Vous êtes sur la page 1sur 156

SPECIAL FEATURES

OUTLOOK 2014
ENVIRONMENTAL
OPPORTUNITY CRUDE PROCESSING
AUTOMATION & CONTROL
REFINING
GAS PROCESSING
PETROCHEMICALS
PETROLEUM TECHNOLOGY QUARTERLY
ptq
Q1 2014
cover and spine copy 18.indd 1 10/12/2013 13:02
1
2
1
1
_
e
www.engineering-solutions.airliquide.com
Whatever the impurity, whatever the composition, Air Liquide Global E&C Solutions
has the right treatment.
Does your raw natural gas contain hydrogen sul de,
carbon dioxide, mercaptans or more?
The composition of natural gas varies
tremendously: almost every source
contains a dierent blend of impurities.
The options for treatment are almost
as diverse. Thats why oering a solution
specically designed for your gas eld
is crucial. We as your partner of choice
provide solutions for all types of natural gas,
including associated and unconventional
gas, from a single source. Customised
and e cient.
AIR_2362_008_HydrocarbonEngineering_0413_RZ.indd 1 07.03.13 17:00 air liquide.indd 1 11/12/2013 19:53
2014. The entire content of this publication is protected by copyright full details of which are available from the publishers. All rights
reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic,
mechanical, photocopying, recording or otherwise without the prior permission of the copyright owner.
The opinions and views expressed by the authors in this publication are not necessarily those of the editor or publisher and while every care
has been taken in the preparation of all material included in Petroleum Technology Quarterly the publisher cannot be held responsible for any
statements, opinions or views or for any inaccuracies.
3 A window half-open
Chris Cunningham
5 Outlook 2014

17 ptq&a

29 Processing Trends: Whose responsibility is reliability?
Robin Peatey Foster Wheeler

39 HS-FCC for propylene: concept to commercial operation
Nicolas Lambert Axens
Iwao Ogasawara JX Nippon Oil & Energy
Ibrahim Abba Saudi Aramco
Halim Redhwi King Fahd University of Petroleum & Minerals
Chris Santner Technip Stone & Webster Process Technology

47 Optimising distillation column product quality
Jose Bird and Darryl Seillier Valero Energy Corporation
55 Extending the treatment of highly sour gases: cryogenic distillation
Franois Lallemand, Gauthier Perdu and Laurent Normand Prosernat
Claire Weiss Total
Julia Magne-Drisch and Sebastien Gonnard IFP Energies nouvelles
63 Optimising hydrogen sulphide scavenging
Giacinto Grande and Alessandra Berra Chimec
69 Overcoming high conductivity in process condensate
Tapas Kanti Samaddar and Gregory M Lilburne The Bahrain Petroleum Company BSC(c)
Calum Mcintosh and Mileta Babovic Johnson Matthey
81 Mercury treatment options for natural gas plants
Satyam Mishra UOP, a Honeywell Company
87 Role of FCC catalyst in renery protability
Ivan Chavdarov, Dicho Stratiev, Ivelina Shishkova and Rosen Dinkov Lukoil Neftohim Burgas JSC
Vladimir Jegorov Grace Catalysts Technologies
Petko Petkov University Prof Dr Assen Zlatarov Burgas
93 Field experience with a Claus furnace checker wall
Jeffrey Bolebruch Blasch Precision Ceramics
Mossaed Y Al-Awwad Saudi Aramco
Meng-Hung Chen CPC Corporation
101 Characterising and tracking contaminants in opportunity crudes
Robert Ohmes KBC Advanced Technologies
115 New process arrangements for upgrading heavy oils and residua
Sepehr Sadighi, Reza Seif Mohaddecy and Kamal Masoudian
Research Institute of Petroleum Industry
123 Increasing conversion and run length in a visbreaker
Matteo Virzi ISAB
Marco Respini Baker Hughes
133 Troubleshooting renery equipment with multiphase CFD modelling
Grant Niccum and Steve White Process Consulting Services
139 High-delity operator training simulators
Martin Sneesby APESS
147 What your transmitter may not be telling you
Eric Moore Swagelok Capital Projects Company
Part of a major upgrade and expansion project, the coker at Repsols Cartagena renery in Spain enables the site to
process heavy and sour crudes. Photo: Repsol
Q1 (Jan, Feb, Mar) 2014
www.eptq.com
p
t
q
Y L R E T R A U Q Y G O L O N H C E T M U E L O R T E P
ed com copy 3.indd 1 12/12/2013 14:45
C
M
Y
CM
MY
CY
CMY
K
KBC Adv - PTQ Q1 2014.pdf 1 12/10/2013 9:29:32 AM
KBC.indd 1 10/12/2013 16:06
T
he European Union has arguably
been the global leader in biodiesel
production and use, with overall
biodiesel production increasing from 1.9
million tonnes in 2004 to nearly 10.3 million
tonnes in 2007. Biodiesel production in the
US has also increased dramatically in the
past few years from 2 million gallons in
2000 to approximately 450 million gallons
in 2007. According to the National Biodiesel
Board, 171 companies own biodiesel
manufacturing plants and are actively
marketing biodiesel.1. The global biodiesel
market is estimated to reach 37 billion
gallons by 2016, with an average annual
growth rate of 42%. Europe will continue to
be the major biodiesel market for the next
decade, followed closely by the US market.
Although high energy prices,
increasing global demand, drought
and other factors are the primary
drivers for higher food prices, food
competitive feedstocks have long
been and will continue to be a major
concern for the development of biofu-
els. To compete, the industry has
responded by developing methods to
increase process effciency, utilise or
upgrade by-products and operate
with lower quality lipids as
feedstocks.
Feedstocks
Biodiesel refers to a diesel-equivalent
fuel consisting of short-chain alkyl
(methyl or ethyl) esters, made by the
transesterifcation of triglycerides,
commonly known as vegetable oils or
animal fats. The most common form
uses methanol, the cheapest alcohol
available, to produce methyl esters.
The molecules in biodiesel are pri-
marily fatty acid methyl esters
(FAME), usually created by trans-
esterifcation between fats and metha-
nol. Currently, biodiesel is produced
from various vegetable and plