three categories.
Fabrication procedures of the different structural components.
Workmanship of the unions to the structure.
Testing of the materials used on site to ensure compliance with the design
hypothesis.
3.1.6.1 Check-points during installation
Materials.
- Compliance with the characteristics required by the project
- Quality certificates
- Ensure that the handling and delivery of the materials comply with the
recommendations made by both the designer and manufacturer.
Surface treatment
- Ensure that the concrete surface has been prepared in accordance with
the project specifications.
- Check for a regular surface
- Always try to obtain rounded corners (with a minimum radius of 3cm
where practicable)
- Check that the FRP surface has been properly treated.
Adhesive
- Determine the right amount of adhesive
- Have the different adhesive components been properly mixed
- Make sure that the application operations are completed while the
adhesive is workable. The temperature at which the procedure is carried
out will have a strong influence.

44
- See that the adhesive layer thickness is the correct one.
- Verify the level of vibration of the structure while the adhesive is curing.
Plates
- Separation between plate and surface
- Plate installation. Check fibre orientation.
- Try not to move the plates once they have been bonded to the concrete.
- Check for the existence of metallic areas in contact with the carbon fibre
plates.
Fabrication
- Amount of resin used. Try to comply with the design amounts.
- Check that the mixing proportions are the correct ones.
- Following compaction, avoid rough or irregular surfaces during site
installation.
- Try not to move the materials after the application
- Verify the quality of each plate installed before placing the next one.
Curing
- Verify whether the FRP curing has been properly done
Tests
- Preparation of specimens in accordance with the specifications
Inspection
- Assess the best installation method, vacuum, countermold, air drying or
other methods.
Filing of documents
- All the operations shall be documented, or digitally recorded, if deemed
necessary for a better understanding of the works performed.

45
3.1.6.2 Site tests
The following table (Table 3-5) describes the possible site tests to be conducted on the
materials used on the project in order to check the manufacturers specifications, which
allow us to assess the feasibility on the on-site element installation.
Table 3-5. Site test specification
Property to be checked Standard Requirements
Youngs modulus in deflection EN ISO 178 > 2000 N/mm2
Youngs modulus in compression EN 13412 > 2000 N/mm2
Shear strength EN 12188 > 12 N/mm2
Maximum overall expansion for any point EN 12617-1
En 12617-3
< 0.1%
Glass transition temperature EN 12614 > 45C, or 20C above
ambient temperature

3.1.6.3 Inspection recommendations
Even though workmanship standards can be very high, there are always factors that are
not taken into account, from the workmanship point of view, and that, after some time,
they come to the surface.
For this reason, it shall be necessary to set up a number of inspections on the completed
work, in order to ascertain that the reinforcements are fulfilling their function and to be
able to correct, in time, any eventual problem that might arise.
The ideal inspection frequency could be set up as follows:
Table 3-6. Inspection recommendations
Structure Inspection Detailed Inspection
Bridge Every year Every 6 years, or less
Buildings Every year Every 10 years, due to different usage or
structural change
Other structures Depending on use Every 10 years or Depending on use


46
Composite materials, by their very nature, hardly need any maintenance while in
operation. Even so, it shall be necessary to conduct a visual inspection to check for
cracks, breakages or delaminations that could indicate the level of damage. It shall also be
necessary to evaluate eventual damages to the material caused by accidents.
Should the strengthening material have been covered with a protection, this should never
be removed. The control, therefore, shall focus on an assessment of the tests conducted
on those materials.
In order to prevent future mistakes, checks shall be made on the identification of the
materials.
In the event of the covering being of the ultraviolet protection type, said protection shall
be monitored in order to control its condition and proceed to their replacement where
appropriate. It shall be up to the manufacturer of the materials to determine the best
replacement method and its timing.
Detailed inspection
The peeling of the polymer composites fibre from the concrete can be determined either
through pressure tests or thermography tests. Nowadays, there are no simple tests that
allow us to determine the condition of a joint and its adhesive. The best way to determine
this condition is through the pull-off test, when controlling the specimens at regular
intervals. However, this requires a number of very frequent tests and a testing period after
the execution of the joint.
Another task that has to be performed is the placing of instruments on the bridge to be
able to compare the data expected at the beginning to those finally and actually obtained.
Should the difference between these data be very great, it shall be necessary to find out
why and, on the basis of the results obtained, assess the resisting diagram. Above all, the
idea is to try to find out the reason why the model and reality are so different, so that, in
this way, eventual similar situations can be assessed in the future.
If during some of the inspections we find out that the condition of the strengthening is
not up to expectations, we shall proceed to inject resin or to place reinforcement on top.
Should the affected area be a large one, we shall have to remove the FRP and proceed to
its reinstatement, always in accordance with all the procedures described so far. In this
case, it shall be necessary to install a sufficiently big strengthening to ensure that loads are
transferred from the old laminate to the new one we have just laid. The newly installed
material should have similar characteristics to those of the old one, particularly with
regard to the direction of the fibres, strength, strain and thickness.

47
3.2 INTEGRATION OF KNOWLEDGE ON FRP RETROFITTED STRUCTURES
3.2.1 Introduction
The use of carbon fibre reinforced polymers (CFRP) for retrofitting a damaged structure
is first reported for the strengthening of the Ibach bridge in 1991, Lucern, Switzerland
[Meier, 1995]. Since this first application, the use of this technology has increased
exponentially, becoming one of the main applications of composite materials in building
and civil engineering. In the case of seismic loads, the improvement provided by CFRP
reinforcements in the structural capacity makes them a good solution to increase the
ductility of the structure, preventing structural damage in an earthquake situation.
Most of the existing knowledge about structural reinforcement and/or retrofitting of
reinforced concrete (RC) structures with fibre reinforced polymers (FRP) is based on
experimental simulations, supported and complemented by analytical calculations; and,
when the problem is treated using a numerical approach, material nonlinearities are
usually linearized and the FRP composite is considered as a single material (i.e.
[Rabinovitch and Frostig, 2001]).
On the other hand, the study of composite materials has been one of the major objectives
of computational mechanics in the last decade. The numerical simulation of composite
materials has been done, traditionally, using orthotropic materials with average properties
of their constituents. With this approximation, no model has been found able to work
beyond the constituents elastic limit state. Thus, these procedures limit the numerical
computation to elastic cases. Different theories have been proposed to solve this
problem, taking into account the internal configuration of the composite to predict its
behaviour. The two most commonly used are herein remarked.
i) Homogenization theory: This method deals with the global problem of composite
material in a two-scale context. The macroscopic scale uses the composite materials to
obtain the global response of the structure; composites are treated as homogeneous
materials in this scale. The microscopic scale corresponds to an elemental characteristic
volume in which the microscopic fields inside the composite are obtained; this scale deals
with the component materials. Homogenization theory assumes a periodical
configuration of the composite material to relate these two scales [Sanchez-Palencia,
1987; Oller et al., 2005].
ii) Mixing theory: The first formulation of the mixing theory corresponds to Trusdell
and Toupn [Trusdell and Toupn, 1960] and it is based in two main hypothesis: 1. All
composite constituents have the same strains. 2. Each constituent collaborates to the
composite behaviour according to its volumetric participation. The main problem of the
mixing theory is the iso-strain condition, which forces a parallel distribution of the

48
constituents in the composite. Some improvements to the original formulation can be
found in [Car et al., 2000].
Despite of all existing studies on both subjects, experimental tests of FRP reinforcements
and numerical characterization of composite materials, few researches are found in which
the structural reinforcement of RC structures using FRP is approached using a numerical
point of view. Thus, the main goal of this work is to join both fields, developing a
numerical procedure able to compute RC structures reinforced with FRP. The developed
formulation based on the finite element method that allows obtaining the structural
performance of existing structures when they are reinforced and/or retrofitted with FRP.
This structural behaviour is obtained taking into account the material non-linearities. The
code also provides the performance of each constituent material in the structure (for
example, it is possible to know the stress state of the composite reinforcement when the
structure collapses). The code can be used to study the same structure using different
FRP configurations, in order to obtain the most suitable for the case considered. It also
allows to apply the reinforcement to already damaged structures, reproducing with more
accuracy the conditions found in real applications.
In next section, the numerical procedures used to simulate RC structures with carbon
fibre reinforced polymers (CFRP) reinforcements are described briefly. Afterwards some
simulation examples are exposed, these examples show the validity of the formulation
proposed as well as the improvement obtained in the structural capacity when it is
reinforced with CFRP. Finally, some conclusions are exposed.
3.2.2 Formulation used to simulate RC structures reinforced and/or retrofitted
with CFRP
3.2.2.1 Simulation of Composite Materials
The structural problem of RC structures reinforced with carbon fibre reinforced
polymers is solved with PLCd [PLCd, 1991]. This is a finite element code, developed at
CIMNE (International Center for Numerical Methods in Engineering) and UPC
(Politechnical Univeristy of Catalonia), that works with two and three dimensional solid
geometries. It can deal with kinematics and material nonlinearities. It contains various
constitutive laws to predict the material behaviour: Von-Mises, Mohr-Coulomb,
improved Mohr-Coulomb, Drucker-Prager, etc. [Malvern, 1968; Lubliner et al., 1989]. It
uses different integration algorithms to simulate the material: Elastic, visco-elastic,
damage, damage-plasticity, etc. [Oller et al., 1990]. Dynamic analysis are developed using
the Newmark method, [Barbat et al., 1997].

49
To deal with composite materials, of the two main existing theories that take into account
composites internal configuration: homogenization and the mixing theory, the one used
and presented in this project is based in the last one. The election of the mixing theory
instead of a homogenization theory is based in the final aim of the code developed,
compute real structures reinforced and or retrofitted with CFRP. A homogenization
theory requires a micro-model for each point of the structure that becomes non-linear.
The resolution of a real structure with this procedure generates such a big amount of
degrees of freedom that the calculation is beyond the computation capabilities of
nowadays personal computers or small servers. On the other hand, the mixing theory
does not increase the degrees of freedom of the problem, as it is only present in the
constitutive section of the finite element code.
a. Classical mixing theory
The classical rule of mixtures was originally developed by Trusdell and Toupin [1960]. It
considers that the interaction between the components in a material point of the
composite is done according to the following hypothesis: (i) each infinitesimal volume of
the composite contains a finite number of material components; (ii) each component
contribution to the global behaviour of the composite is proportional to its volumetric
participation; (iii) all components suffer the same strains (closing equation); (iv) the
volume of each component is significantly smaller than the composite volume.
The third hypothesis, in the case of small strains, can be written as:
1 2 c n
ij ij ij ij
= = = = (3.2.1)
Where,
c
ij
is the strain tensor for the composite and
k
ij
is the strain tensor for
component k of the composite. According to second hypothesis, the stresses of the
composite can be computed as the proportional (according to the volumetric
participation) addition of each component stresses, thus:
1, 1, 1,
k k k k S k k k S c
ij ij ijkl ij ijkl ij
k n k n k n
k k C k C
= = =
= = =

(3.2.2)
where parameter
k
k is the volumetric participation of component k in the composite,
defined as:
0
k k
dV
k
dV
= (3.2.3)

50
A more detailed explanation of this theory, as well as the extension to the large
deformations case and its numerical implementation can be found in [Car, 2000]
b. Serial/Parallel mixing theory
The main problem of the classical mixing theory is the iso-strain condition, which forces
a parallel distribution of the constituents in the composite (Figure 3-1, iso-strain case).
The serial/parallel rule of mixtures (SP RoM) improves the classical mixing theory
replacing the iso-strain hypothesis for an iso-strain condition in the fibre direction and an
iso-stress condition in the transversal directions, allowing to simulate all components
distribution in the composite shown in Figure 3-1. This theory has been developed by
Rastellini and an extensive description of it can be found in [Rastellini, 2006] and in the
Deliverable 49 of the LESSLOSS project ( www.lessloss.org ).
Iso- strain Iso- stress Mixed case Iso- strain Iso- stress Mixed case

Figure 3-1. Different distribution of components in a composite material
The main hypothesis in which is based the numerical model of the Serial/Parallel mixing
theory are:
i) Composite is composed by two component materials: fibre and matrix
ii) Component materials have the same strain in parallel (fibre) direction
iii) Component materials have the same stress in serial direction
iv) Composite material response is in direct relation with the volume fractions of
compounding materials
v) Homogeneous distribution of phases is considered in the composite
vi) Perfect bounding between components is also considered
The equations that define the stress equilibrium and establish the strain compatibility
between components arise form the analysis of the model hypothesis. Thus,
Parallel behaviour:
c m f
P P P
c m m f f
P P P
k k

= =
= +
(3.2.4)

51
Serial behaviour:
c m m f f
S S S
c m f
S S S
k k

= +
= =
(3.2.5)
where,
P
and
S
are the parallel and serial components of the stress tensor respectively,
P
and
S
are the parallel and serial components of the strain, the superscripts c , m
and f represent the composite, matrix and fibre materials and
m
k and
f
k is the
volumetric participation of fibre and matrix in the composite.
In order to verify the compatibility equations, the Serial/Parallel mixing theory proceeds
in the following way.
1. First thing to be done is to split the component strain tensor into its serial and
parallel parts.
2. Afterwards, a first prediction of the matrix serial strains is computed. With this
initial prediction and the strain relation of equation
(3.2.3), the fibre serial strains can be also obtained.
3. Using each constituent constitutive law, the stresses for matrix and fibre can be
calculated, as well as the actualization of each material internal variable.
4. The stress tensors computed in step 3 are split into their serial and parallel parts
and the iso-stress condition defined in equation
(3.2.3) is checked. If it is verified, the matrix strain prediction was correct and the
composite stress can be obtained using equations (3.2.2) and
(3.2.3). If the iso-stress condition is not verified, the initial strain prediction
must be corrected and the process must continue in step 3.
The main problem of the Serial/Parallel mixing theory is that the composite material can
be composed, only, by two material components. However, this inconvenient is solved
with the laminate procedure exposed in the following section.
c. Laminate Composites
Usually fibre reinforced polymers are composed by different layers with different fibre
orientations. The orientation of the fibre can be defined by the engineer in order to
obtain the better performance of the composite according to its application. In example,
if the CFRP application is a flexural reinforcement, fibres will be disposed in the beam
axis direction while, if the CFRP is applied as a column wrapping, fibres will be oriented
in the cross section plane, following the column perimeter.
Under this scope, the limitation of the Serial/Parallel mixing theory (SP RoM) to only
two materials is not such, as the composites used are usually defined by multiple layers
composed by only two components: fibre and matrix. Thus, the SP RoM formulation can

52
be applied to each layer of the composite and, afterwards, compute the composite
behaviour composing the performance of each constituent layer. To obtain the laminate
behaviour, the classical mixing theory is applied to each lamina.
Applying the classical mixing theory to the different layers of the composite implies the
assumption that all laminas of the laminate are under the same strain conditions. This
assumption can be considered correct, as the different laminas usually have fibre
orientation distributions disposed in such a way that provide the laminate with an in-
plane homogeneous stiffness.
At this point, it must be said that not all the layers have to behave according to the S/P
RoM hypothesis. Sometimes, between fibre oriented layers, a randomly oriented one is
disposed. Having random oriented fibres, the layer behaves as an isotropic material and
the best theory to be applied to predict its behaviour is the Classical mixing theory. Thus,
the composite performance is obtained according to the following procedure:
1. For a given strain of the composite, the strain of each composite layer is obtained
applying the Classical mixing theory.
1 2 c L L Ln
= = = =
(3.2.6)
2. Each layer stress is obtained using the Serial/Parallel mixing theory in case of fibre
oriented layers or the Classical Mixing theory in case of randomly oriented layers (or
in other cases of iso-strain behaviour such as single materials layers).
Lk Lk
; using Classic mixing theory or SP RoM (3.2.7)
3. The stresses of the composite are obtained composing the stresses obtained in each
composite layer.
1
n
c Lj Lj
j
k
=
=
(3.2.8)
3.2.2.2 Other Formulations Developed to Simulate CFRP Reinforcements
Having defined the main formulation frame, in which the code works and deals with
composite materials, there are other formulations in it that are used to obtain a better
performance of the finite element code and that are required to obtain a better
approximation of the mechanical behaviour of the RC structures reinforced with CFRP.
In this chapter, all of them are briefly described.

53
a. Two dimensional and three dimensional elements
The original version of PLCd only dealt with two dimensional problems. Due to the
complexity of some geometries required to study CFRP reinforcements, one of the
required implementations was to extend all formulations included in PLCd to the three
dimensional case. The way in which a finite element code deals with two dimensional and
three dimensional cases can be studied in detail in [Oate, 1995; Zienkiewicz and Taylor,
1991]
b. Anisotropy using a mapping space theory
This theory is based on the transport of all the constitutive parameters and the stress and
strain states of the structure, from a real anisotropic space, to a fictitious isotropic space.
Once all variables are in the fictitious isotropic space, an isotropic constitutive model can
be used to obtain the new structure configuration. This theory allows considering
materials with high anisotropy, such as composite materials, using all the techniques and
procedures already developed for isotropic materials.
All the anisotropy information is contained in two fourth order tensors. One of
them,
ijkl
A
, relates the stresses in the fictitious isotropic space (

ij
) with the stresses in
the real anisotropic space (
ij
) and the other one,
ijkl
A
, does the same with the strains.

The relation of both spaces for the strains and the stresses is exposed in equations (3.2.9).
:
:
ij ijkl ij
ij ijkl ij
A
A

=
=
(3.2.9)
A more detailed description of this methodology, the extension to large strains and its
numerical implementation can be obtained in [Car et al., 2001 and Car, 2000]
c. Fibre-matrix debonding
The apparition of matrix cracks in a composite material is usually followed by a relative
movement between the fibres and the matrix. This lost of adherence implies a stiffness
reduction in the composite material. This phenomenon is introduced in the elastic limit
of the material as a modification of its yield surface criterion. The new fibre elastic limit
becomes:
[ ] { }
fib mat fib
N
mat
N
fib
N
fib
R
r f f f f / ) ( 2 ; ) ( ; ) ( min ) (
= (3.2.10)

54
Where
fib
R
f ) ( is the new fibre strength,
fib
N
f ) ( is the nominal fibre strength,
mat
N
f ) ( is the matrix nominal strength and
mat fib
N
f
) ( is the fibre-matrix interface

nominal strength. Equation (3.2.10) shows that the debonding happens when one of the
composite constituents reaches its nominal strength (considering the fibre-matrix
interface as a constituent). The nominal resistance values are obtained from the material
properties. The numerical implementation of this phenomenon is described in [Car, 2000
and Oller, 2002]
d. Construction stages algorithm
Retrofit a structure implies the addition of the structural reinforcement once the original
structure is already damaged. The Construction Stages Algorithm implemented in PLCD
permits running the numerical simulation during the desired load cases, with only some
structural elements active on the structure. At a certain load case, new elements can be
added without stopping the calculation process. These new elements are free of strains
and stresses when they are activated.
The algorithm requires having all elements defined in the structure. The elastic strains are
divided in two components, and active and a non-active, equation (3.2.11). If the element
is not present in the structure, all strains correspond to the non-active part, equation
(3.2.12), while, if the element is active, all strains corresponding to the non-active
situation will be removed, equation (3.2.13), from the total strain. Stresses are computed
considering only the active elements, equation (3.2.14).
e
NA
e
A
e
+ = (3.2.11)
0 ; = =
e
A
e e
NA

(3.2.12)
e
NA
e e
A
=
(3.2.13)
e
A
e e
C =
(3.2.14)
This procedure is explained with more detail in [Martinez et al., 2006; CIMNE, 2006]
e. Compression strength of composites
Although CFRP reinforcements are not recommended to be used to resist compressive
forces [Rabinovich, 2002], there are many situations in which this load state can be found.
This aspect is of special relevance in the case of structures subjected to seismic loads,
where the sign of the load is reversed as the earthquake evolves. Thus, a procedure to
obtain the compression strength of CFRP composites is required in order to take into

55
account this sort of situations. The main failure cause of compressed CFRP is the fibre
buckling phenomenon. Fibres are very slender elements and their second order effects
are avoided by the matrix elastic restrain. However, as damage in matrix evolves, fibre
restrain becomes weaker and fibre buckling occurs.
In all existing bibliography, the fibre buckling problem is solved by using different
methodologies to obtain a general expression for the composite limit compression stress.
This expression is only valid for the composite material and is obtained taking into
account different characteristics of the composite components. Alternatively, in the
present work, the composite material is modelled using the S/P rule of mixtures. Thus,
the fibre buckling problem must be solved in terms of the composite components,
considering their interaction, and not in terms of the composite by itself.
The interaction between fibres and matrix appears when the composite is compressed:
matrix restrain fibres avoiding their transversal movement. Under this approach, matrix
can be considered as an elastic restriction of fibres, and the fibre-matrix system can be
represented as it is shown in Figure 3-2. In this figure the movement of the fibres in case
of compression has been represented in dashed lines. Fibre behaviour is analogous to the
response that is obtained in a curved bar under unilateral restrain. This analogy is used to
formulate the fibre buckling problem. A detailed explanation of the formulation related
to this problem can be found in [Martinez et al., 2007]

Figure 3-2. Fibre-matrix system. Fibre behaviour when the composite is compressed
3.2.2.3 Efficiency Improvement of the Developed Code
Beside the different formulations included in PLCd code to deal with the CFRP
reinforcement of RC structures, it is necessary to pay special attention to the code
performance in order to improve its computation efficiency. The finite element
simulation of a complete structure requires a large number of elements, which requires a
lot of memory and computing time. Hence, all effort put in obtain a more efficient code
will be recovered when the simulations are performed. Two different code improvements
have been developed in this field

56
a. Tangent constitutive stiffness tensor
Depending on the constitutive equation used in a composite constituent material, the
tangent constitutive tensor cannot be obtained analytically. One solution is to use, in this
material, the initial stiffness matrix, which will lead to the equilibrium state but will
require a large amount of structural iterations. Thus, in order to obtain a fast and reliable
algorithm, the expression of the tangent constitutive tensor is required. To obtain it,
when no analytical expression exists, a numerical derivation using a perturbation method
is performed. According to Crisfield [Crisfield, 91], the definition of the tangent
constitutive tensor is:
:
t
= C
(3.2.15)
where,
[ ] [ ]
1 2 1 2
11 1
1
; and
n n
t t
n
t
t t
n nn
c c
c c
= =

=

C
(3.2.16)
The definition of the tangent constitutive tensor, equation (3.2.15), shows that the
variation of stresses due to an increment in the value of the j element of depends on
the values of the j column of
t
C . Thus, writing the j column of
t
C as,
1 2
T
t t t t
j j j nj
c c c =

c (3.2.17)
the stress variation is:
j t
j j
= c (3.2.18)
being
t
j
c the unknown
The perturbation method consists of applying a small perturbation to the strain vector
and, using the constitutive equation of the material, determines the variation that will be
obtained in the stress tensor due to this perturbation. At this point, the j column of the
tangent constitutive tensor can be computed as:
j
t
j
j
= c

(3.2.19)

57
The smaller the value applied to the perturbation, the better the approximation obtained
for the tangent constitutive tensor. Having defined a perturbation value,
j
, the
perturbed stress is computed using the constitutive equation of the material applying the
following input strain:
1
T
j j n
= +

(3.2.20)
And the stress variation due to the perturbation is obtained subtracting the original
converged stress from the computed one:
j
= (3.2.21)
This procedure must be repeated for all strain components in order to obtain the
complete expression of the tangent constitutive tensor. Hence, the numerical cost of
using a perturbation method is rather high. However, this procedure allows obtaining an
accurate approximation to this tensor for any constitutive equation used, ensuring the
convergence of the numerical process in few steps.
b. Improvement of the database system
One of the main advantages of the mixing theory compared to a homogenization theory
is that the computational cost is much lower, as the number of degrees of freedom of the
problem is not modified by taking into account the composite components. However,
taking into account all composite components increase significantly the amount of
information that has to be stored for each gauss point.
The information stored, in a finite element code, for each gauss point are the strains, the
stresses and the internal variables. If the strain and stress tensors have dimension n and
the number of state variables is m, the amount of memory required for each gauss point
if only one material is considered in it is: 2n+m real values.
If now, instead of having one single material, the material defined in the gauss point is a
composite made by three different layers, and each layer containing two different
components, it will be required to save for each layer the strains and stresses and for each
component of each layer the strains, stresses and internal variables. This leads to:
3 layers require 3 x (2n) real values = 6n real values
Each layer has two components, each one requiring 2n+m real values
= 3 x (2 x (2n+m)) = 12n + 6m real values
Thus, in total, the amount of real values required for the gauss point is: 18n + 6m

58
This example shows how, with only a three layer composite, the amount of variables that
must be stored increase nearly by nine, which increases nearly by nine the amount of
memory required to solve the problem.
Being aware of this problem, an optimization of the variables stored by PLDCd code has
been performed. PLCd was first developed as a finite element code to compute small
structures made of just one material. Thus, the code was not optimized and a lot of
useless variables were stored for each gauss point. All these variables have been removed.
It have also been removed all variables that, being previously stored, can be computed
again when necessary. With these two main measures, the code requirement of memory
has been reduced nearly a 70 % in the case of the concrete frame that will be shown in
section 3.2.3.3 of present document.
3.2.2.4 Making the Code More User Friendly
a. Pre and post processing using GiD
When solving the structural problem of a RC structure reinforced with CFRP, the
amount of information required by the finite element code PLCd is large and complex to
define. In the case of the three dimensional framed structure that is exposed in section
3.2.3.3, the input data file has more than 6500 lines. In these cases, is necessary to have a
friendly environment to define the input data file.
GiD [GiD, 2007] is a pre and post processor, developed at CIMNE, which can be
customized to interact with any existing finite element code. This customization includes
two main features, one corresponds to the data that can be defined over the structure and
the second one is the writing of the input data file in a format able to be recognized by
the finite element code. Both features have been included in the PLCd problem type
definition, which allows defining completely the input data file to be used by PLCd code
using GiD.
Following figures show some screen captures obtained during the definition of the
concrete frame used to obtain the CFRP reinforcement behaviour in a push-over analysis
(section 3.2.3.3 of the present document).
Another improvement related to the interaction of PLCd with GiD corresponds to the
ouput result file. Initially, PLCd exported each composite material in a layer and each
component material was exported in a different layer inside the first layer. This is, if two
elements had different composite materials (comp1 and comp2, in example), and both
composites have, as a constituent material, polymeric matrix, matrix results of composite
comp1 will be stored in the layer mat-comp1 and matrix results of comp2 will be stored in

59
mat-comp2. Hence, to see the stresses in matrix materials, different layers must be activated
and the results cannot be seen at once.
This way of storing the results has been changed and now all simple materials are stored
in the same layer. Doing so it is possible to see the stresses of all matrix material at once,
being easier to have a global comprehension of the material behaviour of the structure
using the graphical interface provided by GiD. Also, this way of storing results reduces
significantly the dimension of the output file.

Figure 3-3. Assignation, to the different volumes composing the concrete frame to be modelled, of
the different composite materials and construction stages

60

Figure 3-4. Definition of the different composite materials existing on the structure by the number
material constituents and their volumetric participation. Beneath can be seen some
material data defined for the simple materials.
b. Creation of an ANSYS usermat with the mixing theory
The commercial finite element code ANSYS [ANSYS, Inc, 2007] allows the user
definition of material constitutive equations or, as are called by the code, usermats. With
this option, ANSYS code becomes more versatile, as it can be customized for some
particular applications. With the aim of increase the applicability of the formulation
developed to deal with RC structures reinforced with CFRP, a usermat containing the
lamination formulation using the rule of mixtures (both, the Classical and the
Serial/Parallel) has been programmed.
ANSYS usermat is called when the code is at the gauss point level. The subroutine receives
the strains of the gauss point and has to return to the main code the stresses of the gauss
point, the tangent constitutive tensor and the internal variables actualized to the new
configuration [ANSYS, Inc, 2005].
This application has been already developed and, at this stage of the project, is being
tested in order to assure the correct behaviour of the material definition in all possible
situations related with composite materials and, more precisely, with the reinforcement of
structures using carbon fibre reinforced polymers (CFRP).

61
3.2.3 Numerical examples of the formulation proposed. CFRP reinforcements of
RC structures
3.2.3.1 Code Validation: Bending Reinforcement of a RC Beam
In this section, a numerical simulation of a bending reinforced beam is presented and
used to validate all the formulations proposed and exposed in previous section. This case
shows the efficiency of the serial/parallel rule of mixtures to deal with this sort of
structural problems, as it is able to reproduce the complex mechanical behaviour found in
the beam with an acceptable computational cost. The numerical results are validated with
experimental values. The studied beam is defined in the paper by Spadea [Spadea et al.
1998]. Its geometry and the reinforcements applied to it are shown in Figure 3-5.

Figure 3-5. Geometry and reinforcement of the beam studied
The red (thick) line displayed in the bottom of the beam corresponds to the FRP
reinforcement. This is made of carbon fibres embedded in a polymeric matrix. The
content of fibres is 60 % and the composite thickness is 1.2 mm. The finite element
model developed to simulate the beam reinforcement is shown in Figure 3-6. This is a 3D
finite element made with linear hexahedrons.
The usage of the S/P rule of mixtures allows considering all the reinforcement details
found in the beam using a coarse mesh. In Figure 3-6 it is also included the composite
materials composition. As can be seen, a single finite element contains, in this particular
case, up to three different component materials. The steel reinforcements are considered
as fibres, whose orientation is defined by the bar direction. The FRP reinforcement has
been included adding new finite elements to reproduce with more accuracy its position in
the beam.

62

MAT-01: Concrete (100 %)
Longitudinal Steel (42 %)
Vertical Steel (1 %)
Horizontal Steel (1 %)
Horizontal Steel (1 %)
MAT-06: Polymeric Matrix (34 %)
Long.Carbon Fibr. (66%)
Figure 3-6. Finite element model developed to realize the numerical simulation
The results obtained with this simulation are compared with the experimental results
reported by Spadea [Spadea et al., 1998]. Figure 3-7 shows the capacity curve of the beam
for the numerical and the experimental simulations. This is, vertical displacement of the
beam, in the point where the force is applied, against the load value. This figure
demonstrates the agreement between numerical and experimental results, which proves
the ability of the method to perform this sort of simulations. Figure 3-7 also includes the
results obtained with a numerical simulation of the same beam without FRP
reinforcements. The comparison between the results obtained for the reinforced and for
the non-reinforced beam shows the improvement obtained in the beam performance
when it is reinforced with FRP.

Figure 3-7. Force-displacement graph comparing the experimental and the numerical results
One of the main advantages of the proposed finite element formulation is that it allows
obtaining the structural behaviour of all its components, their failure causes, their strain-
stress state, etc. In Figure 3-8 some results maps, corresponding to the final computed
step, are represented. These show the most relevant information obtained from the
numerical simulation. Figure 3-8a displays the plastic damage in concrete, which shows

63
that the failure cause of the beam are the tensile stresses in concrete at midspan. In Figure
3-8b and Figure 3-8c it is depicted the plastic damage obtained for the longitudinal steel
reinforcement and for the polymeric matrix, respectively. These two figures show that
both materials have reached their yield stress when the beam failure occurs. Finally,
Figure 3-8d shows the stresses in carbon fibres. As it can be seen, they are at less than a
half of their load capacity (the fibre elastic limit stress defined is: 2300
e
= MPa)

FIGURE 3-8a. Plastic damage in concrete

FIGURE 3-8b. Plastic damage in long. steel reinf.

FIGURE 3-8c. Plastic damage in polymeric matrix

FIGURE 3-8d. Long. stress in carbon fibres [kp/cm
2
]
Figure 3-8. Results maps obtained with the finite element model of the beam
3.2.3.2 Code Validation: CFRP Retrofitting of a Beam
Two different numerical models have been developed to study the effect of retrofitting a
structure, depending on the existing damage in the beam when the CFRP reinforcement
is applied to it. These are:
Sp3D-Rt2: The CFRP reinforcement is applied when damage starts in concrete
material.
Sp3D-Rt3: CFRP reinforcement is applied when steel starts yielding.
Results obtained with these two models are compared with those obtained when the
beam is not reinforced (Sp3D-R0 model) and when the beam is reinforced from the
beginning of the loading process (Sp3D-R1 model). The capacity curve obtained for each
model is shown in Figure 3-9:

64

Figure 3-9. Comparison between the CFRP reinforcement and retrofitting by using capacity curves
These results show that the structural stiffness does not depends on when the
reinforcement is applied to the structure. The structure stiffness obtained when the CFRP
reinforcement is applied after the steel yielding (Sp3D-Rt3 model) does not differ
significantly form the structure stiffness obtained after steel yielding in the reinforced
model (Sp3D-R1). However, retrofitting a structure implies that, when the CFRP
reinforcement starts collaborating, the deformation and damage of the structure is larger
than if it had been reinforced from the beginning. Damage reduces the load capacity of
the beam while deformation can make the serviceability state unacceptable (i.e. when a
load of 25 kN is applied to the structure, beam deformations are 45 % larger in the
retrofit model, Sp3D-Rt3, than in the reinforced one, Sp3D-R1).
3.2.3.3 Reinforcement of a Framed Structure using CFRP
This structure was already presented in Deliverable 49 ( www.lessloss.org ). However, the
simulation presented in this Technical Report is new, as now the problem is solved with
the Serial/Parallel mixing theory. This theory provides a most stable and robust code
which provide more reliable simulations.
The main objective of the present simulation is to use the developed formulation to
verify the capability of CFRP reinforcements to increase the strength of concrete frames
when seismic loads are applied to them. Concrete framed structures are common in
building and civil engineering; one of the most stressed zones of these structures, under
seismic loads, are the connecting joints between beams and columns. In many occasions,
these joints show a lack of strength that can be improved with CFRP. The developed
models reinforce the frame joint with two different CFRP configurations, in order to
study the ability of the reinforcements to increase the frame strength and to find out
which configuration offers better results.

65
a. Model description
The concrete frame to be studied is defined with the geometry and the steel
reinforcement commonly used in buildings. Figure 3-10 shows the geometry considered
and Figure 3-11 shows the steel reinforcement and the description of the CFRP
reinforcement that will be applied to the frame joint. The beam cross section, as well as
its steel reinforcement, is dimensioned in such a way to assure the structural failure in the
beam, near the joint, in order to increase the effects of CFRP in the frame joint. The
structure is loaded with a horizontal force P applied in the middle of the frame joint (see
Figure 3-10).

Figure 3-10. Geometric definition of the framed structure considered in the simulation

Figure 3-11. Reinforcements applied to the concrete frame
Two dimensional and three dimensional models have been developed for the concrete
frame. 2D models have been used to calibrate the mesh to be used, as they require less
computational effort than 3D models. Results obtained with 2D models are compared
with the 3D ones to validate the accuracy of each simulation. Three different structures
are considered to study the effect of CFRP reinforcements on the frame joint. These are:

66
2D-noR and 3D-noR: Two and three dimension concrete frame without CFRP
reinforcements
2D-R and 3D-R: Two and three dimension concrete frame with the upper
and lower CFRP reinforcements defined in Figure 3-11
2D-LR and 3D-LR: Two and three dimension concrete frame with the upper,
lower and lateral CFRP reinforcements
All composite materials existing in the concrete frame are defined by combination of four
different basic materials, which are defined in Table 3-7. CFRP reinforcement is 1.2 mm
thick and is composed of 66 % of carbon fibres and 34 % of polymeric matrix. In the
case of the upper and lower reinforcements, the fibres are oriented following the
structure longitudinal axis. In lateral reinforcements, two layers of CFRP are applied to
the frame, in which the fibres are oriented at +0 and +90 respect the horizontal.
Table 3-7. Mechanical characteristics of the constituent materials used to define the composite
materials existing in the framed structure.
Yield Stres[MPa]
Fracture Energy
[kPam]
Material Yield
criterion
Young
Modulus
[ MPa ]
Poisson
Modulus
Compr. Tensile Compr. Tensile
Concrete
Mohr-
Coulomb
2.510
4
0.20 30.0 3.0 50.0 0.5
Steel
Reinf.
Von-Mises 2.110
5
0.00 270.0 270.0 2000.0 2000.0
Polymeric
matrix
Mohr-
Coulomb
1.210
4
0.20 87.5 29.2 36.0 3.0
Carbon
Fibres
Von-Mises 1.510
5
0.00 2300.0 2300.0 2000.0 2000.0

b. 2D results
The structural behaviour of the frame joint for the different reinforcements applied to the
structure is studied using the capacity curves obtained for each model (Figure 3-12). The
displacement considered corresponds to the horizontal displacement suffered by the
point where the load is applied. This displacement depends on the column, beam and
joint stiffness. As the column and the beam are not modified in the different models, if
the joint stiffness is increased with the different applied reinforcements, the force--
displacement graph will reflect this increment.

67

Figure 3-12. Capacity curves obtained with the 2D models
Figure 3-12 shows that the upper and lower CFRP reinforcements (2DF-R model) do not
improve significantly the frame behaviour. This improvement is only found when lateral
reinforcements are applied to the concrete frame. All three curves present a region where
the load reduces, to start increasing again afterwards. These points correspond to the
development of a plastic hinge in the structure. At this load step the structure adopts a
new strength mechanism and can increase its load capacity. Comparing the load applied
to the structure until the development of the plastic hinge, the lateral reinforcements
(2DF-LR model) increase in a 25 % the structural load capacity when compared with the
non--reinforced model (2DF-noR). This increment is only a 4 % if the structure is
reinforced only with upper and lower CFRP.
A better comprehension of the effects of each reinforcement can be obtained studying
the points where the plastic hinges are formed. Figure 3-13 shows the longitudinal strains
of each model at their last computed step. The cross sections where the plastic hinges are
formed are the ones with larger strains.
Figure 3-13 shows that applying only the upper and lower CFRP reinforcements the
plastic hinge moves from the beam to the inner part of the joint, where no
reinforcements are applied. Thus, the presence of CFRP does not modify substantially
the beam behaviour and, once the hinge has been formed, both structures behave
similarly (as it can be seen in Figure 3-12). On the other hand, when the lateral
reinforcement is applied to the structure, it restrains damage in the frame joint and moves
the plastic hinge to the cross section where no CFRP reinforcement is applied, what
allows the structure to increase its load capacity and its stiffness.

68
( a ) ( b )
( c )
Figure 3-13. Plastic hinges in the concrete frame. 2D models. a) model without CFRP reinforcement,
b) model with upper and lower CFRP, c) model with upper, lower and lateral CFRP
c. 3D results
The study by means of three dimensional models is also performed, as in the case of 2D
models, by means of the capacity curves (Figure 3-14). The main difference found when
comparing these results with the 2D ones is that the 3D models are stiffer and can reach
larger loads than the 2D models. This is because the concrete confinement is better
reproduced in this case, as steel stirrups are modelled taking into account their 3D
distribution and not only in one of their directions. Hence, concrete can reach larger
stresses, which increase the stiffness and strength of the structure.
Figure 3-14 shows that plastic hinges in the non-reinforced model (3DF-noR) and in the
reinforced model (3DF-R) appear for the same load and displacement as in the reinforced
beam model (2DF-LR), as a consequence of the increment of concrete strength.
However, in the three dimensional simulation, plastic hinges appear before in the
reinforced model than in the non-reinforced one. The explanation of this effect is shown
in Figure 3-15 (maximum strains in the unreinforced beam model (3DF-noR) before and

69
after the formation of the plastic hinge) and in Figure 3-16 (the same results in the case of
the 3DF-R model).

Figure 3-14. Capacity curves obtained with the 3D models

Figure 3-15. Crack evolution in the 3DF-noR model (model without CFRP reinforcement)

Figure 3-16. Crack evolution in the 3DF-R model (model with upper and lower CFRP)

70
According to these figures, the cross section in which the plastic hinge develops is nearly
the same in both models. But, as this cross section is closer to the initial damage in the
reinforced model than in the non-reinforced one, it is easier to simulate the plastic hinge
when the beam is reinforced. Thus, even if the CFRP reinforcement increases the joint
stiffness, in this case the frame plastic hinge appears for lower loads when this
reinforcement is applied than when the joint is not reinforced.
More differences are found when comparing the three dimensional model with the two
dimensional one, in the case in which lateral CFRP reinforcements are applied to the
frame joint (3DF-LR model). The first difference is that the formation of a plastic hinge
is not visible in the capacity curve. This is because no section is completely damaged
when the algorithm loses its convergence.
However, the main difference is found when looking at the most damaged section. The
strains in the lateral sections of the frame joint (Figure 3-17a) have a similar behaviour to
that of the 2D case: strains are larger in the cross section where the CFRP reinforcement
finishes than in the frame joint. But when the strains in a longitudinal section of the
structure are studied (Figure 3-17b), they show that the plastic hinge is developed in the
frame joint, as happens in the 3DF-R model. Two dimensional models consider the
CFRP reinforcement applied along the whole cross section while actually it is applied
only to the lateral surfaces. Thus, the reinforcement can avoid structural cracks on the
surface of the structure but cannot restrain them inside the joint. This effect can be
observed with more accuracy in Figure 3-18, where a zenith view of the strains in the
column section just below the frame joint is displayed.

Figure 3-17. Plastic hinge in the 3DF-LR model. Lateral view
These last figures show that the structure presents the same structural failure
independently of the CFRP reinforcement configuration applied to it. Thus, it can be
concluded that the only effect of the lateral CFRP reinforcement over the frame joint is
to delay the apparition of cracks in it and the following plastic hinge. However, this delay
is enough to allow a load 20 % larger in the frame when the horizontal displacement in it

71
is of 3.0 cm, increment more than sufficient to consider this reinforcement typology the
best option to improve concrete framed structures seismic strength.

Figure 3-18. Elements with larger deformations in the 3DF-LR model. Zenith view
The results also show the necessity of using three dimensional elements depending on the
simulation to be performed. When the distribution of CFRP reinforcements is not
uniform along the whole cross section, 2D simulations can provide incorrect results.
However, even if the effectiveness of lateral reinforcements is reduced in the 3D model,
the structural strength improvement is significant enough to consider this reinforcement
configuration as the best option to reinforce the column-beam joint of RC frame
structures.
3.2.4 CONCLUSIONS
In this technical report have been exposed the different numerical procedures developed
to solve in a reliable and efficient way the problem of reinforcement and/or retrofitting
of reinforced concrete (RC) structures using fibre reinforced polymers (FRP). Due to the
complexity of the problem to be solved, efforts have been directed not only to the
numerical procedures that allow performing the structural simulation but also to the
efficiency of the code and the way the user interacts with it.
Of all the formulations developed to solve the problem of FRP reinforcement of RC
structures, is of special relevance the general formulation of the mixing theory to deal
with laminated composites. This is, the division of the composite in its different
components until reaching the constituent material, which will be the one that will
provide the structural behaviour of the composite. This formulation can be understood as
a manager of constitutive equations and is the one that allows dealing with the
reinforcement problem, taking into account all its particularities, without increasing the
problem numerical size beyond the computational limits. The good performance of this
theory has been proved with the different simulations presented in section 3.2.3 of the
present document.

72
Besides the mixing theory formulation, other numerical procedures have been developed
to deal with the particular case of FRP reinforcements. The most relevant of them are the
anisotropy using a mapped space theory and the fibre matrix debonding. Special attention
must be paid in the construction stages algorithm, which allows performing simulations
of retrofitted structures (section 3.2.3.2). These simulations have shown that, even the
FRP performance do not vary significantly if it is applied as a reinforcement or as a
retrofit, the structural deformations and the stresses are larger if the FRP is applied when
the structure is already damaged. Finally, it has to be said that the compression strength
formulation has not been tested yet in a FRP reinforcement simulation, however, first
results obtained permit be optimistic about its performance and validity.
To improve the code efficiency it has been revised in detail, deleting all data and variables
not required to solve the problem. It has also been implemented a numerical derivation
procedure, used to obtain the constitutive tangent matrix, in order to reach the structural
equilibrium in the fewer number of iterations possible.
The code is now more user friendly as all the problem can be defined using GiD pre-
processor, and all output results can be studied with GiD post-processor. Also, to make
the developed formulation more accessible, the mixing theory developed to deal with
laminate composite will be soon accessible as a usermat for ANSYS users.
The conclusion that all all formulations exposed are valid and useful to solve the FRP
reinforcement problem has been obtained form the study of the different simulations
performed with them. However, the objective of these simulations was not only to verify
the formulations implemented in the finite element code, but also to study in which way
the simulations of structural reinforcements of CFRP must be developed to obtain
accurate results. The simulation that has provided more information in this aspect has
been the framed structure one. This simulation has shown the necessity to work with
three dimensional elements to obtain an adequate behaviour of the structure. Two
dimensional elements suppose a constant distribution of the materials along the whole
cross section. When the structure has not the same composition along its cross section,
the supposition that these properties can be extended to the whole section can
overestimate the structural performance, as happens with lateral CFRP reinforcements in
the frame simulation.
Finally, all simulations have proven the improvement obtained in the structural
performance when it is reinforced or retrofitted with fibre reinforced polymers. This
improvement depends on the type of reinforcement applied and on the level of existing
damage in the structure when the reinforcement is applied. Both aspects can be
controlled now, with the formulation described in this work, in order to obtain the best
FRP configuration to reinforce the structure considered.

73
3.3 EXPERIMENTAL DATA ON DURABILITY AND FATIGUE
RESISTANCE
This section will deal with a summary on the main factors that have an influence on the
durability and the fatigue resistance of composites materials. Thus, the section is divided
into two clearly distinct parts, one of them dealing with durability and the second one
dealing with the fatigue resistance.
3.3.1 Durability
The durability characteristics of FRP (Fibre Reinforced Polymers) materials have been,
and are, the subject of intense review, with special emphasis on the three most common
types of fibres, currently used in polymeric composites: glass, carbon and aramid fibres.
However, due to their wide application and utilization in Civil Engineering, most of the
available information focuses on glass fibre degradation. An up-to-date summary of the
main characteristics is given below.
3.3.1.1 Fibres environmental degradation
a. Glass fibre
Glass fibres are the most commonly used in Civil Engineering applications, because they
are the cheapest (Jamond 2000) and, therefore, it is important that we should understand
their behaviour when exposed to hostile environments. Although it is common
knowledge that glass fibres degrade in the presence of water and of acid, alkaline or
saliferous solutions, the most serious occurs in alkaline solutions (as reported by research
work on the subject). A good example is the alkaline water coming from concrete pores,
which turns this environment into one of the most critical for glass fibres, due to the
hydration of the cement. Therefore, many durability surveys are aimed at determining the
alkali resistance of GFRP (Glass Fibre Reinforced Polymers). Additionally, there are
debates still raging about the severity of saliferous environments on this type of fibres,
which have proved to be more serious than watery and acid environments, but as serious
as the degradation in alkaline environments (Chin 1997).
The different glass compositions available on the market read as follows: (1) E-Glass
(Electric Grade Glass) is the cheapest and more widely used for general purpose jobs;
(2) S-Glass (Strength Glass) has greater mechanical properties than E-Glass, but costs
more; (3) AR-Glass (Alkali Resistant Glass) is an improved version of E-Glass, to
withstand alkaline attacks through the addition of zirconium.

74
b. Aramid fibre
Aramid fibres are particularly susceptible to water, as these are organic polymeric fibres
and they are known to absorb water (Bank 1995). As happens with glass fibres, the
Absorption of water can cause the total degradation of the aramid fibres, mainly thanks
to the greater stresses caused by expansion. In addition, solutions such as sodium
hydroxide and hydrochloridic acid cause accelerated hydrolysis in the most common
aramid fibres (Kevlar 49), especially when temperature and stresses combine (Hunston
2000).
c. Carbon fibre
Studies on carbon fibre reinforced composites are numerous due to their extensive use in
the aerospace industry (Bank 1995). Carbon fibres offer higher rigidity and specific
strength than glass fibres. In addition, they are resilient, even in chemical environments
and do not absorb water. Machida (1995) points out that they are highly resistant to acid,
alkaline and organic environments. Due to these characteristics, carbon fibres are being
increasingly used in applications where light, high strength, structures are called for.
However, carbon fibres exhibit two drawbacks: (1) they are galvanically inactive and,
therefore, the cathodic reaction can play an important role in the degradation of the
composite, between the polymer and the graphite, that is to say in the interface (Tucker
1990). This can get worse when the carbon fibres come into contact with any metal, such
as those present in external reinforcements, or with saliferous water; and (2) they are a lot
more expensive than glass fibres.
Many studies have been undertaken to observe the typical degradation, both physical and
chemical, of carbon fibres in different environments (acid, alkaline and distil water), but
no deterioration mechanism has been found and so none is available (Sen 1998).
3.3.1.2 Accelerated Ageing Models
The accelerated tests main purpose is to predict the system service life; to achieve this,
three essential items of information are required: 1) data showing the degradation of the
properties of the material, which are going to be predicted in the service life; 2) an
analytical or statistical model to extrapolate the data for long term purposes; and 3) a
proper definition of the failure mechanism of the test samples; in the event of different
failure modes being obtained in the simulated accelerated models, the predictions shall
not be valid.
Katsuki & Uomoto (1997) apply Ficks Law to simulate the degradation of the GFRP
rods quantitatively, by predicting the reduction of the tension strength in accelerated
tests, where the results coincided well with the measured results. This Law is deemed

75
adequate to predict the loss of strength of the tendons or rods. Arrhenius Law can be
used to determine the diffusion rate in non-accelerated temperatures and,
consequently, the strength reduction rate under non-accelerated conditions.
Valter Dejke et al (2001) in his thesis, assumes that the service life at different
temperatures can be derived from Arrhenius Law. Using this approximation, it is
possible to transform the exposure time under accelerated conditions into real application
times, using the time factor (Time Shift Factor, TSF). For example, 1.5 years at 60C
approximately equals 50 years in open air conditions in southern Sweden (which means a
yearly temperature of 7C). In addition, it provides mathematical expressions to calculate
the weight gain and the concentration of the profiles of the materials when they are
penetrated by a liquid, which can be water, an alkaline solution or any fluid medium. All
the above equations assume that the diffusion of the liquid inside the material is
distributed in accordance with Ficks Law.
Different experts have obtained, from several surveys, numerous data about the
degradation characteristics of the FRP. However, what the industry needs is a forecast of
the service life. The behaviour of the tested materials, together with the degradation
characteristics, by themselves, is not sufficient to perform service life predictions. That is
the reason why numerous models have been developed to interpret the results from
accelerated tests. These models consider the information from the tests as input data and
predict the system service life.
a. Degradation models
As the degradation model becomes more complex, so does the model used to describe it.
That is the reason why it is so difficult to develop a model of degradation of FRP with
very complex durability mechanisms. The lack of knowledge about the true meaning and
about the use of the accelerated tests may lead to incorrect conclusions about the tested
product and even result in high costs for the development of the product (Nelson 1990).
Although there exist, in the literature, several proposed models, which are widely used,
each has its own limitations and assumptions. It must be born in mind that there are no
analytical models that take into account the degradation mechanisms and that accurately
estimate the service life of the materials subjected to accelerated tests. On the other hand,
it is very difficult to simulate in the accelerated tests the actual environmental conditions
to which the materials are exposed when they are in operation. Because of this and due to
all the aforesaid uncertainties, the service life predicting models must be used with
extreme care.

76
b. Constant Rate Models
These have been established analytical models which link all the variables involved
(temperature, saliferous solutions, etc.) under the application of a constant acceleration
factor. The close formulation of these models is very hard to obtain, because the
degradation models are not well known in most cases. In addition, it is possible that
deviations in the model might occur over prolong periods of time.
c. Power Law ModelsEquation Chapter 3 Section 3
The Power Law models are the most widely used among the simple speed models of
accelerated tests. These models have the degradation rate represented by a Power
function of a degradation factor as shown by equation (3.3.1). It must be born in mind
that the long-term degradation rate is always the same in these models, as their implies
(Nelson 1990).
= ' V (3.3.1)
Where:
' : The degradation rate.
: A constant of the material.
V : Degradation factor.
d. Exponential Models
These models are very similar to those used with Powers Law. The degradation rate is
represented by an exponential function of the stress, as shown in equation (3.3.2) (Nelson
1990)
= ' exp( ) V (3.3.2)
Where:
' : Rate of degradation.
, : Constants of the material.
V : Accelerated stress.

77
e. Eyring Model
The general formulation of the Eyring model contains two environmental load variables,
one of which is the temperature. The other is the load. This can be the mechanical load,
the humidity or the existing density. The general formulation of the Eyring model is
shown in equation (3.3.3).

=

' exp
V
V
T T
(3.3.3)
Where:
, , , : Constants of the material and of the degradation mechanism.
T : Temperature in K.
f. Arrhenius Stabilised-Status Temperature Acceleration Model
This model can be used when the temperature is the only acceleration factor and the
main concern are the chemical reactions (Caruso 1998). Arrhenius degradation rate can
be formulated like equation (3.3.4) (Nelson 1990)

=

' ' exp
E
A
kT
(3.3.4)
Where:
E : Activation energy of the chemical reaction in electron volts.
k : Boltzmanns constants, 8.617x10
-5
electron volts per C.
T : Absolute temperature in K.
A: A constant of the test conditions and failure of the material.
The acceleration factor for this model can be written like equation (3.3.5) to relate the
degradations in accelerated environments to those under service conditions (Nelson
1990):

78

= =

1 1
exp
' '
n
E
AF
k T T
(3.3.5)
Where:
AF : Acceleration factor.
n
E : Activation energy for the degradation mechanism and for the material.
: life-cycle at the reference temperature T(K).
' : life-cycle at elevated temperature T(K).
k : Boltzmanns constant.
Arrhenius Law models are still being debated and have not been internationally accepted
yet. They can only be used to draw comparisons between those materials whose
degradation can be simulated with this Law.
The Time Shift (TS) technique is used to represent the ageing effect, using a long-term
horizontal Time Shift (Gates 1997). Moreover, by representing the temperature effects
along with this technique in accelerated tests, a description is given of the level of
acceleration of an environmental exposure which is obtained when the temperature is
increased (Valter 2001). However, it does not properly represent ageing, except for short
periods of time.
3.3.2 Fatigue of composites
Fatigue is the condition for which a material cracks or it fails as a result of repeated effort
(cyclic loading). From an engineered point of view, it should be defined like the
permanent, located and progressive structural change that takes place in a material subject
to repeated or fluctuating deformations. In general, these deformations are presented to
smaller efforts that the last tensile strength of the material in a static test, and they
frequently appear for smaller efforts that the fluency limit of the material. The fatigue can
affect practically all the engineering materials subject to cyclic efforts. Cyclic efforts
include repetitive external loads and thermal efforts that result from heating and
alternated cooling.
The fatigue life of a component or material is defined as the total number of necessary
effort cycles to cause failure. It is an important concept that one frequently studies in the
laboratory, for what the obtained information can apply to designs and components in

79
real service. The most common method of studying the fatigue life is to use cyclic loads
of constant width and to register the number of cycles to failure.
Tests data of the fatigue life are generally presented in the form of the curves S-N where
the alternating efforts, S, - the maximum applied tension - are represented in terms of the
number of cycles, N, to failure. In Figure 3-19 the S-N curve of a quasi-isotropic carbon-
epoxy system can be observed. The study of these curves reveals that when the level of
the applied stress is diminished, the quantity of cycles until failure increases. The fatigue
limit is defined as the effort level beyond which failure does not take place due to fatigue.
However, most materials do not have this limit, so that frequently the fatigue limit of
many materials is specified as the effort that will not produce failure of the material below
10
6
cycles. Furthermore, composites generally present a higher resistance to the fatigue
that translates itself in a superior fatigue limit and requires special treatment. The first fact
to highlight is that the S-N curve, which appears in Figure 3-19, seems to follow a direct
line when it is represented in a logarithmic scale. Then it can be represented using
experimental data by means of an exponential law in which the parameters are easily
estimated by means of a simple lineal regression. This characteristic, commonly accepted
below one million cycles, is not the case above that limit. For applications of high cycles
it is recommended to use a law that includes more adjustment parameters.

Figure 3-19. S-N curve of a Carbon T300/Epoxi 5208 [0/90/45]s laminate
3.3.2.1 Factors affecting the fatigue life. Damage mechanisms
Several factors affect the fatigue life, but they can be classified in three main groups:
mechanical, microstructural and environmental factors. The first two are related to the
applied stress and stress concentrations. Hence, important improvements in the fatigue
life can be obtained by means of a proper design and a careful attention of the
mechanical factors that occur in a particular fatigue situation. Microstructural factors play
an important role, considering that the microstructural orientation is very pronounced in

80
an anisotropic material. As for the environment, they can be considered as special cases
of fatigue and include thermal effects, contact fatigue and the effects of corrosion.
In general, the effect of fatigue is the reduction of the strength or the residual stiffness,
and the possible failure, after applying a finite number of load cycles. These load cycles
are smaller in magnitude than the necessary load to produce failure in a single cycle.
Therefore, some process of fatigue damage has to occur that reduces the material
strength under load.
In all ways, this damage is not uniform, that is, not all the volume of the material element
suffers the same stress reduction. Generally, the damage process consists of a series of
discrete events that cause a non-uniform or inhomogeneous response of the material. In
fact it is probable that the most universal characteristic in fatigue is the inhomogeneous
deformation. According to above-mentioned consequence, the fatigue effect is composed
of a series of micro-geometric contributions, such as microcracking that produces local
concentrations, which in turn generate additional damage.
A basic physical principle behind any cycle-dependent behaviour is the appearance of
non-conservative deformation that modifies the internal nature or the geometry of the
material, and its capacity to respond to histories of continuous load. Non-conservative
deformation implies that the part of the energy absorbed by the material through the
stress or the applied displacements is not stored as energy of deformation, but is
dissipated as the driving force of internal processes, as the formation and growth of
microcracks, as thermodynamic events, as atomic or molecular redistribution and as
chemical events. Although not all the non-conservative processes produce fatigue in the
sense that they do not reduce the residual strength, the stiffness or the service life, it is
known that the damage processes for fatigue are not conservative; this implies a
dependence of the behaviour of the materials on the load history.
For composites of interest in civil engineering applications, especially those with fibres of
high modulus, the different micro-events that contribute to the development of the
damage process are classified in the following categories:
Microcrack formation
Chemical damage
Plastic deformation
Delamination
Separation of fibres and matrix
Interface failure

81
Generally, a reinforcement material is chosen to provide a great strength and stiffness to
the composite, while the matrix material is chosen by its ability to transfer the load in a
microscopic level, its durability and its processability.
There are two special characteristics of composite laminates that have a major influence
on the manner in which mechanics representations of damage development are set.
Those characteristics are inhomogeneity and anisotropy. Of the two, inhomogeneity has
the greater influence on laminate response to cyclic loading. From the behavioural
standpoint, inhomogeneity presents a dichotomy. In many respects, it is the single most
important factor contributing to the generally superior resistance of laminated composite
materials to fatigue damage development. At the same time, it is almost certainly the
greatest contributing factor in the initiation of damage at the micro-level. Local damage
events play the role of damage initiators. As we have seen earlier, matrix cracking,
debonding, and ply separation are examples of these events. However, damage growth is
severely inhibited by their different properties and response, and also because of micro-
separations between constituent materials. In fact, this constraint on damage growth
forms the basis for a progressive localization and intensification of damage in a
successively smaller volume of material as loading cycling continues. Finally,
inhomogeneity contributes to the complexity of the stress states associated with damage
development, even before damage develops.
It is also clear that anisotropy has a major influence on damage development. The most
obvious source of this influence is the complex stress state associated with reinforcement
phases that may be aligned or arranged in some geometric fashion that results in a
directional dependence of the mechanical properties of the composite system. Among
other things, these complex stress states may result in coupling; for example, coupling of
axial extension to shear behaviour.
3.3.2.2 Comparison of the fatigue behaviour in the different types of composites
Figure 3-20 shows the behaviour of different composites under one million of cycles. S-N
curves, which show in the chart, are normalised with respect to the initial static strength
of the laminate, so a comparison can be carried out. In view of these curves the excellent
behaviour of carbon-epoxy systems against fatigue can be deduced. In the case of high
modulus carbon fibre laminates, the material strength hardly decreases a 10 % in the first
million of cycles. Hence, the use of this system is recommended in components where
the fatigue plays an important role. However, the carbon-polyester system loses 50% of
its initial strength. Epoxy matrix systems reinforced with boron or aramid fibres show
good fatigue behaviour. In the case of boron, the response of this composite is similar to
the high strength carbon. Aramid fibres ensure its survival when work with safety
coefficients greater than two. Relative to the glass fibre, the lack of strength is greater

82
than the 60%. Matrix influence is observed again, because polyester resin shows a worse
fatigue behaviour than the epoxy matrix systems. Anyway, the survival of this system in
applications which have to support a million of cycles requires large safety factors, at least
equal to three.

Figure 3-20. Normalised S-N data for different composites systems
Glass-Polyester/Glass-Epoxy systems
Figure 3-21 shows the behaviour of several glass/polyester systems. Here it is seen that
unidirectional laminates with fibres in the direction of the load are more affected by
fatigue (we can see later that in carbon laminates the effect is the opposite).

Figure 3-21. S-N curves for 0, 90 and 45 glass/polyester laminates.
S-N data, which appear in Figure 3-22, shows again the sensibility of glass fibre laminates
to the fibre orientation. Figure 3-22 shows the greater slope of a unidirectional laminate
than a 45 laminate system. The strength fall in the first one is more than 80% of the

83
initial static strength, whereas in the 45 laminate is around 30%. Figure 3-22 shows the
influence of the fibre orientation in different laminates. Even when the initial strengths
are very different, the fatigue causes that all systems reach a strength between 100 and
200 MPa, if the number of cycles overcome the million. Another fact to highlight is when
there is 0 plies the strength lack is more significant in the first stages and smooth to a
high number of cycles.

Figure 3-22. S-N curves for Glass S2/Epoxy 5280 laminates to [0]3 and [45]2s
Carbon-epoxi systems
Figure 3-23 compares the behaviour of three laminates: 0, 90 and 45. The reduction
of strength in the unidirectional laminate is not more than 30%. Hence unidirectional
carbon laminates support fatigue much better than the glass fibre ones, in which the
strength loss can reach 80% when the applied load exceeds one million of cycles. Figure
3-23 shows that the presence of 0 laminates diminishes the slope of the S-N curve,
contrary to the case of glass fibre laminates.

Figure 3-23. S-N curves for carbon T300/epoxy [0]6,, [45]8t and [90]15 laminates

84
3.3.2.3 Conclusions
In this section a large number of S-N data for different fibre reinforced polymer systems,
likely to be used in structural components or in FRP strengthening of existing structures
are presented. These systems employ epoxy or polyester matrix and glass or carbon
fibres. The stress decreasement upon the different types of laminates (constitutive
materials) and orientation of the fibres are summarized, shown and discussed.
3.4 COMPUTATION OF THE RESISTANCE OF STRUCTURAL ELEMENTS
CONSIDERING STEEL & FRP
3.4.1 Introduction
Many research works were carried out on strengthening of existing structural members
with these materials.
Toutanji [1999], Fam and Rizkalla [2001] and Becque et al. [2003] developed extensive
experimental data for cylinder specimens, for a variety of fibre types, orientations and
jacket thickness, either for FRP jacketed concrete or concrete filled FRP tubes. Demers
and Neale [1999] and Ilki et al. [2004a] tested reinforced concrete columns of circular
cross-section with FRP jackets. Rochette and Labossiere [2000], Wang and Restrepo
[2001], Shehata et al. [2002] and Ilki et al. [2004b], tested square and rectangular concrete
columns confined by FRP composites.
Xiao and Wu [2000] investigated the effect of concrete compressive strength and
thickness of CFRP jacket and proposed a simple bilinear stress-strain model for CFRP
jacketed concrete. Tan [2002] tested half scale reinforced concrete rectangular columns
with a section aspect ratio of 3.65 under axial loads and investigated the effects of fibre
type and configuration and fibre anchors on the strength enhancement of the columns.
Ilki and Kumbasar [2002] tested both damaged and undamaged cylinder specimens,
which were externally confined with different thickness of CFRP jackets, under
monotonic increasing and cyclic compressive stresses. Based on experimental results they
proposed simple expressions for ultimate strength and corresponding axial strain of
CFRP jacketed concrete.
Lam et al. [2004] also performed monotonic and cyclic compression tests on CFRP
confined concrete cylinders. Then the tests results were compared with a monotonic
stress-strain model proposed by Lam and Teng [2003a], and a cyclic stress-strain model
proposed by Shao [2003], for FRP confined concrete. Lam and Teng [2002] carried out
an extensive survey of existing studies on FRP confined concrete and proposed a simple
model based on a linear relationship between confined concrete strength and lateral
confining pressure provided by FRP composites, which was quite similar to the model

85
proposed by Ilki and Kumbasar [2002] before. Ilki and Kumbasar [2003], after testing
CFRP jacketed concrete specimens with square and rectangular cross-sections, modified
the expressions that they have proposed before to cover non-circular cross-sections. Lam
and Teng [2003a, 2003b] proposed design oriented stress-strain models for both
uniformly and non-uniformly confined concrete members. Ilki et al. [2004c] tested FRP
jacketed low strength concrete members with circular and rectangular cross-sections, and
stated that when the unconfined concrete quality was lower, the efficiency of the FRP
jackets was higher.
De Lorenzis and Tepfers [2003], stated that none of the available models could predict
the strain at peak stress with reasonable accuracy. Bisby et al. [2005] evaluated and
modified available analytical confined concrete models to provide the best fit to the
experimental data base. Matthys et al. [2005] tested large scale cylinder columns and
observed that most of the available stress-strain models based on small cylinder tests
seem to predict the ultimate strength fairly accurately.
Tastani and Pantazopoulou [2004] studied the structural behaviour of FRP confined
corrosion damaged square columns with reinforcement details representative of pre-
1980s. In the light of the experience obtained on FRP retrofitting techniques, a number
of design guidelines and bulletins such as ACI 440.2R-02 [2002], CAN/CSA-S806 [2002],
EN 1998-3 [2003] and fib bulletin no.14 [2001] have been published. Most of the
experimental studies are on small size cylinders without longitudinal and transverse
reinforcement, and the unconfined concrete strengths of almost all of the tested
specimens are in the range of medium to high. In this study, uniaxial compressive
behaviour of reinforced concrete columns jacketed with CFRP sheets was investigated.
The average standard cylinder concrete strength at the time of testing was 15.92 MPa for
LSR series, while other specimens had the average standard cylinder concrete strength of
31.00 MPa (NSR series). The main parameters of the experimental work were the
thickness of the CFRP jacket, cross-section shape, unconfined concrete strength and
corner radius. Additional specimens were included in the test program for determining
the effects of existence of pre-damage, cyclic loading, the bonding pattern and anchorage
details. After the tests of unconfined and CFRP jacketed specimens, it was observed that
the contribution of the CFRP jackets on deformability and/or strength enhancement was
remarkable for specimens with circular and non-circular cross-sections. While the
strength enhancement was more pronounced for the specimens with circular cross-
section, specimens with square and rectangular cross-sections exhibited larger ultimate
axial deformations without a substantial loss in strength. Although original specimens
without adequate internal transverse reinforcement experienced premature buckling of
the longitudinal bars, this phenomenon was delayed significantly after CFRP jacketing.
Since the required confining pressure for a certain increase of compressive strength is
relatively less in the case of lower strength concrete, CFRP jackets were more effective in

86
the case of low strength concrete. Rectangular columns with larger corner radius reached
relatively higher strengths, while no specific effect of corner radius was observed on
deformability. For understanding the effect of pre-damage on the behaviour of retrofitted
columns, some of the columns were loaded until the axial strain of 0.003~0.004 before
retrofitting. Test results showed that such pre-damage did not have an adverse effect on
the performance of the jacketed columns. According to the test results, cyclic loading did
not have a remarkable effect on the behaviour of circular and square columns. However,
the compressive strength of rectangular columns was slightly less in the case of cyclic
loading. For equivalent amount of CFRP, continuous jackets and straps provided similar
performances. Application of additional CFRP anchorages into the concrete for
increasing the effectively confined area did not improve the behaviour of the retrofitted
columns. The compressive strengths and the corresponding axial deformations of the
columns were also predicted by the empirical equations proposed before, [Ilki et al.
2004c]. During these predictions, both the contributions of transverse steel reinforcement
and CFRP jacket were taken into account. After comparison of the predictions and
experimental data, it was seen that these predictions were in reasonable agreement with
experimental results. Although satisfactory results were obtained after comparing
experimental and predicted strength and ultimate deformation values, another
comparative study was also carried out for further verification of the methodology used
for computation of strength and deformability of FRP jacketed reinforced concrete
members. For this purpose, around 280 specimens of different shapes and concrete
strengths tested by different researchers from all over the world were taken into account.
This comparative study also showed that the proposed computation methodology gave
fairly accurate results.
3.4.2 Testing program
3.4.2.1 Outline of The Characteristics of The Tested Specimens
List of the specimens tested are given in Table 3-8 with main characteristics. The
diameter of cylinder columns were 250 mm, and cross-sections of rectangular columns
were either 250 mm 250 mm or 150 mm 300 mm. The 28 days standard cylinder
strength f
c
was 10.94 and 23.86 MPa, respectively for low and medium strength concrete.
Note that unconfined concrete strength of the member f
co
was assumed to be 85% of
the standard cylinder strength at the time of testing. Longitudinal reinforcement ratio
l

was around 0.01 and the clear concrete cover was 25 mm for all specimens. For
longitudinal reinforcement 610, 414 and 412 bars were used for specimens with
circular, square and rectangular cross-sections, respectively. The reinforcement details are
given in Figure 3-24. Only plain bars were used both for longitudinal and transverse
reinforcement for representing the columns of existing relatively older structures in
developing countries. Yield strength f
y
, yield strain
y
and tensile strength f
s,max
were 367

87
MPa, 0.0018 and 523 MPa for 10 mm diameter bars, 339 MPa, 0.0017 and 471 MPa for
12 mm diameter bars and 345 MPa, 0.0017 and 477 MPa for 14 mm diameter bars,
respectively. Yield strength and yield strain of 8 mm diameter bars were 476 MPa and
0.0024, respectively. The specimens were tested after being jacketed with 1, 3 or 5 plies of
CFRP sheets. The tensile strength f
*
fu
, elasticity modulus E
frp
, ultimate rupture strain
fu

and nominal thickness t
f
of CFRP were 3430 MPa, 230 GPa, 1.5% and 0.165 mm,
respectively. These properties are taken from the specifications of the manufacturer. To
prevent stress concentrations around the corners of the non-circular columns, corners
were rounded with a radius of 40 mm, while corners of four specimens were rounded to
10 and 20 mm to examine the effects of corner radius on the behaviour. Four pre-
damaged specimens and four specimens with additional anchorage details were included
in the test program, as well as nine specimens tested under cyclic loads. Four specimens
were wrapped with hoop and spiral type CFRP straps to investigate the effects of the
different wrapping patterns. Note that, all these specimens coded LSR-C-3-a-H, LSR-R-
2-3-40a-H (with hoop type straps) and LSR-C-3-a-S, LSR-R-2-3-40a-S (with spiral type
straps), were wrapped with 50 mm wide straps with 50 mm clear spacing. The general
characteristics of the specimens are given in Table 3-8.
Table 3-8. Characteristics of tested specimens
Specimen f
cj
L.R. T.R. n
f
co

LSR-R-1-0-40a 15.6 414 8/200 0 13.27
LSR-R-1-0-40b 15.9 414 8/200 0 13.53
LSR-R-1-1-40a-A 15.9 414 8/200 1 13.53
LSR-R-1-1-40b-A 15.9 414 8/200 1 13.53
LSR-R-1-1-40a 15.9 414 8/200 1 13.53
LSR-R-1-1-40b 15.9 414 8/200 1 13.53
LSR-R-1-3-40a 15.9 414 8/200 3 13.53
LSR-R-1-3-40b 15.9 414 8/200 3 13.53
LSR-R-1-3-10a 15.9 414 8/200 3 13.53
LSR-R-1-3-10b 15.9 414 8/200 3 13.53
LSR-R-1-3-20a 15.9 414 8/200 3 13.53
LSR-R-1-3-20b 15.9 414 8/200 3 13.53
LSR-R-1-5-40a 15.9 414 8/200 5 13.53
LSR-R-1-5-40b 15.9 414 8/200 5 13.53
LSR-R-2-0-40a 14.2 412 8/175 0 12.09
LSR-R-2-0-40b 14.2 412 8/175 0 12.09
LSR-R-2-1-40a-A 15.9 412 8/175 1 13.53
LSR-R-2-1-40b-A 15.9 412 8/175 1 13.53

88
Specimen f
cj
L.R. T.R. n
f
co

LSR-R-2-1-40a 15.9 412 8/175 1 13.53
LSR-R-2-1-40b 15.9 412 8/175 1 13.53
LSR-R-2-3-40a- 15.9 412 8/175 3 13.53
LSR-R-2-3-40b- 15.9 412 8/175 3 13.53
LSR-R-2-3-40a 15.9 412 8/175 3 13.53
LSR-R-2-3-40b 15.9 412 8/175 3 13.53
LSR-R-2-3-40a-H 15.9 412 8/175 3 13.53
LSR-R-2-3-40b-S 15.9 412 8/175 3 13.53
LSR-R-2-5-40a 15.9 412 8/175 5 13.53
LSR-R-2-5-40b 15.9 412 8/175 5 13.53
LSR-C-0-a 14.3 610 8/145 0 12.22
LSR-C-0-b 14.3 610 8/145 0 12.22
LSR-C-1-a 15.1 610 8/145 1 12.84
LSR-C-1-b 15.4 610 8/145 1 13.14
LSR-C-3-a-PD 15.1 610 8/145 3 12.84
LSR-C-3-b-PD 15.4 610 8/145 3 13.14
LSR-C-3-a 15.6 610 8/145 3 13.27
LSR-C-3-b 15.9 610 8/145 3 13.53
LSR-C-3-a-H 15.9 610 8/145 3 13.53
LSR-C-3-a-S 15.9 610 8/145 3 13.53
LSR-C-5-a 15.1 610 8/145 5 12.84
LSR-C-5-b 15.9 610 8/145 5 13.53
NSR-R-1-050-0- 27.5 414 8/50 0 23.44
NSR-R-1-100-0- 27.5 414 8/100 0 23.44
NSR-R-1-200-0- 27.5 414 8/200 0 23.44
NSR-R-1-050-3- 27.5 414 8/50 3 23.44
NSR-R-1-100-3- 27.5 414 8/100 3 23.44
NSR-R-1-200-3- 27.5 414 8/200 3 23.44
NSR-R-1-050-5- 27.5 414 8/50 5 23.44
NSR-R-1-100-5- 27.5 414 8/100 5 23.44
NSR-R-1-200-5- 27.5 414 8/200 5 23.44
NSR-R-1-000-3- 27.5 - - 3 23.44
NSR-R-2-050-0- 27.5 412 8/50 0 23.44
NSR-R-2-100-0- 27.5 412 8/100 0 23.44
NSR-R-2-175-0- 27.5 412 8/175 0 23.44
NSR-R-2-050-3- 27.5 412 8/50 3 23.44

89
Specimen f
cj
L.R. T.R. n
f
co

NSR-R-2-100-3- 27.5 412 8/100 3 23.44
NSR-R-2-175-3- 27.5 412 8/175 3 23.44
NSR-R-2-050-5- 27.5 412 8/50 5 23.44
NSR-R-2-100-5- 27.5 412 8/100 5 23.44
NSR-R-2-175-5- 27.5 412 8/175 5 23.44
NSR-R-2-175-5- 27.5 412 8/175 5 23.44
NSR-C-050-0 27.5 610 8/50 0 23.44
NSR-C-100-0 27.5 610 8/100 0 23.44
NSR-C-145-0 27.5 610 8/145 0 23.44
NSR-C-050-3 27.5 610 8/50 3 23.44
NSR-C-100-3 27.5 610 8/100 3 23.44
NSR-C-145-3 27.5 610 8/145 3 23.44
NSR-C-050-5 27.5 610 8/50 5 23.44
NSR-C-100-5 27.5 610 8/100 5 23.44
NSR-C-145-5 27.5 610 8/145 5 23.44

5
0
0
1
4
5
250
8
R
4
0
8
250
2
5
0
250
5
0
0
2
0
0
R
4
0
12
300
1
5
0
300
5
0
0
1
7
5
14
1
0
2
5
0
2
0
2
0
T
E
S
T

Z
O
N
E
2
5
0

Figure 3-24. Cross-sections and Reinforcement Details of the Specimens

90
3.4.2.2 Loading and Data Acquisition Setup
The specimens were tested under monotonic or cyclic uniaxial compressive loads by
using an Amsler universal testing machine with a capacity of 5000 kN. Two different
gauge lengths were used for measurement of average axial strains by displacement
transducers. For this purpose, four transducers in the gauge length of 270 mm and six
transducers in the gauge length of 500 mm were used. Axial and transverse strains at mid-
height were also measured by surface strain gauges with a gauge length of 60 mm for all
of the specimens. For specimens with circular cross-section, two vertically and two
horizontally bonded strain gauges were used with 180 degree intervals around perimeter,
for non-circular specimens vertical strain gauges were bonded on two opposite sides. For
specimens with square cross-sections two horizontal strain gauges were bonded on two
opposite sides. For rectangular specimens two horizontal strain gauges were bonded on
short and long sides. To obtain the deformations of the longitudinal and transverse
reinforcement, strain gauges were used with gauge lengths of 5 and 3 mm, respectively.
For data acquisition a 50 channel TML-ASW-50C switch box and a TML-TDS-303 data
logger were used.
3.4.2.3 Test Results and Discussions
The test results are presented in Table 3-9. As seen in this table for the LSR specimens
with circular, square and rectangular cross-sections, the average ratios of confined
concrete compressive strength to unconfined concrete member strength f
cc
/f
co
were 2.3,
1.5 and 1.2 for 1 ply, 4.3, 2.9 and 2.3 for 3 plies and 6.7, 3.7 and 3.3 for 5 plies of CFRP
jackets, respectively. For the NSR specimens with circular, square and rectangular cross-
sections, f
cc
increased 3.2, 1.9 and 1.7 times for 3 plies and 4.0, 2.4 and 2.3 times for 5
plies CFRP jackets, respectively. Note that unconfined concrete strength of member f
co

was assumed to be 85% of the standard cylinder strength at the time of testing f
cj
. Test
results showed that, CFRP jackets were more effective in the case of low concrete
strength and strength enhancement was more significant for circular specimens, while
only slight difference was observed between the strength enhancements of square and
rectangular specimens. The axial strains corresponding to CFRP jacketed concrete
strengths
cc
for the LSR specimens with circular, square and rectangular cross-sections
were 11.5, 10.5 and 8.8 times that of unconfined concrete
co
for 1 plies, 24.5, 30.3 and
31.8 for 3 plies and 32.8, 38.5 and 46.5 for 5 plies of CFRP jackets, respectively. For the
NSR specimens with circular, square and rectangular cross-sections,
cc
increased 16.0,
17.2 and 18.2 times for 3 plies, 21.8, 23.8 and 24.7 times for CFRP jackets of 5 plies,
respectively. The axial strain corresponding to unconfined concrete strength
co
was
assumed to be equal to 0.002. For the LSR specimens the transverse strains on CFRP
jackets at failure
ch
were between 0.007 and 0.018 independent of the jacket thickness,
with an average value as 0.012. For the NSR specimens
ch
were between 0.011 and 0.017,
with an average value as 0.013. These average values are about 80-87% of the value given

91
by the manufacturer. Note that only the strain gauges that could work until failure were
taken into consideration. Generally, the failures were sudden fracture of CFRP sheets at
the mid-heights of the specimens. For circular, square and rectangular specimens, rupture
of CFRP sheets was only one vertical cut with the height of 200-250 mm for the jackets
of one ply, 3-4 vertical cuts with the heights of 40-70 mm for the jackets of three plies
and more vertical cuts of lower heights for jackets of 5 plies. For square and rectangular
specimens, the rupture of sheets was around the corners just after the rounded portion.
Damage patterns were similar for LSR and NSR specimens. The axial stiffness of the
specimens did not show a tendency to increase with increasing jacket thickness. Note that
all stress-strain relationships are given with the non-dimensional vertical axis and while
determining the stress-strain relationship of confined concrete, the contribution of
longitudinal reinforcing bars was subtracted by considering the stress-strain relationships
of these bars. The stress-strain relationships were adapted from coupon tensile tests and
strain hardening was also taken into account.
It should also be noted that although the major contribution on strength and
deformability enhancement was provided by the external CFRP jacket, the internal
transverse reinforcement had also some influence. The contribution of internal transverse
reinforcement is minimal for the specimens with square and rectangular cross-section.
Table 3-9. Test Results
Specimen
f
cc exp
(MPa)
cc exp

ch

long side
ch

short
side f
cc
/f
co

cc
/
co

LSR-R-1-0-40a - - - - - -
LSR-R-1-0-40b - - - - - -
LSR-R-1-1-40a-A 19.02 0.013 0.008 - 1.4 6.5
LSR-R-1-1-40b-A 20.64 0.014 0.007 - 1.5 7.0
LSR-R-1-1-40a 20.58 0.017 0.011 - 1.5 8.5
LSR-R-1-1-40b 20.33 0.025 0.013 - 1.5 12.5
LSR-R-1-3-40a 37.36 0.055 0.016 - 2.8 27.5
LSR-R-1-3-40b 40.92 0.066 0.014 - 3.0 33.0
LSR-R-1-3-10a 24.74 0.049 0.009 - 1.8 24.5
LSR-R-1-3-10b 26.32 0.078 0.011 - 1.9 39.0
LSR-R-1-3-20a 29.25 0.055 0.014 - 2.2 27.5
LSR-R-1-3-20b 30.83 0.072 0.010* - 2.3 36.0
LSR-R-1-5-40a 51.65 0.069 0.015 - 3.8 34.5
LSR-R-1-5-40b 47.77 0.085 0.012* - 3.5 42.5

92
Specimen
f
cc exp
(MPa)
cc exp

ch

long side
ch

short
side f
cc
/f
co

cc
/
co

LSR-R-2-0-40a - - - - - -
LSR-R-2-0-40b - - - - - -
LSR-R-2-1-40a-A 20.58 0.014 0.008 0.006 1.5 7.0
LSR-R-2-1-40b-A 20.88 0.014 0.007 0.005 1.5 7.0
LSR-R-2-1-40a 18.38 0.018 0.007 - 1.4 9.0
LSR-R-2-1-40b 11.70 0.017 0.009 0.010* 0.9 8.5
LSR-R-2-3-40a- 31.77 0.058 0.007 0.008 2.3 29.0
LSR-R-2-3-40b- 40.11 0.080 0.018 0.016 3.0 40.0
LSR-R-2-3-40a 34.91 0.062 0.009 - 2.6 31.0
LSR-R-2-3-40b 26.58 0.065 0.015* 0.014 2.0 32.5
LSR-R-2-3-40a-H 20.05 0.031 0.012 0.017 1.5 15.5
LSR-R-2-3-40b-S 18.81 0.021 0.010 0.014 1.4 10.5
LSR-R-2-5-40a 50.81 0.086 0.009* 0.017 3.8 43.0
LSR-R-2-5-40b 35.98 0.100 0.011* 0.015* 2.7 50.0
LSR-C-0-a - - - - - -
LSR-C-0-b - - - - - -
LSR-C-1-a 29.04 0.022 0.010 - 2.3 11.0
LSR-C-1-b 30.66 0.024 0.008* - 2.3 12.0
LSR-C-3-a-PD 55.64 0.047 0.012 - 4.3 23.5
LSR-C-3-b-PD 57.24 0.047 0.012 - 4.4 23.5
LSR-C-3-a 56.12 0.044 0.005* - 4.2 22.0
LSR-C-3-b 60.04 0.054 0.004* - 4.4 27.0
LSR-C-3-a-H 35.06 0.028 0.012 - 2.6 14.0
LSR-C-3-a-S 32.46 0.032 0.012 - 2.4 16.0
LSR-C-5-a 84.88 0.065 0.013* - 6.6 32.5
LSR-C-5-b 92.61 0.066 0.004* - 6.8 33.0
NSR-R-1-050-0-40 - - - - - -
NSR-R-1-100-0-40 - - - - - -
NSR-R-1-200-0-40 - - - - - -
NSR-R-1-050-3-40 44.55 0.035 0.014 - 1.9 17.5
NSR-R-1-100-3-40 45.68 0.035 0.015 - 1.9 17.5
NSR-R-1-200-3-40 41.97 0.033 0.014 - 1.8 16.5
NSR-R-1-050-5-40 56.89 0.046 0.017 - 2.4 23.0
NSR-R-1-100-5-40 57.23 0.045 0.016 - 2.4 22.5

93
Specimen
f
cc exp
(MPa)
cc exp

ch

long side
ch

short
side f
cc
/f
co

cc
/
co

NSR-R-1-200-5-40 58.26 0.052 0.015 - 2.5 26.0
NSR-R-1-000-3-40 41.08 0.024 0.013 - 1.8 12.0
NSR-R-2-050-0-40 - - - - - -
NSR-R-2-100-0-40 - - - - - -
NSR-R-2-175-0-40 - - - - - -
NSR-R-2-050-3-40 40.15 0.036 0.012 0.005 1.7 18.0
NSR-R-2-100-3-40 42.84 0.039 0.012 0.013 1.8 19.5
NSR-R-2-175-3-40 40.63 0.034 0.013 0.011 1.7 17.0
NSR-R-2-050-5-40 59.99 0.069 0.014 0.014 2.6 34.5
NSR-R-2-100-5-40 53.63 0.045 0.014 0.014 2.3 22.5
NSR-R-2-175-5-40 47.81 0.034 0.012 0.012 2.0 17.0
NSR-C-050-0 - - - - - -
NSR-C-100-0 - - - - - -
NSR-C-145-0 - - - - - -
NSR-C-050-3 77.46 0.034 0.013 - 3.3 17.0
NSR-C-100-3 72.51 0.029 0.011 - 3.1 14.5
NSR-C-145-3 71.83 0.033 0.012 - 3.1 16.5
NSR-C-050-5 95.07 0.044 0.012 - 4.1 22.0
NSR-C-100-5 94.88 0.043 0.011 - 4.0 21.5
NSR-C-145-5 94.25 0.044 0.013 - 4.0 22.0

In Figure 3-25 the stressstrain relationships of specimens with different cross-section are
presented. As seen in these figures, all specimens experienced significant enhancement in
strength and deformability after being retrofitted with 5 plies of CFRP jacket. While the
strength enhancement was more remarkable for the specimens with circular cross-
section, specimens with square and rectangular cross-sections exhibited higher axial
deformations. In Figure 3-26 the comparison of LSR and NSR specimens are presented.
As seen in these figures, CFRP jackets are more effective in the case of low concrete
strength, in terms of strength and deformability.

94
0
2
4
6
8
-20000 0 20000 40000 60000 80000 100000 120000
Deformation (microstrain)
A
x
i
a
l

S
t
r
e
s
s

(
c
/
f
'
c
o
)
LSR-C-5-a
LSR-R-1-5-40a
LSR-R-2-5-40a
S i 4

Figure 3-25. Axial Stress-Strain Relationships for Different Cross-sections
0
2
4
6
8
0 20000 40000 60000 80000 100000 120000
Axial Deformation (microstrain)
A
x
i
a
l

S
t
r
e
s
s

(
c
/
f
'
c
o
)
LSR-C-5-a
NSR-C-5
LSR-R-1-5-40a
NSR-R-1-5-40
LSR-R-2-5-40a
NSR-R-2-5-40

Figure 3-26. Axial Stress-Strain Relationships for NSR and LSR Specimens
The following conclusions are derived after axial loading tests of CFRP jacketed low and
medium strength circular, square and rectangular brittle reinforced concrete columns.
Both for LSR and NSR series CFRP jackets enhanced the compressive strength and
corresponding axial strain of the columns with circular, square and rectangular cross-
sections. While the strength enhancement was more pronounced for circular cross-
sections, deformability enhancement was more for square and rectangular cross-sections.
Although in the case of unstrengthened specimens, longitudinal bars buckled

95
prematurely, in the case of CFRP jacketed specimens, the premature buckling of the
longitudinal reinforcement was prevented. The measured maximum transverse
deformations of CFRP jackets for LSR and NSR specimens were between 0.007~0.018
and 0.012~0.015, respectively. The average values of measured transverse strains were
about 80-93% of the value given by the manufacturer. These values were not dependent
on jacket thickness. The CFRP jackets were more effective in the case of low concrete
strength in terms of strength and deformability enhancements. The pre-damage of the
specimens, cyclic loading and additional anchorages applied for increasing the efficiently
confined cross-sectional area did not have a remarkable influence on the behaviour.
While the increase in corner radius of rectangular specimens provided higher axial
strengths, no clear influence was observed on deformability.
3.4.3 Computation of strength and deformability
A large number of existing buildings in earthquake prone regions of the world need to be
strengthened for reducing the potential live losses and economical losses. Towards this
target, difficulty of retrofitting procedure and disturbance of the occupants are among the
most common obstacles, as well as financial and legal matters. The difficulty of
application of retrofitting can be reduced using new technology materials like FRPs, at
least for some specific structural deficiencies, like lack of adequate ductility or shear
strength due to insufficient transverse reinforcement. At this point, lack of codes or
standards related with this new technology materials can become a very important
obstacle in front of design engineers. Although there are some codes, draft codes and
recommendations, these were not sufficient at the time of the proposal of this FP 6
project. So, a basic purpose of this task in the project was to propose a computation
methodology for the strength of FRP jacketed reinforced concrete members. During the
project duration several important achievements were obtained in the field of FRP
retrofitting. Interesting achievement in Europe is the publication of Italian Code on FRP
retrofitting and publication of Draft Turkish Earthquake Resistant Design Code, which
would be officially finalized within a few months and which is to include FRP retrofitting.
The draft Turkish Code on FRP retrofitting includes some formulations, which are
proposed by the research team of this task and presented in this section. For prediction
of compressive strengths and corresponding axial strains of CFRP jacketed columns, the
empirical equations proposed by Ilki and Kumbasar [2004c] were used, Equations. (3.4.1)
and (3.4.2). These equations were proposed based on experimental results on the
specimens with circular and rectangular cross-sections. In these equations, f
cc
and
cc
are
the confined concrete strength and corresponding axial strain, b and h are the width and
depth of the rectangular member, f
l
is the effective lateral confinement stress and
co
is
the strain corresponding to unconfined concrete strength and it is assumed as 0.002 in
this study.

96
cc l
co co
CFRP
f f
f f
1.2
1 2.4

= +

( 3.4.1 )
cc l
co co
CFRP
f h
b f
0.5
1 20

= +

( 3.4.2 )
Effective lateral confinement stress, f
l
can be obtained by Equation (3.4.3). In this
equation,
a
is the efficiency factor that is to be determined based on the section
geometry as the ratio of effectively confined cross-sectional area to the gross cross-
sectional area and can be calculated as given by ACI 440 [2002] or Wang and Restrepo
[2001]. E
frp
and
f
are the tensile elasticity modulus and ratio of wrapping material to the
concrete cross-section, respectively. Ultimate tensile strength of FRP
h,rup
is assumed to
be 70% of the ultimate strain corresponding to tensile strength of FRP
fu
. According to
Lam and Teng [2003b], FRP ruptures around 60% of its ultimate tensile strain
determined by the direct tension tests when wrapped around concrete members.
a f h rup frp
l
E
f
,
2

=
(3.4.3 )
While predicting the confined concrete strength and corresponding axial strain, the
contribution of internal transverse reinforcement (ITR) was also taken into account. For
this purpose, the empirical equations proposed by Mander et al. [1988] were used for
strength and ultimate axial strain, respectively, Equations. (3.4.4) and (3.4.5). After
obtaining the strength and deformability enhancements provided by external CFRP jacket
and internal transverse reinforcement, the total enhancement in strength and
corresponding strain was calculated by using Equations (3.4.6) and (3.4.7). The
predictions of confined concrete strength and corresponding axial strain for all specimens
are presented in Table 3-10.
cc l l
co co co
ITR
f f f
f f f
7.94
1.254 2.254 1 2

= + +

(3.4.4 )
cc l
co co
ITR
f
f
1 5 1

= +

(3.4.5 )
cc co cc cc
co co co
TOTAL CFRP ITR
f f f f
f f f
' ' ' '
' ' '
1 1

= +

(3.4.6 )
cc co cc cc
co co co
TOTAL CFRP ITR
1 1

= +

(3.4.7 )

97
Table 3-10. Axial Resistance and Deformability Predictions
Specimen
f
cc analy
(MPa)
cc analy
analy
cc
cc
f
f
'
exp
'

analy
cc
cc
exp

LSR-R-1-0-40a - - - -
LSR-R-1-0-40b - - - -
LSR-R-1-1-40a-A 18.20 0.019 1.05 0.68
LSR-R-1-1-40b-A 18.20 0.019 1.13 0.74
LSR-R-1-1-40a 18.20 0.019 1.13 0.89
LSR-R-1-1-40b 18.20 0.019 1.12 1.32
LSR-R-1-3-40a 27.72 0.030 1.35 1.83
LSR-R-1-3-40b 27.72 0.030 1.48 2.20
LSR-R-1-3-10a 22.38 0.025 1.11 1.98
LSR-R-1-3-10b 22.38 0.025 1.18 3.15
LSR-R-1-3-20a 24.36 0.027 1.20 2.04
LSR-R-1-3-20b 24.36 0.027 1.27 2.67
LSR-R-1-5-40a 38.72 0.038 1.33 1.82
LSR-R-1-5-40b 38.72 0.038 1.23 2.24
LSR-R-2-0-40a - - - -
LSR-R-2-0-40b - - - -
LSR-R-2-1-40a-A 18.90 0.036 1.09 0.39
LSR-R-2-1-40b-A 18.90 0.036 1.10 0.39
LSR-R-2-1-40a 18.90 0.036 0.97 0.50
LSR-R-2-1-40b 18.90 0.036 0.62 0.47
LSR-R-2-3-40a- 29.23 0.060 1.09 0.97
LSR-R-2-3-40b- 29.23 0.060 1.37 1.33
LSR-R-2-3-40a 29.23 0.060 1.19 1.03
LSR-R-2-3-40b 29.23 0.060 0.91 1.08
LSR-R-2-3-40a-H 21.26 0.043 0.94 0.72
LSR-R-2-3-40b-S 21.26 0.043 0.88 0.49
LSR-R-2-5-40a 41.15 0.076 1.23 1.13
LSR-R-2-5-40b 41.15 0.076 0.87 1.32
LSR-C-0-a - - - -
LSR-C-0-b - - - -
LSR-C-1-a 23.72 0.025 1.22 0.88
LSR-C-1-b 23.72 0.025 1.29 0.96
LSR-C-3-a-PD 39.40 0.039 1.41 1.21
LSR-C-3-b-PD 39.40 0.039 1.45 1.21

98
Specimen
f
cc analy
(MPa)
cc analy
analy
cc
cc
f
f
'
exp
'

analy
cc
cc
exp

LSR-C-3-a 39.40 0.039 1.42 1.13
LSR-C-3-b 39.40 0.039 1.52 1.38
LSR-C-3-a-H 27.31 0.029 1.28 0.97
LSR-C-3-a-S 27.31 0.029 1.19 1.10
LSR-C-5-a 57.51 0.049 1.48 1.33
LSR-C-5-b 57.51 0.049 1.61 1.35
NSR-R-1-0-40 - - - -
NSR-R-2-0-40 - - - -
NSR-C-0 - - - -
NSR-R-1-3-40 38.94 0.022 1.08 1.50
NSR-R-1-5-40 48.56 0.028 1.20 1.86
NSR-R-2-3-40 38.93 0.043 1.04 0.79
NSR-R-2-5-40 49.37 0.054 0.97 0.74
NSR-C-3 49.80 0.028 1.44 1.18
NSR-C-5 65.65 0.035 1.44 1.29

The empirical equations, proposed by the authors before, predicted the compressive
strength and corresponding axial strains of the specimens with a reasonable accuracy.
3.4.4 Verification of proposed computation methodology
For further verification of the proposed equations, more than 200 specimens tested by
other researchers were examined and their compressive strength and corresponding
deformations were determined using the proposed equations. The comparisons of
predictions with experimental data are presented in Figure 3-27.

99
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0
cc
/
co
(Analytical)
c
c
/
c
o

(
E
x
p
.
)
Circular-Manufacturer
Rectangular-Manufacturer
Circular-Coupon
Rectangular-Coupon
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
20.0
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0
f'
cc
/f'
co
(Analytical)
f
'
c
c
/
f
'
c
o

(
E
x
p
.
)
Circular-Manufacturer
Rectangular-Manufacturer
Circular-Coupon
Rectangular-Coupon

Figure 3-27. Comparison of predictions with available experimental data
3.4.5 Conclusions
The objective of this sub-task is obtaining experimental results of FRP jacketed
reinforced concrete members concerning the effects of different variables such as
thickness of the FRP jacket, cross-section shape, unconfined concrete strength, corner
radius, loading type, bonding details, pre-damage, etc. and study the methodology of
determining the axial resistance of reinforced concrete members considering the
contributions of both internal steel reinforcement and external FRP jacket. Naturally, an
analytical work is carried out in parallel fashion for predicting the strength and
deformability of FRP jacketed members considering the dual effect of internal steel and
external CFRP jackets. Totally, 68 specimens were tested under concentric compressive
stresses, either by monotonic increasing loading or repeated loading. 40 of the specimens
were constructed using low quality of concrete and inadequate transverse reinforcement,
while 28 specimens were constructed using medium strength concrete and various
amounts of internal transverse reinforcement. Most of the specimens were tested under
concentric compression after being jacketed with CFRP, while some specimens were
tested as original for representing the reference behaviour. An analytical study is carried
out for determining the combined confinement effect of internal and external transverse
reinforcement. During this analysis, the contribution of internal reinforcement was taken
into account using different models proposed by different researchers, including the
authors of this section. The confinement effect of external FRP reinforcement was also
taken into account using a model proposed by the authors of this section before. The
predicted analytical results were then compared with the experimental data obtained at

100
the end of this study (for 16 circular and 27 rectangular columns), as well as previous
studies of the authors of this section (for 17 circular and 32 rectangular columns). The
reason of difference between the number of specimens tested in this study and specimens
taken into account for comparison with analytical predictions is; to exclude the
specimens, which had special features like additional anchorages, pre-damage or
specimens loaded cyclically. Otherwise, it was impossible to propose simple and
applicable equations for strength and deformability. For further verification of the
proposed equations, more than 200 specimens tested by other researchers were examined
and their compressive strength and corresponding deformations were determined using
the proposed equations. At the end of this extensive comparison, it was seen that the
proposed methodology and design equations are satisfactorily representing the behaviour.
The proposed equations for the strength and deformability of FRP jacketed columns,
verified with the tests and analytical comparisons carried out during the present project,
are included in the draft version of the Turkish Seismic Design Code (2006).
3.5 URBAN REHABILITATION USING FRP
3.5.1 Introduction
Seismic risk assessment and hazard mitigation for urban infrastructures located in seismic
regions is a challenge shared by many countries around the world. The variability and
associated uncertainty makes it more difficult to assess and mitigate seismic hazards, and
requires development of rapid and cost efficient assessment tools for effective
prioritization and timely mitigation. In urban areas, where there is a large stock of existing
buildings, most of which have inferior seismic capacities, determining the vulnerable
buildings within the existing building stock is a high priority task for taking the necessary
remedial measures for disaster mitigation prior to potential seismic events. Limited time,
scarcity of funds, and the high variability in seismic resistance of structures in these areas
have led to a concentrated research effort towards development of approximate seismic
assessment tools that can potentially provide a rapid, efficient, and reliable mitigation
strategy. In the following sections, the methods developed for seismic assessment of large
building stocks are provided, followed by their implementation on a large group of
buildings located in Zeytinburnu, Istanbul (Chapter 1).
Feasible retrofitting solutions have been developed for selected residential mid-rise
apartment buildings under high seismic risk in Istanbul. A feasible solution is considered
as the optimal combination of cost, downtime, disturbance, technical applicability and
social impact. The basic elements of conventional system retrofitting are concrete infilled
shear walls (Chapter 2). Recently, new techniques such as the existing masonry partition
walls enhanced by applying FRP sheets on their surfaces have been development.
Guidelines for the application of this new technique are provided.

101
3.5.2 Seismic Performance Assessment Procedures
A three tier approach has been developed for seismic vulnerability assessment of
buildings located in urban areas. The purpose of three level assessments is to identify
likely highly vulnerable buildings by reducing the large building stock into a manageable
size so that the mitigation strategies can be investigated. These procedures follow
essentially three main stages: (1) walk-down evaluation, (2) preliminary evaluation and (3)
detailed evaluation. The first stage, walk-down evaluation, does not require any analysis
and it relies on the past performance of similar buildings. The goal in this stage is to rank
highly vulnerable buildings that require further investigation. The second stage,
preliminary investigation stage generally covers the buildings that are designated to be
inadequate in the first stage. In this stage simplified analysis is performed to attain a more
detailed and reliable assessment. The final evaluation stage, which is more case oriented,
covers an in-depth evaluation through sophisticated structural analysis methods. The
need and the scheme for rehabilitation is determined based on the detailed assessment
results. These procedures are explained next briefly and the reader is referred to
LESSLOSS deliverable reports 47 and 107 for the details ( www.lessloss.org ).
3.5.2.1 The Walkdown Evaluation Procedure
A fast and simple seismic risk assessment procedure is proposed for vulnerable urban
building stocks (Sucuolu et al. 2007). It is basically a sidewalk survey procedure based on
observing selected building parameters from the street side, and calculating a
performance score for determining the risk priorities for buildings.
Recent earthquakes in urban environments revealed that building damages increase with
the number of storeys when the buildings lack the basic seismic resistant design features.
Other factors that have significant contribution to damage are also well established.
These are the presence of severe irregularities such as the soft storeys and heavy
overhangs, other discontinuities in load paths, poor material quality, detailing and
workmanship. It is usually difficult to quantify the sensitivity of damage to each
parameter analytically, however statistics can be used. Some of the important parameters
stated above that influence damage significantly can be determined quite easily, by visual
observation.
According to the procedure proposed here, structural parameters that have to be
observed during the field surveys are the number of storeys, soft storeys, heavy
overhangs, pounding, topographic effects and the overall apparent quality of the building
reflecting the quality of construction. Once the vulnerability parameters of a building are
obtained from walkdown surveys and its location is determined, the seismic performance
score PS can be calculated by using Equation (3.1.1). The base scores, BS, the

102
vulnerability scores, VS, and the vulnerability score multipliers, VSM, to be used in
Equation (3.1.1) are defined in Table 3-11 and Table 3-12, respectively.
= +
( ) ( ) ( ) PS BS VSM x VS (3.1.1)
The weight of each building vulnerability parameter was evaluated by statistical
procedures, based on the Duzce database.
Table 3-11. Base Scores and Vulnerability Scores for Concrete Buildings
Base Score (BS) Vulnerability Scores (VS) Number of
storeys
Z
o
n
e

I

Z
o
n
e

I
I

Z
o
n
e

I
I
I

S
o
f
t

s
t
o
r
e
y

H
e
a
v
y

O
v
e
r
h
a
n
g
s

A
p
p
a
r
e
n
t

Q
u
a
l
i
t
y

S
h
o
r
t

C
o
l
u
m
n

P
o
u
n
d
i
n
g

T
o
p
o
g
r
a
p
h
i
c

E
f
f
e
c
t
s

1-2 100 130 150 0 -5 -5 -5 0 0
3 90 120 140 -15 -10 -10 -5 -2 0
4 75 100 120 -20 -10 -10 -5 -3 -2
5 65 85 100 -25 -15 -15 -5 -3 -2
6-7 60 80 90 -30 -15 -15 -5 -3 -2

3.5.2.2 Preliminary Assessment
The preliminary assessment procedures are intended to be applied to the vulnerable
buildings identified by the walkdown assessment procedures, so the buildings are
subjected to a relatively more detailed assessment that require further data. In this
context, the discriminant analysis technique (a statistical method) was used to develop a
preliminary evaluation methodology for assessing seismic vulnerability of existing low- to
medium-rise reinforced concrete buildings (Yakut et al. 2006). The main objective of the
procedure outlined below is to identify the buildings that are highly vulnerable to damage.
The procedure is developed for structures whose structural system is formed by frames
or frames and structural walls. The procedure is applicable for cases where the number of
storeys is less than or equal to 7. The structures described above comprise the majority of
the existing RC buildings in Turkey.


103
Table 3-12. Vulnerability Parameters, (VSM)
Parameter Presence
S
o
f
t

s
t
o
r
e
y

H
e
a
v
y

O
v
e
r
h
a
n
g
s

A
p
p
a
r
e
n
t

Q
u
a
l
i
t
y

S
h
o
r
t

C
o
l
u
m
n

P
o
u
n
d
i
n
g

T
o
p
o
g
r
a
p
h
i
c

E
f
f
e
c
t
s

Does not exist 0 0 0 0 0
Exists 1 1 1 1 1
Good 0
Moderate 1
Poor 2
Considering the characteristics of the damaged structures and the huge size of the
existing building stock in Turkey, the following parameters were chosen as the basic
estimation parameters of the proposed method:
number of storeys (n),
minimum normalized lateral stiffness index (mnlstfi),
minimum normalized lateral strength index (mnlsi),
normalized redundancy score (nrs),
soft storey index (ssi),
overhang ratio (or)
These parameters take into account the major architectural irregularities in plan and in
elevation (ssi, or, nrs), the lateral strength (mnlsi) and stiffness (mnlsi) of the buildings.
Lateral strength and stiffness of the building are determined from the sizes of the
columns, shear walls and masonry infill walls comprising the structural resistance of the
building.
These parameters are calculated for each building to obtain a damage index that is
compared with a cutoff value to determine its expected performance. The damage index
is computed considering two performance objectives, namely life safety and immediate
occupancy. The damage index or the damage score corresponding to the life safety
performance classification (DI
LS
) is computed from the discriminant function given in
Equation (3.1.3).
= + + 0.620 0.246 0.182 0.699 3.269 2.728 4.905
LS
DI mnlstfi mnlsi nrs ssi or (3.1.3)

104
In the case of immediate occupancy performance classification (IOPC), the discriminant
function, where DI
IO
is the damage score corresponding to IOPC, based on these
variables is:
= + + 0.808 0.334 0.107 0.687 0.508 3.884 2.868
IO
DI n mnlstfi mnlsi nrs ssi or (3.1.4)
In the proposed classification methodology, buildings are evaluated according to both
performance levels. The steps to be followed are listed below.
1. Calculate DI
LS
and DI
IO
scores by using Eq. (3.1.3) and Eq. (3.1.4), respectively.
2. Determine the cutoff values for each performance classification by using the
relationships given in Eq. (3.1.5).

The LS
CVR
and IO
CVR
values in these relationships are obtained based on the number of
storeys above the ground level. The CM values are adjustment factors, which introduce
the spatial variation of the ground motion in the evaluation process. These values are
given in LessLoss Deliverable 107 (www.lessloss.org )
= =
LS CVR IO CVR
CV CM LS , CV CM IO (3.1.5)
By comparing the CV values with associated DI value calculate performance grouping of
the building for life safety performance classification (LS
PC
) and immediate occupancy
performance classification (IO
PC
) as follows:
If DI
LS
> CV
LS
take PG
LS
=1
If DI
LS
< CV
LS
take PG
LS
=0
If DI
IO
> CV
LS
take PG
IO
=1
If DI
IO
< CV
LS
take PG
IO
=0

By examining the performance groupings of the building for LSPC and IOPC, make a
decision on the probable expected performance level of the building. In this process, if
both performance groupings yield PG=0 then the building is classified as having low risk,
if PG=1 in both cases then it is classified as having high risk, otherwise it is considered as
intermediate needing further analysis.
3.5.2.3 Detailed Assessment Procedure
As mentioned earlier, detailed assessment procedures are applied to individual buildings
to determine the need for retrofit as well as verifying the selected retrofit option. A
detailed seismic performance assessment procedure has been developed for reinforced
concrete frame buildings with masonry in-fill walls (Erduran and Yakut, 2007). The

105
procedure is based on member damage functions developed for the primary components;
columns, beams and in-fill walls. Analytical investigations carried out to determine the
influence of a number of parameters on the damageability of components were combined
with existing experimental data to develop component damage functions expressed in
terms of member deformations. The component damage states are combined taking into
account their importance to determine the storey and building level damage states that are
then used to assess the global performance. The procedure has been calibrated and
compared with other procedures by predicting the observed performance of the buildings
exposed to recent earthquakes in Turkey. It was observed that the damage experienced by
most of the components of these buildings was predicted satisfactorily, and that the
observed building damage states were captured.
a. Outline of the Procedure
The procedure used in this project is a detailed vulnerability assessment procedure carried
out for an individual building under a given ground motion or design spectrum. The
outline of the procedure is explained in this part. The response parameters that have been
used as pertaining to seismic damage are the interstorey drift ratio for columns and infill
walls and the chord rotation for beams.
The steps involved in the procedure developed are summarized below.
Step 1 Data Collection:
The developed methodology requires a nonlinear analysis, either static or dynamic, of the
given building. For this, first of all, some data must be collected about the building at
hand. This data includes the design drawings, as-built dimensions of the building, the
condition of the building and the material properties, preferably obtained from in-situ
tests.
Step 2 Nonlinear Analysis & Determination of the Member End Deformations:
The computer model developed may be a two dimensional or a three dimensional model,
based on the choice of the user. Similarly, the user chooses the type of nonlinear analysis
(nonlinear static analysis or nonlinear time history analysis) that will be used. If a
nonlinear static analysis is carried out, the capacity curve obtained as a result of this
analysis must be used to determine the performance point of the building under the
prescribed ground motion or response spectrum, using the procedures available in
literature such as the Capacity Spectrum Method summarized in ATC-40, the
Displacement Coefficient Method of FEMA-356 or the Constant Ductility Spectrum
Method. The member end deformations at this performance point will be recorded and

106
used in the subsequent steps. If a nonlinear time history analysis is carried out, then the
maximum member end deformations will be recorded.
Step 3 Determination of the Member Damage Scores:
The maximum member end deformations obtained as a result of the nonlinear analysis
will be input in the damage functions developed to compute the damage score of each
member.

Step 4 Determination of the storey and Building Damage Scores:
Once the damage score for each member is determined, then the weighted average of
these damage scores is computed to determine the damage score of each storey and
finally of the entire building. The weighing coefficients used here depend on the
contribution of each member in resisting the seismic forces and named as component
importance factors. Approximate values for the component importance factors were
developed and are given in a tabular format.
The final step of the procedure is the determination of the performance of the building
based on the computed building damage score.
The details of these steps are given elsewhere (Erduran and Yakut 2004a, 2004b).
b. Component Damage Functions
For each type of component, numerical analyses and experimental test results have been
used to develop damage curves. The functional form for the damage curves is the one
given in Eq.3.1.6.

=

( ) 1 ( )
b
a
Damage e g (3.1.6)
where

=

( ) 0.5 1 cos g
c
if c

107
g()=1 if >c (3.1.7)
In these equations, represents the interstorey drift ratio and a, b, and c are the equation
parameters. In case of beams, is replaced by which represents the end rotation of the
beam. The values of these parameters are given in Table 3-13 and Table 3-15 for
columns, beams and infill walls. The beams and columns are grouped into theree
depending on their ductility level. Infill walls are classified into four based on their
material and geometrical properties. To illustrate the shape of damage curves Figure 3-28
displays them for columns.
Table 3-13. Values of Equation Parameters for Damage Curves of RC Columns
Parameter Low
Ductility
Moderate
Ductility
High
Ductility
a 0.0119 0.0170 0.0205
b 1.4206 1.1021 0.9859
c 0.0093 0.0123 0.0144

0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
0 0.02 0.04 0.06 0.08 0.1
Drift Ratio
D
a
m
a
g
e
Low
Moderate
High
Shear

Figure 3-28. Damage Curves Developed for RC Columns

108
Table 3-14. Values of Equation Parameters for Damage Curves of RC Beams
Parameter Low
Ductility
Moderate
Ductility
High
Ductility
a 0.0118 0.0172 0.0340
b 2.8000 1.5000 1.0500
c 0.0100 0.0140 0.0140
Table 3-15. Equation Parameters for the Damage Curves Developed for Brick Infills
Parameter Group 1 Group 2 Group 3 Group 4
a 0.0030 0.0042 0.0055 0.0070
b 7.0000 10.000 12.000 15.000
c 0.0020 0.0020 0.0020 0.0020

c. Component Importance Factors
Once the damage scores for each member of the building is determined, these damage
scores must be combined to determine the damage score of the building in order to
assess its vulnerability. The most appropriate way to combine the component damage
scores seems to be taking their weighted average. The main problem in the weighted
average procedure is the determination of the weighing coefficients for each member
which are called Component Importance Factors. These importance factors must
reflect the importance or contribution of each component in resisting the seismic forces.
To determine these importance factors, a procedure is developed and applied on several
buildings to propose approximate values for the importance factor of each component.
The contribution of each component to the energy dissipation capacity of the entire
structure was chosen as the criterion to evaluate the importance of that component.
A more detailed summary regarding the development of importance factors was reported
elsewhere (Erduran and Yakut, 2004c).
d. Performance Evaluation of the Building
According to the damage criterion set in this procedure, there are mainly four damage
levels: negligible, light, moderate and heavy. In addition to this, the performance of the
buildings under a given earthquake is mainly grouped into three as immediate occupancy
(IO), life safety (LS) and collapse prevention (CP). The first two of the damage levels
used in this study correspond to the immediate occupancy performance criterion. The

109
moderate damage state corresponds to the life safety performance criterion whereas the
heavy damage level corresponds to the collapse prevention. Recalling the damage scores
assigned to the four damage levels, the damage scores corresponding to the performance
levels are summarized in Table 3-16.
Once the damage score of each storey and the entire building is computed, their
performance levels are evaluated using Table 3-16. However, to be able to take the local
failures that may exist in a single storey of a building such as soft storey, an additional
criterion was also set. According to this criterion, if the damage score of a storey exceeds
70%, then the performance level of the building is accepted to be collapse prevention
regardless of the damage score of the entire building.
Table 3-16. Damage Scores and the Corresponding Performance Levels
Damage Score Performance
0% - 10% Immediate Occupancy (IO)
10% - 50% Life Safety (LS)
50% - 100% Collapse Prevention (CP)

3.5.3 Application to Zeytinburnu
Earlier studies revealed that Zeytinburnu, a district in stanbul, possesses a high seismic
hazard and with its existing building stock it is under a very severe seismic risk.
Therefore, the three-tier seismic safety assessment methodology developed by the METU
was applied to all buildings in Zeytinburnu.
3.5.3.1 Walkdown Survey
An immediate result of the walkdown survey yielded the complete inventory of the
building stock in the Zeytinburnu district. The buildings were identified in terms of their
structural systems, their number of storeys and their type of use. Walkdown survey results
indicated that there are currently 16,030 buildings (13,885 RC buildings) in Zeytinburnu.
This observation further indicated that within the past 5 years the building stock of
Zeytinburnu grew nearly 3.4 percent, which corresponds to an average growth of 0.7
percent per year between 2000 and 2005.

110
For Zeytinburnu, a worst-case scenario earthquake of magnitude 7.5 is assumed to take
place in the Main Marmara Fault. The seismic hazard used in this study was calculated for
an earthquake having 50 percent probability of exceedance within fifty years.
The walkdown survey yielded a preliminary seismic performance grading of the existing
RC buildings in Zeytinburnu relative to each other. As described earlier this procedure
helps the decision makers to establish the priority levels for buildings that require
immediate intervention. So a bigger percentage of the buildings with low performance
scores are expected to perform poorly when compared to those having higher
performance scores.
The calculated performance scores of the RC buildings are given in Table 3-17. This table
shows that the calculated performances scores of the buildings are inversely proportional
with the number of storeys. Meanwhile, this response is in good agreement with the
observations made in the surveillance studies made in the aftermath of the recent
earthquakes in Turkey.
Table 3-17. Calculated performance scores of RC buildings having 7 storeys or less
Performance Scores Total Number of
storeys
PS30
30<PS
60
60<PS10
0
100<PS

1-2
0 0 248 1,716 1,964
3 0 41 1,005 409 1,455
4 28 989 1,563 119 2,699
5 638 2,652 972 0 4,262
6 1,625 593 86 0 2,304
7 848 202 0 0 1,050
Total 3,139 4,477 3,874 2244 13,734

3.5.3.2 Preliminary Assessment
In this stage buildings with a seismic performance score of 30 or less were given priority.
There were 3,139 buildings in this group of which a representative sample of 2,397
buildings has been studied. In addition, 639 buildings with performance scores greater
than 30 were also included in the analyses. The main reason of this inclusion was to
assess the correlation between the methods used in the first and the second stage
analyses.

111
In the second stage 3,036 buildings were analyzed. The field surveys were completed
within the first three months of the study. During this phase field teams gathered specific
information about the structural system of each building under investigation.
After calculating the parameters defined earlier the preliminary assessment procedure was
applied. As shown in Table 3-18, the buildings were classified in high, moderate and low
risk groups according to their estimated seismic performance levels.
Table 3-18. Results of the Preliminary Assessment Method
Number of
storeys
High risk
group
Moderate
risk group
Low risk
group
Total
3 10

18 71 99
4 180 81 71 332
5 713 170 46 929
6 808 262 77 1,147
7 387 119 23 529
Total 2,098 650 288 3,036

Table 3-18 illustrates that 2,098 buildings out of 3,036 (69.1 percent) were classified in the
high seismic risk group. Moreover, as can be revealed from Table 3-18, while a smaller
percentage of the low-rise buildings were in the high seismic risk group, bigger portions
of the mid-rise buildings (i.e. 4-7 storey buildings) were placed in this category. This trend
agrees well with the observations made during the reconnaissance studies performed in
the aftermath of the recent earthquakes.
3.5.3.3 Detailed Assessment
The procedure developed has been used for assessment of some buildings selected from
the Zeytinburnu building inventory. In the assessment of the buildings, the elastic
response spectrum proposed in the National Earthquake Hazards Reduction Program
(NEHRP) document published in 2000 was used to represent a ground motion with a
probability of exceedance of 50% in 50 years. The major parameters in this response
spectrum are the spectral acceleration at short periods (SDS) and the spectral acceleration
at the period of 1 sec (SD1).

112
Within the scope of this project, the detailed assessment of five mid-rise buildings (4-6
storeys) with variable material properties was carried out. Table 3-19 presents the SDS
and SD1 values determined for each building based on their location.
The target displacement of each building under the specified ground motion was
computed using the displacement coefficient method summarized in FEMA 356 (2000).
Once the displacement demand is determined, the assessment procedure summarized
above was applied on each building to determine expected performance of each member,
each storey and the entire building under the given ground motion. Table 3-19
summarizes the target roof drift ratio under the given elastic spectrum (
t
), maximum
storey damage score, building damage score and the expected performance of each
building.
Table 3-19. Building Assessment Results
Building
ID
S
DS
(g) S
D1
(g)
Max storey
Damage
(%)
Building
Damage
Score (%)
Expected
Damage
Expected
Performance
BLD1 0.767 0.454 10.85 5.60 Light IO
BLD2 0.692 0.412 28.22 26.06 Moderate LS
BLD3 0.698 0.417 54.20 25.16 Moderate LS
BLD4 0.729 0.432 90.05 42.67 Heavy/Collapse CP
BLD5 0.873 0.692 77.12 40.88 Heavy/Collapse CP

As shown in Table 3-19 one of the buildings assessed was found to be immediately
occupiable under the given ground motion. Two buildings were found to suffer moderate
damage and the remaining two were found to experience either heavy damage or collapse.
The results of the assessment shows that the buildings with favourable material
properties will probably not suffer heavy damage or collapse, while the ones with poor
material properties are highly vulnerable to devastating earthquakes.
3.5.4 Analysis and Design of FRPs for Seismic Retrofit
Surveys made in several cities in Turkey indicated that the number of seismically deficient
reinforced concrete structures is tremendously high. Although strengthening of RC
frames by introducing RC infills to selected frame bays in both directions proved to be an
effective, practical and economical seismic rehabilitation technique; the construction
work involved is tremendously demanding. Furthermore, this procedure requires
evacuation of the building for several months; therefore its applicability in the
rehabilitation of the existing structures, which are currently in use, is neither feasible nor

113
practical. In order to overcome these shortcomings, alternative retrofit schemes are
needed.
These observations forced the researchers to work on developing rapid and effective
rehabilitation techniques. In a similar attempt a research project, which was carried out
and recently completed in the leadership of the Middle East Technical University
(METU) addressed this issue. The main idea was to develop strengthening methods
which would not require evacuation of the building. It was intended to convert the non-
load bearing existing masonry walls and partitions into structural elements which would
form a new lateral load resisting system by strengthening them with CFRP fabrics and
integrating them with the existing structural system. Thus the rehabilitated structure
would have adequate lateral stiffness and lateral load carrying capacity.
The main premise of this practical retrofit scheme is to limit inter-storey deformations by
FRP strengthened infill walls that are integrated to the boundary frame members through
FRP anchors. Based on failure modes and measured deformation limits obtained from a
large database of experiments, a diagonal compression-strut and tension-tie model was
developed. The estimated envelope load deformation response of frame specimens was
compared with the envelope of reversed cyclic test results. A good agreement was
observed between analytical and experimental strength and deformation capacity of
specimens. Finally a simplified model was proposed for use in displacement based design
of FRPs for deficient RC frame buildings. The performance of the FRP retrofit scheme
was demonstrated by analyzing an actual deficient RC building with HCT infills. It was
observed that prior to FRP retrofit about 75% of the columns are in a total collapse limit
state meaning that their deformation capacities are significantly exceeded. Upon FRP
retrofit it was observed that a drift control was provided and deformation demands on
the structure were significantly reduced.
3.5.4.1 Strengthening with FRP
The idea of the FRP retrofit scheme is to reduce inter-storeydeformation demands by
using FRPs to act as tension ties similar to a steel cross-brace configuration. In order to
achieve this, diagonal FRPs bonded on the infill wall is tied to the framing members using
FRP anchors as shown in Figure 3-29. The tension tie formed in this way provides
additional contributions to the load carrying capacity and the lateral stiffness of the
existing RC frame structure, in addition to the strength and stiffness provided by the
compression strut formed along the infill diagonal. Special embedded fan type FRP
anchors formed by rolling FRP sheets are connected in the corner region in order to
achieve efficient use of FRP materials (Figure 3-29). To eliminate premature debonding
of FRP from plaster surface anchor dowels are used along the thickness of the infill wall
(Figure 3-29).

114

FRP anchors
FRP
dowels
FRP sheet

beam
column
nfill wall
f
w

Hollow clay
tiles
plaster
50mm
Forming FRP anchors

Figure 3-29. FRP strengthening method for reinforced concrete frames with infill walls
Experiments conducted on FRP strengthened reinforced concrete frames with infill walls
revealed that there are two dominant failure modes. First mode initiates with the failure
of the FRP anchors in the form of a combined pull-out and slip failure. As soon as the
anchors fail, the load carried by the FRP is transferred to the diagonal compression strut
and failure of the infill wall occurs due to corner crushing. When three CFRP anchors
with a depth of about five times the hole diameter is used per corner on each side of the
infill, it has been observed that anchor failure occurs at an average effective diagonal FRP
strain of about 0.002-0.003. Second possible failure mode occurs because of FRP
debonding from the infill wall surface upon using sufficient number of FRP anchors.
After FRP debonding, previously formed horizontal cracks start to open and the tie
action of FRP is lost, followed by sliding shear failure of the infill wall. First failure mode
is mainly due to insufficient anchor depth and can be avoided by increasing the depth and

115
number of anchor dowels. However, the second failure mode marks the limiting strength
of the strengthened infill. Tests have shown that at a strain level of about 0.004-0.006,
FRP debonding took place resulting in a sliding shear failure of the infill followed by a
sudden drop of strength.
3.5.4.2 Proposed Analytical Model
The analytical structural model of a strengthened frame proposed in this study is shown
in Figure 3-30. Frame elements (beams and columns) are modelled using elastic elements
with predefined plastic hinge regions at their ends. The cross sections of the plastic hinge
regions are discretized into a number of fibres with appropriate uniaxial stress-strain
behaviour for different materials. Unconfined concrete fibres are modelled using
Hognestad stress-strain curve with a linear descending branch up to a strain of 0.004 at
zero stress. Modified Kent and Park model is implemented for core concrete fibres
confined with transverse reinforcement. Steel reinforcement is modelled with an elastic
perfectly plastic material model. The advantage of fibre models is the consideration of
axial load moment interaction during analysis and avoiding the need of performing
sectional analysis separately. Plastic hinge length, which is the length of the region where
inelastic action is expected, is taken equal to the depth of the member. Effective cracked
stiffness equal to the 75% of the gross section properties and modulus of elasticity of
concrete are used between the plastic hinge regions to model the elastic portion of the
frame elements.

Stress
Stress

Strain
crt
f
ut
f
eff f ,
crt
tu
Strain
us
f
so
fs
crs
E
Infill Strut Model

FRP Tie Model

sm
FRP FRP tie tie
Infill Infill strut strut
Plastic Plastic hinge hinge
element element with with fiber fiber
discretization discretization
elastic elastic frame frame
elements elements
OPENSEES Platform
FRP FRP tie tie
Infill Infill strut strut
Plastic Plastic hinge hinge
element element with with fiber fiber
discretization discretization
elastic elastic frame frame
elements elements
OPENSEES Platform

Figure 3-30. Modelling of infill struts and FRP ties

116
3.6 DESIGN OF FRP REINFORCEMENT OF MASONRY INFILL WALLS AGAINST
TRANSVERSE MOVE
3.6.1 Scope of research
Many buildings in Europe today are of unreinforced masonry (URM) construction and
have features, such as infill panels, that could threaten lives under extreme events like
high-winds or earthquakes. There is a need to find cheap and effective means of
strengthening such structures. This report investigates one particular reinforcing
technique, namely the application of fibre reinforced polymer materials to strengthen
URM infill panels subjected to out-of-plane loading. There has been an unprecedented
growth in research in the area of composites in the last 30 years (Bakis et al. 2002),
resulting in a wide range of high-strength, light-weight and environmentally resistant
materials available at low costs. Numerous investigators have explored the potential use
of fibre reinforced polymer (FRP) strengthening on concrete and on masonry structures
under in-plane or out-of-plane monotonic loading. Previous experimental studies have
shown increases in capacity and ductility of the FRP reinforced structures (Tikalsky et al.
1995, Velasquez-Dimas et. al. 2000, Tumialan et al. 2003, Krevaikas and Triantafillou
2005). The behaviour under seismic loading has been addressed very little experimentally
and mostly tests under simulated earthquake loading (large number of cyclic tests) have
been reported (Ehsani et al. 1999). Simple analytical methods to assess the out-of-plane
behaviour of URM or FRP-reinforced panels were developed based on the flexural
theory of masonry present in the building codes (Velasquez-Dimas et. al., II, 2000,
Hamilton and Dolan, 2001). The general trend has been to adapt the monotonic loading
assessment methods for seismic loading as well, but the developed analytical models have
not been validated through dynamic experimentation. A need to evaluate the mechanical
behaviour under truly seismic conditions has emerged. The present study is meant to fill
the gap in the existing knowledge on the out-of-plane dynamic behaviour of FRP
reinforced masonry. Four full-scale quasi-static tests and four full-scale shaking table tests
on both unreinforced and reinforced specimens were performed. The experimental
observations on pre-cracking and post-cracking behaviour extend understanding on the
reinforcement requirements associated with a seismically-sound system. A software tool
(CRAMP) capable of predicting the mechanical response of FRP-reinforced masonry
walls was developed.
3.6.1.1 Objectives and strategy of the present studies
The research programme at Bristol University investigated the seismic behaviour of FRP-
reinforced masonry panels subjected to seismic loading. The specific features and the
objectives associated with the tests are presented in Table 3.6.1

117
Table 3-20 Summary of experimental tests
No. Ref. Reinforcement
info.
Loading Objectives
1 STU1 Unreinforced

3 cycles of
monotonic loading
1. Control specimen for static tests
2. Investigate cracking mechanisms
and compare data with Taylor(1998)
results and CRAMP simulations
3. Load, deflection, arching force
data
2 STU2 Unreinforced

3 cycles of
monotonic loading
1. Improvement of boundary
conditions
2. Compare data with STU1 results
and CRAMP simulations
data
3 STR1 Reinforced
(100%)
Continuous
jacket on tensile
face
3 cycles of
monotonic loading
1. Investigate quasi-static response
2. Compare response with URM
response
3. Inspiration for future FRP layout
design
4. Comparison with CRAMP
simulations
5. Load, deflection, strain, arching
force data
4 STR2 Reinforced
(100%)
Continuous
jacket on tensile
face
FRP fabric
attached to the
soffit of the top
and bottom
beams to prevent
sliding
3 cycles of
monotonic loading
1. Improvement of boundary
conditions
2. Investigate pseudostatic response
and data reproducibility
3. Compare response with URM and
STR1 response
design
5. Comparison with CRAMP
simulations
force data

118
info.
Loading Objectives
5 SEU1 Unreinforced

Exploratory tests
Seismic tests

1. Investigate the modal parameters
of URM
2. Investigate seismic response of
URM
3. Compare data with Taylor (1998)
data
6 SER1 Reinforced
(100%)
Continuous
jacket on both
faces
Exploratory tests
Seismic tests

of FRP-URM
2. Investigate seismic response
3.Compare data with SEU1 and
CRAMP simulations
force, frequency response data
design
7 SER2 Reinforced
(60%)
Vertical strips on
both faces
Exploratory tests
Seismic tests

1. Lower the stiffness to achieve
cracking and better cost / strength
ratio
& seismic response
4.Compare data with SER1, SEU1
5. Load, deflection, arching force,
frequency response data
design

119
info.
Loading Objectives
8 SER3 Reinforced
(40%)
Vertical strips on
both faces
Exploratory tests
Seismic tests

1. Lower the stiffness to achieve
cracking and better cost / strength
ratio
& seismic response
4.Compare data with SER1, SER2,
SEU1 and CRAMP simulations
5. Load, deflection, arching force,
frequency response data
6. Conclusions and
recommendations for FRP layout
design

Note: Test reference codes: STU (static unreinforced panel), STR (static reinforced
panel), SEU (seismic unreinforced panel), SER (seismic reinforced panel).
Although both the masonry and the FRP have a brittle type of failure, their combination
was expected to show increased load-carrying capacity and ductility favoured by the
elastic deformation of fibres and the shear transfer in the resin connection layer. Beside
the strength benefits, a slowly-progressing mode of failure has also been pursued. The
sudden collapse of unreinforced masonry or the explosive failure of the FRP needed to
be substituted by a slow and energy dissipative process that could be the result of the
FRP delamination. It was believed that the layout of the FRP and the reinforcement
ratios could be used as control factors in triggering the failure mechanisms.
Four full-scale quasi-static tests and four full-scale seismic tests were carried out. The
research concentrated on the simplest case of a panel supported at its top and bottom
only, with the vertical sides being unrestrained. All specimens were single-wythe, top and
bottom supported panels (size 3000 x 2000 x 100 mm) using high density concrete bricks.
The latter were chosen so as to maximise the inertia-driven out-of-plane forces.
The tests were intended to give information on the following aspects of mechanical
response:
i) mechanisms of the failure modes
ii) crack patterns and strain profiles

120
iii) influence of the FRP lay-out and reinforcement ratio on the mechanical response
iv) influence of the loading characteristics on the mechanical response
v) evolution of panel stiffness and modal parameters under seismic loading
3.6.2 Experimental programme
3.6.2.1 Materials and panel configuration
The following material properties were investigated prior to the masonry panel
construction:
i) brickwork density
ii) brick units compression capacity
iii) mortar compression capacity
iv) compressive strength of the brick mortar assemblage (tests on three-brick
prisms)
v) modulus of rupture of the brick-mortar assemblage
vi) stress-strain behaviour of the brick-mortar assemblage.
A summary of the material characteristics is given below:
i) concrete facing bricks: to BS6073, size: 215x 102 x 65, density: 2200 kg/m
3
(given by
the manufacturer), compressive strength: 25 MPa (measured).
ii) mortar: general purpose mortar for masonry, corresponding to group III acc. to
BS EN 998-2:2003, composition 1:1:6 (cement: lime: sand) (given by the
manufacturer), compressive strength: 6 MPa(measured).
iii) masonry: max. compressive stress = 18 MPa, max. tens. strain = 0.6 %
(measured).
a. Masonry panels
The panels were built within a pin-jointed steel frame fabricated from 254x254 universal
column sections. The sections were stiffened by thick plates welded longitudinally
between the flanges on both sides of the beams in order to prevent the flange flexure.
The specimens for the pseudo-static tests were located parallel to the laboratory reaction
wall where the hydraulic actuator was mounted. A load distribution beam consisting of a

121
steel square plate and two steel rollers was employed to distribute the load in the central
area of the wall.
The specimens for the seismic tests were built off the shaking table and left to cure using
standard practices. Each masonry wall was built in two lifts and was top and bottom
supported. Mortar was packed on the top side of the wall to fill the gap caused by
shrinkage. The walls were left for 21 days to cure. When the specimens were cured, they
were bolted down onto the shaking table and instrumented.
b. FRP materials
When selecting the FRP materials for reinforcement of panels, the main parameters that
are generally taken into account are the specific strength (tensile strength/specific gravity)
or the specific stiffness (modulus of elasticity/specific gravity). The carbon fibre
composites exhibit better specific strength and specific stiffness values than their
competitors. However, when the seismic behaviour of the retrofitted system comes into
play, the systems ductility and capacity of energy absorption become important. Fibre
ductility increases from carbon to aramid and further to glass. After weighing the benefits
and the disadvantages of a large set of FRP materials (both plates and wraps), a glass-fibre
fabric system was selected for reinforcement (manufacturer: EXCHEM LTD-UK). Its
mechanical properties are given in Table 3.6.2. In particular, the elastic modulus of the
applied fabric is closest to that of URM, and therefore strain compatibility and ductility
performance are likely to be better than a stiffer fibre material. In addition, glass fibres are
significantly cheaper than alternatives.
Table 3-21 FRP system employed in the tests
Fibre Tensile Strength
(MPa)
fibre/ laminate
Modulus of
elasticity
(GPa)
fibre/
laminate
Elongation

(%)
Thickness

mm
Weight

g/m
2

Wrap Type E
(unidirectional
fabric, wet
application)
3450

/1099
73

/42
4.5 0.167 432

Discrete strips or continuous jackets of FRP fabric were used. Samples using various
reinforcement ratios were tested. Preliminary observations on the first tests led to the

122
selection of reinforcement ratios in further experiments. The lay-out of the FRP was
chosen in relation to the boundary conditions of the panel. The URM tests (STU1, STU2)
showed a horizontal cracking line at the midheight of the panel. In order to prevent the
onset of strains at midheight due to arching in vertical plane, the FRP fabric was mounted
with its fibre on the vertical direction.
The FRP mounting procedure started with the preparation of masonry surface for primer
application. Attention focused on cleaning the joints and on removing of excessive
mortar from the wall surface. The surface was cleaned with a wire brush and was checked
for tool marks and other surface variation problems. A dual system that could be used as
a primer and as a laminating resin was employed (Selfix MPA 22 Laminating Resin,
Exchem Ltd.) A foam roller was used to apply a thin layer of primer on the wall surface
(average primer consumption 1 kg/5 m
2
). The primer was used to close the existing
porosity of the wall surface and to provide a good bonding substrate for the resin. Pre-cut
FRP strips of 300 mm by 1960 mm were saturated with resin (average resin consumption:
1 kg/3 m
2
) while lying on a horizontal surface. They were then applied to the wall surface
using a foam roller and by applying hand-pressure. The change of fabric colour from
white to transparent yellow was used as an indication that saturation was reached. A final
layer of laminating resin was applied on the FRP fabric for full saturation, protection and
instrumentation purposes. The FRP layouts are shown in Figure 3-31 and Figure 3-33.
The FRP fabric was mounted at 21 days after the building of the wall. The testing took
place at a minimum of 10 days after the FRP installation.
3000
2000
3000
2000

Figure 3-31 Layout of FRP fabric on the masonry panel (STR1, STR2, SER1)

123
3000
300
2000
3000
300
2000

Figure 3-32 Layout of FRP fabric on the masonry panel (SER2)
3000
300
2000
3000
300
3000
300
2000

Figure 3-33 Layout of FRP fabric on the masonry panel (SER3)
A photograph of the SER2 panel (60% reinforcement) is shown in Figure 3-34.

Figure 3-34 SER2 panel ready for testing (60% reinforcement)

124
3.6.2.2 Instrumentation
CELESCO type PT101 displacement transducers measured the out-of-plane deflection.
The transducers were placed at mid-height and at positions symmetrically-located about
the centre line. Four load cells were incorporated in the four bolts attaching the top beam
to the frame columns. The load cells measured the arching forces on the wall. The top
beam was separated from the top of the column by washers around the instrumented
bolts. Strain gauges recorded the FRP tensile strain at mid-height and at locations
symmetrically located about the centreline. The strain gauges were mounted on the FRP
fabric in areas located on top of a brick unit. For the monotonic tests only, one 200 kN
load cell was mounted on the same axis with the hydraulic actuator to measure the
applied load. Three SETRA type 141A accelerometers were installed on the shaking table
(X, Y, Z directions) while another three were installed on the panel to measure out-of-
plane accelerations. A seventh accelerometer was installed on the frames top beam also
to measure out-of-plane accelerations.
3.6.2.3 Seismic tests input motions
a. Exploratory tests
The exploratory tests were used for measuring the modal parameters of the specimen.
The resonant frequency of the first mode of vibration and the viscous damping of the
specimen were the main parameters of interest. The exploratory tests used broadband (0-
100) Hz random noise applied on the out-of-plane axis only (Y axis).
b. Seismic tests
An elastic response spectrum for soil type B acc. to Eurocode 8 was used in the tests
(Table 3.6.3 and Figure 3-35). Acceleration, velocity and displacement time histories were
generated to match the Eurocode 8 response spectrum. The displacement time history
was fed into the shaking table control software.
Table 3-22 Parameters of design response spectrum
Ground type S T
B
(s) T
C
(s) T
D
(s)
B 1.2 0.15 0.5 2

125
0
0.5
1
1.5
2
2.5
3
3.5
0.5 2 6.6 40 100
frequency (Hz)
S
e
/
a
g

Figure 3-35 Required response spectrum as per Eurocode 8 (Soil B)
3.6.2.4 Quasi-static tests experimental observations
a. STU1 and STU2 (quasi-static tests on unreinforced panels)
The URM wall cracked along an horizontal line situated at the midheight of the wall
(Figure 3-36) This cracking pattern had been observed by other researchers and it is
considered typical for the top-and-bottom-supported masonry panels. The main
resistance mechanism was arching of the panel with increasing of contact forces between
panel and top and bottom supports.

Figure 3-36 Experimental rig used for the static tests (STR1, STR2)

126
b. STR1 and STR2 (quasi-static tests on reinforced panels)
The presence of the FRP brought changes to the following aspects of mechanical
response:
i) sequence of occurrence of the failure mechanisms.
ii) masonry cracking pattern
iii) lateral deflection profile
iv) distribution of strains in the masonry substrate
v) mechanical parameters: loading capacity, stiffness, pseudo-ductility
The FRP was a unidirectional fabric with vertical glass fibres oriented perpendicular to
the wall supports. The orientation of fibres (which had a very high tensile strength:
1090 MPa) prevented the initiation of vertical strains in the masonry, therefore no
horizontal crack was formed in the initial stage of loading. The low compressive strength
of masonry (approx. 18 MPa as measured in the material tests) caused an early vertical
crack to occur at the mid-width of the panel. The interlaminar resin gave way and caused
the fabric to tear in the vertical direction (Figure 3-37). The strain gauge situated at
midheight was damaged following the propagation of rupture in the FRP along the
vertical direction.

Figure 3-37 Crack pattern in reinforced walls subjected to out-of-plane monotonic loading (STR1)
An important experimental observation was that the vertical FRP strips changed the way
the panel arched. The wall flexure in the vertical plane was no longer possible and the
panel arched in horizontal plane.

127
The main failure modes identified during the STR1 and STR2 tests were: the flexural
failure, the shear failure and the FRP debonding. Flexural failure was triggered by the
crushing of the masonry in compression and by the rupture of the FRP in tension. Initial
cracking occurred at the interface of mortar and masonry. The initial cracking was
delayed by the presence of the FRP. Since the tensile strength of the masonry was lower
than that of the epoxy resin (55 MPa), the failure line appeared first in the masonry. The
vertical crack in the masonry caused the wall to slide while a differential displacement in
the shear plane developed. The shear failure mechanism resulted in the FRP debonding
alongside the masonry crack. The FRP fabric situated above the opening crack of the
masonry started to deform and broke in tension vertically. Debonding of the FRP
laminate from the masonry substrate was triggered by the central vertical crack in the
masonry and propagated at first vertically, then multiaxially towards the final stages of
loading. It was noticed that part of the concrete brick faceshell remained attached to the
laminate. This showed that the resin-masonry connection was stronger than the tensile
strength of the masonry. End-debonding of the FRP was expected as a main failure
mode, but this did not occur.
The strength and stiffness of the STR walls increased dramatically in comparison to the
STU specimens. A 2.6 times increase in maximum strength was recorded (see Figure
3-38).
0 10 20 30 40 50 60 70 80
0
5
10
15
20
25
30
35
40
midheight deflection (mm)
o
u
t
-
o
f
-
p
l
a
n
e

l
o
a
d

(
k
N
)

R2
R1
R3
R4
U2 U-collapse
U1
R-collapse

Figure 3-38 Loading curves during static testing (STU1 -unreinforced wall, STR1 reinforced wall)

128
0 10 20 30 40 50 60 70 80 90 100
0
500
1000
1500
2000
deflection (mm)
h
e
i
g
h
t

(
m
m
)
red - R
sliding collapse
R4 (37 kN)
black - U
U2 (13 kN)
R2 (29 kN)
R3 (35 kN)

Figure 3-39 Comparison of deflection profiles for reinforced (STR1) and unreinforced (STU1) walls
The elastic behaviour of the FRP under loading was obvious. The first loading cycle (R1)
was almost a straight line. There was very little degradation of stiffness due to repeated
loading, again because the FRP behaved elastically up to failure. Overall, the FRP
presence caused an increase in capacity (from 13.5 kN to 35 kN) with confinement of the
lateral deflections from 30 mm (STU1) to 24 mm (STR1) (Figure 3-39).
In the STR2 test the FRP fabric was folded and glued to the soffit of the top and bottom
beams in order to prevent sliding. Arching took place in horizontal plane. Cracking
occurred at midwidth on the vertical direction. The FRP broke at the interface between
the wall and the top beam due to wall sliding. The cracking pattern was a vertical line
situated at the middle of the panel. The main observation coming out from this test was
that by attaching the FRP jacket to the soffit of the top beam the sliding of the wall could
be prevented. This created uniform boundary conditions for the top and bottom
interfaces which lead to a vertical failure line situated at the midwidth of the panel.
3.6.2.5 Seismic tests
The dynamic properties of a structure are defined by its mode shape, damped natural
frequency and damping for each mode of vibration over some frequency range of
interest. The exploratory tests carried out prior and during the seismic testing revealed
information about the above parameters of interest for URM and reinforced panels in
intact or cracked form. An example of damping ratio measurement is given in Table 3.6.4
for an URM wall (SEU1) and in Table 3.6.5 for a reinforced wall (SER1). Both panels
were intact when the mentioned measurements were taken.

129
The evolution of panel stiffness during testing takes place in parallel with a measurable
shift in panels natural frequency. The natural frequency of vibration increased from 12
Hz for the URM panel (SEU1) up to 30.31 Hz for the reinforced panel with maximum
reinforcement ratio (SER1). Lowering the reinforcement ratio from 100% to 40% led to a
shift in the natural frequency from 30.31 Hz to 15 Hz (Table 3.6.6).
Table 3-23 Damping corresponding to main modes of vibration in SEU1 (intact wall)
Natural
Freq (Hz) Damping ratio (%)
12.00 4.30
19.94 1.53
23.38 0.32
29.46 3.81
37.62 4.09
Table 3-24 Damping corresponding to main modes of vibration in SER1 (intact wall)
Natural
Freq (Hz)
Damping ratio (%)
30.31 4.87
53.58 2.68

Table 3-25 Evolution of panels natural frequency during seismic testing
Test Reinforcement
information
Uncracked
natural
frequency
(Hz)
Cracked
natural
frequency
(Hz)
Cracking table
acceleration or max. table
acceleration
SEU1 unreinforced 12.00 7.50 cracking: 1.175 g (SEU1_60),
collapse: 2.4g
SER1 100% reinf. 30.31 uncracked max : 3.53g (SER1_130)
SER2 60% reinf. 28.25 uncracked max : 2.28g (SER2_120)
SER3 40% reinf. 15.00 8.20 cracking:2.44g (SEU3_80)


130
3.6.2.6 Seismic tests experimental observations and numerical simulations
Particular attention has been given to tests SEU1_65 that led to the URM wall cracking
and SEU1_110 that led to its collapse. The table acceleration reached 2.44g in SEU1_110
(Figure 3-40).
The lateral deflection results collected at midheight on the panel are shown in Figure
3-41.
A simulated evolution of wall displacement in time is presented in Figure Figure 3-42.
The measured and the simulated arching forces are shown in Figure 3-43 and Figure 3-44.
An interesting part of SEU1 test was the SEU1_65, when the imposed seismic input led
to cracking. The input motion lasted for 40 s and accelerations up to 1.175 g were
recorded by the Y-axis accelerometer mounted on the shaking table (Figure 3-45).
10 15 20 25 30
-3
-2
-1
0
1
2
3
time(s)
s
u
p
p
o
r
t

a
c
c
e
l
e
r
a
t
i
o
n
(
g
)
Expriment, SEU1-110

Figure 3-40 Shaking table acceleration in test SEU1_110

131

Figure 3-41 Midheight deflection measured in test SEU1_110 (left general view, right detail)
10 12 14 16 18 20 22
-40
-30
-20
-10
0
10
20
30
40
Simulation, SEU1-110
time (s)
m
i
d
h
e
i
g
h
t

d
e
f
l
e
c
t
i
o
n

(
m
m
)

Figure 3-42 Simulated midheight deflection in test SEU1_110

Figure 3-43 Arching forces measured in test SEU1_110 (left -south face, right north face)

132
10 15 20 25 30
-50
0
50
Simulation, SEU1-110
time (s)
a
r
c
h
i
n
g

f
o
r
c
e
s

(
k
N
)

Figure 3-44 Simulated arching forces for test SEU1_110

Figure 3-45 Shaking table and wall accelerations in test SEU1_65

Figure 3-46 Midheight deflection in test SEU1_65 (left- measured, right-simulated)

133
The experimental results were compared to simulated data obtained in CRAMP. The
simulations used the same table motion input used in the experiments.
The main features of CRAMP simulation: wall parameters: height: 2 m, thickness: 0.1 m,
width: 3 m, mass density: 2200 kg/m
3
, compressive strength: 4.8 MPa, max strain: 0.006,
ultimate strain: 0.0065, damping ratio: 4.3 %. Top and bottom supports of very high
stiffness, no sliding allowed.
a.Reinforced panels (SER1, SER2, SER3)
The 100% reinforced SER1 panel withstood accelerations of up to 3.53g (Figure 3-47)
without cracking. The panel behaved like a highly-stiff rigid block and no significant
lateral deflections or arching forces were recorded. The imposed shaking table
displacement history was amplified from 2% onwards. The shaking table could not cope
to amplification factors higher than 130% and the testing was stopped.
The 60% reinforced SER2 panel withstood accelerations of up to 2.28g without cracking.
It is worth mentioning that the URM panel collapsed at 2.4g. Therefore, the presence of
the FRP led to a large capacity increase. The SER2 test did not lead to any cracks in the
panel and it was stopped at a table displacement amplification factor of 120%. Minor
deflections with the average ranging from +2 to -2 mm and max peak of 8.4 mm were
recorded (Figure 3-48). Arching was limited by the vertical strips of the FRP and low
contact forces were recorded.
The first cracks in the 40% reinforced SER3 panel appeared at accelerations of around
2.44g (Figure 3-49). During the SER3_80 test, the first cracks appeared in the
unreinforced areas of the panel as step-like lines propagating diagonally from the corners
of the FRP strips.

Figure 3-47 Support acceleration and wall deflection in test SER1_130 (100% reinforcement)

134


3.6.2.7 Conclusions from the experimental programme
The URM walls cracked along a horizontal fracture line, typical for their boundary
conditions. The crack pattern for the URM walls was the same in both the static and the
seismic tests. Arching in vertical plane was the main resistance mechanism in the URM
walls.
The FRP application procedure proved to be a simple one and it did not require highly-
skilled personnel to accomplish. The entire reinforcement activity took on average 5 hrs
for each panel.
The FRP reinforced panel subjected to monotonic out-of-plane loading withstood forces
that were 2.7 times larger than the maximum load taken by the URM walls.
A summary of relevant data from seismic testing is presented in Table 7. The reinforced
panels exhibited increased load-carrying capacity. A layout of vertical FRP strips

135
prevented the panel from arching in the vertical plane. A relatively low reinforcement
ratio of 40% was sufficient to bring sufficient strength and ductility to the reinforced wall
(SER3_80, a = 2.44 g, first cracks) in order to cope with seismic forces that previously
caused the collapse of the URM wall (SEU1_110, a = 2.4 g).
Higher reinforcement ratios (60% and 100%) led to a dramatic increase of panels
stiffness and natural frequency of vibration. The arching was completely prevented and
the walls systems behaved like rigid blocks exhibiting very small lateral deflections.
Table 3-26 Summary of relevant results from seismic testing
Test Reinforcement
information
Cracking table acceleration
or max/col. table
acceleration
Cracking Max.midheight
deflection
SEU1 unreinforced cracking: 1.175 g (SEU1_65),
collapse: 2.4g (SEU1_110)
-cracked
- brought to
collapse
8 mm (SEU1_65)
42 mm (SEU1_110)
SER1 100% reinf. max : 3.53g (SER1_130) -uncracked 7.1 mm (SER1_130)
SER2 60% reinf. max : 2.28g (SER2_120) -uncracked 8.1 mm (SER2_120)
SER3 40% reinf. cracking:2.44g (SEU3_80) -cracked 26 mm (SER3_80)

3.6.3 Analytical studies
The numerical modelling part of this study was based on the assumption that a panel,
whether unreinforced or reinforced, would crack in out-of-plane flexure early in the
earthquake loading and then behave dynamically as a non-linear assembly of parts.
However, as demonstrated by the experimental programme, the FRP reinforced panels
did not behave in this way, remaining intact and essentially elastic for reinforcement ratios
of 60% and above. In view of this, the numerical modelling developments proved
essentially redundant and will only be described here briefly in overview for
completeness.
The modelling of the FRP reinforced panels built upon an existing model for predicting
the seismic behaviour of URM panels (Taylor 1998). The URM model is embedded in the
CRAMP program (Cracked Response Analysis of Masonry Panels) written in MATLAB
6.5. The main features of the CRAMP model are given below:
i) The masonry panel is considered a SDOF system which exhibits residual
strength due to arching action (simple case: rigid top and bottom supports,
horizontal crack along the wall length) (Figure 3-50).

136
ii) Strip analysis technique (each vertical strip consists of two segments that join at
the location of the crack).
iii) Non-linear dynamic analysis. Equation of motion solved using the linear
acceleration time-stepping formulation.
iv) Seismic capacity is expressed in the form of fragility curves.
v) Panel segments are assumed to behave non-linear elastically, with the masonry
following a simple parabolic stress-strain model. Crushing at the faces of the
panel is allowed for.
CRAMP enables multiple time history simulations to be run for a given panel
configuration. The outputs of these simulations are then assembled in the form of
fragility curves from which the probability of failure relative to, for example, peak input
acceleration may be found.
As the CRAMP model involves several significant simplifying assumptions, and the
problem is highly non-linear and therefore sensitive to parameter and loading variability,
it is unlikely to yield a very close match at a time history level with any given experimental
result. However, previous experience with earlier CRAMP models (Taylor, 1998) gave
confidence that probability estimates based on CRAMP outputs were meaningful.
arch thrust
arch thrust
p
height con_length
equivalent
compression spring
arch thrust
arch thrust
p
height con_length
equivalent
compression spring

Figure 3-50 Modelling of top and bottom supported URM panels

137

Figure 3-51 Fragility curve for a panel height of 4m (t=0.229m, mass density =1950 kg/m,
f
c
=4.8 MPa,
max
=0.4%,
ult
=0.45%, viscous damping =5%)
Since the FRP material will never reach its max tensile strength (1099 N/mm
2
), it is fair
to say that the FRP membranes behave elastically during testing and that no rupture of
fibres takes place.
Figure 3-51 shows a typical fragility curve from CRAMP.
3.6.4 Conclusions
This study has shown the efficacy of a relatively simple method of reinforcing
unreinforced masonry infill panels with glass fibre FRP laminates. Provided the
reinforcement covers at least 60% of the surface area of the panel (preferably 100%), and
the reinforcement overlaps the surrounding frame by at least 75mm, the reinforced panel
gains a significant amount of strength to out-of-plane seismic loads. So great are the
strength and stiffness increases that, for typical panel configurations, it appears unlikely
that a strengthened panel would suffer significant distress during an earthquake due to
out-of-plane loads. (Note, however, that this study has not considered combined in-plane
and out-of-plane loading.) The evidence derived from this study suggests that this kind of
FRP strengthening can simply be specified, rather than requiring detailed engineering
calculations for its justification.
The method of application of the FRP reinforcement is very simple, requiring minimal
training. It is within the ability of a semi-skilled builder, subject to appropriate
supervision. The chosen glass fibre based materials are relatively cheap.

4.SEISMIC DESIGN AND RETROFIT OF STRUCTURES
USING DISSIPATIVE DEVICES
4.1 RC STRUCTURES
4.1.1 Introduction
Conventional seismic design practice permits designing precast and cast in place
reinforced concrete (PRC and RC, respectively) structures for forces lower than those
expected from the elastic response on the premise that the structural design assures
significant energy dissipation potential and, therefore, the survival of the building when
subjected to severe earthquakes. Normally, energy dissipation during seismic actions
occurs in critical zones of the structure specially designed to admit large ductility
demands. Frequently, these zones are located near the beam-column joints and during
earthquakes, several structural members can suffer a great amount of damage with
irreversible degradation of the mechanical properties of the materials. Even if a limited
level of structural damage dissipates part of the energy offering certain level of protection
against seismic actions (Mata et.al. 2006, Lessloss Deliverable report 54, 2005
(www.lessloss.org)), the large displacements required for developing hysteretic cycles in
dissipative zones can cause severe damage in those members, but this situation is
generally considered economically acceptable if life safety and collapse prevention are
achieved.
In the last decades, new concepts for the design of building, based on the manipulation
of the energy dissipation, have improved the seismic behaviour of the RC and PRC
structures providing higher levels of safety for the occupants, buildings and nonstructural
components.
The new techniques are based on adding devices to the buildings with the main objective
of dissipating the energy demand imposed by the earthquake alleviating the ductility
demand on primary structural elements and decreasing the acceleration response (Soon
and Dargush, 1997; Handson et.al. 1993). The purpose is to control the seismic response of
the buildings by means of a set of dissipating devices which constitutes the control system,
adequately located in the structure.

140
The passive energy dissipating devices (EDD) work dissipating the energy of the earthquake,
localizing and concentrating the damage of the structure in the devices without the need
of an external energy supply (Soon and Dargush 1997). Today, the passive control of
structures is widely accepted by the engineering community.
By one hand, passive EDDs, also called supplemental dampers, are devices located
throughout the structure to absorb and dissipate an important part of the energy input
induced in the structure by earthquakes. The effectiveness of the implementation of
EDDs in RC structures can be analyzed starting from the energy balance equation as:
d D S K I
E E E E E + + + = (4.1.1)
Where E
I
is the absolute earthquake energy input, E
K
is the absolute kinetic energy, E
S
is
the elastic strain energy, E
D
is the energy dissipation due to inelastic behaviour in the
structure (including viscous effects) and E
d
is the contribution of energy dissipation due
to the addition of EDDs.
Using the assumption that the term E
d
has no influence on E
I
, it is possible to see from
Eq. (4.1.1) that increments of the contribution of E
D
+ E
d
implies reductions of E
K
+E
S

and, therefore, lower responses are obtained when extra energy dissipation is provided
(Aiken, 1996; Soon and Dargush, 1997). The main objective of designers when applying
passive control in improving the seismic behaviour of PRC structures, is defining
appropriately the properties of the EDDs in such way that the inelastic demand on
primary structural members E
D
be transferred to the term E
d
. After a sever earthquake,
EDDs can be replaced by new ones if necessary. EDDs can be classified according to the
nature of their dissipative mechanism in displacement dependent, e.g. friction, metallic and
extrusion devices, velocity dependent e.g. viscous, mixed e.g. viscoelastic and others such as
tuned liquid and tuned mass dampers. A great amount of works comparing the ability of
different passive EDDs in controlling the seismic response of structures is available in
the literature (Luck et.al., 1996).
For Aiken (1996) presents the contribution of the extra energy dissipation due to EDDs
as an equivalent damping added to the linear bare structure and displacement reduction factors are
given as a function of the damping ratio added to buildings. Lin and Copra study the
accuracy in estimating the dynamic response of asymmetric one-storeybuildings equipped
with EDDs, when the dissipating devices are replaced by their energetic equivalent linear
viscous dampers; a correction factor is provided estimating the maximum forces in the
EDDs.
Today, only a few countries have design codes for RC buildings with EDDs; in United
States, the US Federal Emergency Management Agency (FEMA) proposes code provisions and

141
standards pertaining to the design of EDDs devices for use in buildings. The document
includes Prestandard and Commentary for the Seismic Rehabilitation of Buildings (FEMA 356,
2000) and NEHRP Recommended Provisions for Seismic Regulations for New Buildings and Other
Structures (FEMA 368, 2000).
In the case of Europe, efforts have been focused mainly in developing codes for base
isolation. In the case of EDDs the developments have been limited to guidelines rather
that codes or official standards.
A great part of the design methods for RC and PRC structures are based on supposing
that the behaviour of the bare structure remains elastic, concentrating the energy
dissipation demands on the control system. However, even though this assumption can
be useful for a preliminary design, experimental and theoretical evidence show that
inelastic behaviour will occur in the main structural elements during severe strong motion
as noted by Shen and Soong (2005). It is widely recognized that nowadays it is possible to
carry out nonlinear time-history analysis, which can provide a most precise evaluation of
the response of structural systems incorporating EDDs (see Mata et.al. 2006; Valles et.al.
1996). But it also requires relatively large amounts of computational time. It is also
recognized that equivalent linear dynamic procedures are not able to represent the fully
complexity of the nonlinear dynamic behaviour of RC and PRC structures.
Independently of the type of analysis chosen, it is clear that the nonlinear time history
analysis has gained space in the passive control of RC and PRC structures subjected to
earthquakes, therefore, sophisticated numerical tools became more necessary for both
academics and practitioners. During the last decades, great efforts have been done in
developing numerical formulations and their implementation in computer codes for the
simulation of the nonlinear dynamic response of RC and PRC structures, for example a
recent state of the art review for the case of concrete structures can be found in reference
(Spacone and El-Tawil, 2004).
The engineering community agrees with the fact that the use of general fully 3D
numerical techniques, such as nonlinear finite elements (FE), constitutes the most precise
tools for the simulation of the behaviour of RC buildings subjected to earthquakes (Soon
and Dargush, 1996; Spiliopoulos and Lykidis, 2005). However, usually the computing
time required when using full models of real structures became their application
unpractical. Considering that most of the elements in RC and PRC buildings are columns
or beams, one-dimensional formulations for structural elements, appear as a solution
combining both numerical precision and reasonable computational costs (Mata et.al.,
2007). Experimental evidence (Bayrak and Sheikh, 2001) shows that nonlinearity in beam
elements can be formulated in terms of cross sectional forces and moments (Coleman

142
and Spacone, 2001). Some formulations of this type have been extended to take into
account geometric nonlinearities (Sivaselvan et.al., 2002; Vignjevic, 1997).
An additional refinement is obtained considering inhomogeneous distributions of
materials on arbitrarily shaped beam cross sections [Kumar et.al., 2004; Valles et.al., 1996].
Specific numerical models based on a secondary discretization of the beam cross sections
have been developed allowing to include multiple materials. Using this approach the
mechanical behaviour of beams constituted by complex combinations of materials, such
it is the case of RC beams, can be simulated (Hajjar, 2000). In general, the engineering
community agrees with the fact that although these models are computationally more
expensive, they allow estimating more precisely the nonlinear response of RC structures
(Davenne et.al., 2003; Taucer et.al., 1991). Formulations of this kind considering both
constitutive and geometric nonlinearity are rather scarce (Driemeier et.al., 2005);
moreover, most of the geometrically nonlinear models for beams are limited to the elastic
range of materials, as it can be consulted for example in references (Reissner, 1972; Simo,
1985) and the treatment of inelasticity has been mainly restricted to plasticity. Recently,
Mata et.al. (2007) has extended the geometrically exact formulation for beams due to
Reissner-Simo (Reissner, 1972; Simo, 1985; Simo and Vu-Quoc, 1986 and 1988) for
considering and arbitrary distribution of composite materials with the cross sections for
the static and dynamic cases.
From the numerical point of view, the nonlinear behaviour of EDDs usually has been
described in a global sense by means of forcedisplacement or moment-curvature
relationships that intend to capture appropriately the force/moment level and the energy
dissipating capacity existing in the devices. That is to say that, a rather simplified
description appears to be enough for the mechanical characterization of the EDDs.
The inclusion of EDDs in a software package for the seismic analysis of RC structures is
frequently done by means of using link elements equipped with the mentioned nonlinear
relationships (Soon and Dargush, 1996). In this way, the link elements connect the
different points of the model of structure which represent the anchorage point of the
EDDs in the real buildings. During the seismic event the relative displacement and/or
rotation between the anchorage points activates the dissipative mechanisms of the
devices.
Nowadays, there are several numerical codes available for the study of the nonlinear
seismic response of RC structures with EDDs (Valles et.al., 1996, Lessloss Deliverable
report 54, 2005 ( www.lessloss.org )). Commercial software has incorporated different
kind of inelastic analysis including beam element with plastichinges and specific link
elements for EDDs.

143
In this report, a specific geometric and constitutive formulation for beam structures with
EDDs used for the determination of the seismic response of RC and PRC buildings is
presented. In this work, the general nonlinear constitutive behaviour is included in the
dynamic version of the geometrically exact formulation for beams due to ReissnerSimo
considering an intermediate curved reference configuration between the straight reference
beam and the current configuration. The displacement based method is used for solving
the resulting nonlinear problem. Plane cross sections remain plane after the deformation
of the structure; therefore, no cross sectional warping is considered, avoiding to include
additional warping variables in the formulation or iterative procedures to obtain corrected
cross sectional strain fields. An appropriated cross sectional analysis is applied for
obtaining the cross sectional forces and moments and the consistent tangential tensors in
the linearized problem. Thermodynamically consistent constitutive laws are used in
describing the material behaviour, which allows to obtain a more rational estimation of
the energy dissipated by the structures. The simple mixing rule is also considered in
modelling materials which are composed by several simple components.
EDDs are considered as beam element without rotational degrees of freedom. A brief
presentation of damage indices capable of estimate the remaining load carrying capacity
of buildings is also given. The discretization by means of the FEM method and the
numerical implementation in a computer program is also explained.
4.1.2 Nonlinear analysis of beam structures
The three dimensional nonlinear analysis of beam structures has captured the interest of
many researchers during the past decades. Many contributions have been focused on the
formulation of geometrically consistent models of beams undergoing large displacements
and rotations, but considering that the material behaviour remains elastic and, therefore,
employing simplified linear constitutive relations in terms of sectional forces and
moments. On the contrary, the constitutive nonlinearity in beams has been described by
means of concentrated and distributed models, both of them formulated, in the most cases,
for small strain and small displacement kinematics hypothesis. Only a few works have
been carried out using a fully geometrical and material nonlinear formulation for beams,
but they have been mainly focused on plasticity.
One of the most invoked geometrically exact formulations is that of Simo (Simo 1985;
Reissner, 1972), which generalize to the three-dimensional dynamic case the formulation
originally developed by Reissner for the plane static problem. This formulation allows to
consider finite shearing and extension. Posteriorly, Simo and Vu-Quoc (Simo and Vu-
Quoc, 1986 and 1988) implemented the numerical integration of the equations of motion
of rods in the context of the finite element (FE) framework for the static and dynamic
cases.

144
Works on constitutive nonlinearity have progressed based on a different approach, thats,
lumped and distributed plasticity models. The lumped plasticity models consider linear
elastic structural elements equipped with plastic hinges at the ends. This method is
recommended by certain authors due to its numerical efficiency when compared with the
full threedimensional (3D) formulation of the problem.
In the case of distributed plasticity models, the constitutive nonlinearity is evaluated at a
fixed number of cross sections along the beam axis, allowing to obtain a distributed
nonlinear behaviour along the structural elements. These sections are divided into a
number of control points corresponding to fibres directed along the beams axis.
Therefore, this approach is frequently referred as fibre approach (Shao et.al., 2005, Lessloss
Deliverable report 54, 2005 (www.lessloss.org)). The employment of fibres allows
predicting a more realistic strainstress state at the cross section level, but it requires the
definition of uniaxial constitutive laws for each material point.
Both approaches are affected by the strain localization phenomenon when materials with
softening behaviour are employed. In any case, the whole structural response becomes
mesh dependent if no appropriate corrections are considered. Hanganu et.al. (2002) and
Barbat et.al. (1997) regularize the energy dissipated at material point level, limiting its
value to the specific fracture energy of the material (Oliver et.al, 1990). This method
ensure that the whole structural response remains objective, but the length of the zone
where softening occurs is still mesh dependent.
One of the most common limitations of the distributed formulations lies in the fact that
inelasticity is defined for the component of the strain acting in the direction normal to the
face of the cross section and, therefore, the shearing components of the stress are treated
elastically. This assumption does not allows to simulate the nonlinear coupling between
different stress components at constitutive level, resulting in models where sectional
shear forces and torsion moments are transmitted elastically across the elements
(Davenne et.al., 2003; Neuenhofer and Filippou, 1997).
Works considering both constitutive and geometric nonlinearity are scarce and they have
been mainly restricted to plasticity (Saje et.al., 1998). Outstanding works considering
warping of arbitrary sections made of rate dependent and rate independent elasticplastic
material are proposed by Simo et.al. (1984) and Gruttmann et.al. (2000), respectively.
4.1.2.1 Finite deformation initially curved beams
The formulation of Simo and Vu Quoc for beams that can undergo large deformations in
space is expanded considering an intermediate curved reference configuration between
the straight reference beam and the deformed beam in the current configuration.

145
The straight reference beam is defined simply by means of straight fixed curve
parameterized in terms of its arch-length coordinate. Beam cross-sections are described
by means of a local coordinate system contained in the plane of the section. The curved
reference beam is defined by means of a spatially fixed curve
0
. Additionally, each point
on this curve has rigidly attached an orthogonal local frame, {t
} (=2,3), constructed by
means of an orientation tensor
0
contained in SO(3). where the rotation manifold is
denoted by SO(3). The beam cross-section, A, is defined considering this local coordinate
system. The planes of the cross sections are normal to the vector tangent to the reference
curve.
The motion deforms the centroid line of the curved reference beam adding a translational
displacement and the local orientation frame is simultaneously rotated together with the
beam cross section. In general, the normal vector to the cross section does not coincide
with the tangent to the current curve because of the shearing. The position vector of any
material point on the current reference beam, x, is:
=
+ =
3
2

t x (4.1.2)
where curve is the current centroid curve, and

is the current rotation tensor.

Figure 4-1. Configurational description of the beam
Eq. (4.1.2) implies that the current beam configuration is completely determined by the
pairs (,).
By the other hand, the deformation gradient (tensor) is defined as the gradient of the
deformation mapping and determines the strain measures at any material point of the
beam cross section. The deformation gradients of the curved reference beam and of the
current beam referred to the straight reference configuration are denoted by F
0
and F,

146
respectively. The deformation gradient F
n
= FF
0
1
relating the differential arch length
elements of the curved reference configuration with the current placement is responsible
for the development of strains, and can be expressed as
(4.1.3)
Where (.),S denotes partial differentiation with respect to the arc-length parameter S,
n
and
n
are the spatial curvature tensor and the rotation tensor relative to the curved
reference beam. The material representation of F
n
is obtained as F
m
n
=
T
F
n
0
.
Removing the rigid body component from F
n
, it is possible to construct the strain tensor
n
= F
n

n
, conjugated to the asymmetric First Piola Kirchhoff (FPK) stress tensor P,
which is referred to the curved reference beam. P
j
is the corresponding FPK stress vector
acting on the deformed face in the current beam corresponding to the normal t
0j
in the
curved reference configuration. The spatial strain vector acting on the current beam cross
section relative to an element of area in the curved reference beam is obtained as
n
=
n
t
01
. The spatial form of the stress resultant n and the stress couple m vectors can be estimated
from the stress vector P
1
as:
(4.1.4)
According to the developments presented by Antman, (1991), the classical form of the
equations of motion of the Cosserat beam for the static case are:
(4.1.5)
Where n
,S
and m
,S
are the external body force and body moment per unit of reference length at
time t. A
0
, S
0
and I
0
are the mass density, the first mass moment density and the
second mass moment density per unit of length of the curved reference beam,
respectively. The terms D
1
and D
2
appear due to the initial curvature and an
inhomogeneous distribution of the material properties; and their contribution to the
dynamics can be neglected or added to the external forces and moments.
Considering a kinematically admissible variation, h=(,), of the pair (,), taking the
dot product with Eq. (4.1.5) and integrating over the length of the curved reference

147
beam, we obtain the nonlinear functional G(x,h) corresponding to the virtual work principle,
with the following form:
(4.1.6)
4.1.3 Nonlinear constitutive models
Frequently, the material properties have been assumed hyper elastic, isotropic and
homogeneous (Ibrahimbegovic, 1995) and, therefore, the reduced constitutive equations
became very simple. In this case, realistic studies involve constitutive nonlinearities as
well as geometric effects for representing appropriately an adequate combination of
strength, ductility and stiffness (Elnashai and Mwafi, 2002). Material points on the beam
cross sections are considered as formed by a composite material corresponding to a
homogeneous mixture of different simple components, each of them with its own
constitutive law (see Figure 4-2). The behaviour of the composite is obtained by means of
the mixing theory described in the following sections.

Figure 4-2. Cross section showing the composite associated to a material point
Two kinds of nonlinear constitutive models for simple materials are used in this work: the
damage and plasticity models. They correspond to a particular cases of more general
formulations, (Barbat et.al. 1997), but formulated in a way such that it is possible to
include them in the geometrically exact formulation for beams previously described. The
constitutive models are formulated in terms of the material form of the FPK stress and
strain vectors, P
m
1
and
n
, respectively.

148
4.1.3.1 Degrading materials: damage model
The behaviour of most of the degrading materials is presented attending to the fact that
microfissuration in geomaterials occurs mainly due to the lack of cohesion between the
particles, however a large amount of other processes are also involved as it can be
consulted in (Hanganu et.al. 2002). The damage theory employed in this work is based on
a special damage yielding function, which differentiates the mechanical response for
tension or compression components of the stress vector. In this case, fissuration is
interpreted as a local effect depending on the evolution of a set parameters of the material
and the corresponding evolution equations. The progress of the damage is based on the
evolution of a scalar parameter, which affects all the components of the elastic
constitutive tensor and in this sense; it constitutes an isotropic damage model (Oliver et.al.
1990).
a. Constitutive Equation
In the case of thermally stable problems, considering the unconditional fulfilment of the
ClausiusPlank inequality and applying the Colemans principle, we have the following
constitutive relation for the material form of the FPK stress vector acting on each
material point of the beam cross section:
(4.1.7)
Where C
me
=Diag[E
0
,G
0
,G
0
] is the material form of the elastic constitutive tensor, with E
0
and
G
0
the Young and shear undamaged elastic modulus and C
ms
=(1 d)C
me
is the material
form of the secant constitutive tensor. Eq. (4.1.7) shows that the FPK stress vector is obtained
from its elastic (undamaged) counterpart by multiplying it by the degrading factor (1 d).
The internal state variable d measures the lack of secant stiffness of the material.
b. Damage yield criterion
By analogy with the developments presented in (Hanganu et.al., 2002; Oller and Barbat,
2006), the damage yield criterion denoted by the scalar value F is defined as a function of
the undamaged elastic free energy density and written in terms of the components of the
material form of the undamaged principal stresses,

149
(4.1.8)
where P
m
is the equivalent (scalar) stress and the parameters r and n given in function of
the tension and compression strengths f
c
and f
t
and the parts of the free energy density
developed when the tension or compression limits are reached, (
0t
)
L
and (
0c
)
L
,
respectively.
Other kind of damage yield criteria can be used in substitution of P e.g., MohrCoulomb,
DruckerPrager, Von Mises etc, according to the mechanical behaviour of the material. A
more general expression equivalent to that given in Eq. (4.1.8) is the following:

(4.1.9)
where G(P) is a scalar monotonic function to be defined in such way to ensure that the
energy dissipated by the material on a specific integration point is limited to the specific
energy fracture of the material (Oliver et.al., 1990).
c. Evolution of the damage variable
The evolution law for the internal damage variable d is given by:
(4.1.10)
A damage yield condition and consistency condition are defined analogously as in
plasticity theory. Finally, the Kuhn-Thucker relations have to be employed to derive the
unloadingreloading conditions. The following expression is employed for the function G
(Oliver et.al., 1990):
(4.1.11)

where the term G() gives the initial yield stress for certain value of the scalar parameter
. The parameter of Eq. (4.1.11) is calibrated to obtain an amount of dissipated energy
equal to the specific fracture energy of the material when the entire deformation path is

150
followed. For a uniaxial tension process with a monotonically increasing load, the
following expression is obtained for :
(4.1.12)
where it has been assumed that the equivalent stress tension is equal to the initial damage
stress f
c
. The values of the maximum dissipation in tension
maxt
is a material parameter
equal to the corresponding fracture energy density g
f
, which is derived from the fracture
mechanics as g
f
= G
f
/l
c

where G
f
the tensile fracture energy and l
c
is the characteristic length of the fractured
domain employed in the regularization process (Lubliner et.al., 1989).
d. Tangent constitutive tensor
After several algebraic manipulations, which can be reviewed in (Hanganu et.al., 2002,
Lessloss Deliverable report 54, 2005 (www.lessloss.org)), we obtain that the material form
of the tangent constitutive tensor C
mt
can be calculated as where I is the identity tensor. It is
worth noting that C
mt
is nonsymmetric and it depends on the elastic FPK stress vector.
(4.1.13)
4.1.3.2 Plastic materials
In case of materials which can undergo nonreversible deformations the plasticity model,
assuming a thermally stable process and small elastic and finite plastic deformations,
formulated in the material configuration is used for predicting their mechanical response.
a. Constitutive equation
Following analogous procedures as those for the damage model i.e. employing the CP
inequality and the Colemans principle (Lubliner, 1985; Malvern, 1969), the secant
constitutive equation and the mechanical dissipation take the following forms
(4.1.14)
where the material description of the secant constitutive tensor C
ms
coincides with the
elastic one and the super script P is used to denote plastic strains.

151
b. Yield and plastic potential functions
Both, the yield function, F
p
, and plastic potential function, G
p
, for the plasticity model, are
formulated in terms of the material form of the FPK stress vector and the plastic damage
internal variable k
p
as:
(4.1.15)
where P
p
is the (scalar) equivalent stress, which is compared with the hardening function f
p

depending on the damage plastic internal variable k
p
and the current stress state, and K is
a constant value [38,46].
According to the evolution of the plastic damage variable, k
p
, it is possible to treat
materials considering isotropic hardening. However, in this work k
p
constitutes a measure
of the energy dissipated during the plastic process and, therefore, it is well suited for
materials with softening. In this case k
p
is defined (Lubliner et.al., 1989) as:
(4.1.16)
where G
Pf
is the specific plastic fracture energy of the material in tension and lc is the
length of the fractured domain defined in analogous manner as for the damage model.
The integral term in Eq. (4.1.16) corresponds to the energy dissipated by means of plastic
work and, therefore, k
p
constitutes a measure of the part of the fracture energy that has
been consumed during the deformation.
c. Evolution laws for the internal variables
The flow rules for the plastic strain vector and k
p
are defined as usual for plastic models
defined in the material configuration according to
(4.1.17)
where is the plastic consistency parameter and is the following hardening vector

152
(4.1.18)
n is the plastic dissipation and r are as for the damage model. It is interesting to note that
the proposed evolution rule allows to differentiate between tensile and compressive
properties of the material.
As it is a standard practice in plasticity, the loading/unloading conditions are derived in
the standard form from the Kuhn-Tucker relations formulated for problems with
unilateral restrictions. By other hand, starting from the plastic consistency condition and
considering the flow rules it is possible to deduce the explicit form of as:
(4.1.19)
d. Tangent constitutive tensor
The material form of the tangent constitutive tensor is calculated taking the time
derivative of Eq. (4.1.14), considering the flow rules, replacing the plastic consistency
parameter and after several algebraic manipulations, it is obtained as:
(4.1.20)
where
p
is the so-called hardening parameter.
4.1.3.3 Mixing theory for composites
Each material point on the beam cross is treated as a composite material according to the
mixing theory considering the following assumptions: (i) Each composite has a finite
number of simple components. (ii) Each component participates according to its
volumetric participation. (iii) All the components are subjected to the same strain field.
Therefore, the interaction between all the components, defines the overall mechanical
behaviour of the composite at material point level. The assumption (i) implies that the N
different components coexisting in a generic material point are subjected to the same
material strain and, therefore, we have the following closing equation:
(4.1.21)

153
which imposes the strain compatibility between components. After applying the CP
inequality and the Colemans principle, it is possible to obtain the material form of the
FPK stress vector for the composite in analogous way as for simple materials, i.e.
(4.1.22)
where the FPK stress vector is obtained as a weighted sum, according to the volumetric
fraction, of the material form of the stress vectors corresponding to each one of the N
components. The material form of the secant constitutive equation, the secant and
tangent constitutive tensors, C
ms
and C
mt
, for the composite are obtained as (Oller et.al.,
1996)
(4.1.23)
where (C
ms
)
q
, and (C
mt
)
q
are the material form of the secant and tangent constitutive
tensors, respectively. Having calculated the material form of the FPK stress vector, the
stress resultant and couples are then calculated employing Eqs. (4.1.4).
The mechanical behaviour of some advanced composed materials are based on a main
matrix component which is reinforced with oriented fibres, e.g. epoxy based materials
with glass or carbon fibres or even reinforced concrete, where the steel reinforcing bars
and stirrups can be seen as embedded reinforcing fibres. The mixing rule provides an
appropriated framework to simulate these kinds of composites, associating the one-
dimensional version of the described constitutive laws to the reinforcements.
4.1.3.4 Energy Dissipating Devices
The energy dissipating devices are simulated using a rod element with only one Gauss
integration point. The rotational degrees of freedom are released in both ends of the
beam to obtain only relative displacements in the device. The constitutive law employed
for dissipating devices corresponds to a bilinear plasticity, but any other one dimensional
description can be employed, for example in Mata et.al. (2006), (Lessloss Deliverable
report 54, 2005 (www.lessloss.org)), a constitutive description for elastomers to be
employed in energy dissipating devices is given.

154
4.1.4 Numerical implementation
In order to obtain a Newton type numerical solution procedure, the linearized form of
the weak form of Eq. (4.1.6) is required, which can be written as L[G(,, h)] =
G(,,h)+DG(,,h) p where L[G] is the linear part of the functional G(,,h) at the
configuration defined by (,) and p (, ) is an admissible variation. The term
G(,,h) supplies the unbalanced force and the differential DG( ,,h) p, the tangential
stiffness, which is calculated as
(4.1.24)
The numerical implementation in a FEM code of the internal, external and inertial terms
of G in Eq. (4.1.24) follows standards procedures and it can be consulted in reference
(Ibrahimbegovic and Mikdad, 2000).
The employment of nonlinear constitutive equations at material point level implies that
the global structural damping is added to the system in the term G
int
(x; h) of Eq. (4.1.24)
by means of the reduced stress resultant and couples.
4.1.4.1 Tangential stiffness tensors
The discrete counterpart of the term DG(,,h) p of Eq. (4.1.24) leads to the tangential
stiffness matrices, which can be obtained by means of the linearization of the different
contributions due to internal, external and inertial terms. The linearization of the internal
and external terms and its numerical implementation has been already presented in (Simo,
1986) for the case of linear elastic materials. For the case of an arbitrary distribution of
materials within the cross section the linearization of the internal term produces
(4.1.25)
Where the operator [B] relates the linearized increment h and the linearized form of the
co-rotated strain measures. The linearized form of the cross sectional force and moment

155
vectors takes into account the linearized strain-stress relations existing between P
1
m
,
obtained using the mixing theory, and the strain and strain rate vectors and is obtained as
(4.1.26)
Where C
sv
ij
is the spatial forms of the reduced tangential constitutive tensors, which are
calculated simply replacing C
se
within each material point by the tangential C
sv
and
viscous C
se
constitutive tensors in spatial description, and integrating over the cross
section according the procedure described in (Kapania and Li, 2003) for the linear case. It
is interesting to note that in the present formulation the reduced tangential and viscous
constitutive tensors are rate dependent. Finally, it is possible to rewrite Eq. (4.1.25) as
(4.1.27)
Where [K]
G
gives the geometric part of the tangent stiffness matrix, [K]
M
and [K]
V
give
the constitutive and viscous parts, which consider the contribution of the linearized
stresses corresponding to each compounding material.
The inertial contribution of G(x,h) can be rewritten in terms of this material quantities
and its linearized form is DG
ine
h as
(4.1.28)
where


156
(4.1.29)
And then it is possible to rewrite the linearized form of the spatial description of the
inertial contribution term as
(4.1.30)
The terms [M], [C
gyr
] and [K
cent
] of Eq. (4.1.35) correspond to the mass, gyroscopic and
centrifugal inertial tangential matrices of the discrete problem. The resulting inertial matrix
is nonsymmetric and configuration dependent on the rotational degrees of freedom, as it
can be verified from the gyroscopic and centrifugal terms. It is worth to note that all
these terms appear coupled in Eq. (4.1.33). Finally, Eq. (4.1.29) can be rewritten as
(4.1.31)
where [K]
P
corresponds to the part of the tangent stiffness tensor, which is dependent on
the loading and can be consulted in (Simo, 1986). The discretization of Eq. (4.1.35) by
using the FEM method follows standard procedures and can be reviewed in
(Hibrahimbegovic, 1996).
4.1.4.2 Cross sectional analysis
The cross section analysis is carried out expanding each integration point on the beam
axis in a set of integration points located on each fibre on cross section. In order to
perform this operation, the beam cross section is meshed into a grid of quadrilaterals,
each of them corresponding to a fibre oriented along the beam axis (see Figure 4-3). The
estimation of the average stress level existing on each quadrilateral is carried out by
integrating the constitutive equations of the compounding materials of the composite
associated to the corresponding quadrilateral and applying the mixing rule. The geometry

157
of each quadrilateral is described by means of normalized bi-dimensional shapes
functions and several integration points can be specified in order to estimate more
precisely the value of a given function according to a selected integration rule. In the case
of the average value of the material form of the FPK stress vector acting on a
quadrilateral we have
(4.1.32)
where A
c
is the area of the quadrilateral, N
p
and N
q
are the number of integration points
in the two directions of the normalized geometry of the quadrilateral, P
m
1
(y
p
, z
q
) is the
value of the FPK stress vector existing on a integration point with coordinates (y
p
, z
q
)
with respect to the reference beam axis, which is obtained from the corresponding
material strain vector using the constitutive laws and the mixing rule, J
pq
is the Jacobian of
the transformation between normalized coordinates and cross sectional coordinates and
W
pq
are the weighting factors. The coefficients of the tangent constitutive tensors can be
estimated in an analogous manner as in Eq. (4.1.32) but replacing P
m
1
(y
p
,z
q
) by C
mt
(y
p
,z
q
).
Finally, having obtained the stress level on each quadrilateral, the cross sectional forces
and moments are obtained by means of the discrete form of Eqs. (4.1.6) as
(4.1.33)
were N
fibre
is the number of quadrilaterals of the beam cross section, (A
c
)
k
is the area of
the k quadrilateral, (P
m
1
)
k
is the average value of the material form of the FPK stress
vector and k = (0, y
k
, z
k
) are the coordinates of the gravity centre of the kth quadrilateral
with respect to the local beam reference frame.

Figure 4-3. Discrete fibre like model of the beam element

158
4.1.5 Numerical examples
4.1.5.1 Nonlinear Seismic Response of Planar Frame
In this work the seismic nonlinear response of a typical concrete industrial building is
studied. The building has a bay width of 20 m and 24 m of interaxes length. The storey
high is 10 m. The concrete of the beam is H50, (50 MPa, ultimate compression), with an
elastic modulus of 35.000 MPa for the beam and H30 for the concrete of the columns.
It has been assumed a Poison coefficient of 0.2 for both concretes. The steel bar
reinforcements considered in the study and the fibre discretization of the sections are
those shown in Figure 4-4. The ultimate tensile stress for the steel is 510 MPa. The
dimensions of the columns are 60 x 60 cm
2
. The beam has a variable section with an
initial high of 80 cm on the supports and 120 cm in the middle of the span.

Figure 4-4. Columns and beam reinforcements and fibre model of the sections
The permanent loads considered are 2000 N/m
2
and the weight of upper half of the
closing walls, that is, 270,000xN. The employed acceleration record is the NS
component of the El Centro earthquake, 1940. The section of the energy dissipating
devices was designed for yielding with an axial force of 300.000 N and for a relative
displacement between the two ending nodes of 1.0 mm. The length of the dissipating
devices is of 2.5 m (see Figure 4-5).
In Figure 4-6 it is possible to see the contribution of the dampers to reduce the
displacements response. The obtained reduction is the order of 51 % minimizing the P
effects. The maximum acceleration shows a reduction of the order of 30 % compared
with the case where no devices are incorporated.

159

Figure 4-5. Precast industrial building without and with dissipaters

Figure 4-6. Displacements time history
4.1.5.2 3D Precast concrete building
The nonlinear seismic response of a precast building constructed in urban areas is
studied. See Figure 4-7. The building has one bay and two storeys of 6 and 3 m width,
respectively. The concrete of the structure is H25, (25 MPa, ultimate compression), with
an elastic modulus of 25,000 Pa. It has been assumed a Poison coefficient of 0.2. The
steel bar reinforcements considered in the study are those corresponding to the 8 % of
the sectional area for elements near to the joint (25 % of the column or beam length), and
4 % for elements in the middle part of the span. The ultimate tensile stress for the steel is
510 MPa. The dimensions of the columns are 30 x 30 cm
2
. The beams have a section of
15 x 30 cm
2
.
The permanent loads considered are the weight of the concrete floors, a live load of 2500
N/m
2
and the weight of the closing walls, (432,000 N). The employed acceleration record
is the same as before in the direction X and the same record scaled by 0.3 in the
orthogonal direction.

160
No accidental or structural eccentricities were considered in this work, but it is possible to
do it modifying the mass density of the beams or adding another structural element in the
same of the planar frames.
1 2
3
4

Figure 4-7. (1) 3D Frame. (2) Dissipating devices incorporated. (3), (4): Column and beams
sections.
Several numerical simulations were carried out to obtain an optimized combination of the
characteristic of the energy dissipation devices for knowing what is the yielding level, F
y
,
stiffness, K, and yielding displacement, D
y
, which give a biggest protection level to the
structure. The properties of the employed energy dissipating devices are summarized in
Table 4-1.
Table 4-1. Parameters of the energy dissipation devices
Device Characteristics
F
y
, (N)
D
y
, (mm)
1000
1.250
F
y
, (N)
D
y
, (mm)
2000
1.250
F
y
, (N)
D
y
, (mm)
3000
1.250
F
y
, (N)
D
y
, (mm)
4000
1.250
F
y
, (N)
D
y
, (mm)
5000
1.250
F
y
, (N)
D
y
, (mm)
1000
2.500
F
y
, (N)
D
y
, (mm)
2000
2.500
F
y
, (N)
D
y
, (mm)
3000
2.500
F
y
, (N)
D
y
, (mm)
4000
2.500
F
y
, (N)
D
y
, (mm)
5000
2.500
F
y
, (N)
D
y
, (mm)
1000
5.000
F
y
, (N)
D
y
, (mm)
2000
5.000
F
y
, (N)
D
y
, (mm)
3000
5.000
F
y
, (N)
D
y
, (mm)
4000
5.000
F
y
, (N)
D
y
, (mm)
5000
5.000

The results of the simulation are expressed in terms of maximum top and middle floor
displacements; base shear and over-tuning moment are presented simultaneously as
function of the type of employed device in Figure 4-8. From this figure it is possible to
see that even when the biggest benefits in terms of the selected global variables are
attained for different device characteristics, the more advantageous characteristics are
related with flexible devices (K = 8000 N/mm) with a medium yielding displacement
(approx. 2.5 mm) and yielding force around the 4000 N.
Therefore, the selected properties of the dissipaters were: Plastic yielding for a axial force
of 4000 N, relative yielding displacement between the two ending nodes of 2.50 mm. The

161
length of the dissipating devices is of 6.7 m. The dissipating devices only were
incorporated in the direction where the strongest ground acceleration record is applied.
1 2
3
4

Figure 4-8. Maximum response for each energy dissipating device. 1: Overtuning Moment. 2: Top
floor displacement. 3: Middle floor displacement. 4: Base shear.
4.1.6 Conclusions
The geometrically exact formulation due to Reissner and Simo for 3D beams is
considered in the context of initially curved beams and extended to include arbitrary
distribution of composite materials in the cross sections. The resulting formulation is
used for studying the constitutive and geometric nonlinear behaviour of framed
structures in the static case.
Constitutive laws for the simple materials are based on thermodynamically consistent
formulations allowing to obtain more realistic estimations of the energy dissipation in the
nonlinear range. The simple mixing rule for composites is used for modelling complex
material behaviours at material point level.
A detailed cross sectional analysis, consistent with the kinematics hypothesis is presented.
The proposed method, even when inexact from the point of view of the elasticity theory,
gives a computationally convenient way of approximating the strainstress distribution in
the section. A mesh independent response is obtained by means of the regularization of

162
the energy dissipated at constitutive level considering the characteristic length of the
volume associated to a specific integration point and the fracture energy of the materials.
The seismic behaviour of two typical precast concrete structures is studied employing a
numerical code, which incorporate a geometrically exact finite strain formulation for rods
using appropriated constitutive laws for materials. The fibre beam model presented in
this work provides a useful tool to simulate the earthquake effects on structures. A
specific plastic energy dissipating device element is employed in the simulations. The
advantages of employing dissipating devices to protect and improve the seismic
behaviour of flexible and low damped precast structures with non ductile connecting
joints is studied for the 2D and 3D cases presented here. From the results it is possible to
see that several numerical simulation are required to validate the best choice for selecting
the mechanical characteristics of the control devices to ensure the biggest improvements
in the seismic response of the controlled structure.
4.2 PRECAST CONCRETE PORTAL FRAMES
4.2.1 Post-earthquake surveys
Typical damages are observed on precast reinforced concrete structure caused by
earthquake: in many cases, beams simply fall down from their support, due to lack of
resistance and energy dissipation capacity at the beam-column connections.

Figure 4-9. General collapse of a precast structure for a one-storey industrial building (Kocaeli
earthquake, Turkey, 1999) [Toniolo (2002)]

163

Figure 4-10. Damaged half joint between a column and a simple supported beam of precast
structure [Toniolo (2002)]
According to Toniolo [2002], the principal cause of collapses of precast buildings is the
loss of bearing of the beam on the column, together with the overturning of beams,
under the combined effects of horizontal and vertical actions of earthquake. Bad design
corresponds to scanty joint connections with dry bearing supports of beams without
adequate structural tie or to a poor detailing of steel reinforcement of critical zones
(mainly the pattern and spacing of stirrups).
In the Adana earthquake post survey [AFPS], an industrial collapsed building is analyzed.
It seems that the causes of collapse are the underdesign of the dowel connections
between beams and columns and the bad implementation of the grouted mortar around
these dowels.
Concrete structure prefabricators use habitually pinned connections and particularly the
dowel connection (simple to realise) as beam-to-column connection for one-storey
structures.


164

Figure 4-11. Example of pinned connections: bolted and dowel connections [FIP, 1994] [Collinet,
2004-2005]
It is clear that the causes of collapse are complex and result of several parameters. The
weak point of precast concrete structures made of linear elements is clearly that they have
concentrated connection zones, where the risk of failure is very high if not well designed.
These are generally non redundant connections, sensitive to bad evaluation of the action
such as additional axial force in the beam - if differential movement of the columns
foundations is present for example.
4.2.2 INERD pin connection
The INERD pin connection consists of two external eye-bars welded or bolted to the
adjacent member. It was originally developed in the context of the steel structures and it
is presented with detail in section 4.3. of this report. INERD pin connection can be used
to solve connection problems in structures made of any material, if capacity design is
used in such a way that energy dissipation takes place in the connection, while the
structural elements, beams and columns, remain elastic. However the peculiarities of
application to other structures than steel structures need to be assessed. This is done in
the following for precast concrete structures made of linear elements.
4.2.3 Bracings using INERD pin connections
To reduce the vulnerability of existing precast concrete portal frames, we propose to add
2 short bracings using INERD pin connections at beam-column joints. These bracings
should be dissipative. The detailed study on the use of bracings using INERD pin
connections in precast reinforced concrete structure is presented in deliverable D55 of
the LESSLOSS Project ( www.lessloss.org ).

165
Beam
Column Column
Bracing using
INERD pin
connections
Bracing using
INERD pin
connections

Figure 4-12. Precast reinforced concrete structures braced with INERD pin connections
The use of the system is suited for new construction as well as for upgrading existing
precast RC structures, which is the focus of subproject 7 of the LESSLOSS project. In
the case of upgrading existing precast RC structures, bracings using 2 INERD pin
connections must be adapted to the characteristics of the beam and column elements:
dimensions, position of the reinforcements, type of beam-column connection.
The configuration of the bracing system using 2 INERD pin connections is sketched at
Figure 4-13.
Front view
Left view

Column
Dowel
connection
INERD Pin
Connection
Beam

166
Figure 4-13.Implementation of the INERD dissipative connection in a one-storey industrial frame
This dissipative system could be applied to concrete structures with beam-column
connections other than dowel connections.
4.2.4 Design Model for the systems
The study of the bracings using 2 INERD pin connections considers a reference precast
reinforced concrete structure using dowel beam-column connections and analyses the
influence of adding bracings using INERD pin connections in a concrete structure. Only
one dimension of pin is considered.
4.2.4.1 Definition of the model
In order to evaluate the possible dissipative behaviour of the INERD pin connection, 2
structures are compared:
a reference structure without bracings (see Figure 4-14)
a structure with bracings using INERD pin connection (see Figure 4-15)
Top view

3D view


167

Figure 4-14. Reference structure without bracings

Figure 4-15. Studied structure with bracings using INERD pin connection
The 2 structures have the same dimensions and same characteristics of elements.
In the reference structure, the beam-column connections are realized with non dissipative
dowel connection. The dowel connection is modelised assuming a perfect hinge
6 m
Beam
(L x w x h: 17 m x 30 cm x 40/80 cm) cm)
Column
(L x w x h: 6 m x 40 cm x 40 cm)

Column
(L x w x h: 6 m x 40 cm x 40 cm)
17 m
Plastic
hinges
Plastic
hinges
1 m
1 m
1 m
Bracings using
INERD pin
connection
Bracings using
INERD pin
connection
1 m
6 m
Beam
(L x w x h: 17 m x 30 cm x 40/80 cm)
Column
(L x w x h: 6 m x 40 cm x 40 cm)
Column
(L x w x h: 6 m x 40 cm x 40 cm)
17 m
Plastic hinges
Plastic hinges

168
behaviour as dowel connection that cannot transmit bending moments between column
and beam, but can transmit beam axial force to the column. The second structure has
also a dowel connection at the beam-to-column joint to which we add a bracing with an
INERD pin connection. The bracing provides a resisting moment at the beam to column
connection. Once the INERD pin connection is activated, it can dissipate earthquake
energy by hysteretic loops.
Figure 4-15 shows the implementation of the INERD pin connections system in a portal
frame structure. Each brace includes 2 INERD pin connections but we consider in the
model only one INERD pin connection by bracing with a capacity of deformation twice
that of a single pin.
The formation of plastic hinges at the column bases is expected early during the loading.
The model integrates the possibility to form plastic hinges at the column base by
concentrated plastic hinges at that place.
4.2.4.2 Plastic hinges at column bases
FEMA 273 [1997] gives modelling parameters for non linear procedures in the case of
reinforced concrete columns. We use the modelling parameters of columns controlled by
flexure. 3 levels of ductility are defined (see Figure 4-16) corresponding to 3 plastic
rotation capacities:
low ductility : a = 0,005 rad, b= 0,005 rad
average ductility : a = 0,01 rad, b= 0,015 rad
high ductility : a = 0,02 rad, b = 0,03 rad
The ductility of the flexural plastic region of the column depends on the design of
stirrups in this column base region, which influence the concrete containment.

169
0
50
100
150
200
250
0 0.01 0.02 0.03 0.04
Rotation [rad]

M
o
m
e
n
t

M

[
k
N
.
m
]
Low ductility
(FEMA-273)
Average ductility
(FEMA-273)
High ductility
(FEMA-273)
Low ductility
(used in the model)
Average ductility
(used in the model)
High ductility
(used in the model)

Figure 4-16. Concrete non linear material law at the column base used in the model (3 ductilities :
low, average, high) (FEMA Publication 273 [1997])
Figure 4-16 shows the 3 levels of ductility of FEMA and the 3 equivalents levels that are
used in our models. Contrary to FEMA Publication, we do not consider a residual
resisting moment after plastic rotation angle capacity a for numerical resolution
reasons. It should be noted that the model is rigid-plastic. The rigidity is infinite until
obtaining the column plastic moment M
pl
= M
Rd,column
= 190 kN.m.
4.2.4.3 Design of INERD pin connectionEquation Chapter 4 Section 2
The INERD pin connection law is determined from design formulae defined in 4.3. In
this work, we study feasibility of bracings using INERD pin connection in the case of
precast concrete frame. Thus we chose only one the mechanical characteristics chosen for
the pin connection are compatible in dimension with the moment capacity of the beam
and column elements of the frame structure:
Rd,c o lum n
Rd,c o nnec tio n
M
M 152 kN.m
1, 25 =
(4.2.1)
This capacity design relation expresses that INERD pin connection should be designed
as a dissipative fuse which cannot be stronger than the others structural elements
(capacity design).
a
b

170

Figure 4-17. Detail of the beam-to-column connection with bracings using INERD pin connections
With the dimensions indicated at Figure 4-17, one can derive:
P
u
.d = M
Rd, connection
(4.2.2)
with d m
2
2
=
Rd connection
u
M
P kN
d
,
2.152
215
2
= (4.2.3)
We can now assign characteristics to the pin in order to obtain this ultimate resistance P
u
..

bracings using
INERD pin
connections
1 m
1 m
d
beam
column
P
u

M
Sd


171
Table 4-2. Characteristics of the studied INERD pin connection
Force Displacement
Force-displacement behaviour
of pin connections
) 1 , 1 / (
2
a
M
P
p
y
= =
97,44 kN
a
M
P
p
u
=
4
=
214,375 kN
( )
4 3
6
5 , 1
2

=
I E
M
p
y
=
2,68 mm
a
II
= 2 , 0 = 20 mm
a = 8 , 0
lim
= 80 mm
Geometrical and material data of the
pin connection
Geometrical data of the pin connection
y pl p
f W M = = 5,36 kN.m
a = = 0,33
h = 35 mm
b = 35 mm (Pin square shape section)
= 300 mm (equal to width of the beam)
a = 100 mm
f
y
= 500 N/mm
W
pl
= 10718,75 mm
I = 125052 mm
4

E = 200000 N/mm
a a
l
t
ext
t
ext
h
t
int
t
int
b

P
u

P
y

P
II
I
II
y
lim


172
-250
-200
-150
-100
-50
0
50
100
150
200
250
-80 -60 -40 -20 0 20 40 60 80
[mm]
P

[
k
N
]
P [kN]

Figure 4-18. Force-displacement behaviour of the studied INERD pin connection
-200
-150
-100
-50
0
50
100
150
200
-0.15 -0.05 0.05 0.15
= / d [rad]

M

=

P
.
d

[
k
N
.
m
]
M [kN.m]

Figure 4-19. Equivalent moment-rotation behaviour of the beam column connection with the
addition of the studied INERD pin connection

173
4.2.4.4 Static non linear analysis (Pushover analysis)
In order to evaluate the effectiveness of bracings connected with INERD pin
connections, we consider the 2 structures (see Figure 4-14 and Figure 4-15: structure
with and without the studied INERD pin connection) integrating non linear material
behaviours (non linear material law for concrete at the column base in all cases and
INERD pin connection law whenever it is necessary), and realize several pushover
analyses. Considering different levels of ductility (rotation capacity) at column bases (see
Figure 4-16.).
The loading is monotonic with the effects of the cyclic behaviour and load reversals being
estimated by using a modified monotonic force-deformation criteria and with damping
approximations. Static pushover analysis is an attempt by the structural engineering
profession to evaluate the real strength of the structure. In our analysis, the increasing
load is a horizontal force F applied at the top of the column (near the centre of gravity of
the structure) and the monitoring displacement is the top column displacement d
r
. The
pushover is realized with an increasing displacement and a measure of the horizontal
force corresponding to this displacement because SAP 2000 requires that the monitored
parameter function (displacement in this case) must be strictly monotonous during the
pushover. With the increase in the loading, weak links and failure modes of the structure
should be found.

Figure 4-20. Pushover analysis: increasing force F and monitored displacement dr
F

d
r

F/2
F/2

d
r
d
r,pl
d
r,el
d
r,pl
Centre of gravity
of the structure

d
r,el

174
The spectral acceleration a is determined from the horizontal force F by the relation:
F ma = (4.2.4)
a F m / = (4.2.5)
with m = M
mmf
= total moving mass of the frame structure = 33 tons.
The spectral acceleration a is thus obtained from the applied force in an opposite way to
the equivalent static method (equivalent analytical model). This method is valid if the
horizontal force F is applied at the centre of gravity of the structure where the total mass
moving is considered. It is almost the case in the portal frames studied here (see Figure
4-20), as the centre of gravity of the structure is almost on the level of the beam where we
monitor the displacement d
r
. We calculate the elastic displacement S
de
(T) ( top column
displacement d
r
) from the period T of the structure for different levels of accelerations (a
g

= 0,05 g to 1 g). This elastic displacement S
de
(T) represents the horizontal displacement of
the gravity centre of the structure under a given acceleration level. The structure should
be able to achieve this displacement before failure is observed. This is why it is referred to
as the target displacement which the structure should be able to reach. S
de
(T) is
determined by the relation:
De e
T
S T S T
2
( ) ( )
2

=

(4.2.6)
where S
e
(T) = elastic response spectrum
T = period of the structure
In our case, we consider the type 2 elastic response spectrum S
e
(T) for ground D (soil D)
of Eurocode 8. The 2 periods of the 2 structures are:
T = 1,117 s in the frame structure without bracings before formation of plastic
hinges at columns base.
T = 2,022 s in the frame structure with bracings and INERD pin connections, after
formation of plastic hinges at columns base.

175

Figure 4-21. T = 1,117 s in frame structure without bracings before formation of plastic hinges at the
base of the column

Figure 4-22. T = 2,022 s in frame structure with bracings and INERD pin connection, after
formation of plastic hinges at the base of the column
Results
Table 4-3. Values of target displacement S
de
(T) corresponding to different Peak Ground
accelerations (PGA)
Reference structure
Structure with bracings and
INERD pin connection
T [s] = 1,117 s T [s] = 2,022 s
PGA
S
e
(T) [m/s] S
De
(T) [m] S
e
(T) [m/s] S
De
(T) [m]
a
g
= 0,10 g 1,2086 0,0382 0,3962 0,0410
a
g
= 0,20 g 2,4172 0,0764 0,7925 0,0821
a
g
= 0,30 g 3,6258 0,1146 1,1887 0,1231
a
g
= 0,40 g 4,8344 0,1528 1,5849 0,1641
a
g
= 0,60 g 7,2516 0,2292 2,3774 0,2462
a
g
= 0,80 g 9,6688 0,3056 3,1699 0,3283
a
g
= 1 g 12,0859 0,3820 3,9623 0,4104
T = 2,022 s
T = 1,117 s

176
Pushover analysis
0
0. 05
0. 1
0. 15
0. 2
0. 25
0. 3
0. 35
0. 4
0 0. 1 0. 2 0. 3 0. 4 0. 5
Displacement at the top of column d
r
[m]
S
p
e
c
t
r
a
l
e

a
c
c
e
l
e
r
a
t
i
o
n
/
g

[
m
/
s
]
Low ductility reference structure
Average ductility reference structure
High ductility reference structure
a
g
= 0,8 a
g
= 1 g
a
g
= 0,4
a
g
= 0,3
a
g
= 0,2
a
g
= 0,1 a
g
= 0,6

Figure 4-23. Pushover curves of reference structure (3 ductility classes), with indication of target
displacements corresponding to different a
g
(PGA)

177
Pushover analysis
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0 0.1 0.2 0.3 0.4 0.5
Displacement at the top of column d
r
[m]
S
p
e
c
t
r
a
l
e

a
c
c
e
l
e
r
a
t
i
o
n
/
g

[
m
/
s
]
Low ductility structure with INERD Pin connection
Average ductility structure with INERD Pin connection
High ductility structure with INERD Pin connection
a
g
= 0,8 a
g
= 1 g
a
g
= 0,4
a
g
= 0,3
a
g
= 0,2
a
g
= 0,1 a
g
= 0,6

Figure 4-24. Pushover curves of structure with bracings using INERD pin connection (3 Ductility
class of column) with indication of target displacement corresponding to different a
g
(PGA)
4.2.4.5 Dynamic non linear analysis (Time history analysis)
With an aim of validating the pushover analysis, dynamic analysis are made with different
levels of peak ground accelerations (a
g
= 0,05 g to 0,4 g). 3 accelerograms are used at each
PGA. For each levels of acceleration, the maximal top column displacement d
r,max
and the
maximal plastic rotation
max
(see Figure 4-20) obtained with the 3 accelerogram are
reported in the table of results (see Table 4-4) together with the results of the pushover
analysis.


178
Results
Table 4-4. Results of dynamic non linear analysis (time history analysis)

Reference structure
Structure with bracings using
INERD pin connections

low ductility
average
ductility
high
ductility
low ductility
average
ductility
high
ductility

u
= 0,005
rad
u
= 0,01
rad
u
= 0,02
rad
u
= 0,005
rad
u
= 0,01
rad
u
= 0,02
rad
d
r, max
[m] 0.0206 0.0206 0.0206 0.0164 0.0164 0.0164
a
g
= 0,05 g
max
[rad] 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
d
r, max
[m] 0.0411 0.0411 0.0411 0.0328 0.0328 0.0328
a
g
= 0,10 g
max
[rad] 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
d
r, max
[m] 0.0617 0.0617 0.0617 0.0492 0.0492 0.0492
a
g
= 0,15 g
max
[rad] 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
d
r, max
[m] 0.0795 0.0795 0.0795 0.0627 0.0627 0.0627
a
g
= 0,20 g
max
[rad] 0.0028 0.0028 0.0028 0.0021 0.0021 0.0021
d
r, max
[m] 0.1114 0.1114 0.0705 0.0705 0.0705
a
g
= 0,25 g
max
[rad]
Failure of
the
structure 0.0080 0.0080 0.0035 0.0035 0.0035
d
r, max
[m] 0.1391 0.1099 0.0931 0.0931
a
g
= 0,30 g
max
[rad]
Failure of
the
structure
Failure of
the
structure 0.0124 0.0197 0.0075 0.0075
d
r, max
[m] 0.1639 0.1138 0.1250 0.1168
a
g
= 0,35 g
max
[rad]
Failure of
the
structure
Failure of
the
structure 0.0165 0.0200 0.0232 0.0115
d
r, max
[m] 0.1214 0.1473 0.1291
a
g
= 0,40 g
max
[rad]
Failure of
the
structure
Failure of
the
structure
Failure of
the
structure 0.0222 0.0256 0.0136
d
r, max
[m] 0.0933 0.1241 0.1850 0.7912 0.7915 0.7900
Pushover
max
[rad] 0.005 0.01 0.02 0.1354 0.1361 0.1352


179
4.2.4.6 Analysis of the results
From Figure 4-23 which shows the pushover curve of reference structures, it appears
that:
low ductility reference structures can sustain a
g
up to 0,24 g;
medium ductility reference structures can sustain a
g
up to 0,3 g
high ductility reference structures can sustain a
g
up to 0,45 g,
Figure 4-24 shows the pushover curves of structures in which bracings with INERD pin
connections are added. It appears that with these bracings, structure of any level of
ductility at the column base will survive to all possible a
g
.
Under a peak ground acceleration or spectral acceleration a
g
0,2 g, there is no failure,
even in structures with low plastic rotation capacity at column base. In low seismic area
(a
g
0,2 g), bracings with INERD pin connections only bring rigidity at the structure
and thus reduce rotation
max
and displacement d
r,max
(SLS state). Looking the ultimate
limit state, bracings using INERD pin connections are not needed in precast concrete
structures in low seismicity area
For higher peak ground acceleration a
g
> 0,2 g, bracings with INERD pin connections
show all their effectiveness, especially for low plastic rotation capacity at column bases.
The system ensures stability (ULS) and reduces deformations of the structure (SLS).
In a structure without bracings (Figure 4-23), when the plastic rotation capacity at column
bases is reached, columns bases behave as real hinges and can be modelled like that. So,
the structure becomes a mechanism and the failure of the structure takes place. Then, the
structure collapse depends only of the column resisting moment and of the ductility of
column at its bases.

180

Figure 4-25. Frame structure without INERD pin connections after formation of plastic hinges at
the column bases = Mechanism
In a structure with bracings and INERD pin connections (Figure 4-24), after the
formation of plastic hinges at columns bases, the structure is still statically acceptable.
The bracings provide an additional strength, rigidity and ductility to the structure.
Pushover curves at Figure 4-24 show the contribution of each structural component
(concrete column and bracings with INERD pin connections). In the first part of the
curve, the behaviour of the column bases, because of its rigidity, control the structural
behaviour. In the second part, it is the behaviour of bracings with INERD pin
connections which control. This is an important advantage of installing bracings with
INERD pin connection system.

Figure 4-26. Frame structure with INERD pin connections after the formation of plastic hinges at
the column bases= statically acceptable structure
By comparing pushover curves corresponding to structures with and without INERD pin
connections at same level of ductility, we observe that not only the deformation capacity
of the structure is increased but also its rigidity. Thus plastification of column bases will
occur for greater horizontal forces. Bracings with INERD pin connections also reduce
damages in the structure.
Statically
acceptable
structure
F

Mechanism
F


181
As said before, the pushover analysis has been realized considering only one INERD pin
connection characteristic in the bracing system (see Table 4-2). So, the second part of the
pushover curves represents the behaviour of the INERD pin connection. By modifying
the INERD pin connection characteristics, we can modify directly the behaviour of the
structure (rigidity, ductility of the structure (rotation capacity), strength).
In the line of the preceding point, we see through the pushover curves that, at the
formation of plastic hinges at column bases of the structure, the INERD pin connection
is still in its elastic behaviour. To improve structural behaviour, it would be better to have
the INERD pin connection working in the plastic domain so as to dissipate in a more
effective way the earthquake energy. To achieve it, we can adjust the pin connection
characteristics; for example: to reduce distance a (see Table 4-2) and thus reduce
,
lim
.
The results from the dynamic non linear analyses confirm results obtained by the
pushover analyses
A disadvantage of adding bracings in a precast concrete structure is that it adds shear in
the columns and beam. These elements should be checked for their resistance to shear.
This is particularly true if bracings are added to an existing structure in order to reduce its
vulnerability.
Globally, the study has demonstrated the possibility to reduce the vulnerability of existing
precast concrete portal frames by means of added bracings. These bracings must of
course be dissipative. The use of the INERD pin connections in these bracings is one
easy-to-apply solution which has the advantage of putting the designer in actual control
of plastic capacity.

182
4.3 STEEL FRAMES WITH CONCENTRIC BRACINGS. CONNECTIONS
4.3.1 Reasons for using dissipative connections in frames with concentric
bracings and purpose of the research activity in LESSLOSS.
Seismic resistant structures are designed for stiffness, strength and ductility.
Stiffness requirements are imposed in order to limit non-structural damage in case of
minor to moderate earthquakes and limit instability effects. Strength is needed to ensure
the capacity of the structure to resist safely the action effects. Ductility is a way to
dissipate part of the seismic input energy through inelastic deformations and therefore
reduce the cost of the structure in a comparison to a non ductile one.
All conventional frames have certain disadvantages in respect to the above design criteria.
(Table 4-5). As dissipative connections necessarily bring to a structure more flexibility
than rigid connections designed for overstrength, their use should not be envisaged in
structures which are flexible by nature. It can be shown in the case of moment resisting
frames that the loss of rigidity brought by use of dissipative connections must be
compensated by deeper beam and column sections.
Table 4-5. Structural typologies and main characteristics for Steel Frames
Moment resisting
Frames
(MRF)
Concentric Braced
Frames
(CBF)
CBF with
dissipative
connections
Stiffness Low High High
Ductility High Low High
Dissipative
zones
Beam-ends Diagonals Connections

On the contrary, frames with concentric bracings, which are naturally stiff, are a field of
application for dissipative connections. Dissipative connections can be designed to be the
weak point in the bracing system, while diagonals can be designed to avoid buckling; in
that way, both tension and compression diagonals do contribute to the earthquake
resistance and are present in the simple models used for design.
Dissipative connections however require research to demonstrate their ability to provide
stable cyclic behaviour and to assess their potential for practical applications in structures.
In the INERD project Plumier et al, 2006 - such dissipative connections have been

183
developed and evaluated experimentally. One of them, the pin connection, was
demonstrated to possess the required qualities of strength, stiffness and resistance. Their
positive input for a good global behaviour of low rise structures with X bracings has been
shown in numerical studies. Furthermore, due to the character of localised fuse of such
connection, the possibility is open to better tune resistance to demand, bringing a more
homogeneous and, consequently, more ductile global behaviour of structures.
Further evaluations were needed to assess the practical range of application of frames
with bracings and dissipative connections in the range of tall buildings. Those evaluations,
which are presented in detail in Deliverable D56 of the LESSLOSS project
(www.lessloss.org), are summarised here, together with design guidance for such
structures.
4.3.2 The INERD pin connection geometry and properties
An INERD pin connection connect a brace and a columns by means of plates (eye-bars)
and a pin running through them (Figure 4-27). The design procedure developed by Vayas
& Thanopoulos 2005 is explained in details in Plumier et al. 2006. The behaviour of the
pin connection can be modelled by a tri-linear curve. The design expressions are
presented in Table 4-6 here under.
Inner plates
Outer plates
Dissipative pin
Column
Diagonal bracing

Figure 4-27. Perspective and plan views of INERD pin connection. Definition of geometry.

184
Table 4-6. Design formulae for the connection with 2 internal plates
Eye-bars in Force Displacement
Point I
yielding y
Compression
p
y Rk
M
P
a
,
2
( /1, 1)
= ( )
p
y
M
EI
2
1, 5 3 4
6
=
Point II
ultimate u
Compression
p
u Rk
M
P
a
,
4
( /1, 1)
=
II
a 0, 2 =
Points I and II

Tension 90% of above
values
for P
y
and P
u

Over-strength for
capacity design checks
30% beyond P
u

Deformation
capacity

a
lim
0, 8 =

Symbols.
p pl y
M W f = a = f
y
= yield stress of pin
= pin length (axial distance between external eye-bars)
a = clear distance between internal and external eye-bars
W
pl
= plastic modulus of pin I = moment of inertia of pin
E = elastic modulus of pin s material
For rectangular pins, W
pl
= bh/4 and I = bh/12 h = pin height b = pin width

4.3.3 Code rules for braced frames with pin INERD-connections
Frames with bracings and dissipative connections may be designed according to the rules
of Eurocode 8, with some modifications justified by the fact that energy dissipation takes
place in the connections and not in the tension braces. Table 4-7 presents the original
Eurocode 8 rules for frames with dissipative concentric bracing and the modified rules if
dissipative INERD pin connections are used.

185
Table 4-7. Eurocode 8 rules for frames with bracings left, standard right, with dissipative INERD
pin connections.
Paragraph Dissipative diagonals Dissipative INERD pin connections
6.7.1
Design
criteria
(1)P Concentric braced frames shall
be designed so that yielding of the
diagonals in tension will take place
before failure of the connections and
before yielding or buckling of the
beams or columns.
(1) P Concentric braced frames with pin
INERD connections shall be designed so
that:
(a) yielding of the pins will take place
before compression failure of the diagonals
and before yielding or buckling of the
beams or columns.
(b) failure of the connections precedes
buckling failure of the diagonals
(c) yielding of the pins is limited at the
damage limitation state
(d) the other connection elements shall be
designed with adequate overstrength
relevant to the pins.
(2)P The diagonal elements of
bracings shall be placed in such a way
that the structure exhibits similar load
deflection characteristics at each
storey in opposite senses of the same
braced direction under load reversals.
(2)P The diagonal elements of bracings
and their connections shall be placed in
such a way that the structure exhibits
similar load deflection characteristics at
each storey in opposite senses of the same
braced direction under load reversals.
6.7.2
Analysis
(2)P The diagonals shall be taken
into account as follows in an elastic
analysis of the structure for the
seismic action: in frames with
diagonal bracings, only the tension
diagonals shall be taken into account;
(3)
(2)P The diagonals shall be taken into
account as follows in an elastic analysis of
the structure for the seismic action:
in frames with diagonal bracings, both
the tension and compression diagonals
shall be taken into account;
(3) Does not apply
6.7.3
Diagonal
members
(1) to (9) (1) to (9) Do not apply, except (8)
6.7.4
Beams
and
columns
(1) (1) as in the original text, with:
is the minimum value of
i
= P
u,Rd,i
/N
Ed,i
over all the connections
of the diagonals of the braced frame
system; where P
u,Rd,I
is the ultimate
strength of the pin INERD connection.
4.3.4 Practical design procedure
For practical applications, the following design steps are recommended:

186
Step 1. Selection of the pin dimensions, according to the following requirements:
DH
a
lim
cos
0.8
2
= ( 4.3.1 )
=
,
, ,
y Rk
y Rd E ser
ser
P
P N ( 4.3.2 )
=
,
,
0
u Rk
u Rd Ed
P
P N ( 4.3.3 )
a,
lim
, P
y,Rk
, P
u,Rk
: see Table 4.3.3 and Figure 4.3.1.
D = lateral drift ratio H = storey height
= angle of inclination of diagonal (from horizontal to diagonal)
0
and
ser
partial safety factor of resistance (=1,0)
N
Ed
design force of the diagonal
N
E,ser
design force of the diagonal at the damage limitation state
=
,
Ed
E ser
N
N ( 4.3.4 )
reduction factor for lower return period of seismic action associated with damage
limitation
Step 2. Verification of the brace dimensions
, b,Rd u Rk
P N ( 4.3.5 )
where: N
b,Rd
buckling resistance of the diagonal
Step 3. Dimensioning of eye-bars, welds etc
All connection elements (eye-bars, bolts, welds etc.) shall be verified for the capacity
design force, equal to:
=
,
1, 3
Ed u Rk
P P ( 4.3.6 )

187
The thicknesses of the eye bars shall additionally check:
t
ext
0,75 h t
int
0,5 t
ext
a h
With h = pin height, t
ext
= thickness of external eye bars, t
int
= thickness of internal eye
bars
Steel quality of the eye-bars has to be equal to or higher than that of the pin.
4.3.5 Application of dissipative connections to a tall office building with X
bracings
4.3.5.1 Design stage
The structure chosen as reference is 15 storey high (Figure 4-28).
52,5
10,5
10,5
10,5
10,5
10,5
3,5
10 10 10
30
28
10 10 10
30
28x28
10
1
0
10
1
0
10
1
0
30
3
0

Figure 4-28. Reference structure. Elevation and plan.
The study explores parameters. Two interstorey drift limits are considered. Structure
design 1 has ductile non structural elements and the interstorey drift limit is 0.01 h.
Structure design 2 has fragile non structural elements and the interstorey drift limit is
0.005 h. Design with and without dissipative connections are made. The possibility to

188
tune resistance to demand in order to bring a more homogeneous and ductile global
behaviour of structures has also been considered: design without and with variable pin
connections are made.
The set for comparison is:
One reference structure designed for wind.
Structure 1: interstorey drift limit 0.01 h without dissipative connection
with dissipative connection constant pins
with dissipative connection variable pins
Structure 2: interstorey drift limit 0.005 h without dissipative connection
with dissipative connection constant pins
with dissipative connection variable pins
A complete design of each structure is made which gives ground to comparison. The
design Peak Ground Acceleration is a
g
= 4 m/s .The fundamental periods obtained are
shown at Figure 4-29. Structures with pin connections are more flexible than those
without dissipative connections, but all structures are flexible with periods largely in the
descending branch of the spectrum, so that there is no significant difference in the
earthquake action applied to each structure.
In structures Type 1, the column sections are smaller with dissipative connections, but
some diagonals are slightly bigger because of the non buckling criteria. The earthquake
resistant structures with dissipative connections are 18 % lighter.
In structures Type 2, the interstorey drift criterion under Damage Limit State governs the
design, so that no reduction in the sections of columns and diagonals is brought in by the
use of dissipative connections. However, a high steel grade is necessary for the structure
without dissipative connections (f
y
355 MPa), while a lower steel grade (f
y
= 235 MPa)
can be used in the design with dissipative connections.
The only design fulfilling all the Eurocode 8 criteria are those in which the resistance of
the connections was tuned to the required resistance (structure 1 and 2 with variable
pins).


189
0
1
2
3
4
5
6
0 0.5 1 1.5 2 2.5 3 3.5 4
Periods in seconds
Pseudo-
acceleration
(m/s)
s
t
r
u
c
t
u
r
e

2

w
i
t
h

I
N
E
R
D

p
i
n

c
o
n
n
e
c
t
i
o
n
s
s
t
r
u
c
t
u
r
e

1

w
i
t
h

I
N
E
R
D

p
i
n

c
o
n
n
e
c
t
i
o
n
s
s
t
r
u
c
t
u
r
e

2

w
i
t
h
o
u
t

d
i
s
s
i
p
a
t
i
v
e

c
o
n
n
e
c
t
i
o
n
s
t
r
u
c
t
u
r
e

1

w
i
t
h
o
u
t

d
i
s
s
i
p
a
t
i
v
e

c
o
n
n
e
c
t
i
o
n
s
t
r
u
c
t
u
r
e

a
f
t
e
r

w
i
n
d

d
e
s
i
g
n

Figure 4-29. Periods of the analysed structures.
4.3.5.2 Pushover analysis
The structures without dissipative connections are modelled with diagonals in tension
only.
The pushover is performed by increasing a triangular distribution of forces, based on the
assumption of predominance of the first mode of vibration.
A target displacement is assessed for the different structures on the basis of the FEMA
273 (1997):
target(roof)
= C
0
C
1
C
2
C
3
S
De
(T) ( 4.3.7 )
The target displacements for the different structures for a
g
= 4 m/s are given at Table
4-8. The pushover curves of the structures are presented in Figure 4-30 with indication of
the design base shear.


190
Table 4-8. Target roof displacements for pushover analyses

Structure
T
(s)
S
e
(T) S
De
(T)
target(roof)
Structure 1 without pins 1,70 3,52 0,26 0,46
Structure 2 without pins 1,22 4,96 0,19 0,34
Structure 1 with pins 1,78 3,36 0,27 0,49
Structure 2 with pins 1,28 4,64 0,19 0,35

a)
0
1000
2000
3000
4000
5000
6000
7000
0 0.2 0.4 0.6 0.8 1
Top displacement (m)
B
a
s
e

r
e
a
c
t
i
o
n

(
k
N
)
without pins
with variable pins
target
displacement
design base shear
teta = 0.1
teta = 0.2
failure in the pin co

b)
0
1000
2000
3000
4000
5000
6000
7000
0 0.2 0.4 0.6 0.8 1
B
a
s
e

r
e
a
c
t
i
o
n

(
k
N
)
without pins
with variable pins
with constant pins
target
displacement
design base shear
teta = 0.1
teta = 0.2
failure in the pin co

Figure 4-30. Pushover curves of structures (a) drift limit = 0.01h (b) drift limit = 0.005h

191
Target displacements are also reported, as well as displacements corresponding to values
of the P- effect parameter = 0.1 and 0.2 and to pin connections failure.
Table 4-9 reproduces those drifts at = 0.1; they are greater for structures with variable
pin connections demonstrating a lower sensitivity to second order effects of structures
with connections of resistance tuned to demand.
Table 4-9 also indicates the behaviour factors q deduced from the pushover curves,
computed as: q = ductility = d
u
/d
y

For structures with dissipative connections, d
y
is found at the intersect of the elastic
behaviour line with the design shear, in order to avoid the influence of the tri-linear curve
used to model pin behaviour and to approach the value which would be found
considering a bilinear behaviour. For classical design, d
y
is found at the intersect of elastic
and plastic behaviour. For both type of structures, d
u
is conventionally defined as the
point corresponding to the attainment of =0.1 in the structure.
Figure 4-31 shows the impact of design conditions on the deformed shapes at = 0.1.
The pushover curves indicate that some yielding takes place in structures with dissipative
INERD pin connections for loads lower than the design base shear. This is due to the tri-
linear force-displacement curve used to describe the behaviour of pin connections. In
spite of this early yielding, criteria on deflections under service load are satisfied. This
design cannot be criticized, because it is similar to what is currently done in the modelling
of moment resisting frames: the M- curve used to characterize the behaviour of beams
in bending is represented by a bilinear law, which is a simplification of a real M- curve.
The design using dissipative connections takes fully advantage of the maximum
behaviour factor for frames with concentric bracings (q = 4), while classical design result
in an overstrength of about 2, so that the value of q really used is in fact 0.5 times the
code value. This trend is clearly demonstrated in Table 4-9 in which computed qs are
given.
The only design which generates a global plastic mechanism is the one with pins of
variable strength. In the other design, plastic deformations are localized in the bottom
bracing.


192
Table 4-9. Drifts at which = 0,1 and behaviour factor q from the pushover analysis
Structure Drift (%) q
Structure 1 without pins 0,9 1,7
Structure 2 without pins 1,2 3,3
Structure 1 with pins 1,6 5,4
Structure 2 with pins 0,9 4,0
Structure 2 with variable pins 1,5 6,4
a) b)
Figure 4-31. Deformed shape at failure (a) with variable pins (b) all other structures
4.3.5.3 Dynamic Non linear Time History Analyses
Time history analyses are realized with 3 artificial accelerograms, constructed for a type 1
earthquake on soil B, with a PGA of 0.4g and a duration of 15 s. Figure 4-32.
-6
-4
-2
0
2
4
6
0 15
time (sec)
a
c
c
e
l
e
r
a
t
i
o
n

(
m
/
s
)
Accelerogram 1

Figure 4-32. One artificial accelerogram used in the non linear dynamic analysis.

193
In the structures without dissipative connections, where the bracings are supposed to
work in tension only, both bracings are modelled with a dissymmetric behaviour law. The
diagonal in tension has its full plastic capacity while in compression, the diagonal are
supposed buckled and keep 33 % of their buckling resistance.
The main result observed is the reduction of shear force at the base of the structures
when using dissipative connections. Figure 4-33. In structures with ductile non structural
elements, the reduction in base shear is 25 %. In structures with fragile non structural
elements, the reduction is nearly 50 % .Table 4-10.
The parameter has been assessed in the dynamic analyses. Its maximum values are given
in Table 4-11. In the structures were the resistance in the pins is adapted to the needed
resistance, the values of are quite constant at the different levels. For the structures with
overstrength in the upper levels (classical design and structure 2 with constant pins), is
small for the upper levels and is big for the bottom level.
-6000
-4000
-2000
0
2000
4000
6000
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
B
a
s
e

s
h
e
a
r

(
k
N
)
without pins
with pins
-6000
-4000
-2000
0
2000
4000
6000
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
B
a
s
e

s
h
e
a
r

(
k
N
)
without pins
with pins

(a) structure 1 (drift < 0.010h) (b) structure 2 (drift < 0.005h)
Figure 4-33. Base shear-top displacement curves under dynamic analyses.
Table 4-10. Maximum base shear in non linear dynamic analyses
Max base shear Without pin kN With pin kN Ratio With / without %
Structure 1 Drift < 0,010 h 4105 3169 75
Structure 2 Drift < 0,005 h 5972 3329 53


194
Table 4-11. Maximal values of the parameter in non linear dynamic analyses
max
(at level 1) Without pin With pin
Structure 1
Drift < 0,010 h
0.11 0.05
Structure 2
Drift < 0,050 h
0.05 0.07

The great overstrength of the diagonal sections in the upper levels in a classical design is
confirmed. Figure 4-34. At level 1, both classical diagonals and dissipative connections
yielding and dissipate energy, but at level 4, the diagonals of the classical design remain
elastic while variable pin connections dissipate energy by plastic deformations.
-4000
0
4000
8000
-0.15 -0.05 0.05 0.15
deformation [m]
f
o
r
c
e

[
k
N
]
level 4
without pins
with variable pins
-4000
0
4000
8000
-0.15 -0.05 0.05 0.15
deformation [m]
f
o
r
c
e

[
k
N
]
level 1
without pins
with variable pins

Figure 4-34. Force-deformation curves of diagonals.
The results of the dynamic non linear analysis indicate that some yielding takes place
under the service earthquake both the classical design and the one with dissipative
connections.
The tri-linear curve used to represent the behaviour of pin connections is one cause
already mentioned. The contribution of higher vibration modes, affecting a design made
by the lateral force method (1 mode), is another possible explanation of some yielding, as
some peaks of the accelerograms are 25 % above the 4 m/s PGA.
Plastic deformations are essentially localized in the bottom bracing of the frame, except in
the design with pins of variable strength. All design do satisfy the design criteria.

195
4.3.5.4 Conclusions of the application of dissipative connections to a tall office
building with X bracings
Weak points of the classical design of frames with bracings have been set forward. In
high rise structure, the Serviceability Limit State (interstorey drift limits under service
earthquake) criteria governs the design of braces, leading to a heavy structure. The search
for stiffness concludes into large diagonals and capacity design extend large overstrength
to the non dissipative elements, columns and beams. The need for heavy diagonal
sections is also in opposition with the lower bound limit of slenderness in Eurocode 8
(stockiness limit!). The high stiffness of the structure also results in a higher seismic
design action.
Structures designed with dissipative connections of the diagonals eliminate the problem
of buckling of the diagonals and the associated uncertainties on the real stiffness of the
structure to consider in the design process. The designer is in control of the real plastic
resistance of the structure, since the dissipative devices can be calibrated to the needed
resistance. The fact that dissipative connections are calibrated devices eliminates the
uncertainty on the real yield strength limit of the plastic zones and deletes the need to
consider material overstrength factors in the design of non dissipative elements adjacent
to the dissipative ones.
All designed structures are demonstrated to be able to reach the target displacement.
Structures using INERD pin connections with variable resistance deform plastically at all
storey levels. This more uniform distribution of deformations results in an ability to reach
greater displacements for a given value of the parameter (P- criterion). This means
less sensitivity to P-D effects and more available global ductility with the use of
dissipative connections. Structures using INERD pin connections can be designed to
provide the required resistance and have little overstrength. Classical design always
conclude in a significant overstrength, which is of at least 2: resultant base shear is 3000
kN in a design using pin connections; it is 6000 kN in a classical design, which
correspond also to much higher forces applied to the foundations and to much higher
foundation costs.
The structures analyzed are tall and limitation of deflections governs the design; for this
reason, the use of pin connections does not result in a significant reduction of the
necessary beam, column and diagonal sections. It results in the possibility to use a lower
steel grade: S235 instead of S355. Previous evaluations made on low rise structures have
shown that a weight reduction of structural steel framework leading to increased
competitivity for low rise structures (Plumier et al. 2006). The assessment presented here
confirms that trend.

196
4.3.6 Application of dissipative connections to a tall industrial building with V
bracings
4.3.6.1 Reference structure and design conditions.
The reference bracing structure is part of an industrial building. It is 41 m high and the
dimensions in plan are 9.5x 9.4 m, with 3 bays in each direction. The central bay has a 3,5
m span and is dedicated to the bracing. Figure 4-35. All bracing elements are supposed
hinged at both ends.

Figure 4-35. Perspective of the studied structure with its original X bracings (left) and with the
studied inverted V bracings (right).
Gravity loading G= 17,2 kN/m is a uniform load / m of span applied on all beams at all
14 storeys. It includes the dead load and the permanent loads.
The earthquake is near field or type 2 with PGA a
g
S = 2 m/s
2
.
As the bracings are V types, the behaviour factor q is considered equal to 2, which is the
Eurocode 8 value for dissipative frames with V bracings of class DCM.

197
Five different structures are designed and analysed in order to evaluate the influence of
the design conditions (Table 4-12).
Table 4-12. The 5 design options analysed
Structure 1
Classical design of dissipative structure.
Yielding in tension diagonal. No dissipative connections.
Comply with all Eurocode 8 requirements. q = 2
Structure 2
Skip Eurocode 8 requirement
max
min
1, 25
q =2
Structure 3
Skip Eurocode 8 requirement 2 q = 2
Structure 4
Dissipative connections.
All Eurocode 8 requirements fulfilled. q = 2
Structure 5
Non dissipative structure. q = 1,5

Comparison of periods and spectral ordinates.
Figure 4-36 shows the periods of 1st modes of each structure and the spectrum ordinates.
One can see that the differences due to the design criteria considered are huge.
The ratio for periods between structures 1 (T1 = 0,78 s) and structures 3, 4 and 5 (T1 =
1,5 s, 2,6 s and 2,64 s respectively) is around 3 and the ratio for spectrum ordinates
around 5. It may seem strange that limits of deformation are still respected in those
flexible structures, but this is a result of a better distribution of storey drifts in those
flexible design, in comparison with classical design, as it is shown later with more details.

198

Figure 4-36. Results in terms of periods and spectral ordinates.
4.3.6.2 Comparison of mass of designed structures.

Figure 4-37 shows the total mass of each primary resisting structure and the distribution
between the various structural elements.
The design process set forward the fact that two criteria of Eurocode 8 lead to huge
sections in comparison to wind design:
criteria 1: 2 which leads to
min
1
2
y
f
L
i
E
( 4.3.8 )
and prevent the use of sections with a radius of gyration smaller than 18,5 mm
criteria 2 :
max
min
1, 25
Period T(s)

199
The elements which are mostly penalised are diagonals and columns, which represent
80% of the structure weight. Not respecting one of the criteria brings a reduction in mass
of the earthquake resisting structure up to 90 %. The most requiring criteria is the limit in
slenderness. The simultaneous compliance to criteria 1 and 2 is related to the choice of
designing a dissipative structure (but with q = 2 only in this case, which does not bring
much return on the investment). This choice can be questioned and a non dissipative
design (structure 5, q = 1,5) is indeed a more economical option. However, this is
normally allowed only in low seismicity area, which is not the case here (a
g
S = 0,2 g > 0,1
g).
Structure 4 design with dissipative pin connections fulfil all Eurocode 8 criteria and
also brings a 90% reduction in mass of steel in comparison to a classical dissipative design
without dissipative connections complying with all EC8 criteria (structure 1). This
impressive influence results from several factors:
reduction of base shear due to a lower stiffness of the structure
reduction of overstrength coefficient (1,00 instead of 3,96 and 1,76 in structure 1)
no unbalance between forces in the tension and the compression diagonals, resulting
in no additional action effects in beams
The use of pin connections does not create problems with P- effects: the parameter
max
remains smaller than 0,10.
Up to now, the comparison bears on superstructures, but one knows that foundations
also have a cost and that aspect of the problem should be envisaged. This is studied in the
next paragraph.


200

Structure 1 2 3 4 5
Diagonals 3258 1186 318 1527 368
Ext. Columns 7523 3752 1980 1980 1980
Beams 17908 10321 4739 1716 5132
Int. Columns 103635 17904 4865 7397 6852

Figure 4-37. Mass of structures 1 to 5. Distribution between diagonals, ext. columns, etc, is from top
to bottom in the graph.
4.3.6.3 Influence of superstructure design on the dimensions of foundations.
The assessment considers only the reaction in tension at the base of the bracing. It allows
a comparison between various design options for what concerns the anchorage required.
The unit of comparison is the isolated foundation pad presented at Figure 4-38, which
has to resist tension. Two volumes of soils are considered:

201
The volume which it is necessary to dig out to realise the foundation pad. Figure
4-38.
The volume which contributes to the stability.
The mass per unit volume of soil is taken equal to 19 kN/m3.

Figure 4-38. Volume of soil to dig out to realize the foundation pad.
4.3.6.4 Results of the analysis.
The results are presented at Table 4-13. They show that the classical dissipative structure
1 requires foundations which are by far the most expensive one. Very interesting is the
difference between design 4 and 5, with 4 involving 25% less soil digging than for
structure 5.

Table 4-13. Applied action effect at the foundation, dimension of a pad and volumes of soils to dig
out.

F (kN) a (m) Volume of soil to dig out (m3)
structure 1 3030 4,2 359,03
structure 2 1440 3,28 192,27
structure 3 693 2,57 106,53
Volume of soil
to dig out to
realize the
foundation pad

202
structure 4 1080 2,97 150,70
structure 5 1350 3,21 182,29

4.3.6.5 Conclusions of the design stage.
A classical design of dissipative frames with V or bracings generates a very heavy
structure. This is due to the fact that the design, though it is qualified as dissipative is
affected by a very low behaviour factor q = 2, while at the same time one keeps paying
the price of two criteria on slenderness of diagonals, which generates rather heavy
sections, coupled to the capacity design of beams and columns, which on the basis of
heavy diagonals generates very heavy beams and columns. On the whole, the impact of
design conditions on structure 1 weight is high. From the above observations, it results
that two design options can be envisaged:
a non dissipative design.
a design using dissipative connections.
Those design options results in very important weight reduction by a factor above 5 in
comparison to the classical design of a dissipative structure 1.The design made for
structures 4 and 5 indicate similar weight for both solutions, but non dissipative design
should normally not be used in moderate and high seismic zones. On the contrary, design
using dissipative connections can be used in all zones.
The analysis of foundations of the different design of superstructures shows that the
classical dissipative structure 1 involve foundations which are by far the most expensive
one, requiring above 2 times more work in foundations that design making use of
dissipative connections
4.3.6.6 Push-over analysis
The push-over curves of the various structures are presented at Figure 4-39 which shows
the important difference in resistance between structure 1 (classical dissipative design
complying with Eurocode 8 requirements) and structures 2, 3, 4 and 5. This set forward
that there is for structure 1 an important difference between the design shear and the
design resistance: the overstrength is high.


203
0
200
400
600
800
1000
1200
1400
1600
1800
2000
0 0,5 1 1,5
dplacement en tte (m)
c
i
s
a
i
l
l
e
m
e
n
t

e
n

b
a
s
e

(
k
N
)
courbe pushover
structure1
structure2
structure3
structure4
structure5

Figure 4-39. Push-over curves of structures 1 to 5.
In structures 2 to 5, the difference between required and provided resistance is much
smaller, which indicates much better design. Structure 4 is characterised by a progressive
decrease of stiffness as a result of the trilinear behaviour curve of pin connections .
Figure 4-40 shows the deformed shapes of structures 1 to 5 at = 0,1. One can notice
that there is no plastic redistribution in structures 1, 2, 3 and 5. Their failure mechanism is
restricted to yielding at one level, which goes together with a low global ductility of the
structures. On the contrary, structure 4 with variable dissipative connections offers a
global plastic mechanism. Its deformed shape is continuous and regular, which explains
why the parameter is smaller than in the other design. Structures design making used of
dissipative connections can be much less sensitive to P effects.
S
h
e
a
r

B
a
s
e

F
o
r
c
e

(
k
N
)


204

Figure 4-40. Deformed shapes of structures 1 to 5 at = 0,1
industrial building with V bracings
A classical design of frames with V or bracings in which energy dissipation takes place
in the diagonals is not economical. This is due to the low behaviour factor q allocated to
such structures in Eurocode 8, which is justified by the fact that such design cannot really
be very ductile. The analysis shows that yielding tend to concentrate into the first yielded
member, which involve two negative consequences: the plastic capacity of that member
defines one limit for the structure, which may not be so high; the drift is concentrated in
the storey where the diagonal is yielding, which creates a high P effect in that storey,
high enough to be the cause of a global failure.
Designing a non dissipative structure is an option which generates reduction of the
structure cost, but it is in principle restricted to low seismicity areas.


205
industrial building with V bracings
A classical design of frames with V or bracings in which energy dissipation takes place
in the diagonals is not economical. This is due to the low behaviour factor q allocated to
such structures in Eurocode 8, which is justified by the fact that such design cannot really
be very ductile. The analysis shows that yielding tend to concentrate into the first yielded
member, which involve two negative consequences: the plastic capacity of that member
defines one limit for the structure, which may not be so high; the drift is concentrated in
the storey where the diagonal is yielding, which creates a high P effect in that storey,
high enough to be the cause of a global failure.
Designing a non dissipative structure is an option which generates reduction of the
structure cost, but it is in principle restricted to low seismicity areas.
The use of variable dissipative connections (the provided resistance meets the required
resistance) leads to the best design: a real global plastic mechanism is formed, local
exhaustion of plastic capacity of a member is avoided, as well as huge P- effects due to
concentration of in one single storey over the height. This design option is the most
economical in mass of the earthquake primary resisting structure (90% lighter than a
classical design) and in terms of foundation costs (25% less than for a non dissipative
structure).
4.3.7 General conclusions on the use of dissipative connections in frames with
bracings
Structures designed with dissipative connections of the diagonals eliminate the problem
of buckling of the diagonals and the associated uncertainties on the real stiffness of the
structure to consider in the design process. The designer is in control of the real plastic
resistance of the structure, since the dissipative devices can be calibrated to the needed
resistance. The fact that dissipative connections are calibrated devices eliminates the
uncertainty on the real yield strength limit of the plastic zones and deletes the need to
consider material overstrength factors in the design of non dissipative elements adjacent
to the dissipative ones.
Structures using INERD pin connections with resistance calibrated to demand
(variable pins) deform plastically at all storey levels. It is not the case with classical
dissipative design. The more uniform distribution of deformations achieved results in an
ability to reach greater displacements for a given value of the parameter (P- criterion).
This means less sensitivity to P- effects and more available global ductility with the use
of dissipative connections. Structures using dissipative connections can be designed to
provide the required resistance and have little overstrength. On the contrary, classical

206
dissipative design always conclude in an overstrength which is of at least 2 and
correspond to bigger steel sections and higher forces applied to the foundations, resulting
in higher foundation costs.
The flexibility in tuning resistance to a demand pattern is of interest in new design as well
as in the retrofitting of existing structures. Such flexibility can only result from the use of
specific industrialised components having resistances which are precisely under control.

5. SEISMIC UPGRADING OF STRUCTURES USING BASE
ISOLATION
5.1 DISPLACEMENT BASED DESIGN MODELS FOR BASE ISOLATED HISTORICAL
BUILDINGS
5.1.1 Introduction
Recent earthquakes have shown the particular vulnerability of old masonry structures and
the need to develop adequate retrofitting methodologies.
In the definition of the retrofit interventions to improve the seismic capacity of the
structures is important to respect the balance between the structural safety and the
architectural preservation. It is in this frame that base isolation has been suggested as a
valid retrofit strategy for old masonry buildings.
The use of base isolation to reduce the vulnerability of old masonry structures is still very
limited in number of applications. Since the emphasis on the seismic design of structures
is changing from strength to performance, the objective of this study is to adapt the
Performance Based Methodologies to the analysis of base isolated structures, in particular
low ductility structures as old masonry historical buildings.
The objective of this document is to present a methodology for the seismic analysis of
old masonry structures with base isolation, based on the concept of Performance Based
Design.
In this document is also presented the application of the proposed methodology in the
study of the base isolation of an old masonry structure, the Capelinhos Lighthouse.
5.1.2 The proposed methodology
The proposed methodology is based on the concept that the base isolation decouples the
horizontal movement of the structure from the soil movement. This behaviour means
that the base isolation system acts like a filter, reducing the seismic input to the structure.
This global behaviour is as more effective as the ratio between the fixed base frequency

208
and the isolated frequency increases. Ratio values higher than 3 are recommended in
order to obtained the desired effect [Skinner et al., 1993].
Since this methodology is supposed to be applied in existent structures is assumed that
the behaviour of the structure when base fixed is already known and well defined. In this
case is possible to consider a base isolation solution with the characteristics necessary to
assure the right isolation.
The vibration mode configurations typical of isolated structures show that, apart the first
mode, the other modes present deformation patterns similar to the fixed base modes
(Figure 5-1). In the first isolated mode the deformation is concentrated on the isolation
system with reduced structural deformation. The deformation on the structure, and
consequent stresses are due to the higher modes, with deformation similar to the fixed
base structure but with much less intensity since the mode response is reduced due to the
isolation.

Figure 5-1 Vibration modes comparison
The assessment of the structure response can be realized using the capacity spectrum
definition [e.g. Priestley, 2000]. In Figure 5-2 is presented a typical capacity spectrum for a
fixed base structure with non ductile behaviour. In this figure is also represented a typical
behaviour curve for an isolated structures. Since the base isolation is consider almost
linear and the structural behaviour of the isolated structure is assumed to remain elastic,

209
the behaviour curve of the isolated structure is a radial line starting at the origin, with the
slope characteristic of the isolated period.
0.0
1.0
2.0
3.0
4.0
5.0
6.0
0.00 0.02 0.04 0.06 0.08 0.10
Displacement (m)
A
c
c
e
l
e
r
a
t
i
o
n

(
m
/
s
2
)
= 5%
= 15%

Figure 5-2 Example of capacity spectrum with base isolated structure
It is important to refer that to obtain the capacity curve of the structure is necessary to
define the equivalent single degree of freedom oscillator, since the structure behaviour is
simplified to a deformation versus an acceleration (or force).
The single degree of freedom behaviour is expressed through effective displacement and
mass values. The effective mass definition is used to establish the relation between the
base shear and the mass acceleration and the effective displacement is a measure of the
structure deformation. According to this consideration and based on the analysis of the
mode configurations presented on Figure 5-1, one can assume that the behaviour of the
structure on the top of the base isolation layer is similar to the behaviour of the fixed
base structure and will be governed by the same capacity curve. Two different equivalent
single degree of freedom systems must be considered: one considering the global
deformation of the structure, including the base isolation displacement; and a second
system considering only the structural deformation.
The total structural deformation of the base isolated structure (D
T
) must be subdivided in
two components: the base isolation deformation (D
I
) and the structures deformation
(D
S
). For each type of deformation, effective displacement and effective mass are defined
using the following equations [Priestley, 2000]:

Fixed
Base
Isolated

210

1) Global deformation

n
i Ti
i
I
n
i Ti
i
m D
m D
2
1
1
( )
( )
=
=
=
(5.1.1)

n
i Ti
i
I
I
m D
m
1
( )
=
=
(5.1.2)
2) Structure deformation

S T I
D D D =

(5.1.3)

n
i Si
i
S
n
i Si
i
m D
m D
2
1
1
( )
( )
=
=
=
(5.1.4)

n
i Si
i
S
S
m D
m
1
( )
=
=
(5.1.5)
To evaluate the behaviour the base isolated structure is proposed the following
methodology. The first step must be the definition of the performance of the isolated
structure (Point 1 in the Figure 5-3). This point is defined by the spectral displacement
(S
d
) and the spectral acceleration (S
a
) of the point where the capacity curve of the isolated
structure crosses the design capacity spectrum. Using the effective mass value evaluated
for the global deformation model (m
I
) is possible to calculate the base shear force (V
I
).

I I a
V m S = (5.1.6)
Since the global base shear force on the isolating system is the same as the structure base
shear, is possible to estimate the equivalent acceleration on the equivalent structure model
and define the performance point of the structure itself (Point 2).

S I equiv
V V ms Sa ( ) = =

(5.1.7)


211

I
equiv
S
m
Sa Sa
m
( ) = (5.1.8)
0.0
1.0
2.0
3.0
4.0
5.0
6.0
0.00 0.02 0.04 0.06 0.08 0.10
Displacement (m)
A
c
c
e
l
e
r
a
t
i
o
n

(
m
/
s
2
)
= 5%
= 15%

Figure 5-3 Capacity spectrum with methodology presentation
This procedure can be used also in the reverse order. Starting with the definition of the
desired structure performance state (similar to Point 2 in Figure 5-3), it is possible to find
a base isolation solution with the necessary behaviour (Point 1 definition).
In the following sections is presented the application of this methodology to a non-
ductile masonry structure, the Capelinhos Lighthouse.
5.1.3 The Capelinhos lighthouse
5.1.3.1 Introduction
The Capelinhos lighthouse is a landmark of the Fayal Island in Azores (Figure 5-4). The
lighthouse tower is around 30 m height and is made of stone masonry. At the base there
is a two storey building with rectangular shape. The first storey of this building is buried
in the ashes of the 1957/58 Capelinhos Volcano eruption. During the eruption the
landscape around the lighthouse had a profound modification with the elevation of a
mountain from the water just in front of the lighthouse. The lighthouse itself was also
affected by the ashes and, since then, it was abandoned. Recently the local authorities
decided to rehabilitate the structure and several studies had been conducted. The
information collected due to those studies allowed the development of the present work.
(1)
(2)

212

Figure 5-4 The Capelinhos Lighthouse
The tower cross section is octagonal with a face length of 2.32 m at the base and a face
length of approximately 2.00 m at the top. The windows and doors are located in just two
opposite faces of the octagon. The structure is composed by regular stone masonry and,
in result of an intervention that happened before the Volcano eruption, there is an inside
reinforced concrete reinforcement with 0.20 m depth. The RC reinforcement is in poor
conditions and is obvious the lack of continuity of the vertical steel bars.
The first step on the process of modelling the Lighthouse was the definition of the 3D
elastic model of the fixed base structure. The geometry of the model was based on copies
of the original drawings of the structure. For the complete definition of the model it was
necessary to assume some characteristics of the materials, such as the weight for unit
volume () and the Modulus of Elasticity (E).
As a result of the observation of the masonry type and according to previous experiences
with similar materials it was assumed a value of 20 kN/m
3
for the weight per unit volume
(). The definition of the Modulus of Elasticity value was based on the results of a
dynamic identification.
The dynamic identification was conducted in order to get the values of the natural
frequency of vibration of the tower. Those results were important to calibrate the elastic
characteristics of the model. These elastic characteristics were not only the Modulus of
Elasticity of the material but also the definition of the restraining at the base, imposed by
the soil and by the walls of the structure that exists at the tower base.
The identified frequencies were 2.15 Hz, for the North-South direction, and 2.25 Hz for
the East-West direction. The difference of frequency on the two directions is related with

213
the presence of the openings (windows and doors) and the main directions of the
restrictions at the base of the tower.
The calibration the model characteristics was performed through the comparison of the
measured frequencies with those obtained with a linear elastic model.
To simulate the Lighthouse tower behaviour, three-dimensional solid elements were used.
The element used was an 8 nodes isoparametric one with 3 degrees of freedom at each
node. To simulate also the interior reinforced concrete layer, two different materials were
defined, corresponding one to the masonry and other to the concrete.
The model considers all the lighthouse height, including the base that is, at the present
moment, covered with ashes. To simulate the effect of the ashes layer that surrounds the
base and the restraining effect of the masonry walls of the structure at the base, linear
springs were used.
For the reinforced concrete it was established a weight per unit volume of 25 kN/m
3
, and
a Modulus of Elasticity of 25 GPa (weak concrete). For the masonry it was decided to
assume a specific weight of 20 kN/m
3
, value considered valid for the type of masonry of
the tower. After the complete geometric definition of the model some values of masonry
Modulus of Elasticity were tested in order to get a good match between the frequencies
obtained with the model and those recorded in the field tests. With this procedure it was
obtained a value of 4.4 GPa for the Modulus of Elasticity of the masonry, value that was
considered acceptable since the masonry was not deteriorated. On Table 5-1 are referred
the values of the first five modes of vibration.
Table 5-1 Modal frequencies of fixed base solution
Mode Frequency (Hz) Observations
1 2.148 1
st
mode in X direction
2 2.243 1
st
mode in Y direction
3 9.970 2
nd
mode in X direction
4 10.121 2
nd
mode in Y direction
5 10.823 1
st
torsion mode

After the complete definition of the fixed base model it was possible to build the isolated
model. It was decided to locate the isolation system at the base of the tower, at the level
of the top of the two storey building that exists on the base. With this option it was

214
decided to only isolate the tower, maintaining the base building rigidly connected to the
ground.
To define the stiffness of the isolation system the option was to assume an isolated
frequency of 0.50 Hz. This frequency is approximately one fourth of the rigid base
frequency, ratio that is usually pointed as an optimal value. It must be stressed that the
natural frequency of the fixed base model is little above 2.00 Hz and the most important
vibration modes involve essentially the tower, element that is going to be isolated.
According to the developed model, the total mass above the isolation level is around 650
tons. With this value and targeting an isolated frequency of 0.50 Hz it was assumed a
global horizontal stiffness on the bearings of 6400 kN/m. It was decided to consider a
total of 8 bearings, with a horizontal stiffness of 800 kN/m each.
This isolation solution, with a frequency of 0.50 Hz, would be the reference solution in
the study. Other solutions, with different horizontal stiffness values would be also
analysed.
After the modal analysis of the isolated model there were obtained the results presented
on Table 5-2.
Table 5-2 Modal frequencies of base isolated solution
Mode Frequency (Hz) Observations
1 0.475 1
st
torsion mode
2 0.486 1
st
mode in Y direction
3 0.487 1
st
mode in X direction
4 2.879 2
nd
mode in Y direction
5 2.938 2
nd
mode in X direction

5.1.3.2 The Capelinhos Lighthouse non linear model
Since the objective of the study is the analysis of the structure behaviour including the
non linear range it was necessary to develop a non linear model of analysis. This model
will allow evaluating the seismic behaviour of both the fixed base and isolated situations.
The only difference would be the introduction of a horizontal flexible layer at the
isolation level on the isolated model.

215
In order to allow a large number of numerical analysis it was decided to define a
simplified model. This model is composed of vertical linear frame elements connected by
joints with nonlinear behaviour. In the model, the nonlinear behaviour would be
concentrated on the nodes. The nonlinear characteristics of the joints were obtained
through the analysis of fibre models representative of each reference cross section.
To consider the non linear behaviour it was developed a fibre model for each analyzed
cross section. There were defined nonlinear cross sections separated at a maximum
distance of 1 metre and also at all the irregular cross sections, such as sections with
openings. There were considered a total of 31 different cross sections. On Figure 5-5 are
presented two different fibre model cross sections as examples.
On the fibre model definition it was admitted the hypothesis of no tension on the
masonry and a maximum compression value of 4.0 MPa. For the moment deformation
relation definition, there were applied crescent deformation to the section and evaluated
the resulting moment. The procedure was conducted at a constant axial force. The axial
force considered at each section corresponds to the effect of the dead weight of the
tower above the cross section in analysis. With this procedure, the moment-deformation
characteristics of all the nonlinear joint elements considered in the model were obtained.
On Figure 5-6 is presented an example of moment-deformation relation obtained in the
analysis. The relation represented in the figure is representative of the behaviour for just
positive moments. In most of the sections the behaviour for negative moments is similar
due to the symmetry of the sections. The only exceptions are related with the sections
with openings.

Figure 5-5 Examples of cross section fibre models

216
0
2500
5000
7500
10000
12500
15000
17500
0.0E+00 5.0E-04 1.0E-03 1.5E-03 2.0E-03
Curvature
M
o
m
e
n
t

(
k
N
m
)

Figure 5-6 Moment-curvature relation example
5.1.3.3 Capacity curve definition
For the capacity curve definition there were conducted seismic nonlinear analysis based
on the presented simplified model. To test the evolution of the structure behaviour, the
seismic action was applied in different levels, starting on 25 % of the reference peak
ground acceleration (Pga) and finishing at a maximum of 200 % the reference Pga. It was
considered as reference Pga the peak ground acceleration defined in the Portuguese code
[RSA, 1983] for the seismic action used in the study.
The result obtained at each level corresponds to the average result of the analysis with all
the accelerograms considered in the study. The reference values were: the shear force at
the isolation level, the top horizontal displacement, the horizontal displacement at the
isolation system and the horizontal relative displacement between the top and the
isolation level.
In the analysis of the fixed base situation it was tried the following 6 different seismic
levels: 25 %, 50 %, 75 %, 100 %, 125 % and 150 %. Since it was verified that the tower
could not support a 150 % seismic level there were conducted analysis at 110 %, 120 %,
130 % and 140 %. It was also verified that the 140 % level was also too strong for the
structure. For that reason were considered only the results of the analysis of seismic
intensity from 25 % Pga to 130 % Pga.
In the analysis of the isolated solution the study was conducted, without any structural
problem, until a 200 % seismic level (200 % of the code defined Pga). In order to obtain
more information for the study it was decide to consider 2 different additional isolation
solutions, corresponding one to a isolated frequency of 0.40 Hz (more flexible solution)

217
and other to a isolated frequency of 0.60 Hz (more rigid solution). In the study the
original isolation solution is referred as Solution 1, the softer additional solution as
Solution 2 and the other as Solution 3. It was also considered different damping
coefficients values for each solution. The basic characteristics of the base isolation cases
considered in the study are presented on Table 5-3.
Table 5-3 Base isolation solution characteristics
Solution Frequency (Hz) Damping Coefficient (%)
1 0.50 12.5
2 0.40 15.0
3 0.60 10.0

In the global behaviour there are two main points in analysis: the maximum base shear
force on the tower and the maximum displacement at the isolation level. The first is
related to the structure capacity and the second to the base isolation system limit state.
On Figure 5-7 are presented the evolution of the shear base with the seismic intensity, for
all the tested situations. According to the results the base fixed structure exhibits
nonlinear behaviour when the isolated solutions show a linear evolution for the shear
base with the seismic intensity, revealing that the structure was always in the elastic range,
as expected.
0
200
400
600
800
1000
1200
0.00 0.50 1.00 1.50 2.00
( x PGA )
B
a
s
e

S
h
e
a
r

F
o
r
c
e

(
k
N
)
Fixed Base
Solution 1
Solution 2
Solution 3

Figure 5-7 Vulnerability functions for shear force
On Figure 5-8 the obtained results are represented over the capacity spectrum for the
seismic action used in the analysis. To mark the seismic response of the fixed base
structure and the response of all the three isolation solutions it was necessary to evaluate

218
the effective displacement and the effective mass associated to each case. On Table 5-4
the values of the equivalent single degree of freedom systems are presented. In the Table
the effective displacement is presented as a percentage of the top displacement, and the
effective mass as a percentage of the total mass.
Table 5-4 Equivalent single degree of freedom characteristics
Solution Effective Displacement Effective Mass
Fixed 0.57 D
TOP
0.65 M
TOTAL

Isolated 0.95 D
TOP
1.00 M
TOTAL

The results presented on Figure 5-8 show that the fixed base structure could not resist to
the seismic action present on the current version of the Portuguese code. With all the
base isolation solutions it is possible to sustain a seismic action stronger than the code
defined.
As expected, the behaviour of the base isolated situations corresponds to a straight radial
lines, each one with the slope characteristic of each isolated period. The Solution 3, that
corresponds to the more flexible solution (T=2.5 sec.) presents a higher slope than the
other two lines (T=2.0 sec. and T=1.67 sec.).

Figure 5-8 Capacity Spectrum with all the base isolation solutions

219
To confirm the presented methodology is necessary to check if the behaviour of the
structure on the top of the base isolation follows the global capacity curve of the fixed
base structure. To perform this confirmation it was necessary to evaluate the effective
mass (m
S
), related only to the structural deformation, using equation 5.1.5. The
calculations confirmed a result similar to that obtain for the fixed base structure, 65 % of
the total mass.
To establish the comparison between the results three pairs of points were marked over
the capacity curves diagram. There were chosen three different situations, one for each
base isolation solution, corresponding all to the 100 % Pga seismic action level.
The results can be observed on Figure 5-9, where the square marks, (on the right side)
refer the performance points on the base isolated structure, and the circle marks (on the
left side) represent the corresponding performance points over the capacity curve of the
structure. The points marked over the capacity curve of the structure presents the same
base shear force of the corresponding points marked over the isolated structure. The
higher values of equivalent acceleration result from the lower value of effective mass
characteristic of the structure deformation when compared with the base isolated
situation.
The presented results show a good agreement between the points obtained from the
proposed methodology and the capacity curve obtained directly from the fixed base
analysis. It is also possible to conclude about the efficiency of the base isolation solution
on the protection of structures.

Figure 5-9 Capacity Spectrum with all the base isolation solutions

220
5.1.4 Conclusions
The main objective of this document was the description of a methodology proposed to
adapt the displacement based design concept (DBD) for the analysis of old masonry
buildings.
According to the obtained results we can conclude that the proposed methodology may
be a useful tool on the design of seismic base isolation to protect an existen structure.
The fixed base capacity curve can be used to characterize the seismic behaviour of the
isolated structure over the isolation surface.
The base isolation is an efficient solution to provide seismic protection to structures with
non-ductile behaviour.
5.2 NONLINEAR METHOD FOR CONTROL OF AUTO-ADAPTIVE SEMI ACTIVE BASE
ISOLATOR
Civil engineering structures have traditionally been built as passive structures with no
adaptability to uncertain dynamic loads. Indeed, solidity and massiveness have been
considered as a measure of the safety and reliability. In recent decades, new alternative
approaches such as supplemental damping, passive control, semiactive and active control
have been developed to protect structures from earthquakes and severe winds. A detailed
and comprehensive literature review for these new approaches has been given in
Deliverables 43 (www.lessloss.org ).
One challenge in the use of semiactive controllable fluid dampers is to develop
implementable nonlinear control algorithms. In the first annual report of the project, an
implementable nonlinear control algorithm has been proposed and its performance has
been investigated compared to uncontrolled and exact optimal cases. The comparison
results of the proposed control, passive and semiactive cases are given and discussed.
Lastly, a new linear optimal control rule is derived, which can be used to obtain
semiactive control laws.
5.2.1 Structural System
This study examines the earthquake response of seismically isolated two-degree of
freedom (2DOF) structures, in which a controllable fluid device (i.e., an MR damper) is
placed at the isolation level. The base isolation system is assumed to be linear and is
represented by a spring in parallel with a linear viscous dashpot.

221
The masses m
s
and m
i
represent the superstructure of the building and the mass of the
base floor above the isolation system, respectively. The superstructure and the isolation
stiffness and damping coefficients are represented by k
s
, c
s
, k
i
and c
i
, respectively.
Structural relative displacements and the ground displacement are denoted by r
1
, r
2
and z,
respectively. For the investigated structure; the masses, the stiffness and the damping
coefficients are selected as m
i
=m
s
=10
5
kg, k
i
=16.5 kN/cm, k
s
=300 kN/cm, c
i
=1.1
kN/cm/sec and c
s
=0.3 kN/cm/sec.
The MR damper implemented in the isolation level exhibits controllable stiffness and
damping and is modelled by the following algebraic expression with five parameters
MR d d
f f uH u r d r v k r c r
0 1 0 1 0 1 1
( )tanh( / / ) = + + + (5.2.1)
where the constant parameters d
0
and v
0
are used to describe the pre-yield behaviour of
the device, f
0
is a controllable yield force, and k
d
and c
d
describe the post yield behaviour
and the behaviour when u=0. The control force f is changed optimally via the control
decision variable u; H(u) is the Heaviside step function of u. Device parameters used for
this case are f
0
=100 kN, d
0
=5 cm, v
0
=4 cm/sec, c
d
=2 kN/cm/sec and k
d
=3 kN/cm.
These parameters can be used to represent an MR damper, an ER damper, or a
controllable valve damper.
5.2.2 System Dynamcs
The nonlinear system of dynamical equation of a 2DOF base isolated structure with an
MR damper defined by (5.2.1) can be expressed as
t t u t t u u z t t
0
( ) ( ( ), ( ), ) ( , ) ( ), ( ) = = + + = x f x Ax g x H x 0 (5.2.2)
where
i s d s i s d s
i i i i
s s s s
s s s s
r
r k k k k c c c c
r m m m m
r k k c c
m m m m
1
2
1
2
0 0 1 0
0 0 0 1
;

+ + + +

= =

x A
(5.2.3)

222
i
u
f m H u r d r v
0 1 0 1 0
0 0
0 0
( , ) ;
( / ) ( )tanh( / / ) 1
0 1

= =

+

g x H
(5.2.4)
5.2.3 Optimal and Sub-optimal Control
Optimal and sub-optimal control schemes have been given in detail in the first annual
report. Optimal control minimizes
f f
f
t t
t t
t
s s
s t
J L t u t t dt r t z t dt
k r t r t c r t r t
dt
m
0 0
0
2
2
2
1 2 1 2
( ( ), ( ), ) [ ( ) ( )]
[ ( ) ( )] [ ( ) ( )]
= = +
+
=

x

(5.2.5)
where the Lagrangian ) ), ( ), ( ( t t u t L x is the square of the superstructure absolute
acceleration. The sub-optimal control trajectory minimizes the integral cost function
t N t
so so
t
J L u d ( ( ), ( ), )

+
=
x (5.2.6)
5.2.4 Passive Viscous Damping Control And on-off Cases
Using the previously defined quantities, the basic equations of motion of 2DOF base
isolated structure with no control can be modelled by a first-order system of differential
equations with constant coefficients. For the passive viscous damping case in which
f
MR
=0 during the earthquake, viscous damping coefficient of the base isolation system
was changed as c
i
=1.1, 2.0, 5.0, 7.0 and 10.0 kN/cm/sec. The selected viscous damping
coefficients correspond to first-mode damping ratios of 5.8, 10.6, 16.7, 37.5 and 54.2
percent of critical.
In on-off cases, semiactive damper force is

MR d d
f k r c r
1 1
= + (5.2.7)

223
for u=0 and

MR d d
f f r d r v k r c r
0 1 0 1 0 1 1
tanh( / / ) = + + + (5.2.8)
for u=1.
5.2.5 Causal semactve control
The causal semiactive control system for this case uses the MR damper force (5.2.1) in
which the control policy u H(u) is replaced by v, where the delayed control decision v is
given by

v MR
v T v H f r z
1
(1/ )[ (( )( ))] = + (5.2.9)
for bang-bang control case and

v MR
v T v r z H f r z
1 1
(1/ )[ (( )( ))] = + + (5.2.10)
for continuous control case. The constant T
v
is the response time of the controllable
damper. The coefficient is a design parameter for the control rule (5.2.10) which
determines the device force level.
5.2.6 Numerical application
Seismic response of 2- degree-of-freedom low-rise base isolated structure with a
controllable fluid damper under the May 18, 1940 Imperial Valley earthquake El Centro
(NS component) is evaluated in order to examine the performance of the proposed
control in comparison to uncontrolled case and the exact optimal case. The calculations
are performed for the first 10 sec-duration of the excitation, which includes the peak
acceleration values. The damper force is zero for the uncontrolled case. The maximum
response quantities for all the investigated cases are given in Table 5-5 and Table 5-6,
respectively for comparison purposes.
It s clear from the above given results that the optimal control outperforms all the
investigated cases. However, it is known that the optimal control is an ideal case and
cannot be implemented. Among the passive cases, the case in which the base isolation
damping is c
i
=0.5 kN/cm/sec, is the optimal one in terms of the reduction in absolute
acceleration of superstructure and the cost function. It is known that the optimal isolation
damping for each earthquake is different. Since the earthquake characteristics are not
known a priori, it is almost not possible to determine the optimal isolation damping for
that specific earthquake. This is, in fact, one of the reasons of studying the semiactive
systems.

224
Table 5-5. Maximum response quantities for uncontrolled and controlled cases

Table 5-6. Maximum response quantities for passive cases

It is also seen from the tables that the proposed control performs better than all the
passive viscous damping and on-off cases in terms of the reduction in absolute
acceleration of superstructure and the cost function. Provided that the proposed control
is compared to the causal semiactive control rules, it is shown that the proposed sub-
optimal control performs better than the causal semiactive bang-bang control while it
performs almost the same as the causal semiactive continuous control. These analyses
show that the proposed sub-optimal implementable control is promising for the control
Maximum
Values
No
control
(f=0)
Proposed
control
(implementable)
Optimal control
(nonimplementable)
Bang-
bang
control
(BBC)
Continuous
control
(CC)
1
r (cm)
19.1 10.2 5.0 5.49 7.42
1 2
r r (cm)
0.54 0.42 0.3 0.44 0.40
) ( ) (
2
t z t r +
(cm /sn
2
)
163.1 125.5 97.5 134.19 119.55
J 41671 18439 16579 25000 18600
Passive linear viscous damping Passive on-off cases
Maximum
Values
c
i
=1.1
kN/cm/
sec
c
i
=2.0
kN/cm/
sec
c
i
=5.0
kN/cm/
sec
c
i
=7.0
kN/cm/
sec
c
i
=10.0
kN/cm/
sec

u=0

u=1
1
r (cm)
19.10 15.00 8.28 6.28 4.66 11.04 5.07
1 2
r r (cm)
0.54 0.47 0.43 0.50 0.67 0.47 0.74
) ( ) (
2
t z t r +
(cm/sn
2
)
163.13 140.62 129.67 150.32 200.33 139.09 217.09
J 41600 24900 18700 22300 30200 22542 44888

225
of earthquake response of base isolated structures. Proposed semiactive control scheme
can protect the base isolation system without increasing the superstructure response.
5.2.7 Derivation of the Linear Control Law
To give a brief overview of the active control concept, the governing equations of motion
for a general multi-storey shear-beam lumped mass linear building structure are
[ ] t t t f t t t t t t t
1 2 0 1 0 0
, , ; ( ) ( ) + + = + = = MX( ) CX( ) KX( ) D ( ) D U( ) X X 0

(5.2.11)
where X(t)=( x
1
,..., x
n
)
T
is the n-dimensional response vector (T indicates the transpose of
a vector or a matrix) denoting the relative displacement of the each storey unit with
respect to the ground; the superposed dot represents the differentiation with respect to
time; M is the (nxn)-dimensional diagonal constant mass matrix with diagonal elements m
i

= mass of i th storey (i=1,..., n); C and K are the (nxn)-dimensional viscous damping and
the stiffness matrices, respectively; D
1
is the (nx1)-dimensional location matrix of
excitation and given by ) ,..., (
1
T
1 n
m m = D ; D
2
is the (nxr)-dimensional location matrix of r
controllers; U(t) is the r-dimensional active control force vector and described as
U
T
(t)=(u
1
(t),...,u
r
(t)) and scalar function f(t) is the one dimensional earthquake
acceleration. In the state space, equation (5.2.11) becomes
[ ] t t t f t t t t t
0 1 0
( ) ( ) ( ) ( ), , ; ( ) = + + = Z AZ BU D Z 0 (t) (5.2.12)
in which

nxn
1 1

=

0 I
A
M K M C
;
nxr
1
2

=

0
B
M D
;
nx1

=

0
D
;

=

X
Z
X
(5.2.13)
such that 0
nxm
is the (nxm)-dimensional zero matrix; I is the (nxn)-dimensional identity
matrix; = (1,...,1)
T
is the n-dimensional vector.
In the classical optimal control law; the classical integral type quadratic performance
measure

T T
1
t
C C
0
J dt = +
(Z Q Z U R U) (5.2.14)

226
is minimized under the constraint imposed by equation(5.2.12), where [0, t
1
] time interval
is the control time and defined to be longer than that of the external excitation duration;
Q
C
is a positive semi-definite symmetric weighting matrix ; R
C
is a positive definite
symmetric weighting matrix and the subscript C refers to classical linear optimal control.
It is known that an important property of classical linear optimal control solution is that
any selection of the weighting matrices within the constraint of positive semi-definiteness
and positive definiteness will preserve asymptotic stability. Provided that the weighting
matrices Q
C
and R
C
are selected as

nxn nxn
C C nxn
nxn nxn
; ; 0, 0

= = > >

I 0
Q R I
0 0
(5.2.15)
in the numerical applications then they satisfy the above-given stability condition. Since
the derivation of the necessary conditions of optimality is given in detail by Anderson and
Moore,

only the results will be given here. Classical linear optimal control law is obtained
as

T
C
t t
1
1
( ( ) ( )
2
) = U R B P Z t (5.2.16)
where P is the solution of the following nonlinear matrix Riccati equation,

T T
C C
t t t t t t 1
1
1
( ) ( ) ( ) ( ) ( ) 2 ; ( )
2
+ + + = =
.
P P A P BR B P A P Q 0 P 0 (5.2.17)
In this study, to determine the active control force, we introduce the integral type
quadratic functional

t h
T
t h
t
I s s ds
T
,
( ) ( ) ( ) ( ( min
+
= + ) )

U Z Q Z U RU s s (5.2.18)
with two parameters 0 and
0
> h t t , where Q is the (2nx2n)-dimensional positive semi-
definite symmetric weighting matrix and R is the (rxr)-dimensional positive definite
symmetric weighting matrix; h is the time interval between the successive control
instants. The expression given by equation (5.2.18) can be considered as a functional
defined for the solutions ] , [ ), ( h t t s s + Z , of the differential equation

227
[ ] s s s f s s t t h ( ) ( ) ( ) ( ), , ; = + + +
.
Z AZ BU D ) ( | ) ( t s
t s
Z Z =
=
(5.2.19)
with admissible control vectors
[ ]
r
(s ) PC t t h , + U (5.2.20)
In Eq. (5.2.20), [ ] h t t + , PC
r
is the vector space of r-dimensional column vectors, which
have piecewise continuous elements for [ ] h t t + , s . The norm in [ ] h t t + , PC
r
can be
defined, for example, as

r
i
t h
t s t h i
Sup s
,
1
( )
+ =
=
U U (5.2.21)
It is clear that the problem of minimizing the functional (5.2.18) under the conditions
(5.2.19) and (5.2.20) is an optimal control problem with two parameters t and h.
Therefore, the exact optimal solution of this problem can be considered as a function of
[ ] h t t + , s while it also depends on the parameters t and h. We denote the exact optimal
control of the problem defined by equations (5.2.28-5.2.20) by ] , [ ), (
,
h t t s s
h t
+ U . In this
study, we will define the value of the active control U(t) at time t as the value of exact
optimal control ] , [ ), (
,
h t t s s
h t
+ U , at point s=t, that is, as ) ( ) (
,
t t
h t
U U = . For this purpose,
we will first obtain the optimality conditions for the problem (5.2.28-5.2.20). For
calculation of the first variation of the functional (5.2.28), we consider the following
increment of this functional:
t h
T T T T
t h t h t h
t
I I s s s s ds
, , ,
( ) ( ) ( )( ) ( ) ( )( ) ( ) ( )
+
+ = + + + +

U U U Z Q Q Z U R R U U
(5.2.22)
where ) (
,
U
h t
is the remainder term given by

t h
T
t h
t
s s ds
T
,
( ) ( ) ( ) ( (
+
= + ) )

U Z Q Z U R U s s (5.2.23)
It is clear that ) (s Z represents the solution of the system

s t
s s s s s t t h ( ) ( ) ( ) , ( )| , [ , ]
=
= + = + Z A Z B U Z 0
(5.2.24)

228
with [ ] h t t s + , PC ) (
r
U . The solution of equation (5.2.24) is
[ ] s s t t h
( )
( ) ( ) , ,
= +
A
Z B U
s
s
t
e d (5.2.25)
The expression (5.2.23) together with (5.2.25) shows that

t h
t h
,
,
( )
0

U
U
(5.2.26)
when 0
,
h t
U . Consequently, the first variation or differential of the functional
(5.2.18) can be written as

t h
T
t h
t
I s s ds
' T
, 0 0
( )( ) ( ) ( ) ( (
+
= + ) )

U U Z Q Z U R U s s (5.2.27)
where
T
0
Q Q Q + = and
T
0
R R R + = . If we substitute equation (5.2.25) into equation
(5.2.27), one obtains ) ( ,
'
U h t I as follows:

t h
T
t h
t
t h
T
0
t
I s ds
(s ) s d
' ( ) T
, 0 0
( ) T
0
( )( ) ( ) ( ) ( (
( (
+ +

= + ) )

= + ) )

A
A
U U Z Q B U U R U
Z Q B U R U
s
s
t
t h
s
e d s s
e d
(5.2.28)
It is obvious that the necessary condition of optimality for the functional (5.2.28) is
[ ]
t h r
I s PC t t h
'
,
( )( ) 0 , ( ) , = + U U U (5.2.29)
The functional (5.2.18) defined for the solutions of the problem given by equations
(5.2.19-5.2.20) is convex in the sense that
[ ]
t h t h t h r
I I I s PC t t h
'
, , ,
( ) ( ) ( )( ) , ( ) , + + + U U U U U U (5.2.30)

229
This shows that the condition (5.2.29) is also sufficient condition for optimality. So, the
necessary and the sufficient condition of optimality of [ ] h t t s + , PC ) (
r
U is
[ ]
T s
0
( ) s s t t h
( ) T
0
( 0 , ,
+ ) = +
A
Z Q B U R
t h
s
e d (5.2.31)
Note that the optimal state vector [ ] h t t s s + , ), ( Z , corresponding to the exact optimal
control ] , [ ), ( ) (
,
h t t s s s
h t
+ = U U , of the problem (5.2.18-5.2.20) can be easily obtained as
[ ] [ ]
s
s t s
s t f d s t t h
( ) ( )
( ) ( ) ( ) ( ) ; ,

= + + +
A A
Z Z B U D
t
e e (5.2.32)
Consequently, optimal state and control vectors ) ( and ) ( s s U Z can be obtained as the
solutions of the system of integral equations (5.2.31-5.2.32) in the interval [ ] h t t s + , . If
the earthquake acceleration f(s) can be estimated for [ ] h t t s + , , then the optimal state and
control vectors ) ( and ) ( s s U Z can be determined as a unique solution of the equations
(5.2.31-5.2.32).
For the purpose of defining the active control, we use the optimality condition (5.2.31), in
which the integral term on the left hand-side is continuous for [ ] h t t s + , . So, equation
(5.2.31) is meaningful in special case when s=t. Then, we have

T t
0
( ) t
( ) T
0
( 0
+ ) =
A
Z Q B U R
t h
t
e d (5.2.33)
Using the trapezoidal rule of numerical integration, equation (5.2.33) can be rewritten as

T T T
0
h
t h t O h
3
0 0
( ) ( ) ( ( ) 0
2
+ + + ) + =

Z Q KB Z Q B U R t (5.2.34)
where
h
e
A
K = and ) (
3
h O denotes the quantity which satisfies the condition
3
0
3
) ( h C h O with a positive constant C
0
. Then, U(t) is obtained easily from equation
(5.2.34) as

T T
h
t t h O h
1 3
0 0 0
( ( ) ( ) ( )
2
) = + + +

U R B Q Z K Q Z t (5.2.35)

230
It is noted here that the real time application of the resulting control scheme (5.2.35)
requires the estimation of the state vector ) ( h t + Z at time t. For the sake of simplicity, it is
also possible to obtain a simple closed-loop form from equation (5.2.35), which does not
require the estimation of ) ( h t + Z . Due to the space limitations, detailed derivation of
optimal U(t) and Z(t) will not be given, only the results will be given here. Ignoring some
terms with order O(h
2
) and O(h
3
) for symmetric matrices Q and R, optimal U(t) and Z(t)
are derived as follows:

T
h t
1 1
0
( ( )

) = U N R B Q Z t (5.2.36)
and

T
h h
t h f t
1
2
1 1
0
( ( ) ( )
2 2

) = + +

Z I BN R B Q y D t (5.2.37)
where
T
h
2
1
0
4
= + N I R B QB
[ ]
Ah Ah
t h e t h h e t h f t h ( ) ( ) ( 2) ( ) ( ) = + + y Z B U D
5.2.8 Conclusions
Exact optimal control resulting from the numerical solution to Euler-Lagrange equations
is an idealized case and cannot be implemented since it requires a priori knowledge of
earthquake in the control interval. A sub-optimal implementable control rule, which uses
the prediction of near-future ground accelerations, is proposed. To evaluate the
performance of the proposed control, the real-time control of 2- degree-of-freedom low-
rise base isolated structure with a semiactive controllable fluid damper under earthquake
excitation is investigated and the results are compared to uncontrolled and exact optimal
control cases.
Optimal semiactive controls which use a priori knowledge of the investigated earthquake
excitation in the control interval always gives the best performance in terms of the
reduction in all structural responses compared to uncontrolled case and the proposed
implementable semiactive control. This indicates that with better nonlinear control

231
methods semiactive control may provide better performance than is achievable by the
proposed causal control rules used in this study.
Proposed implementable semiactive control method uses only the 10-step ahead
predictions of ground accelerations at the control instant to generate the control forces.
However, they show similar behaviour to optimal semiactive control, which uses the
complete knowledge of the excitation and follow the optimal trajectories closely.
The controllable fluid damper used in conjunction with the proposed semiactive control
can decrease the accelerations and interstorey drifts in the superstructure of base isolated
structures. The effectiveness of the proposed control is also shown by the significant
reduction in the investigated integral cost function.
A simple integral type quadratic functional is also introduced as the performance index
for the purpose of suppressing the seismic vibrations of buildings. The optimal active
control force is derived by using the method of calculus of variations based on the
minimization of the proposed performance index in the nearest control interval.
Although the resulting optimal control law depends on the structural displacements and
velocities in the minimization interval, for simplicity in implementing the control system
and evaluating the control effectiveness, a simpler control law which depends on the
current state of the system is obtained. New semiactive control rules may be developed
based on the proposed linear control law.

6.MITIGATION OF HAMMERING BETWEEN BUILDINGS
6.1 INTRODUCTION
Building collision, commonly called pounding, occurs during an earthquake when, due
to their different dynamic characteristics, adjacent buildings vibrate out of phase and the
at-rest separation is insufficient to accommodate their relative motions. Pounding means
an instance of rapid strong pulsation and sometimes, like hammering, repeated heavy
blows. Because building separations in urban areas are often insufficient to preclude
pounding, there is a need for safe and economical retrofitting methods to reduce
structural pounding.
In the past, major earthquakes affecting large metropolitan areas have induced severe
pounding damage. In some cases, the additional forces generated by the impact
interactions have led to collapse structural problem. In other cases, the buildings
presented minor local damages, but indicating that pounding may be a serious threat to
the structures if a stronger earthquake take place. In recent years, research has being done
to study the pounding phenomenon. Pounding has been included in the list of important
areas to be checked during a seismic evaluation, but in general, the engineer does not
have much information on how to evaluate the effects of pounding, nor how to reduce
them.
The overall goal of the project is to develop strategies for retrofitting closely spaced
buildings to reduce the potential for damage due to seismic pounding, while minimizing
the modifications to the existing structural system. Many possibilities of mitigation exist.
All of them will be explained in details but special attention is carried on connection
between buildings, commonly called Pounding Reduction Devices (PRDs).
More detailed development on the Topic can be found as a Lessloss Project Report listed
as Deliverable 46 Analysis of hammering problems (www.lessloss.org ).

234
6.2 ANALYSIS OF POUNDING BETWEEN BUILDINGS AND MITIGATION BY LINKING
INTRODUCTION TO THE NUMERICAL STUDY
6.2.1 Assumptions and limitations
Pounding is a very complex phenomenon. Moreover, an extended number of possible
cases with pounding can appear: pounding between two or more structures, pounding
between structures with different height, different floor levels, and different types of
structures, different material and so.
The main assumptions and limitations made in this study are summarized below.
Buildings are supposed to be regular in plan. When considering two different
structures, their frames are supposed to be in the same plane. The ground motion is
assumed to occur in one direction in the plan of symmetry of the buildings so that
the problem can be reduced to a two-dimensional problem.
In this work, the spatial variations of the ground motion is neglected because the
total plan dimension in the direction of excitation is not large; only two structures are
studied.
There is some evidence of correlation between occurrences of pounding and soft
foundation soil conditions. In this study, the interaction between soil and structure
are neglected. Neglecting soil-structure interactions limits the applicability of the
results to stiff, firm ground and less restrictively to building whose foundations are
not massive (e.g. footing foundations).
The structures under consideration are designed in respect to DCM of Eurocode 2
and Eurocode 8. No brittleness was supposed to be made during the design or the
construction of the structures.
6.2.2 Design of the structures considered in the analysis
The structures studied are presented at Figure 6-1. They are all designed according to the
Eurocodes [EC8, EC2] for a peak ground acceleration of 0,4g, 0,25g, 0,10g. Buildings are
supposed to be on a ground type C which is deep deposits of dense or medium-dense
sand, gravel or stiff clay with thickness from several centimetres to many hundreds of
meters. The floors are made of a reinforced concrete slab and supposed to act as
diaphragms. The floor diaphragms are rigid in-plane; therefore, the impact forces are
distributed to all the structural elements connected to the floor level.

235
MRF 6
(heavier than MRF1)
MRF 1 W 1 MRF 2
MRF 5
(stiffer than MRF1)
MRF 4
T1=0,66 sec
MRF 3
T1=0,916 sec

Figure 6-1. The structures studied (dimensions in meters)
A small time step is needed to ensure the convergence of each point. The program used
is SAP 2000.
Structural members, beams and columns, are idealized as distinct elements and all
inelastic deformations are concentrated at their two ends using the NEHRP
recommendation. This modelling is known as point hinge models. Reinforced concrete
columns have been modelled using the same models identified for beams, except that
axial force variations under the action of earthquake loading are also taken account.
Proportional damping is specified for dynamic analysis cases. A Rayleigh damping matrix
is calculated as a linear combination of the stiffness matrix scaled by a user-specified
coefficient, and the mass matrix scaled by a second user-specified coefficient. The two
coefficients are computed by specifying equivalent fractions of critical modal damping at
two different frequencies.
The collisions between adjacent buildings are simulated by a special contact elements
activated when the bodies come in contact. Other possible way of modelling are
presented at paragraph 6.3.1. The Kelvin model is represented by a linear spring in
parallel with a damper, as shown in Figure 6-2. Contact points are supposed to be located
at each slab of the buildings.

236
m1
kp
cp
m2
u1 u2
gp

Figure 6-2. Traditional Kelvin model
The stiffness of the impact spring is typically large and represents the local structural
stiffness at the point of impact that will react to the shocks during contact. The constant
of the associated dashpot determines the amount of energy dissipated during impact. The
forces in the contact element may be calculated from:
( ) ( )
p p 1 2 p p 1 2
F =k u -u -g +c u -u
1 2 p
if u -u -g 0 (0.33)
p
F =0
1 2 p
if u -u -g <0 (0.34)
Where k
p
and c
p
are the spring and dashpot constants of the element, u
1
and u
2
are the
displacements of the impacting bodies,
1
u and
2
u are the velocities of the impacting
bodies and g
p
is the static separation between the structures.
6.2.3 On the use of linear or nonlinear analysis to evaluate pounding effects
It is unrealistic to expect that a traditionally designed structure will remain entirely elastic
during a major seismic disturbance. Most of structures go in the non elastic domain when
subjected to earthquake. An important question is posed here: is a structure subjected to
pounding can still be adequately represented by an elastic model?
As illustrated in Figure 6-3, for elastic system, when impact occurs, the transfer of energy
from one structure to the other lead to almost unchanged maximum displacement in the
structure with the higher level of energy and an increase displacement in the other one.

237
F
d
Structure with the
higher level of energy
Without
pounding
With
pounding
F
F
F
d
d
d
F
F
F
F
d d
d
d
Without
pounding
With
pounding
Structure with the
lower level of energy
Structure with the
higher level of energy
Structure with the
lower level of energy
Elastic SDOF systems Inelastic SDOF systems

Figure 6-3. Illustration Force-displacement for an elastic and inelastic SDOF systems
The behaviour of the inelastic system is completely different. There is also an exchange of
energy during impact but the most important effect of pounding is the impulse force.
Due to this impulse forces, the structures are pushed at the side opposite to pounding.
Then, they oscillate around this modified position and might not be able to come into
contact again. This phenomenon is illustrated on the right of Figure 6-3.
This simple example demonstrates that, when pounding occurs, a structure entering in
the plastic domain cannot generally be adequately represented by an elastic model. This
conclusion limits the applicability of the previous work made on the elastic domain.
Moreover, it induces that it is very difficult for practical engineers to take into account the
general effect of pounding. If this evaluation cannot been done, how engineers can use
strengthening as a good way of mitigation?
6.3 OVERVIEW OF POUNDING BETWEEN ADJACENT BUILDINGS
6.3.1 Pounding modelling
Pounding between adjacent structures is a very complex phenomenon. The non-linear
deformations are not easily incorporated into the modelling of pounding. Therefore,
idealizations and assumptions have inevitably been used in theoretical models as
mentioned in paragraph 6.2.1.
The collisions between adjacent buildings are simulated either by means of special contact
elements (of the spring-dashpot type) activated when the bodies come in contact or by
applying the impact laws of mechanics for particles (stereomechanical impact), with a

238
coefficient of restitution (C
R
) for plastic impacts. Table 6-1 summarizes the existing
models and their advantages and disadvantages. The first approach, also called piece-wise
impact or simply contact element method, can provide a better approximation to the real
problem, under the condition that appropriate values of the impact element properties
are used. And while these properties are highly uncertain and hence difficult to determine
with accuracy, it turns out that the response is quite insensitive to wide changes in their
values.
Table 6-1. Summary of pounding model
Piece-wise model
Model Advantage Disadvantage
Linear spring
gp
u1
m1
kp
m2
u2

Easily
implemented in
software
Energy loss
cannot be
modelled
Variation:
non-linear
stiffness
ki kf
Fp
m m2

Stay simple
Some energy
dissipation due to
hysteretic
behaviour
Kelvin-Voight
m1
kp
cp
m2
u1 u2
gp

The constant of
the dashpot
determines the
amount of
energy
dissipated
The viscous
element remains
activated when
the structures
trend to separate
Variation:
Impact Kelvin
element

kp
cp
m1 m2
gp

The viscous
element does
not remain
activated when
the structures
trend to separate
Time consuming

239
Piece-wise model
Hertz Contact
( )
3/2
p p 1 2 p
F =k . u -u -g
More realistic
Energy loss
cannot be
modelled easily
implemented in
software
Hertzdamp
model
( ) ( )
3/2
p p 1 2 p p 1 2
F =k u -u -g +c u -u
The constant of
the dashpot
determines the
amount of
energy
dissipated

General
comment
The piece-wise model can provide a better approximation than stereomechanical
to the real problem, under the condition that appropriate values of the impact
element properties are used
Stereomechanical
Model Advantage Disadvantage
Classical formulation of the problem
No longer valid if the
impact duration is large
( )
( )
2 1 2
1 1 R
1 2
m u -u
u =u - 1+C .
m +m

Permanent deformation is accounted
by the coefficient of restitution
Difficult to implement in
commercial software

An extended review of previous studies on impact between structures is presented in
details in the Deliverables 46 Rev (www.lessloss.org).
6.3.2 Pounding effects
The analyses made highlight important effects of pounding. There are summarized below.
6.3.2.1 Case A: Adjacent buildings with equal height and with aligned floor
levels
The conclusion is based on the results found in the literature and on the analyses realized
in this work. Unfortunately, no general trends in the pounding behaviour could be found.
Sometimes the response of the flexible structure increased and sometimes it was the one
of the stiffer structure. Many parameters, as the periods of structures, their masses,
strength and stiffness and the hypothesis made, interact. However, two distinct types of
behaviour can be differentiated:

240
If the structure examined is adjacent to another flexible one then pounding amplifies
the displacement of both structures at the side opposite to pounding. Figure 6-4 Case
A1 shows this type of behaviour.
When the studied structure is adjacent to a stiffer system, this one acts essentially as a
'stopper' that reduces the motion amplitude of its neighbour (Figure 6-4 Case A2).
Undeformed structure B
Deformed shape of
the structure B
without pounding
Deformed shape of the
structure B with pounding
Case A1 Case A2
Undeformed structure A
Deformed shape of the
structure A with pounding

Figure 6-4. Typical type of behaviour for case A (same building height, aligned floors)
Another remarkable result observed is that the flexible structure rebound in the opposite
direction so much that they might not be able to come into contact again. This type of
behaviour is evidently possible only if the structures are able to act in the plastic domain.
As expected, pounding is found to be more critical for highly out-of-phase systems.
If both structures have the same height and the same floor levels, the designers face
generally two problems:
The displacements are increased in the direction opposite to the side of pounding
and this more especially as the stand-off distance decreases. Designers must take care
of this effect especially if the P- sensitivity coefficient is already high without
considering pounding effect.
The shear action effects in both structures increased, which could lead to shear action
effect higher than the shear resistance and generate brittle failure.


241
6.3.2.2 Case B: Adjacent buildings of unequal height and with aligned floor
levels
For the case B, two buildings having different height and aligned floor levels, from
review of literature, observations and modelling, three kinds of interaction between
buildings can be observed:
Interaction 1: The taller structure pushes the smaller one, imposing upon it, in one
direction, its own higher amplitude and longer period motion, increasing the plastic
deformation of the smaller one, while the displacements of the taller remained almost
unchanged. A result of this behaviour is that the plastic deformations of the small
building increase and a plastic drift occurs. This type of response happens when the
shorter structure is "lighter" than the taller one.
Interaction 2: If the lower building is more massive and stronger, the sway of the taller
building is abruptly restricted by the shorter building and it suffers high storey shear
forces above the pounding location. With its lower half restrained from moving in one
direction, the taller building exhibits a whiplash (the smaller structure blocks the move of
the taller one; the free part of the structure moves like a whip) type of behaviour that
could have catastrophic consequences. The shorter and heavier building response under
pounding is generally less than those from the no-pounding case are. However, near the
pounding location, the interstorey drifts and shear action effects increase. These results
suggest that pounding has relatively more of an adverse effect on taller but lighter
building rather than on the shorter but heavier building.
Interaction 3: The last case is a combination of the two previous ones. The storey shear
of the taller building above the pounding level and the storey shear of the shorter
building at the pounding level increase significantly.
6.3.2.3 Case C: Adjacent buildings of similar or different height and with not
aligned floor levels
It is important to mention that the results presented in this study are based in one case
only. However, the lessons of this detail case give already important informations.
Moreover, the effects of pounding in case C can be view as the superposition of a global
effect, similar to case A or B, and a local effect, the destruction of the impacted
column(s). Whatever the stand-off distance is, a column impacted by a slab is destroyed
by the impact forces (peak of shears that cause a brittle failure).

242
6.3.2.4 Conclusions
One point must be kept in mind: where pounding occurs, neglecting its possible
effects lead to non-conservative building design.
6.4 OVERVIEW OF POUNDING MITIGATION BETWEEN ADJACENT BUILDINGS
Several methods have been proposed to avoid pounding induced collapse of buildings.
The methods may be classified according to their approach to the problem of pounding:
Methods to avoid pounding or limit pounding problems. This method is used in
the majority of the building codes: the joints must have a certain minimum width
(seismic gap) to avoid or limit pounding problems. Increasing the stiffness of one or
both buildings can also avoid impact. The last technique consists of building the
primary structure away from the property limits. This solution conducts to damage of
the lateral facades of buildings but not to the structure itself.
Methods to strengthen structures to withstand pounding. These types of
method consist of inserting in building supplemental energy dissipation or conceiving
alternative load paths.
Techniques to improve the behaviour of structures towards pounding effect.
For example, by installing a device between structures, commonly called pounding
Reduction Device (PRD).
The different solution can be used alone or combined. Typically, the reconnection of two
structures can improve the behaviour but does not solve all the problems. In
combination with strengthening, the reconnection can resolve the pounding problem.
6.4.1 The seismic gap
The first work for pounding prevention is to establish a good and reliable estimate of the
minimum gap required for the design earthquake so that pounding between the structures
will not occur. Providing a sufficient gap has been the commonly accepted strategy
adopted by building codes throughout the world. The value of the separation distance
between two structures that is sufficiently large to prevent pounding is known as the
seismic gap or critical gap, g
cr
. Nevertheless, where it is possible for new buildings to
be designed with sufficient gap width, the problem persists in the case of existing
buildings designed under older building codes with considerably smaller gap widths than
those specified in the current codes.
Although building codes call attention to this problem, building designers are often
reluctant to provide the necessary space between buildings to eliminate the problem,

243
principally because the required space would reduce available square footage in the
building being developed. Various authors have studied different methods to evaluate the
seismic gap that are generally less conservative than the prescriptions of the codes.
The types of foundation and soil have a great importance on the displacements observed
in the structures.
Based on the non-linear time history analyses made, few comments on the proposed
method to evaluate the seismic gap are made:
The method proposed in the Eurocode 8 for buildings belonging to the same
property and having the same floor elevations is nearly always unsecured. This means
that pounding occurs and the buildings must accept the effects produced.
The method proposed by Kasai is the least conservative one. Whereas, Penzien
proposes a method giving an always conservative estimate (in the models studied).
Based on the analyses made, the SPD methods are generally not more accurate than
the ABS or SRSS rules.
When the periods of the structures are closed to each other, using the rules of the
Eurocode (0,7.ABS) gives good results, similar to the SRSS method. Whereas, for
different periods, the ABS rule must be used to obtain secure results.
6.4.2 Increasing the stiffness of one or both buildings
Since the gap between two existing buildings usually cannot be increased, increasing the
stiffness of one or both buildings may reduce the seismic deformations to the point
where impact is precluded with the existing gap. Increasing the stiffness of the building
reduces its period and leads to a decrease of its displacement.
6.4.3 Supplemental energy dissipation
Another method to avoid pounding is the use of supplemental energy dissipation devices
in the buildings. Energy dissipation devices are attractive because they improve the
overall behaviour of the structure by increasing its internal damping through the energy
dissipated by the inelastic deformation of these special devices. The installation in
buildings of devices involves distributing them throughout the structures. The seismic
response of a damped building would be similar to that of a conventional building.
Another advantage of using energy dissipaters in buildings is that high damping case
shows prominent in-phase motion of the two buildings in spite of different periods of the
buildings. This occurs, since out-of-phase motion of the buildings caused mainly by free

244
vibration is damped out due to high damping, and in-phase motion closely following the
earthquake excitation history dominates (forced vibration dominates the response).
Indeed, structural damping may be defined as an internal energy absorption characteristic
of a structural system that acts to attenuate free vibration.
6.4.4 Strengthening
Deficient bending and shear capacity of concrete moment frames can be improved by
jacketing with concrete, steel or fibre-reinforced polymers (FRP) in beam and column
joints. Strengthening columns to prevent the formation of storey mechanisms may be
difficult to accomplish effectively. In addition to strengthening the columns, it is
necessary to strengthen the connection between the beams and columns to allow
development of the larger moments.
6.4.5 Alternative load paths
A solution proposed is to provide alternative load paths for the vertical load-resisting
members that may be damaged or destroyed by the impact. These alternative load paths
would include supplementary columns or vertical shoring to support the floor or roof
systems. These supplementary supports would be installed at sufficient distance from the
vulnerable exterior walls or columns to be protected when the existing elements are
damaged. The highest disadvantage of this solution is the loss of space due to the
alternative load paths. So, they must be located to minimize any adverse effect on access,
egress, or functional circulation within the building.
6.4.6 Strong shear wall
These strong shear walls act as "bumper" elements protecting the rest of the building.
Bumpers place at the points of probable contact can help to absorb the blows of
pounding. The impact of the bumpers would, however, still transfer some degree of
impulsive loading to the structures, which was probably not anticipated in the dynamic
design.
6.4.7 Primary structure away from property limits
Due to the high cost of land in metropolitan areas and the small lot sizes in many cities,
many owners want to build on the overall property. A solution to allow that while
avoiding impact on the earthquake primary resisting structure is by installing the primary
structure not on the limit of the property but at a certain distance. When the structures
pound each other, the facing is destructed but the building is still stable. After the

245
earthquake, only the facing must be removed and replaced. If a partitioning is necessary
then the material has to crash. Its destruction can also dissipate energy.
6.4.8 Reconnection
One method of eliminating dynamic contact between closely spaced buildings is by a
connection. The basic motivation for such connection is to provide a relatively simple
and cost effective modification to prevent the buildings from hitting without introducing
any complexities to the dynamic response or changing the basic characteristics of the
building dynamics.
6.5 RECOMMENDATIONS FOR THE MITIGATION OF POUNDING PROBLEMS
BETWEEN ADJACENT BUILDINGS
The basic idea developed in this work to eliminate dynamic contact between closely
spaced buildings is by a connection which maintains a minimum separation distance at
certain point of the structures. A guidance based on the review of literature and on the
models done is presented below for engineers in charge of solving that type of structural
problems.
6.5.1 Guidance to mitigate pounding with a PRDs
The basic idea developed in this work to eliminate or reduce dynamic contact between
closely spaced buildings is by using a connection that maintains a minimum separation
distance at certain point of the structures. The tuned reconnection is envisaged in
different ways; it can be made of shock absorbing material placed in the gap, hinge-ended
beams, or joint dampers. In some case, installing dampers or damping material in the
joint is found to contribute to a solution.
As a result of this study, simplified design guidelines have been established for retrofitting
such buildings. A practical guide is presented below to assist engineers in the choice of
the appropriate type and properties of the reconnection in view of the characteristics of
the structures submitted to pounding and of their stand-off distances.


246
6.5.1.1 Adjacent buildings of equal height, with aligned floor levels and similar
structural types, in particular their stiffness
Recommended intervention:
Using hinge-ended elastic beams as links between structures. For the number and
location see 6.5.2.3.
Comment:
The idea behind the use of permanent linkages is to provide forces to the structures that
are continuous and following the height of the buildings. These forces are in-line with the
dynamic behaviour of the unlinked frames. The main advantages of these links are the
following :
they prevent the two structures from oscillating out of unison ;
the forces through the connections are small (due to similar dynamic properties). The
requirements on the seismic resistance of the structures are not increased.
Effects:
This system is very effective in some situations, but has some inherent disadvantages.
This connection modifies the dynamic behaviour of one or both structures and possibly
for the worse. This is particularly important since the coupling is always in effect, even
under mild shaking which would not cause pounding of the unconnected structures.
Moreover, if the structures are of an asymmetrical geometry, the linkage could increase
undesirable torsional response.
the hinge-ended beams force the structures to behave as if it was only one:
- the shear action effects V
Ed
tend to decrease in the flexible structure and increase
in the stiffer one,
- the displacements of the stiffer structure increase. A check of the damage
limitation and the P- effect needs to be performed,
the accelerations peaks observed disappear.
Links properties:
the stiffness of the links k
c
must be sufficiently high to preclude pounding, but not
too high, to avoid creating too high restraint forces. A starting point of the design of
connection can be the stiffness of the building, K, evaluated by a concentred forces
applied at the top storey k
c
=K ;

247
an important point to note, and mentioned by Plumier et al. [2005], is that in order to
be able to predict the response (of the connections and the structures connected) it is
desirable to maintain elastic response in the linkage ;
it is also necessary to place a system that will avoid the buckling of the bars used to
connect the structure ;
if the stand-off distance between the buildings is very small, then the stiffness of the
links needs to be higher if in order to preclude pounding. Another possible solution is
that the links can have different properties in the compression or tension side.
Other methods:
Other possible methods to avoid or limit pounding problems are summarized at
paragraph 6.5.
6.5.1.2 Adjacent buildings of equal height, with aligned floor levels and different
structural types
Using dampers and springs as links between structures. For the number and location see
6.5.2.3.
Effects:
The advantage of non-elastic links, beside the possible dissipation of energy, is that it
limits the link forces to levels that are within the range of the existing lateral load of the
adjacent buildings. The use of a spring in addition to the damper prevents the structures
from pounding each other.
Links properties:
It is difficult to define adequate properties for the links. An efficient method found is
proposed in Zhu and Xu [2005] and in Zhu and Iemura [2000] in which analytical
formulas for determining optimum parameters of passive dampers are presented. The
recommanded properties are based on the modelling of each structure as a single-degree-
of-freedom system. A starting point proposed here in the definition of the properties to
use is the computed value found, proposed by those previous researchers, divided by the
number of storeys where the connectors are placed. The choice of the connected storeys
is made by keeping in mind that

248
the upper storeys are more efficient to dissipate energy, due to the higher velocities at
those levels;
the building is kept in service during the retrofitting;
placing the reconnections only where it is necessary allow to make economy.
If the damper stiffness is above an optimal value, its effectiveness deteriorates rapidly.
Indeed, strong damper stiffness reduces the relative velocity and hence the energy
absorbing capacity from the dampers decreases. When the damper stiffness is very high,
the relative displacement and velocity between the adjacent buildings tends to zero so
that the two buildings behave as almost as rigidly connected. As a result, no matter what
value the damper coefficient is the damper totally loses its effectiveness. With the
decrease of damping coefficient from the optimal value, the performance of the damper
deteriorates gradually and as the damping coefficient approaches to zero the two
buildings finally return to the unlinked situation. On the other hand, if the damping
coefficient is above the optimum value, the performance of the damper also declines and
as the damping coefficient becomes very large, therefore the two buildings behave as
though almost rigidly connected.
Special comment:
When the stand-off distance is small, then the stiffness of the links must be substantially
high to prevent pounding.
Other methods:
Other possible methods to avoid or limit pounding problems are summarized at
paragraph 6.4.
6.5.1.3 Adjacent buildings of unequal height, with aligned floor levels and same
structural types
Using joint dampers and springs as links between structures. For the number and the


249
Comment:
Various ways of mitigation have been tested but none of them can completely solve the
pounding problems. Using joint damper links reduce the shear action effects in both
structures. The retrofitting should combine dissipative links with local reinforcement
(Figure 6-5), essentially on ductility of beams, to solve the pounding problem. Concrete,
steel or composite-materials jacket retrofits can enhance flexural ductility.
Reconnection
Increasing the
ductility in
the beams of
this zone

Figure 6-5. Proposed mitigation for case B
Effects:
The joint dampers dissipate energy and can prevent any impact if they have a sufficiently
high stiffness. As the lower part of the tall structure is constrained by the linking to the
smaller building, the seismic action effect can be seriously increased in the upper storeys
of the tall structure and special attention to the formation of plastic hinges in that one is
needed.
Links properties:
As in the paragraph 6.5.1.3, it is suggested to use connectors that can dissipate energy
based on Zhu and Xu [2005] or Zhu and Iemura [2000] in combination with local
reinforcement in one or both structures.
Other methods: Other possible methods to avoid or limit pounding problems are
summarized at paragraph 6.4.

250
6.5.1.4 Adjacent buildings of unequal height, with aligned floor levels and
different structural types
None of the links studied does help. Other type of mitigation must be used to avoid
pounding.
Comment:
Connecting the structures does not improve at all the behaviour, because it imposes
action effects for which the buildings have not been designed.
The alternative can be:
For a new building:
- To separate the buildings with an adequate distance to avoid impact.
- To increase the stiffness of one or both buildings.
- To build the primary structure away from the property limits. This solution leads
to the degradation of the faade, which needs to be replaced but prevents the damage
of the buildings.
For an existing building:
- To insert supplemental energy dissipation.
- To strengthen the structures.
- To use an alternative load path.
All these methods are presented at paragraph 6.4.


251
6.5.1.5 Adjacent buildings of similar or different height, with not aligned floor
levels and similar or different structural types
Two solutions are proposed. The choice depends on the situation met.
The reconnections link each storey level. A reinforced columns connected with a link
to the adjacent floor level
Connector

Connector
Reinforced
column

Commentary:
The connectors must be placed to transfer
the action effects directly from floor to
floor in order to prevent any impact within
each individual column length.
Commentary:
Strengthen columns of one building in
order to increase significantly their shear
and bending resistance. Connect the floor
levels of the other building to the
strengthened columns by means of links.
Effects:
This situation superposes one of the situations described in 6.5.1.1, 0, 6.5.1.3 or 6.5.1.4
to a local potential impact between a slab and a column.
Linking two structures prevent them to pound each other but may also force them to
oscillate together. The check of this new type of behaviour is absolutely needed.
Properties to use:
In comparison to the proposed value used
in 6.5.1.1 and 0, the stiffness needed is
higher due to the inclination of the
connectors.
Properties to use:
The proposed solution is to use the links
proposed in case 6.5.1.1, 0, 6.5.1.3 or
6.5.1.4 (according to the buildings
characteristics).

252
6.5.1.6 Buildings with a small seating length (unseating problems)
Recommended intervention
Restrainers (elastic spring) or non-linear viscous dampers. For the number and the
Comment:
Special attention is needed in the connection design: capacity design of anchorages to
avoid brittle failure.
Properties to use:
The proposed linking is essentially based on the study made for unseating in decks of
bridges. To design strength and stiffness, 2 methods are proposed:
1. Non-elastic dynamic time-history analysis where the strength and stiffness of the
restrainers can be varied until acceptable results are obtained.
2. The second method is based on the hypothesis that the connected frames
respond essentially in-phase. This observation has for consequences that the
maximum tensile force is equal to the difference between the frame overstrength
longitudinal shear capacities.
Effects:
The restrainers reduce the relative displacements between the two parts of the same
structure.
Special comment:
Where locking or reducing the relative move of buildings at joints is undesirable or
impracticable, another solution is to extend the effective seating length of the expansion
joint. The simplest way to increase the support length is by adding corbels or brackets
(Figure 6-6).

253
Concrete Block
Concrete cover
removed
Anchor bar
Injection of
resin

Figure 6-6. Corbels and brackets for bearing supported superstructures
6.5.2 Some practical indications on the design of Pounding Reduction Devices
(PRD's)Choice of the PRD
6.5.2.1 Introduction
Comments are made in this part to describe the problems raised by the installation of
linking devices. The devices are to be placed at some or each storey.
Several criteria need to be examined for screening the most promising PRD. There can
be difficulties in placing devices between two buildings.
6.5.2.2 Requirements in selecting PRD's
Paragraph 6.5.1 defines the PRD's to install for various situations of adjacent buildings.
Several criteria need to be examined to define the PRD type
their ability to sustain large force levels and dissipate large quantities of energy over
short displacements;
their ability to sustain high strain rate;
their ability to sustain many cycles of loading without degradation of mechanical
properties;
the need for predictable and stable mechanical properties over the range of possible
loading amplitudes, displacements and frequencies are needed;

254
the possibility to tune the mechanical properties of the device;
their resistance to weather (if not protected);
their initial and maintenance cost.
A device must comply with its prescribed hysteresis law during the seismic event for
which it has been designed. The device must follow its intended behaviour during the
whole installation period, where it might be subjected to a variety of static and dynamic
loads. Permanent changes of the basic mechanism of the device may occur with time (e.g.
due to creep, crack development,), however, these changes should not alter the
hysteresis law of the device during its life cycle in a substantial manner.
An important point for the effectiveness of the devices is their anchorage. To develop the
predicted forces and maybe dissipate energy, the weak point of the devices should not be
anchoring. Moreover, the peak link force has to be limited to values compatible with the
existing structural system if the installation of the devices is realised without upgrading
the existing lateral load carrying systems of the buildings.
6.5.2.3 Number and location of the devices
Intuitively one would conclude that placing PRD at the bottom of the buildings would
not be necessary. And actually, Zhang and Xu [1999] found that the sensitivity of the
modal frequency and the modal damping ratio of the combined system to the dampers
are very small near the bottom of the buildings. This result indicates that there is no need
to install viscoelastic dampers near the bottom of two 20-storey buildings. Ni et al. [2001]
showed that non-linear hysteretic dampers are effective even if they are placed only on
few floor levels. Bhaskararao and Jangid [2005] studied the effect of insert a friction
damper. They found that it is not necessary to connect two adjacent structures at all
floors but lesser dampers at appropriate locations can significantly reduce the earthquake
response of the combined system. The same conclusions were drawn and applied in a
practical retrofit by Plumier et al. [2005]. The number of reconnected storey can be
reduced but then the magnitudes of the reconnection forces at one storey increases and is
more difficult to transmit.
Although placing only few connectors is appealing, designers have to make sure that
connectors are able to sustain the design forces.
It is necessary to install a system that will prevent the buckling of the bars used to
connect the structure.


255
6.5.3 Models and programs for impact zone
The collisions between adjacent buildings are simulated either by means of special contact
elements (of the spring-dashpot type) activated when the bodies come in contact or by
applying the impact laws of mechanics for particles (stereomechanical impact), with a
coefficient of restitution (C
R
) for plastic impacts.
Both methods express the fact that, during contact of structures, a transfer of energy is
produced. Moreover, an impulse force is transmitted to each structure. This force
generates non-elastic structures to rebound in the opposite direction of pounding. The
effects of pounding are function of the level of energy in each structure when contact
occurs. This is why the use of a gap or the magnitude of the velocities at the moment of
impact should be known in order to model adequately the effects of pounding.
The ability to model either piece-wise or stereomechanical methods is limited to a small
number of programs. This complex modelling of pounding may force designers to use
ad-hoc programming systems with a possible loose of time-efficiency and an increased
risk for errors. Some programming systems offering a possibility to model pounding
phenomenon are :
SAP2000 is integrated software for structural analysis and design; which can be used
as explained in chapter 6.2. This program has the ability to model pounding effects by
the piece-wise method.
DRAIN-2DX is a static and dynamic analysis of non-elastic plane structures. The
element library contains a non-elastic link element that can act in
compression/tension with initial gap or axial force. This element can be used to
model pounding effects by the piece-wise method. For example, this program is used
in the study of pounding by Karayannis at al. [1998].
ANSR-1 is a general purpose program for analysis of non-linear structural response.
It is a non linear program for 3-D structures with a gap-friction element. This element
can be used to model pounding effects by the piece-wise method.
SLAM and SLAM-2, developed by Maison and Kasai [1988], are two micro-
computer pounding analysis programs. These programs contain gap element, which
have the ability to model piece-wise contact.
IDARC, developed at the State University of New York at Buffalo, was first
introduced in 1987 for analyzing earthquake damage in multi-storey, reinforced
concrete buildings. Since then, numerous enhancements have been added, including
the ability to analyze a wide variety of structures, structural materials, and, most
recently, structural damping devices. This program like the entire previous one has
gap element to model piece-wise contact.

256
CASTEM 2000 is a computer code for the analysis of structures by the finite element
method. This code was developed by the Mechanical Department and Technology
(DMT) of the French Police station with Atomic Energy (ECA). The gap element
present in the program permits the modelling of pieces-wise contact.
SEISMOSTRUCT is a Finite Element package capable of predicting the large
displacement behaviour of space frames under static or dynamic loading, taking into
account both geometric non-linearities and material non-elasticity. Concrete and steel
material models are available, together with a large library of 3D elements that may be
used with a wide variety of pre-defined steel, concrete and composite section
configurations. The gap element present in the program permits the modelling of
pieces-wise contact.
6.6 CONCLUSIONS
Analyses of pounding have shown that there is considerable scatter in the amplification
of action effects caused by pounding. This study clearly shows the sensitivity of the
system response to parameters affecting the pounding phenomenon, i.e. characteristics of
buildings, plan layout, structural system and frequency content of the input ground
motions. The results depend on the excitation characteristics and the relationship
between the buildings fundamental period.
The impulse found when pounding occurs increases suddenly the acceleration and the
velocity. These accelerations generated by the impacts may cause significant damage to
the structural components, especially in the contact area of pounding. Non-structural
components (electrical/mechanical units and architectural features) in some buildings are
important to the building's function.
As expected, pounding is found to be more critical for highly out-of-phase systems. For
pounding occurring between adjacent buildings having very different dynamic properties,
the flexible structure is abruptly stopped by the adjacent one. The pounding phenomenon
tends to increase the displacement at the side opposite to pounding. Consequently, the
gravity and the subsequent P- effect may cause the collapse of the building. Another
problem that could appear due to this increase in displacements is the damage limitation
requirement. The interstorey drifts could not fulfil anymore the criterion. Even if this
criterion is not as stringent as the P- effect, it might correspond to damage to non-
structural elements.
For structures having different height but aligned floors levels, the damage is typically
concentrated at the roof level of the shorter building and at the level just above pounding
for the taller one. The pounding produced by the adjacent short structure causes the
shear failure of the columns in the upper storeys of the tall structure and this for each
stand-off distance considered. In addition to this probable shear failure, plastic hinges are

257
formed at the extremities of the upper columns. These hinges conduct to an increase of
the interstorey drift and subsequently to problems with the P- effects or to damages to
secondary elements. The worst case of pounding occurs when columns are impacted by
slabs. Whereas when considering impact between two slabs pounding can sometimes be
allowed, possible impact between a slab and a column must be avoided. The shear action
effect due to impact in columns is so high that it destroys the element.
This study has established mitigation solutions for the typical problems of pounding
between adjacent structures. Each case is different and requires an appropriate way of
mitigation. The only method proposed in codes to mitigate pounding is by providing
sufficient separation between structures. Various other ways of pounding mitigation
exists and are presented in details in Warnotte [2]. The proposed methods can imply one
or both structures. They can preclude impact or strengthen structures to withstand
pounding effects. The proposed mitigation studied in this work is by linking adjacent
structures.
As a result of this study, simplified design guidelines are established for retrofitting
buildings. The proposed simplified design guideline can be adopted by practicing
engineers and can be used for preliminary design of dampers for retrofitting buildings.
This guide indicates mitigation methods which should be successful, with justifications.
Pounding Reduction Devices are efficient to mitigate pounding effects for adjacent
structures having same total height. For example, models realized in this work show that
using hinge-ended beams as connectors between similar structures can mitigate pounding
by preventing them to oscillate out of unison. The connections force the structures to
behave as only one structure, which is an intermediate between the two buildings. This
method penalizes one of the structures in favour of the other by transferring seismic
demands between structures; it might exceed the capacity of the existing lateral system in
any of them and require additional retrofits to the lateral resisting system. For adjacent
structures with the same total height but with different dynamic characteristics, using an
appropriate viscous damper in addition of a spring can mitigate pounding effects. On the
contrary, for adjacent buildings of unequal height, connecting the structures does not
improve at all the behaviour, because it imposes action effects which are surely different
of those for which the buildings were designed. However, linking adjacent structures can
be a good and reliable way of mitigation if used in combination with local reinforcement.

7.METHODOLOGY OF ANALYSIS FOR UNDERGROUND
STRUCTURES IN SOFT SOILS
In this part of the report the seismic behaviour of large underground reinforced concrete
structures in soft soils is analysed. It is shown that code procedures for structures that
develop above ground are inadequate and may lead to the design of unsafe structures. A
new design methodology and conception criteria are proposed for these structures. A
practical example is shown.
7.1 HISTORICAL BACKGROUND
The analysis of damage in large underground structures due to large earthquakes shows
that these structures are in general less vulnerable than structures that develop mainly
above ground (Gomes, 1999, Hashash et al. 2000). However, recent events have shown
that these types of structures may also be vulnerable to earthquake actions. During the
Hyogoken-Nanbu earthquake, that in January 1995 hit the town of Kobe in Japan, 6 out
21 tube stations were strongly damaged (Iwatate, 2000). Figure 7-1 shows the example of
Dakai tube station, in which collapse was triggered by rupture of the central row of
columns. Iwatate et al (2000) attributed this to the large horizontal displacement field
imposed to the structure by the surrounding soil in the transverse direction.

Figure 7-1. Collapse of Dakai tube station


260
7.2 ANALYSIS OF THE SEISMIC VULNERABILITY
Under earthquake actions the deformation of underground structures, such as the tube
stations previously mentioned, is essentially conditioned by the surrounding soil, as the
inertia forces in the soil are much larger than the inertia forces in the structure. Therefore,
the dynamic behaviour of the soil/structure system is essentially controlled by the mass
and stiffness of the soil. For structures embedded in competent soil, the soil
deformations are small and the structures present reduced seismic vulnerability. If the soil
is soft, the soil horizontal displacements can be large. In these situations the seismic
vulnerability of the structures increase and may even lead to collapse if the displacements
imposed by the soil exceed the structure deformation capacity. The seismic behaviour of
these structures is also influenced by their shape in plan. In tube stations of
approximately rectangular shape one dimension is generally larger than the other and two
types of vertical cuts can be distinguished: (i) the ones designated as rigid alignments, that
are close or that contain very stiff elements, such as perimeter walls in their own plan, and
oppose significant resistance to the soil movement and therefore undergo very reduced
displacements, and (ii) the ones away from the zone of influence of the rigid alignments
and that undergo horizontal displacements similar to the soil displacements in the free-
field, designated as flexible alignments. Figure 7-2 helps to distinguish flexible from rigid
alignments.

Figure 7-2. Schematic representation of rigid and flexible alignments
The observation of the seismic behaviour of Kobe tube stations indicates that collapse
was triggered by the columns of the flexible alignments, unable to withstand
simultaneously the permanent loads and the displacements imposed by the surrounding
soil.
surface
soft
soil
structure
stiff soil
R
R
F
structure
PLAN
VERTICAL CUT
R rigid alignment
F flexible alignment

261
7.3 SEISMIC BEHAVIOUR
The seismic performance of an underground structure depends essentially of the ability
of the flexible alignments to withstand the imposed displacements. Therefore, an
adequate performance can be enforced by means of the following strategies: (i) to treat
the soil to control the respective horizontal displacement field along the height during the
design earthquake, or (ii) design the structure to withstand the displacements imposed by
the soil, while maintaining the ability to resist to the permanent loads.
It results from the above that the design of the structure (always referring to the flexible
alignments) aims essentially at providing deformation capacity while it maintains the
ability to sustain the permanent loads. Current code procedures, essentially derived for
structures that develop mainly above ground, assume that the seismic performance of a
structure is a combination of its ability to resist to inertia forces and its ductility and
energy dissipation capacity. Since engineers are essentially used to design structures to
resist to applied forces, code procedures are usually based on the explicit evaluation of
the effects of the inertia forces, the effects of ductility and energy dissipation capacity
being accounted for approximately by means of a global factor (q factor in EC 8). The
application of this procedure to underground structures is inadequate since these do not
need to resist horizontal inertia forces (except for some minor local effects, usually
irrelevant) which can be transferred directly to the soil on the sides of the structure and
do not need to be transferred to the foundation. Within the usual code framework this
would be equivalent to consider the behaviour factor infinite. This highlights the
inadequacy of applying code procedures for structures that develop above ground to
underground structures. This derives from the fact that providing resistance to horizontal
displacements is qualitatively different from the resistance to horizontal forces. The
difference between applying displacements and forces can be illustrated in terms of a
reinforced concrete section in bending, considering two situations: (i) if a bending
moment is applied, the higher the area of flexural reinforcement, the lower will be the
respective stresses, the strains and the curvature, which are output of section analysis, (ii)
if a curvature is applied the higher the area of reinforcement the higher is the associated
bending moment. Note that being the curvature applied by an external source the strains
can be evaluated as a function of section geometry, this is, do not depend on the amount
of flexural reinforcement (assuming that the position of the neutral axis does not change
much with the amount of flexural reinforcement, as generally happens in flexure without
axial force).
Another feature that results from the above is that adding flexural reinforcement is
useless to prevent yielding, as this depends on the fact that the imposed strain reaches the
steel yield strain or not. This clearly contradicts current code concepts according to which
a structure can be designed to remain elastic under earthquake actions by designing it

262
with q=1. This is valid in general in structures that develop above ground, but not for
structures or elements under applied deformations. Reasoning again at section level, for a
given imposed curvature an elastic analysis yields a bending moment. If the amount of
flexural reinforcement necessary to resist to that moment (equivalent to consider q=1) or
slightly more is provided, the section should remain elastic. The contradiction can be
explained with the help of Figure 7-3, as follows: point 1 represents the yield point of a
section in bending without axial force. If for instances the flexural reinforcement was
duplicated maintaining the same distribution, the flexural capacity would increase in such
a way that, if the flexural stiffness was constant as assumed in linear analysis, the new
situation would be represented by point 2. In fact what happens is that the flexural
capacity increases but the curvature almost does not increase (as the neutral axis remains
in the same position and the steel yield strain does not change) and the point
representative of the new situation is point 2 and not 2. Therefore, the increase in the
flexural capacity does not increase the yield curvature and does not avoid yielding.

Figure 7-3. Schematic representation of change in the yield moment and curvature by increasing
flexural reinforcement (N=0)
7.4 CONCEPTION
The conception (in the usual sense of defining the geometry of the structure and its
elements) of large underground reinforced concrete structures in soft soils to withstand
earthquake actions must aim at ensuring that flexible alignments are provided with the
required deformation capacity under horizontal displacement fields without losing the
ability to sustain the permanent loads. For this purpose along the flexible alignments the
structure must be as flexible and ductile as possible, within the limits associated to the
need to resist to all other actions.
Under imposed deformations the larger the cross section dimensions the higher will be
the strains imposed both on steel and concrete. Therefore the dimensions of structural
elements of the main resisting structure (elements whose collapse would induce
irreparable damage or global collapse) in the plan of flexible alignments should be as
M
1
2 2
EI
sec
1
EI
sec
2

263
reduced as possible, in practical terms should be the ones strictly necessary to resist to
permanent and other actions but the earthquake action. Therefore the use of counterforts
or other rigid elements, such as short beams, should be avoided. Besides, rigid elements
tend to generate higher shear forces what makes them more vulnerable to seismic actions,
as shear tends to reduce the available ductility, an effect that is magnified by the cyclic
nature of the load history. In order to minimize section dimensions both concrete and
steel of higher strength should be used. In the case of elongated tube stations in which
one direction corresponds to a flexible alignment and the perpendicular direction to a
rigid alignment, rectangular sections with the lowest dimension in the direction of flexible
alignment are recommended.
In order to maximize the ductility it is necessary to minimize compressive axial forces in
elements where yielding can be expected. Therefore in order to minimize axial forces in
the columns it is recommended that large soil covers on top of the structures are avoided.
Exceptions to this recommendation may be justified in some cases in which this could
lead to higher imposed deformations due to the insertion of the structure in superficial
soil layers of worse characteristics.
Secondary structural elements (whose collapse yields repairable damage), such as stairs, or
others, and non structural elements, such as masonry partition walls, may have a negative
effect by restricting the deformation of the main structure. Therefore their geometry and
location should be chosen avoiding these types of interferences. For instances stairs
should develop preferably in the plan of rigid alignments and should not be supported at
middle height of the main columns.
7.5 DESIGN METHODOLOGY
The evaluation of the deformations imposed to the structure requires the analysis of the
soil/structure system under the design earthquake action. The overall dynamic behaviour
depends essentially on the properties of the soil, which usually presents stiffness and
damping that are highly dependent on the level of distortion. Therefore a common
methodology consists of analysing first the soil alone and estimates a value of an
equivalent damping coefficient and distortion stiffness for the expected distortion
amplitudes under the design seismic action. On a second stage this properties are used as
input for the linear analysis of a soil/structure model. The simulation of the structure
assuming linear behaviour is an approximation with little influence on the result, since the
dynamic behaviour of the soil structure system depends essentially on the soil properties.
The deformations obtained from this analysis can then be imposed on a structure model
to evaluate its effects.

264
The analysis of the deformation capacity of ductile structures involves the analysis of the
behaviour in the post-yield range, this is, a physically non-linear analysis. For this
purpose, it is necessary to know the yield capacity of structural elements and sections,
therefore global structural analysis requires not only the knowledge of the geometry of
the structure and its elements but also knowledge of amounts and details of
reinforcement. These are necessary to allow the definition of confined concrete
constitutive relationships and evaluation of yielding and rupture. The analysis considers
flexural and axial deformations but not shear deformations, which are not relevant in well
conceived structures. Therefore, the analysis requires as input the definition of the
geometry of the structure and its elements and the explicit definition of the flexural
reinforcement, including the full monotonic constitutive relationship until rupture. The
transverse reinforcement does not need to be defined explicitly, but knowledge of
amounts and details of this reinforcement are necessary to evaluate confining stresses and
the constitutive relationships for confined concrete. The detailed characterization of the
structure allows performing the global structural analysis and the evaluation of axial
stresses and strains anywhere in the structure simultaneously. Thus, it allows doing safety
verifications at material level by comparing the maximum strain demands with the
corresponding acceptable limits.
Global analysis of the structures, linear or nonlinear, are usually performed assuming
average material properties. However, safety checkings at element and section level are
usually based on design values of material properties to account for the possibility that at
some locations in the structure the materials properties are worse than average. In order
to follow this analysis and safety verification methodology it is necessary to decouple the
global structural analysis from the section or element analysis using separate models.
However, since there is an interest in performing both the global analysis and safety
verifications with the same model, it is necessary to make a choice of what material
properties to use. Since the action upon the structure is represented by imposed
displacements, the evaluation of curvatures is essentially a cinematic problem whose
result is almost not influenced by material constitutive relationships. Therefore, it is not
relevant what material properties are used in the global analysis. Since design values
should be used for safety verifications at section or material level, these properties will be
used in the analyses that will be presented in the next sections.
Current design practice is usually based on a procedure with the following phases: (i)
conception of the structure, (ii) global structural analysis based on a constant stiffness of
all structural elements, (iii) safety verification at section or material level. In reinforced
concrete structures, the third phase is transformed in the calculation of the amounts of
reinforcement necessary to ensure the prescribed safety verification. In the methodology
that is proposed in this work the third phase is a really verification phase, as the
reinforcement needs to be known before the analysis and safety checkings are done, by

265
comparing strain demands with the corresponding limits. As the structure has to
withstand all other actions, the first phase of design must be the design of the structure,
including the calculation of the necessary amounts of reinforcements to resist all load
combinations in which the seismic action is not the main variable action. The next phase
is the increase of the amounts of reinforcement to increase the ductility of the structure.
This can be considered a second conception phase. The third phase is the analysis and
safety verification. The second and third phases may need to be repeated, if the first
verification does not yield suitable results, yielding an iterative procedure.
7.6 PRACTICAL APPLICATION EXAMPLE
In this section the application of the proposed methodology to an underground structure
with appropriate conception (in terms of geometry and dimensions) is shown,
complemented by the presentation of criteria for the conception of the reinforcement
added (to what is necessary to resist to other actions) to increase its ductility. The
application of the proposed methodology is also compared with code procedures for
structures that develop above ground, namely EC 8 Part 1, both in what regards seismic
performance and economy.
The geometry of the example structure, with a conception considered adequate, is shown
in Figure 7-4. A reduced width of 3.80m of the exterior walls was used in the calculations.
The materials chosen are steel A500 and concrete C35/45.

Figure 7-4. Example underground structure
9m 9m
5m
6m
5m
6.5m
Soft
Soil
Stiff
soil
Top slab - thickness: 1.20m
Bottom slab thickness: 2.00m
Perimeter walls thickness: 1.20m
Columns: 0.7 x (1.4) m
2
Beams: 0.9 x (1.4) m
2
Distance between flexible alignments:
7.0 m
The first dimension of beams and
columns is the one on the plan of the
flexible alignments

266
The seismic action can be simulated by means of applying to the structure horizontal
displacement fields with the profiles represented in Figure 7-5. The profile shown in
Figure 7-5 a intends to represent the effect of a soil with increased stiffness with depth.
The profile shown in Figure 7-5 b consists of a sinusoidal variation of the displacements
along the height, and corresponds to the first mode shape of a soil with constant stiffness
along the height. However, it is not uncommon to find strong variations of soil stiffness
along the height, for instances due to the existence of more than one soil layer. This can
be simulated by a displacement profile as shown in Figure 7-5 c, in which the
deformations are concentrated at an intermediate soft soil layer. The examples shown
next are based on the linear profile; the effects of the other profiles are discussed only
qualitatively. The maximum distortion
max
is used as a measure of the deformation
capacity of the structure.

Figure 7-5. Horizontal displacement profiles
7.6.1 Structure designed according to current code concepts
Following current code procedures, seismic action-effects are obtained dividing the
results of elastic analysis by a behaviour factor (q-factor in EC 8), a procedure that will be
designated as Direct Design. Since EC8 does not cover this type of structures, an
extrapolation of Part 1 will be made, as this is the most likely procedure designers will
adopt. EC 8 Part 1 (referred to as EC 8, from now onwards) considers three main
Ductility Classes in seismic design: Low, Medium and High. Ductility Class Low
structures are designed to resist earthquake effects essentially in the linear range and no
procedures are applied to increase ductility. EC 8 prescribes a q-factor of 1.5 for this type

H /
max
=

A
H
H1
H2
H3
Stiff soil
Stiff soil
Soft soil
B
/
2
max
=
2 /
max
H =
a) Linear b) Sinusoidal c) Concentrated

267
of structures to account for some levels of overstrength that is assumed is always
available in reinforced concrete structures. Structures of Ductility Classes Medium and
High are designed to resist earthquake actions by a combination of their resistance to
inertia forces with their ductility and energy dissipation capacity. This represents an
intermediate type of design between the one associated with Ductility Class Low and the
proposed methodology. Therefore, to highlight the differences to the proposed
methodology the example structure is designed as a Low Ductility Class structure.
Since in the framework of Direct Design applied displacements result in internal action-
effects on the structure (bending moments, shear and axial forces), the maximum
displacement the structure can withstand is restricted by the maximum amounts of
reinforcement that is possible to place in any structural member. Assuming q=1.5 and
that the constant member stiffness assumed in the elastic analysis is half the stiffness of
the gross concrete sections as prescribed in EC 8, the maximum allowable distortion
associated with the linear profile of imposed displacements is
max
=8.2x10
-3
. The
reinforcement corresponding to this distortion is shown in Figure 7-6.
The explicit evaluation of the deformation capacity of this structure was evaluated by
means of a static nonlinear analysis imposing the permanent loads and the linear
displacement profile. It is assumed that proper detailing ensures the anchorage and
effectiveness of all reinforcement, in particular confinement reinforcement after spalling
of the concrete cover. The deformability of the nodes and shear deformations were
disregarded, only flexural deformations were accounted for. The nonlinear behaviour of
concrete and steel were simulated using the constitutive relationships for confined
concrete prescribed in EC 8 Part 2 and constitutive relationships for steel obtained
from a large statistical characterization of the Tempcore steels used in Europe (Pipa,
1993). Figure 7-11 shows the constitutive relationship for steel and an example of
constitutive relationships for confined concrete. Rupture was defined by the attainment
of the maximum axial strain anywhere in the structure. The maximum allowable strain for
steel is
max
=7.5%, corresponding to steel type C and for concrete it depends on the level
of confinement, according to the equation prescribed in Annex C of EC 8 Part 2.
The results of this analysis indicate that the maximum average distortion that the
structure can withstand is
max
=5.0x10
-3
. Figure 7-7 shows the curvature diagrams at this
situation, indicating the yield curvature at some sections and showing that flexural
yielding took place at several locations. The maximum tensile strain is =27.8 , 13
times the steel yield strain (=2.07 ). Note that at this stage the maximum distortion
was 60% of the distortion evaluated according to EC8 (Low Ductility Class), at which the
structure was supposed to be elastic. If the sinusoidal profile had been applied the
ductility demand would be higher at the lower part of the structure and it would
withstand a lower relative displacement () between the top and bottom slabs.

268
The above results show that the design with a low behaviour factor does not prevent
yielding if the action is an imposed displacement field, contradicting widely held views
and basic concepts of current code prescriptions for seismic design of structures that
develop above ground. It also shows that extrapolating those procedures to underground
structures can be unsafe, as lead to an overestimation of the structure deformation
capacity.

Figure 7-6. Reinforcement for maximum displacement according to Direct Design
Section A-A
Section B-B
Section C-C
Section D-D
Section E-E
2x32//0.10
32//0.10 12//0.20
25//0.20+20//0.20
32//0.20 32//0.20+ 32//0.20
1132
1132 1132+1125
2x25//0.20
32//0.20
32//0.10
32//0.10
10//0.20//0.30
10//0.20//0.30
12//0.20//0.20
10//0.20//0.30
A
A
B
B
C C
D
D
E E
1132
1132 1132+1125
1132
1132 1132+1125
25//0.20
2x32//0.10
32//0.10
32//0.20
32//0.20
632
632
632
32//0.20+25//0.20
32/0.20+32//0.20
25//0.20+20//0.20 10//0.30
32//0.20
32//0.10 10//0.30
1632
12//0.25
432
1132+1125
1132
10//0.20
220

269

Figure 7-7. Curvature diagrams at maximum displacement code design [/1000m]
7.6.2 Structure designed according to the proposed methodology
In order to maximize de ductility of the structure, Capacity Design principles must be
applied.
7.6.2.1 Choice of deformation mechanism
The number and location of plastic hinges involves in general the choice of a partial or
global mechanism (structure with fewer connections than necessary to maintain
equilibrium). In structures that develop above ground the mechanism can be chosen by
the designer, but in an underground structure it must compatible with the applied
displacement profile. For the linear, sinusoidal or any other profile reasonably regular
along the height (not the one shown in Figure 7-6 c) two main global mechanisms can be
foreseen, as shown in Figure 7-8.

Figure 7-8. Example structure: global mechanisms

a) b)

270
In what regards the choice of an appropriate mechanism it would be difficult to
formulate standard recommendations for all cases. However, some considerations can be
made, as follows. In nodes where elements with very different dimensions in the plan of
the flexible alignments join, it may not be possible to choose the element in which the
plastic hinge will develop. That is the case where beams or columns join slabs or
perimeter walls. In general, it is very difficult to avoid that the hinges in the vicinity of
this nodes develop in the beams or columns, as it is almost impossible to design these
elements with more flexural capacity than the slabs or walls. This hinges are identified in
Figure 7-8 by the grey colour.
In wall-slab or beam-column connections the location of the hinges is in general a
designers choice. Some criteria to support these choices can be considered. The bottom
slab is usually a very thick element with considerable flexural capacity. It is therefore
easier that at the connection with perimeter walls the hinges develop at the walls. At the
wall-top slab connections the dimensions of both elements usually are not too different
and the designer may be able to choose where to develop the hinges, as of the point of
view of performance (maximization of the global ductility) both options can be
acceptable. Therefore, two criteria can be used: easiness of construction and easiness of
repair after a strong earthquake. The zone where the plastic hinge develops needs to be
confined, what implies placing a large amount of reinforcement perpendicular to wall or
slab faces to provide confining stresses in that direction. The horizontal reinforcement
perpendicular to the thickness of the wall is probably easier to place than vertical
reinforcement in the slabs. And since other plastic hinges develop in the perimeter walls
(at the base and other locations, as will be shown later), the best options appears to be to
locate the hinges in the walls. This allows maintaining the top slab elastic during strong
earthquakes, avoiding the need to repair it afterwards.
A similar option about the location of the plastic hinges has to be done at the beam
column joints. Note that the reasons why EC 8 prescribes the weak-beam/strong-column
mechanism in building frames dont apply to underground structures: there is no need to
avoid the soft storey mechanism since the deformation of the structure is conditioned by
the surrounding soil and therefore large ductility demands and large second orders effects
can not be triggered due to the soft-storey deformability. Another issue related with the
choice of the hinges location at beam-column joints is the shape of the displacement
profile imposed on the structure. If it is a profile similar to the one shown in Figure 7-5.c,
it is impossible to avoid hinging at intermediate levels of the vertical elements, as shown
in Figure 7-9. Note that even though in node 2 the designer can choose to locate the
hinges in the beams or in the columns, in nodes 1 and 3 there is a variation of rotation
between the columns converging on those nodes, which forces column hinging regardless
of beam design.

271

Figure 7-9. Mechanism with unavoidable hinges at intermediate locations of walls and columns
Since column hinging is unavoidable at the extremities and probably also at intermediate
levels, it is the first option to consider and probably the most suitable. Another argument
of practical nature in support of this option is that for the other actions the columns are
essentially under axial compression, while the beams also have to withstand flexure and
shear effects, leading in general to larger dimensions in the bending plan. However if the
beams have similar dimensions to columns and larger aspect ratios it may be possible to
provide more ductility to beams than to columns, leading to a larger deformation capacity
for the structure. Another feature of behaviour highlighted in Figure 7-9 is that unless the
soil characteristics are very uniform in the entire vicinity of the structure yielding can take
place anywhere in the perimeter walls. Therefore, it may be necessary to provide
confinement reinforcement throughout the perimeter walls.
Following the above discussion the example structure was designed to develop the
mechanism shown in Figure 7-8 b and the perimeter walls were confined at all locations
in order that a reasonable curvature ductility is available at any location.
It is worth to emphasize that the different constraints to the choice of the best
mechanism in underground structures as compared to building frames lead to criteria
different from the ones prescribed in EC 8 for those structures.
7.6.2.2 Design of reinforcement
The starting point for this phase is the structure as designed to resist to all other actions
but the seismic action. According to Capacity Design principles the zones chosen to
remain elastic must be designed to do not yield during the development of the plastic
hinges. This implies these zones must be provided with enough reserve strength for that
3
1
2


272
purpose. The plastic hinge zones must be designed for ductility as well as to avoid any
brittle type of failure. Considering the chosen mechanism the main implications for the
different structural elements are as follows:
perimeter walls: it is not necessary to increase the flexural capacity as hinges are
expected to develop at the walls (remind that the proposed methodology is
equivalent to consider q=). It is necessary to increase the available ductility
throughout the walls: for this purpose confinement reinforcement, comprising
horizontal links in the direction of the wall thickness and properly anchored at the
extremities around the vertical reinforcement must be provided. Figure 7-11 shows
the new design of the wall cross sections.
slabs: the design for the other actions ensures that slabs are stronger then the
columns to which they are connected. However the flexural capacity may need to be
increased, particularly in the extremities of the top slab, to be higher than the
maximum moment at the walls hinges, in order to avoid the formation of plastic
hinges at the slab extremities. For this purpose at the extremities the slab is designed
for a bending moment which is Msd
slab
=
0
.Mrd
wall
, with both moments evaluated by
the usual design procedure prescribed in EC 2. A value
0
=1.3, as prescribed in EC 8
for column design, seems appropriate for the first iteration of the proposed design
procedure. Figure 7-10 shows a longitudinal cut of the top slab.
beams: in order to increase the ductility of the extreme sections where plastic hinges
are expected to develop, confinement reinforcement must be provided at these
zones. Flexural reinforcement on the lower face was also added in order to reduce
the size of the compressive zone when the top reinforcement yields at beam
extremities. The effectiveness of this extra reinforcement in increasing the curvature
ductility can be easily evaluated by section analysis. In what regards interior beam
column joints it was decided to develop the plastic hinges in the columns. Therefore
in the first iteration the flexural reinforcement on the beams in the vicinity of these
nodes provide an excess flexural capacity above the sum of the moments of
resistance of the columns converging at the same node of 30%, what also depends
on column design. However the analysis showed this was not enough. Figure 7-10
shows the new design of the beams. The beams were provided with more transverse
reinforcement at the zones plastic hinges are expected to develop to increase the
ductility of confined concrete.
columns: since the columns are essentially under axial compression for all other
loads, can be designed for that purpose with the minimum amounts of flexural and
transverse reinforcement. Since the columns are not intended to remain elastic there
is no need to increase their flexural capacity (q=). However flexural reinforcement
may be useful to decrease the ductility demand because of the following reasons: (i)
to increase column stiffness relatively to the beams, in order to reduce the

273
restrictions that the beams impose to column rotations at beam-column joints, (ii)
because large spacing of vertical reinforcement reduces the effectiveness of
confinement, (iii) because the spacing of confinement reinforcement should be
proportional to the diameter of the flexural reinforcement, therefore this should not
be too small.

Figure 7-10. Details of design according to the proposed methodology
Section A-A
Section B-B
Section C-C
Section D-D
Section E-E
2x25//0.20
2x25//0.20 12//0.10
25//0.20+32//0.20
25//0.20+32//0.20 10//0.30
25//0.20
25//0.20 25//0.20+20//0.20 25//0.20+32//0.20
832
832+832 832
832
832
832
832+832 832
2x25//0.20
32//0.20
32//0.20
2x25//0.20
2.50
12//0.20//0.10
10//0.20//0.30
10//0.20//0.30
A
A
B
B
C C
D
D
E E
25//0.20+32//0.20
25//0.20+
20//0.20
832+232
10//0.20
10//0.20
832+832
832+232
832+232
12//0.10
832
832
220
16//0.10
12//0.10
832+232
832+232
225
16//0.10
32//0.20
32//0.20 10//0.30

274
Besides there is the obvious need to provide confinement reinforcement in the plastic
hinge zones to increase the available curvature ductility in those zones. The efficiency of
the above can be evaluated by section analysis. Figure 7-11 shows the constitutive
relationships for steel, confined and unconfined concrete and the moment curvature
diagrams at the base of the columns before and after the increase in reinforcement,
evaluated considering the axial force at maximum displacement.
[MPa]
0
10
20
30
40
50
0 5 10 15
[/1000]
[MPa]
-600
-400
-200
0
200
400
600
-80 -60 -40 -20 0 20 40 60 80
[/1000]
Moment [kNm]
-6000
-4000
-2000
0
2000
4000
6000
-150 -75 0 75 150
Curvature [/1000m]
Without seismic design
With seismic design

a) Concrete b) Steel c) Column moment-curvature diagrams
Figure 7-11. Material constitutive relationships and moment-curvature diagrams at the column base
section
It should be emphasized that the process of maximizing the overall structural ductility is
an iterative procedure, that starts from the structure as designed to resist to all other
actions. Successive analysis and changes were done in order to improve the overall
ductility. The following examples highlight this procedure: (i) at each analysis the rupture
point and other locations close to rupture were identified and the possibility of increasing
the available ductility at those locations was analysed; this was the case at beam
extremities that initially were all designed with 6 vertical stirrups 12, that the analysis
showed were not enough to prevent rupture at the beams, limiting the overall ductility of
the structure; in the final design, at the extremities the beams were designed with
612+216 vertical stirrups; another change of this type was the use of external stirrups
16 at the three lower column hinges; (ii) column flexural reinforcement was increased in
order to increase its stiffness (according to the concept discussed in section 7.3 and
illustrated in Figure 7-3, the amount of flexural reinforcement influences the member
stiffness) relatively to the beams, to reduce the ductility demand on the columns; note
that the increase in column flexural reinforcement also led to an increase in beam flexural
reinforcement to avoid beam hinging but due to the curtailment of reinforcement, the
stiffness of the beams increased less than the stiffness of the columns, in which there was
no curtailment of flexural reinforcement; (iii) beam overstrength at beam-column joints
was increased far above the initial value of
0
=1.3, because the balance between beam
moments on both sides of the nodes changed in the non-linear range increasing the
moment demand.

275
The above is qualitatively different from current elastic analysis in which the designer
knows the exact procedure that must be followed. The design for ductility leaves the
designer with much more freedom but demands more knowledge and capacity to
anticipate the potential seismic behaviour of the structure in order to decide at each
iteration what are the most adequate changes to the design that resulted from the
previous iteration.
7.6.2.3 Results
The non linear analysis of the structure designed according to the proposed methodology
showed it could withstand a distortion of
max
=14.6x10
-3
, corresponding to a horizontal
relative displacement between top and bottom of the structure of =32.9cm. Figure 7-12
shows the curvature diagrams at this stage.

Figure 7-12. Curvatures at maximum displacement - proposed methodology [/1000m]
The comparison of this results with the ones of the structure designed according to
current code concepts
max
=5.0x10
-3
shows the superior seismic performance of the
structure designed according to the proposed methodology. The comparison between the
curvatures at maximum displacement for both structures (Figure 7-7 and Figure 7-12)
highlights the reasons for this difference: the higher ductility of the structural elements
and the efficient exploration of that ductility throughout the structure designed according
to the proposed methodology. A full comparison of costs cannot be done as the structure

276
was not fully defined, neither was the constructive process. However, in terms of
materials most of the difference regards the amount of steel in the perimeter walls. The
proposed methodology leads to the use of less flexural reinforcement, but needs large
amounts of confinement reinforcement, leading to almost equal total amounts of steel
spent in the perimeter walls. In the slabs the proposed methodology leads to moderate
savings, as the flexural reinforcement is conditioned essentially by the minimum levels
prescribed in EC2. In beams and columns the general trend is similar to that observed in
the perimeter walls, with some savings for the design of the columns according to the
proposed methodology. The above indicates that in general the design according to the
proposed methodology does not has a significant influence on the overall costs, and may
even lead to slight savings in some elements.
7.7 SUMMARY AND CONCLUSIONS
During earthquakes underground structures do not have to resist to horizontal inertia
forces, as structures that develop essentially above ground, but only to withstand the
displacements the soil imposes on them without losing the capacity to resist to permanent
actions. Therefore reinforced concrete structures must be designed to be flexible and
ductile. For instances large underground reinforced concrete structures, such as tube
stations, should be designed in the transverse direction with elements whose dimensions
must be the ones strictly necessary to resist to other actions but the seismic actions. Stiff
elements, such as counterforts or short beams should be avoided, as well as large soil
covers. The interference of secondary or non-structural elements with the deformation of
the main structure should be avoided.
The structure must be designed by stages: first for all load combinations whose main
variable action is not the seismic action; second for the seismic action. Since there are no
inertia forces (equivalent to consider the behaviour factor infinite) the designer must
choose a suitable deformation mechanism and apply Capacity Design principles, this is,
to design the potential plastic hinge zones for ductility and the remaining zones with
excess strength to remain elastic. An application example is shown. The proposed
procedure tends to lead to considerable savings in flexural reinforcement but more
confinement reinforcement. In general terms it leads to structures with better seismic
performance than the extrapolation of code procedures derived for structures that
develop above ground, that may lead to unsafe underground structures. Therefore it is
recommended that EC8 covers explicitly the seismic design of underground structures.

REFERENCES
CHAPTER 1.
Ilki, A., Boduroglu, M.H., Ozdemir, P., Baysan, F., Demir, C., Sirin, S. [2003] Comparison of the
results obtained by seismic index method and structural analysis for existing and retrofitted
buildings Proceedings of 5
th
National Conference on Earthquake Engineering, Istanbul, Turkey, Paper
No. AT-119.
JSIM. [1977]. Standard for evaluation of seismic capacity of existing reinforced concrete buildings. Japan
Building Disaster Prevention Association, Ministry of Construction, Tokyo, Japan
Ozdemir, P., Ilki, A., Boduroglu, M.H., Sirin, S., Demir, C., Baysan, F. [2004] A modified rapid
screening procedure for medium rise rc structures Proceedings of 13
th
World Conference on
Earthquake Engineering,, Vancouver, Canada, Paper No. 1542.
TERDC-98. [1998]. Specification for structures to be built in disaster areas: Earthquake disaster prevention.
Ministry of Settlement and Public-Works, Ankara, Turkey.
TERDC-2006. [2006]. Specification for structures to be built in disaster areas: Earthquake disaster prevention.
Ministry of Settlement and Public-Works, Ankara, Turkey.
CHAPTER 2.
ATC-40, [1996], Seismic Evaluation and Retrofit of Concrete Buildings, ATC, California, USA.
FEMA-440, [2004], Improvement of Nonlinear Static Seismic Analysis Procedures, FEMA, Washington
DC.
TEC-2006 [2006], Earthquake Code of Turkey, Ministry of Construction, Ankara, (in Turkish).
SAP2000 [2006], Structural Analysis Software, Computers and Structures Inc., Berkeley, CA, USA.


278
SECTIONS 3.1 & 3.3
Amar Khennane; Robert E. Melchers. Durability of Glass Polymer Composites Subject to Stress
Corrosion. Journal of Composites for Construction, Vol 7, No. 2, May 2003, 10117.
Bond behaviour of corroded steel reinforcement in concrete wrapped with carbon fibre reinforced
polymer sheets. Khaled Soudki, Ted Sherwood [Journal of composites for construccion,
ASCE, Julio-Agosto. 2003]
Dowling N.E., Mechanical Behaviour of Materials: Engineering Methods for Deformation,
Fracture, and Fatigue, Ed. Prentice Hall,.
Guide for the design and construction of externally bonded FRP systems for strengthening
concrete structures. American Concrete Institute (ACI). [ACI 440.2R-02]
Hahn H.T., Fatigue Behaviour of Composite Laminate, J. Composite Materials, Vol. 10, 1976.
Principe, F., Design Requirements, Engineered Materials Handbook
Reifsnider K.L., Fatigue of Composite Materials, Composite Materials Series, Volume 4, Ed. R.B.
Pipes, 1991.
Tannous, F. E. and Saadatmanesh, H. (1998). Environmental Effects on the Mechanical
Properties E-Glass FRP Rebars, ACI Materials Journal, Vol. 95(2), pp 87100.
Valter Dejke. Durability of FRP Reinforcement in ConcreteLiterature Review and
Experiments. Thesis for the Degree of Licentiate of Engineering, Department of Building
Materials, Chalmers University of Technology; Gteborg, Sweden 2001.
CHAPTER 3.2
ANSYS, Inc [2007] ANSYS products, www.ansys.com
ANSYS, Inc [2005] Guide to ANSYS User Programmable Features, ANSYS Release 10.0, Canonsburg,
PA,, USA
Barbat, A.H., Oller, S., Oate, E., Hanganu, A. [1997] Viscous damage model for Timoshenko
beam structures International Journal of Solids and Structures, Vol.34, No 30, pp. 3953-3976

279
Car, E., Oller, S., Oate, E. [2001] A Large Strain Plasticity for Anisotropic Materials
Composite Material Application, International Journal of Plasticity, Vol. 17, No. 11, pp. 1437-
1463.
Car, E., Oller, S., Oate, E. [2000] An anisotropic elasto-plastic constitutive model for large strain
analysis of ber reinforced composite materials Computer Methods in Applied Mechanics and
Engineering, Vol.185, No 2-4, pp. 245-277
CIMNE [2006] LessLoss Deriverable 49: Integration of knowledge on FRP retrofitted structures, LessLoss
project (GOCE-CT-2003-505488), Sub-Project 7
Crisfield, M.A. [1991], Non-linear finite element analysis of solids and structures. Volume 1. Essentials, John
Wiley & Sons, West Sussex, UK
GiD [2007] GiD 8. The personal pre and post processor, www.gidhome.com
Lubliner J., Oliver, J., Oller S., Oate, E. [1989] A plastic damage model for concrete
International Journal of Solids and Structures, Vol. 25, No 3, pp. 299-326
Malvern, L. E. [1968] Introduction to the mechanics of a continuous medium, Prentice-Hall, Englewood
Cliffs, NJ, USA.
Martinez, X., Oller, S., Barbat, A., Rastellini, F. [2007] New procedure to calculate compression
strength of FRP using the Serial/Parallel mixing theory, Accepted at 8
th
FRPRCS-8, Patras,
Greece
Martinez, X., Oller, S., Barbat, A. [2006] Numerical tool to study structural reinforcement of steel
reinforced concrete (RC) structures under seismic loads using fibre reinforced polymers
(FRP), ECEES-1 Proceedings, Geneva, Switzerland
Meier, U. [1995] Strengthening of structures using carbon bre/epoxi composites Construction
and Building Materials, Vol.9, No 6, pp.341-351
Oller, S., Miquel, J., Zalamea, F. [2005] Composite material behaviour using a homogenization
double scale method Journal of Engineering Mechanics, Vol.131, No 1, pp. 65-79
Oller, S. [2002], Anlisis y clculo de estructuras de materiales compuestos, CIMNE, Barcelona, Spain
Oller, S., Oate, E., Oliver, J., Lubliner, J. [1990] Finite element non--linear analysis of concrete
structures using a plastic-damage model Engineering Fracture Mechanics, Vol. 35, No 1-3, pp.
219-231

280
Oate, E. [1995] Clculo de estructuras por el mtodo de elementos finitos. Anlisis esttico lineal, CIMNE,
Barcelona, Spain
PLCd Manual [1991-now] PLCd. Non-linear thermomechanic finite element code oriented to PhD student
education. Code developed at CIMNE
Rabinovitch, O., Frostig, Y. [2001] Nonlinear high-order analysis of cracked RC beams
strengthened with FRP strips Journal of Structural Engineering, Vol.127, No 4, pp. 381-389
Rastellini, F. [2006] Numerical modelling of the constitutive non-linearity of composite laminates, PhD thesis,
Departament de Resistncia de Materials i Estructures a lEnginyeria (RMEE) UPC,
Directors: S. Oller y E. Oate.
Snchez-Palencia, E.S. [1987], Homogenization techniques for composite media, Springer-Verlag, Berlin,
Germany. Chapter: Boundary layers and edge effects in composites, pp. 121-192
Spadea, G., Benicardino F., Swamy, R. [1998] Structural behaviour of composite RC beams with
externally bonded CFRP Journal of Composites for Construction, Vol.2, No 3, pp. 132-137
Trusdell C., Toupn, R. [1960] The classical field theories, Handbuch der physic iii/I, Edition,
Springer-Verlag, Berlin, Germany
Zienkiewicz, O. C., Taylor, L.R., [1991] The finite element method, McGraw-Hill, London, England
CHAPTER 3.4
American Concrete Institute (ACI) Committee 440. (2002). ACI 440: Guide for the design and
construction of externally bonded FRP systems for strengthening concrete structures,
Detroit.
Becque, J., Patnaik, A.K., and Rizkalla, S.H. [2003] Analytical models for concrete confined with
FRP tubes, J. Compos. Const., Vol. 7, No. 1, pp. 31-38.
Bisby, L.A., Dent, A.J.S., and Green, M.F. [2005] Comparison of confinement models for fibre-
reinforced polymer-wrapped concrete, ACI Structural Journal, Vol. 102, No. 1, pp. 62-72.
Canadian Standards Association (CSA). (2002). CAN/CSA-S806: design and construction of
building components with fibre-reinforced polymers, Ottawa.

281
De Lorenzis, L., and Tepfers, R. [2003] Comparative study of models on confinement of
concrete cylinders with fibre-reinforced polymer composites, J. Compos. Const., Vol. 7, No. 3,
pp. 219-237.
Demers, M., and Neale, K.W. [1999] Confinement of reinforced concrete columns with fibre-
reinforced composite sheets-an experimental study, Can. J. Civ. Eng., Vol. 26, No. 2, pp.
226-241.
Comite Europeen de Normalisation (CEN). (2005). EN 1998-3: Eurocode 8 part 3: assesment
and retrofitting of buildings., Brussels.
Fam, A.Z., and Rizkalla, S.H. [2001] Confinement model for axially loaded concrete confined by
circular fibre-reinforced polymer tubes, ACI Structural Journal, Vol. 98, No. 4, pp. 451-461.
Fdration Internationale du Bton (fib). (2001). Bulletin no.14: externally bonded FRP
reinforcement for rc structures. Technical Report.
Ilki, A., and Kumbasar, N. [2002] Behavior of damaged and undamaged concrete strengthened by
carbon fibre composite sheets, Struct. Eng. and Mech., Vol. 13, No. 1, pp. 75-90.
Ilki, A., and Kumbasar, N. [2003] Compressive behaviour of carbon fibre composite jacketed
concrete with circular and non-circular cross-sections, Journal of Earthquake Engineering, Vol.
7, No. 3, pp. 381-406.
Ilki A., Peker O., Karamuk E., and Kumbasar N. [2004a] External confinement of low strength
brittle reinforced concrete short columns, Proc., International Symposium on Confined Concrete
(CD-ROM), Changsha, China.
Ilki A., Koc V., Peker O., Karamuk E., and Kumbasar N. [2004b] Strengthening of RC Columns
with Inadequate Transverse Reinforcement, Proc., The Second Int. Conf. on FRP Comps. in Civil
Engineering, 8-10 Dec. 2004, Adelaide, Australia, pp. 211-218.
Ilki, A., Kumbasar, N., and Koc, V. [2004c] Low strength concrete members externally confined
with FRP sheets, Struct. Eng. and Mech., Vol. 18, No. 2, pp. 167-194.
Lam, L., and Teng, J.G. [2002] Strength models for fibre-reinforced plastic-confined concrete, J.
Struct. Eng., Vol. 128, No. 5, pp. 612-623.
Lam, L., and Teng, J.G. [2003a] Design-oriented stress-strain model for FRP-confined concrete,
Construction and Building Materials, Vol. 17, pp. 471-489.

282
Lam, L., and Teng, J.G. [2003b] Design-oriented stress-strain model for FRP-confined concrete
in rectangular columns, Jour. of Reinf. Plast. and Comps., Vol. 22, No. 13, pp. 1149-1186.
Lam, L., Teng, J.G., Cheung, C.H., and Xiao, Y. [2004] Behavior of FRP-confined concrete
under cyclic axial compression, Proc., International Symposium on Confined Concrete (CD-ROM),
Changsha, China.
Mander, J.B., Priestley, M.J.N., and Park, R. [1988]. Theoretical stress-strain model for confined
concrete, ASCE Jour. of the Struct. Div., Vol. 114, No. 8, pp. 1804-1826.
Matthys, S., Toutanji, H., Audenaert, K., and Taerwe, L. [2005] Axial load behavior of large-scale
columns confined with fibre-reinforced polymer composites, ACI Structural Journal, Vol.
102, No. 2, pp. 258-267.
Rochette, P., and Labossiere, P. [2000] Axial testing of rectangular column models confined with
composites, ASCE Jour. of Comp. for Cons., Vol. 4, No. 3, pp. 129-136.
Shehata I.A.E.M., Carneiro L.A.V., and Shehata L.C.D. [2002] Strength of short concrete
columns confined with CFRP sheets, Mat. and Struct., Vol. 35, No. 245, pp. 50-58.
Tan, K.H. [2002] Strength enhancement of rectangular reinforced concrete columns using fibre-
reinforced polymer, J. Compos. Const., Vol. 6, No. 3, pp. 175-183.
Tastani, S.P., and Pantazopoulou S.J. [2004] Experimental evaluation of FRP Jackets upgrading
RC corroded columns with substandard detailing, Engineering Structures, Vol. 26, pp. 817-829.
Toutanji, H.A. [1999] Stress-strain characteristics of concrete columns externally confined with
advanced fibre composite sheets, ACI Mat. Jour., Vol. 96, No. 3, pp.397-404.
Wang, Y.C., and Restrepo, J.I. [2001] Investigation of concentrically loaded reinforced concrete
columns confined with glass fibre-reinforced polymer jackets, ACI Structural Journal, Vol. 98,
No. 3, pp. 377-385.
Xiao, Y., and Wu, H. [2000] Compressive behaviour of concrete confined by carbon fibre
composite jackets, ASCE Jour. of Mat. in Civ. Eng., Vol. 12, No. 2, pp. 139-146.


283
CHAPTER 3.5
Applied Technology Council (ATC), [1996], Seismic evaluation and retrofit of concrete buildings, ATC-40,
California
Binici, B., and G. Ozcebe, 2005. Analysis of Infilled Reinforced Concrete Frames Strengthened
with FRPs. Advances in Earthquake Engineering for Urban Risk Reduction, NATO Science Series,
Springer, 455-470.
Erduran E. and Yakut A., [2004a], Drift based damage functions for reinforced concrete
columns Computers and Structures, 82: 121-130
Erduran E. and Yakut A., [2004b], Drift Based Damage Functions for Components of RC
Structures, 13th World Conference on Earthquake Engineering, Vancouver.
Erduran E. and Yakut A., [2004c], Development of Component Importance Factors for RC
Buildings, 6th International Congress on Advances in Civil Engineering, 6-8 October, Bogazici
University, Istanbul, Turkey.
Erduran E. and Yakut A., [2007], Vulnerability Assessment Of Reinforced Concrete Moment
Resisting Frame Buildings ASCE, Journal of Structural Engineering (in press).
Federal Emergency Management Agency (FEMA), [2000], Prestandard and commentary for the seismic
rehabilitation of buildings, FEMA-356, Washington, DC.
Sucuoglu, H., U. Yazgan, A. Yakut [2007], A Screening Procedure for Seismic Risk Assessment
in Urban Building Stocks, Earthquake Spectra (in press), EERI, USA, March.
Yakut, A., G. Ozcebe, M.S. Yucemen, [2006],Seismic Vulnerability Assessment using Regional
Empirical Data, Earthquake Engng Struct. Dyn. 2006; 35:11871202, August.
CHAPTER 3.6
Bakis C.E., Bank L.C., Brown V.L., Cosenza E., Davalos J.F., Lesko J.J., Machida A., Rizkalla S.H.,
Triantafillou T.C. [2002] Fibre-Reinforced Polymer Composites for Construction State-
of-the-Art Review, J. Compos. for Constr., Vol. 6, No. 2, pp. 73-87.
Ehsani M.R., Saadatmanesh H, Velasquez-Dimas, J.I. [1999] Behaviour of retrofitted URM walls
under simulated earthquake loading, J. Compos. for Constr., Vol. 3, No.3, pp. 134-142.

284
Hamilton H.R. , Dolan C.W. [2001] Flexural Capacity of Glass FRP Strengthened Concrete
Masonry Walls, J. Compos. for Constr., Vol. 5, No. 3, pp. 170-178.
Krevaikas T.D. and Triantafillou T.C. [2005] Masonry Confinement with Fibre-Reinforced
Polymers, J. Compos. for Constr., Vol. 9, No. 2, pp. 128-135.
Taylor C. [1998] Seismic performance of masonry panels for Nuclear Electric plc, EERC Project
Report NE395/RP/6, Bristol, United Kingdom.
Tikalsky P.J., Atkinson R.H, Hammons M.I. [1995] Journal of Structural Engineering, Vol. 121, No.2,
pp. 283-289.
Tumialan J.G., Galati N., Nanni A. [2003] Field Assessment of Unreinforced Masonry Walls
Strengthened with Fibre Reinforced Polymer Laminates, Journal of Structural Engineering, Vol.
129, No. 8, pp. 1047-1056.
Velasquez-Dimas J.I., Ehsani M.R., Saadatmanesh H. [2000] Out of plane behaviour of brick
masonry walls strengthened with fibre composites, ACI Struct. J., Vol. 97, No. 5, pp. 377-
387.
Velasquez-Dimas J.I., Ehsani M.R., II. [2000] Modeling Out-of-Plane Behaviour of URM Walls
Retrofitted with Fibre composites, J. Compos. for Constr., Vol. 4, No.4, pp. 172-18.
CHAPTER 4.1
Aiken I. Passive energy dissipation hardware and applications. Proceedings, Lo Angeles County
and Season Symposium on Passive Energy Dissipation Systems for New and Existing
Buildings, Los Angeles, July 1996.
Antman S.S., Nonlinear Problems of Elasticity, SpringerVerlag, 1991.
Barbat A.H., Oller S., Hanganu A., Onate E., Viscous damage model for Timoshenko beam
structures, International Journal for Solids and Structures. 34 (30) (1997) 39533976.
Bayrak O. and Sheikh S. A.. Plastic hinge analysis. Journal of Structural Engineering, 2001;
127(9):10921100.
Coleman J. and Spacone E.. Localization issues in force-based frame elements. Journal of Structural
Engineering, 2001; 127(11):12571265.
Davenne L., Ragueneau F., Mazar J., Ibrahimbegovic A., Efficient approaches to finite element
analysis in earthquake engineering, Computers and Structures. 81 (2003) 12231239.

285
Driemeier L., Baroncini S.P. and Alves.M. A contribution to the numerical non-Linear analysis of
three-dimensional truss system considering large strains, damage and plasticity.
Communications in Nonlinear Science and Numerical Simulation, 2005; 10:515535.
Elnashai A.S., Mwafi A.M., Overstrength and force reduction factors of multistorey reinforced
concrete buildings, The Structural Design of Tall Buildings. 11 (2002) 329351.
Federal Emergency Management Agency. Prestandard and Commentary for the Seismic Rehabilitation of
Buildings, Report 356, Washington, DC: FEMA. 2000.
Federal Emergency Management Agency. NEHRP Recommended Provisions for Seismic Regulations for
New Buildings and Other Structures. Report 368, Washington, DC: FEMA. 2000.
Gluck N., Reinhorn A.M. , Gluck J. and Levy R.. Design of supplemental damping for control of
structures. Journal fo Structural Engineering, 1996; 122(12):13941399.
Gruttmann F., Sauer R. and Wagner W., Theory and numerics of threedimensional beams with
elastoplastic material behaviour, International Journal of Numerical Methods in Engineering.
48 (2000) 16751702.
Hanganu A.D., Oate E., Barbat A.H., Finite element methodology for local/global damage
evaluation in civil engineering structures, Computers and Structures. 80 (2002) 16671687.
Ibrahimbegovic A., On finite element implementation of geometrically nonlinear Reissners beam
theory: three-dimensional curved beam elements, Computer Methods in Applied Mechanics
and Engineering. 122 (1995) 1126.
Ibrahimbegovic A., Al Mikdad M., Quadratically convergent direct calculation of critical points for
3D structures undergoing finite rotations, Computer Methods in Applied Mechanics and
Engineering. 189 (2000) 107120.
Ibrahimbegovic A., Shakourzadeh H., Batoz J.L., Mikdad M.A., Guo Y.Q., On the role of
geometrically exact and secondorder theories in buckling and postbuckling analysis of
threedimensional beam structures, Computers and Structures. 61 (6) (1996) 11011114.
Kapania R.K., Li J., On a geometrically exact curved/twisted beam theory under rigid cross
section assumption, Computational Mechanics. 30 (2003) 428443.
Lubliner J., Thermomechanics of deformable bodies, Department of Civil Engineering, University
of California, Berkeley, 1985.
Lubliner J., Oliver J., Oller S., Oate, A plasticdamage model for concrete, International Journal
of Solids and Structures. 25 (3) (1989) 299326.
Neuenhofer A., Filippou F.C., Evaluation of the nonlinear frame finiteelement models, Journal
of Structural Engineering. 123 (7) (1997) 958966.

286
Shao Y., Aval S., Mirmiran A., Fibreelement model for cyclic analysis of concretefilled fibre
reinforced polymer tubes, Journal of Structural Engineering. 131 (2) (2005) 292303.
Hajjar J.F. Concrete-filled steel tube columns under earthquake loads. Progress in Structural
Engineering Materials, 2000; 2:7281.
Hanson R. D., Aiken I. D., Nims D. K., Ritchter P. J. and Batchman R. E. State of the art and
state of the practice in seismic engineering dissipation. Procedings, ATC-17-1. Seminar on seismic
isolation, passive energy dissipation and active control. Applied Technology Council, San Francisco,
California, March 1993.
Kumar P., Nukala V.V. and White D. W.. A mixed finite element for threedimensional nonlinear
analysis of steel frames. Computer Methods in Applied Mechanic and Engineering, 2004;
193(5):25072545.
Mata P., Oller S. and Barbat A.H. Static analysis of beam structures under nonlinear geometric and
constitutive behaviour. Computer Methods in Applied Mechanic and Engineering, 2007; (Submited).
Mata P., Oller S. and Barbat A.H. Dynamic analysis of beam structures considering geometric and
constitutive nonlinearity. Computer Methods in Applied Mechanic and Engineering, 2007;
(Submited).
Mata P., Boroschek R., Barbat A.H. and Oller S. High damping rubber model for energy
dissipating devices. Journal of Earthquake Engineering, JEE, 2006.
Mata P., Oller S., Barbat A.H., and Boroschek R.. Numerical code for seismic analysis of
structures incorporating energy dissipating devices. First European Conference on Earthquake
Engineering and Seismology, ECEES, Geneva, Switzerland, 3-8 September 2006.
Mata P., Oller S. and Barbat A.H. and Boroschek R. Evaluation of the seismic behaviour of
precast concrete buildings with energy dissipating devices. First European Conference on
Earthquake Engineering and Seismology, ECEES, Geneva, Switzerland, 3-8 September 2006.
Oliver J., Cervera M., Oller S., Lubliner J., Isotropic damage models and smeared crack analysis of
concrete, Proceedings 2nd ICCAADCS, Zell Am See, Austria, Pineridge Press. 2 (1990) 945
958.
Oller S., Barbat A.H., Moment-curvature damage model for bridges subjected to seismic loads,
Computer Methods in Applied Mechanics and Engineering. 195 (2006) 44904511.
Oller S., Oate E., Miquel J., Botello S., A plastic damage constitutive model for composites
materials, International Journal of Solids and Structures. 33 (17) (1996) 25012518.
Reissner E.. On one-dimensional finite-strain beam theory: the plane problem. Journal of Applied
Mathematics and Physics, 1972; 23:795804.
Shen K.L. and Soong T.T. Design of energy dissipation devices based on concept of damage
control. Journal of Structural Engineering, 2005; 122(1):7682.

287
Simo J.C.. A finite strain beam formulation. The three-dimensional dynamic problem. Part I.
Computer Methods in Applied Mechanic and Engineering, 1985; 49:5570.
Simo J.C. and Vu-Quoc L. A three-dimensional finite-strain rod model. Part II: Computational
aspects. Computer Methods in Applied Mechanics and Engineering, 1986; 58:79116.
Simo J.C. and Vu-Quoc L.. On the dynamics in space of rods undergoing large motions - A
geometrically exact approach. Computer Methods in Applied Mechanics and Engineering, 1988;
66:125161.
Simo J.C., Hjelmstad K.D., Taylor R.L., Numerical simulations of elastoviscoplastic response of
beams accounting for the effect of shear, Computer Methods in Applied Mechanics and
Engineering. 42 (1984) 301330.
Sivaselvan M. V. and Reinhorn A. M. Collapse analysis: large inelastic deformations analysis of
planar frames. Journal of Structural Engineering, ASCE, 2002; 128(12):15751583.
Soong T.T. and Dargush G.F.. Passive Energy Dissipation Systems in Structural Engineering, 1997; John
Wiley & Soms Ltda.
Spacone E.and El-Tawil S. Nonlinear analysis of steelconcrete composite structures: state of the
art. Journal of Structural Engineering, ASCE, 2004; 130(2):159168.
Spiliopoulos K.V. and Lykidis G.Ch.. An eficient threedimensional solid finite element dynamic
analysis of reinforced concrete structures. Earthquake Engineering and Structural Dynamics, 2005.
Taucer F. F., Spacone E. and Filipou F. C.. A fibre beam-column element for seismic response analysis of
reinforced concrete structures. Technical report No. UCB/EERC-91/17. Earthquake Engineering
Research Center, College of Engineering, University of California, Berkeley; 1991.
Valles R.E., Reinhorn A.M., Kunnath S.K., Li C. and Madan A. IDARC 2D Version 4.0: A Program
for the Inelastic Damage Analysis of Buildings, Technical Report NCEER-96- 0010. National
center for earthquake engineering research. State University of New York at Buffalo, January
8, 1996.
Vignjevic R. A hybrid approach to the transient collapse analysis of thin walled frameworks II.
Computer Methods in Applied Mechanics and Engineering, 1997; 148:423437.
CHAPTER 4.2
AFPS. Le sisme de Ceyphan-Misis (Adana, Turquie) du 27 juin 1998, Rapport de Mission post-
sismique. ISBN 2-91179-04-7
Applied Technology Council (ATC-40), November [1996], Report N SSC 96-01: Seismic
Evaluation and Retrofit of Concrete Buildings, Volume 1, California, United States.

288
Biondini, F., Toniolo, G. 1-4 May [2002] Probabilistic Calibration of Behavior Factors of EC 8
for Cast-In-Situ and Precast Frames, In the Proc. of the BIBM17th International Congress Of
Precast Concrete Industry, Istanbul, Turkey.
Colinet, G.[2005] Prise en compte de ma dformabilit des diffrents types dappuis dans
lanalyse et le dimensionnement dossatures de btiments en bton prfabriqu, Universit
de Lige, Master thesis.
CEN - prEN1998-1-1:2003. Eurocode 8: Design of structures for earthquake resistance, Part 1:
General rules, seismic actions and rules for buildings.
CEN - prEN1993-1-1:2002. Eurocode 3: Design of steel structures, Part 1: General structural
rules and rules for buildings.
CEN - prEN1993-1-8:2003. Eurocode 3: Design of steel structures, Part 8: Design of joints.
de Chefdebien, A. [1998] Semi-rigidit des assemblages dans les ossatures en bton prfabriqu,
CERIB, Publication Technique n125.
de Chefdebien, A., Ghavamian, Sh. 1-4 May [2002] Seismic Behavior of Precast Concrete
Industrial Buildings, In the Proc. of the BIBM17th International Congress Of Precast Concrete
Industry, Istanbul, Turkey.
Federal Emergency Management Agency (FEMA Publication 273), octobre [1997], NERHP
Guidelines for the Seismic Rehabilitation of Buildings, Washington DC, United States.
FIP (Fdration Internationale de la prcontrainte). [1994] Planning and design handbook on
precast concrete building structures.
FIB (Fdration Internationale du Bton). Seismic design of precast concrete building structures
bulletin 27.
Lebacq, B. [2005] Etude dun assemblage pouter-colonne dans une structure prfabrique en
bton sous charge sismique, Haute Ecole Lonard de Vinci (ECAM), Master thesis.
New Zealand Concrete Society. Guidelines for the use of structural precast concrete in
buildings.
Plumier, A., Doneux, C., Castiglioni, C., Brescianini, J., Crespi, A., DellAnna, S., Lazzarotto, L.,
Calado, L., Ferreira, J., Feligioni, S., Bursi, O., Ferrario, F., Sommavilla, M., Vayas, I.,
Thanopoulos, P., Demarco, T. December [2004] Two Innovations for earthquake

289
resistance design, the INERD Project, Final report of research program CECA-7210-PR-
316. Partners of INERD project: University Degli Studi di Trento, Politecnico di Milano,
National Technical University of Athens, Instituto Superior Tcnico, Profil ARBED
Recherches and Universit de Lige.
Reis, P.R., Lucio, V.J.G. 1-4 May [2002] Beam - Column Connection for Precast Concrete
Structures in Seismic Regions, In the Proc. of the BIBM17th International Congress Of Precast
Concrete Industry, Istanbul, Turkey.
Toniolo, G. 1-4 May [2002] The Seismic Design of Precast Concrete Structures in the New
Eurocode 8, In the Proc. of the BIBM17th International Congress Of Precast Concrete Industry,
Istanbul, Turkey.
Vayas, I. and Thanopoulos, P. Innovative dissipative INERD pin connections for seismic
resistant braced frames, submitted to International Journal of Steel Structures.
Wood, Sh.L. 1-4 May [2002] Development of Precast Concrete Structural Systems for Resisting
Seismic Forces, In the Proc. of the BIBM17th International Congress Of Precast Concrete Industry,
Istanbul, Turkey.
CHAPTER 4.3
CEN - EN1998-1:2004. Eurocode 8: Design of structures for earthquake resistance. Part 1:
General rules, seismic actions and rules for buildings.
FEMA 273 [1997] Federal Emergency Management Agency NEHRP Guidelines for the Seismic
Rehabilitation of Buildings, Washington D.C., USA.
Plumier, A., Doneux, C., Castiglioni, C., Brescianini, J., Crespi, A., DellAnna, S., Lazzarotto, L.,
Calado, L., Ferreira, J., Feligioni, S., Bursi, O., Ferrario, F., Sommavilla, M., Vayas, I.,
Thanopoulos, P., Demarco, T., Two Innovations for earthquake resistance design, the
INERD Project, Final report of research program CECA-7210-PR-316, Report EUR 22044
EN. ISBN 92-79-01694-6. Office for Official Publications of the European Communities,
2006.
Vayas, I., Thanopoulos, P. 2005 Innovative dissipative INERD pin connections for seismic
resistant braced frames, International Journal of Steel Structures


290
CHAPTER 5.1
Priestley, MJN [2000] Performance Based Seismic Design, Proceedings of 12th World

Congress on
Earthquake Engineering, Auckland, New Zealand.
RSA [1983] Portuguese National Code for Actions, Portugal.
Skinner, R.I., Robinson, W.H., McVerry, G.H. [1993] An Introduction to Seismic Isolation, John Wiley
& Sons, Chichester, England.
CHAPTER 5.2
Aldemir, U., Bakioglu, M., and Akhiev, S. S. (2001). Optimal control of linear structures.,
Earthquake Engrg. and Struct. Dyn., 30(6), 835-851.
Aldemir, U., and Bakioglu, M. (2001). Active structural control based on the prediction and
degree of stability., J. Sound and Vibration 247(4), 561-576.
Aldemir, U. and Bakioglu, M. (1997). Application of Maximum Entropy Method and the Kalman
Filtering Technique for the Near-future Prediction of Seismic Excitation, Water and Statistics,
A Satellite Meeting to the 51st International Statistical Institue Session, 203-213, Ankara, Turkey.
Aldemir, U., Bakioglu, M. and Cigizoglu, K. (1997). Prediction of Near-future Earthquake, First
Japan-Turkey Workshop on Earthquake Engineering, 137-147, Istanbul, Turkey.
Bakioglu, M and Aldemir, U. (2001). A new numerical algorithm for sub-optimal control of
earthquake excited structures. International Journal For Numerical Methods in Engineering 50(12),
2601-2616.
Gavin, H. P., and Aldemir, U. (2005). Optimal control of earthquake response using semiactive
isolation. J. Engrg. Mech., ASCE, 131(8), 769-776.
Athans, M., and Falb. P. L. (1966). Optimal Control: An Introduction to the Theory and its Applications,
McGraw-Hill Book Company, New York.
Buckle, I. G., and Mayes, R. L. (1990). Seismic isolation history, application and performance-a
world view. Earthquake Spectra, 6(2), 161-201.
Carlson, J. D., and Weiss, K. D. (1994). A growing attraction to magnetic fluids. Machine design,
(Aug), 61-64.

291
Carlson, J. D., and Spencer, B. F. (1996). Magneto-rhelogical fluid dampers for semiactive seismic
control. Proc., Third. Int. Conf. On Motion and Vibration Control, Vol. III, 35-40.
Dyke, S. J., Spencer, B. F., Sain, M. K., and Carlson, J. D. (1996b). Modeling and control of
magnetorhelogical dampers for seismic response reduction. Smart Mat. and Struct., 5, 565-
575.
Dyke, S. J., Spencer, B. F., Sain, M. K., and Carlson, J. D. (1998). An experimental study of MR
dampers for seismic protection. Smart Mat. and Struct., 7, 693-703.
Ehrgott, R. C., and Masri, S. F. (1994a). Experimental characterization of an electrorheological
material subjected to oscillatory shear strains. J. Vibration and Accoustics, 116 (Jan.), 53-60.
Garrard, W. L. (1972). Suboptimal feedback control for nonlinear systems. Automatica, 8, 219-
221.
Gavin, H. P., Hanson, R. D., and Filisko, F. E. (1996a). Electrorheological dampers, Part I:
Analysis and Design. J. Appl. Mech., ASME, 63, 669-675.
Gavin, H. P., Hanson, R. D., and McClamroch, N. H. (1996b). Electrorheological dampers, Part
II: Testing and Modeling. J. Appl. Mech., ASME, 63, 676-682.
Gavin, H. P. (2001). Control of seismically excited vibration using electrorheological materials
and Lyapunov methods. IEEE Trans. On Control Systems Technology, Vol 9, No. 1, January.
Housner G.W., Bergman L. A., Caughey T. K., Chassiakos A. G., Claus R. O., Masri S. F., Skelton
R. E., Soong T. T., Spencer B. F., and Yao J. T.P. (1997). Structural Control: Past, Present
and Future. J. Engrg. Mech., ASCE, 123(9), 897-958.
Jansen, L. M., and Dyke, S. J. (2000). Semiactive control strategies for MR dampers: Comparative
study. J. Engrg. Mech., ASCE, 126(8), 795-803.
Karnopp, D. (1995). Active and semiactive vibration isolation. Transactions of the ASME, Special
50
th
Anniversary Design Issue, Vol. 117, 177-185.
Kelly, J. M. (1981). Aseismic base isolation: its history and prospects. Proc., First World Congr. on
Joints and Bearings, ACI-SP-70, Vol. 1, 549-586.
Kelly, J. M. (1986). Aseismic base isolation: review and bibliyography. Soil Dynamics and
Earthquake Engrg., 5(3), 202-216.

292
Kelly, J. M. (1993). Earthquake-resistant design with rubber, Springer, London.
CHAPTER 6
Bhaskararao A.V., Jangid, R.S. [2005] "Seismic analysis of structures connected with friction
dampers" Engineering structures, Vol. 28 Issue 5, pp. 690-703.
Karayannis C.G., Fotopoulou M.G. [1998] " Pounding of multi-storey RC structures designed to
EC8 & EC2", Proceedings of the 11
h
European Conference on Earthquake Engineering, Balkena,
Rotterdam, pp.1-11.
Lavelle, F.M., Sues, R.H. [1992] "Seismic pounding retrofits for closely spaced buildings,"
Proceedings of 10
th
World conference on Earthquake Engineering, Madrid, Spain.
NEHRP, October 1997 "Guidelines for the seismic rehabilitation of buildings," Federal Emergency
Management Agency, Washington, DC.
Penzien, J. [1997] "Evaluation of building separation distance required to prevent pounding during
strong earthquakes," Earthquake engineering and structural dynamics; Vol.26,pp. 849-858.
Plumier, A., Denol, V., Sanchez, L., Doneux, C., Warnotte, V., Van Alboom, W. [2005] "Seismic
assessment and retrofitting of a S shape building with expansion joints" Seismic Performance
Assessment and Rehabilitation of Existing buildings International Workshop, pp. 253-267.
Westermo, B.D. [1989] "The dynamics of interstructural connection to prevent pounding,"
Earthquake Engineering and Structural dynamics, Vol. 18, pp. 687-699.
Zhang, W.S., Xu, Y.L. [1999] "Dynamic characteristics and seismic response of adjacent buildings
linked by discrete dampers," Earthquake Engineering and Structural dynamics, Vol. 28, pp. 1163-
1185.
Zhu H., Iemura H. [2000] "A study of response control on the passive coupling element between
two parallel structures," Structural Engineering and Mechanics, Vol. 9, No. 4, 383-396.
Zhu, H.P., Xu, Y.L. [2005] "Optimum parameters of Maxwell model-defined dampers used to link
adjacent structures," Journal of Sound and Vibration, Vol. 279, pp. 253-274.


293
CHAPTER 7
EC 2 [2004] Design of concrete structures- Part 1-1: General rules and rules for buildings, EN
1992-1-1, Brussels
EC 8 [2005] Design of structures for earthquake resistance Part 1: General rules, seismic action
and rules for buildings, EN 1998-1, Brussels
EC 8 [2005] Design of structures for earthquake resistance Part 2: Bridges, EN 1998-2, Brussels
EC 8 [2004] Design of structures for earthquake resistance Part 5: Foundations retaining
structures and geotechnical aspects, EN 1998-5, Brussels
Gomes, R.C. [1999] Behaviour of undeground structures under seismic actions MSc
thesys, IST, Lisbon, Portugal
Hashash, Y.M.A., Hook, J.J., Schmidt, B. [2000] Seismic design and analysis of underground
structures A state-of-the-art Report sponsored by the International Tunneling association,
Working Group N2
Iwatate, T., Kobayashi, Y., Kusu, H., Rin, K., [2000] Investigation and shaking table
tests of subway structures of the Hyogoken-Nanbu earthquake 12
th
Word Conference
Earthquake Engineering, paper 1043, New Zeleand.
Pipa, [1993], Ductility of reinforced concrete elements under cyclic actions. Influence of the
reinforcement mechanical characteristics (in Portuguese), PhD thesis, IST, Lisbon

LIST OF SYMBOLS
CHAPTER 1
A
0

= Effective ground acceleration coefficient
A

= Cross sectional area of wall
B

= Index for regional seismicity
C

= Strength index.
C
c

= Strength index of a column
C
sc

= Strength index of a short column
C
w

= Strength index of a shear wall
k
C

= k
th
vertical members strength index
D

= Depth of column
D

= Structural irregularity index
I

= Building importance factor
I
D
.

= Required seismic index
I
s

= Seismic index of a building
K

= Time dependent deterioration index;
c
M
u
;
w
M
u

= Moment capacity of columns, wall
N
= Axial load of a column
= Axial load of a wall
N
max

= Axial compressive strength of a column
N
min

= Axial tensile strength of a column
P

= Seismic capacity index
P
0

= Index for basic seismic demand

296
P
g

= Seismic capacity index of ground floor
Q
E

= Lateral elastic force
Q
Y

= Lateral yield force;
Q
g

= Lateral load carrying capacity of the ground floor
k
Q
mu

= Flexural strength of k
th
vertical element
k
Q
su

= Shear strength of k
th
vertical element
R

= Ductility index
R
a
(T)

= Seismic load reduction factor
R
c

= Ductility index of a column
R
sc

= Ductility index of a short column
R
w

= Ductility index of a shear wall
S(T)

= Spectrum coefficient
T

= Natural period of the building
U

= Index for usage
V
t

= Total Equivalent Seismic Load (base shear),
Z

= Index for local soil conditions
W

= Total weight of the building
W

= Total weight above the storey concerned
a
1
; a
2
; a
3

= Strength reduction factors due to displacement compatibility
a
h

= Cross sectional area of lateral reinforcement
a
t

= Area of tension reinforcement of a column
= Area of tensile reinforcement at the walls end zone
a
g

= Total area of longitudinal reinforcement
a
w

= Total cross sectional area of a set of stirrups
a
Wy

= Total area of vertical reinforcement of a wall
b

= Width of compressive side of a column
b
e

= Equivalent thickness of wall

297
d

= Effective depth of column
f
c

= Compressive strength of concrete
h
0

= Clear height of column
i

= Number of the storey concerned
l

= Wall length
l
W

= Distance between center of wall and end zones
n

= Number of storeys of a building
s

= Spacing of lateral reinforcement

= Coefficient for inflection height

= Ductility ratio
se

= Equivalent lateral reinforcement ratio
t

= Tensile reinforcement ratio(%)
te

= Equivalent tensile reinforcement ratio of a wall
w

= Shear reinforcement ratio
0

= Axial stress on columns
0e

= Axial stress on wall
y

= Yield strength of longitudinal reinforcement
wy

= Yield strength of lateral reinforcement, stirrups

CHAPTER 2.
A0

= Effective ground acceleration
I

= Building importance factor
PPS-2006

= Performance Point Seeker 2006
T

= Characteristic spectral period
ks

= Soil modulus

298
q

= Allowable soil stress

CHAPTER 3.1 & 3.3
A = Constant of the test conditions and failure of the material
AF = Acceleration factor
CGS

= Global safety coefficient
E = Activation energy
n
E
= Activation energy for the degradation mechanism and for the
material
FRP

= Fibre Reinforced Polymers
GFRP

= Glass Fibre Reinforced Polymers
M
d

= Design moment
M
r

= Section resisting moment
N = Number of cycles
RFI = Resin Film Infusion
S, = Stress
T = Temperature in K
TS = Time Shift
TSF = Time Shift Factor
ULS = Ultimate Limit State
V = Degradation factor
V = Accelerated stress
k = Boltzmanns constant
n = Number of layers
t = Exposure time
' = Degradation rate

299
, = Constants of the material
, = Constants of the degradation mechanism
= Angle of orientation of the fibres
= Life-cycle at the reference temperature T
= Life-cycle at the elevated temperature T

CHAPTER 3.2
ijkl
A

= Fourth order strain tensor which contain the anisotropy
information
ijkl
A

= Fourth order stress tensor which contain the anisotropy
information
c = Composite material superscript
f = Fibre material superscript
fib
N
f ) ( = Nominal fibre strength
mat fib
N
f
) ( = Fibre-matrix interface nominal strength

mat
N
f ) ( = Matrix nominal strength
fib
R
f ) ( = New fibre strength
f
k
= Volumetric participation (parameter) of fibre in the composite
k
k

= Volumetric participation (parameter) of component k in the
composite
m
k
= Volumetric participation (parameter) matrix in the composite
m = Matrix material superscript
P
= Parallel component of the stress tensor
S
= Serial component of the stress tensor
c
ij
= Strain tensor for the composite
k
ij

= Strain tensor for component k of the composite
j

= Strain variation (perturbation) value
ij

= Stresses in the real anisotropic space

300
ij

= Stresses in the fictitious isotropic space
j

= Stress variation
P

= Parallel component of the strain
S

= Serial component of the strain

CHAPTER 3.4
CFRP = Carbon fibre reinforced polymer
E
frp
= Elasticity modulus of FRP
FRP = Fibre reinforced polymer
ITR = Internal transverse reinforcement
L.R. = Longitudinal reinforcement
LSR. = Low strength reinforced concrete specimens
NSR = Medium strength reinforced concrete specimens
T.R. = Transverse reinforcement
b = Width of cross-section
f
c
= Concrete standard cylinder strength
f
cc
= FRP jacketed concrete strength
f
co
= Concrete strength of the member
f
*
fu
= Tensile strength of FRP
f
l
= Lateral stress of FRP jacket
f
s,max
= Tensile strength of steel reinforcement
f
y
= Yield strength of steel reinforcement
h = Height of cross-section
n = Number of plies of FRP
t
f
= Nominal thickness of FRP

301
o
= Axial strain corresponding to unconfined concrete strength
cc
= Axial strain corresponding to jacketed member strength
ch
= Transverse strain of the FRP jacket
*
fu
= Ultimate rupture strain of FRP
h,rup
= 70% of
*
fu

y
= Yield strain of reinforcement
l
= Diameter of longitudinal bars
a

= Cross-sectional efficiency factor

= Pi
f

= Fibre ratio to concrete section
l
= Longitudinal reinforcement ratio

CHAPTER 3.5
BS

= Base score
CFRP = Carbon Fibre Reinforced Polymers
CM
= Adjustement factor, which introduce the spatial variation of the
ground motion in the evaluation process
CV
IO
= Cutoff value for IOPC
CV
LS
= Cutoff value for LSPC
DI
IO
= Damage score corresponding to IOPC
DI
LS
= Damage index or the damage score corresponding to LSPC
FRP = Fibre Reinforced Polymers
IO
CVR

= Immediate occupancy coefficient based on the number of storeys
above the ground level
IOPC = Immediate Occupancy Performance Classification
LS
CVR

= Life safety coefficient based on the number of storeys above the
ground level

302
LSPC = Life Safety Performance Classification
PG
IO
= Performance groupings yield for IOPC
PG
LS
= Performance groupings yield for LSPC
PS = Seismic performance score
SD1 = Spectral acceleration at the period of 1 sec
SDS = Spectral acceleration at short periods
VS

= Vulnerability score
VSM

= Vulnerability score multiplier
a, b, and c = Equation parameters
mnlsi = Minimum normalized lateral strength index
mnlstfi

= Minimum normalized lateral stiffness index
n

= Number of storeys
nrs

= Normalized redundancy score
or

= Overhang ratio
ssi

= Soft storey index

= Interstorey drift ratio
t
= Target roof drift ratio under the given elastic spectrum

= End rotation of the beam

CHAPTER 4.1
A = Beam cross-section
A
c
= Area of the quadrilateral
A
0
= Mass density per unit of length of the curved reference beam

303
C
me
= Material form of the elastic constitutive tensor
C
ms
= Material form of the secant constitutive tensor
C
mt
= Material form of the tangent constitutive tensor
C
se
= Viscous constitutive tensors in spatial description
C
sv
= Tangential constitutive tensors in spatial description
C
sv
ij
= Spatial forms of the reduced tangential constitutive tensors
D
y
= Yielding displacement
EDD = Passive energy dissipating devices
E
d
= Energy dissipation due to the addition of EDDs
E
D

= Energy dissipation due to inelastic behaviour in the structure
(including viscous effects)
E
K
= Absolute kinetic energy
E
L
= Absolute earthquake energy input
E
S
= Elastic strain energy
E
0
= Young undamaged elastic modulus
F = Damage yield criterion (scalar value)
F = Current beam referred to the straight reference configuration
F
0
= Deformation gradients of the curved reference
F
n
= Deformation gradient
F
p
= Yield function
F
y
= Yielding level
G
0
= Shear undamaged elastic modulus
G
f
= Tensile fracture energy
G
p
= Plastic potential function
G
Pf
= Specific plastic fracture energy of the material in tension
G(P)
= Scalar monotonic function to be defined in such way to ensure that
the energy dissipated by the material on a specific integration point
is limited to the specific energy fracture of the material

304
I = Identity tensor
I
0

= Second mass moment density per unit of length of the curved
reference beam
J
pq

= Jacobian of the transformation between normalized coordinates
and cross sectional coordinates
K = Stiffness
N
fibre
= Number of quadrilaterals of the beam cross section
N
p
,N
q

= Number of integration points in the two directions of the
normalized geometry of the quadrilateral
P = Asymmetric First Piola Kirchhoff (FPK) stress tensor
P
j

= Corresponding FPK stress vector acting on the deformed face in
the current beam corresponding to the normal t
0j
in the curved
reference configuration
P
m
, P
p
= Equivalent (scalar) stress
PRC = Prescast Reinforced Concrete
RC = Reinforced Concrete
SO = Rotation manifold
S
0

= First mass moment density per unit of length of the curved
reference beam
W
pq
= Weighting factors
d
= Internal state variable which measures the lack of secant stiffness of
the material
f
c
= Compression strengths
f
t
= Tension strengths
f
p
= Hardening function
g
f
= Fracture energy density
k
p
= Plastic damage internal variable
l
c

= Characteristic length of the fractured domain employed in the
regularization process
m = Stress couple vector
m
,S
= Body moment per unit of reference length
n = Stress resultant vector

305
n = Plastic dissipation
n
,S
= External body force per unit of reference length
n, r
= Parameters function of the of the tension and compression
strengths f
c
and f
t
t = Time
{t
}
= Orthogonal local frame
x = Position vector of any material point on the current reference beam
p

= Hardening parameter

= Current rotation tensor
0

= Orientation tensor
n

= Rotation tensor relative to the curved reference beam
maxt

= Values of the maximum dissipation in tension

(
0c
)
L
= Parts of the free energy density developed when the compression
limits are reached
(
0t
)
L

= Parts of the free energy density developed when the tension limit
are reached

= Scalar parameter

= Current centroid curve
0

= Spatially fixed curve
n

= Strain tensor

= Parameter

= Plastic consistency parameter
n

= Spatial curvature tensor and the rotation tensor relative to the
curved reference beam

= Hardening vector

CHAPTER 4.2
E = Elastic Youngs modulus
F

= Force

306
F
tot
= Total seismic force applied to the frame
I

= Moment of inertia
L

= Length of elements
M

= Moment
M
mmf

= Total moving mass of the frame
M
p
= Plastic moment
M
Rd
= Design value of resisting moment
P

= Force
P
y

= Yield strength of the connection
P
u

= Ultimate strength of the connection
S
de
(T)

= Elastic displacement response spectrum
S
e
(T)

= Elastic response spectrum
T

= Period of Vibration
W
pl

= Plastic modulus
a

= Clear distance between internal and external eye-bars
= Distance between the 2 formed plastic hinges
a = Acceleration
a
g
= Peak ground acceleration (PGA)
b

= Width of the pin connection
d

= Perpendicular distance between INERD Pin connection system and
beam-column node
d
r, el
= Displacement of the top of the column due to elasticity of the
column
d
r,pl
= Displacement of the top of the column due to plastic rotation of
the column
d
r
= Absolute displacement of the top of the column
d
r,max
= Maximum absolute displacement of the top of the column
f
y
= Yield stress
h

= Height of the pin connection

307

= Pin length (axial distance between external eye-bars)
m = Mass
t
ext
= Thickness of the external plates of the pin connection
t
nt
= Thickness of the internal plates of the pin connection

= Ratio of the clear distance between internal and external eye-bars to
the axial distance between external eye-bars

= Overstrength factor
lim

= Deformation capacity of the pin connection
y

= Deformation of the pin connection at yield capacity
II

= Deformation of the pin connection at ultimate capacity

= Pi

= Plastic rotation at the base of the column
max
= Maximum plastic rotation at the base of the column

= Ultimate plastic rotation at the base of the column

CHAPTER 4.3
C
0

= Modification factor to relate spectral displacement to the roof
displacement
C
1

= Modification factor to relate expected maximum inelastic
displacements to displacements calculated for linear elastic
response
C
2

= Modification factor to represent the effects of pinched hysteresis
shape, stiffness degradation and strength deterioration on the
maximum displacement response
C
3

= Modification factor to represent increased displacements due to
P- effects
CBF

= Concentric braced frames
D

= Lateral drift ratio
DCM

= Medium ductility class
E

= Elastic modulus of pins material
FEMA

= Federal emergency management agency

308
G

= Gravity loading
H

= Storey height
I

= Moment of inertia of the pin
INERD

= Innovations for earthquake resistant design
L

= Buckling length of the bracing
M

= Moment
M
p

= Plastic moment
MRF

= Moment resisting frames
N
b,Rd

= Buckling resistance of the diagonal
N
Ed

= Design force of the diagonal
N
Eser

= Design force of the diagonal at the damage limitation state
P

= Vertical load
P
Ed

= Capacity design force
PGA

= Peak ground acceleration
P
u

= Ultimate strength of the pin
P
u,Rk

= Characteristic ultimate strength of the pin
P
y

= Yield strength of the pin
P
y,Rk

= Characteristic yield strength of the pin
S

= Soil parameter
S
d(T)

= Ordinate of the design spectrum for the reference return period
S
De(T)

= Elastic displacement response spectrum
T

= Period
W
pl

= Plastic modulus of pin
a
= Clear distance between internal and external eye-bars
= foundation slab dimension
a
g

= Design peak ground acceleration
b

= Pin width

309
d
u

= Ultimate displacement
d
y

= Yield displacement
f
y

= Yield stress of pin
g

= Acceleration of gravity
h

= Pin height
i
min

= Minimum radius of gyration
l

= Pin length (axial distance between external eye-bars)
q

= Behaviour factor
s

= Second
t
ext

= Thickness of external eye bars
t
int

= Thickness of internal eye bars

= Horizontal displacement

= Overstrength coefficient

= Ratio of the clear distance between internal and external eye-bars to
the axial distance between external eye-bars

Mser

= Partial safety factor of resistance
M0

= Partial safety factor of resistance
lim

= Deformation capacity of the pin connection
II

= Deformation of the pin connection at ultimate capacity
target(roof)

= Target displacement of roof
y

= Deformation of the pin connection at yield capacity

= P- effect parameter

= Non-dimensional slenderness of the element

= Reduction factor for lower return period of seismic action
associated with damage limitation

= Angle of inclination of diagonal (from horizontal to diagonal)


310
CHAPTER 5.1
DBD = Displacement based design
D
I
= Base isolation deformation
D
S
= Structures deformation
D
T
= Total structural deformation of the base isolated structure
E = Modulus of Elasticity

= Weight per unit volume
Pga

= Peak ground acceleration
S
a

= Spectral acceleration
S
d

= Spectral displacement
V
I
= Base shear force
m
I
= Effective mass value evaluated for the global deformation model

CHAPTER 5.2
ER = Electrorheological dampers
H(u) = Heaviside step function of u
J = Cost function
L = Lagrangian
MR = Magnetorheological dampers
T
v
= Response time of the controllable damper
c = Damping coefficient
c
d
= Describe the post yield behavior
d
0

= Constant parameter to describe the pre-yield behavior of the
device
f
0
= Controllable yield force
f
MR
= MR damper force

311
k = Stiffness
m = Mass
r = Structural relative displacements
t = Time
u(t) = Control function
v
0

= Constant parameter to describe the pre-yield behavior of the
device
x(t) = State vector
z = Ground displacement

= Design parameter for the control rule

CHAPTER 6
CR

= Coefficient of restitution
DCM

= Ductility Classes Medium
FRP

= Fibre-reinforced polymers
PRD

= Pounding Reduction Device
K

= Stiffness
SDOF

= Single Degree Of Freedom
VRd

= Design value of shear resistance
VSd

= Design value of shear action effect
cc

= Dashpot constant of the connector modelling
cp = Dashpot constant of the pounding modelling
e

= Stand-off distance

312
gp

= Static separation between structures
kc

= Spring constant of the connector modelling
kp

= Spring constant of the pounding modelling
m1

= Mass of the first building
m2

= Mass of the second building
u1

= Displacement of the first building
u2

= Displacement of the second building
1

= Velocity of the first impacting body
'1

= Final velocity of the first impacting body
2

= Velocity of the second impacting body
'2

= Final velocity of the second impacting body

= Delta

= Interstorey drift sensitivity coefficient

CHAPTER 7
EI = Flexural stiffness
EI
sec
= Secant flexural stiffness
F = Force
H = Height of the structure
M = Moment
M
Sd
= Design bending moment
M
Rd
= Flexural capacity

313
N = Axial force
d
bl
= Diameter of longitudinal reinforcement
q = Behaviour factor (q-factor)
s = Spacing of transverse reinforcement

= Curvature

= Displacement
u
= Maximum displacement
y
= Yield displacement

= Strain

= Distortion
max
= Maximum distortion
o
= Overstrength factor

= Stress

Guideline For Seismic Vulnerability Reduction in The Urban Environment

Transféré par

Informations du document

Description originale:

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

Guideline For Seismic Vulnerability Reduction in The Urban Environment

Transféré par

Droits d'auteur :

Formats disponibles

EDITOR: A.

, relates the stresses in the fictitious isotropic space (

, does the same with the strains.

) ( is the fibre-matrix interface

) ( = Fibre-matrix interface nominal strength

Vous aimerez peut-être aussi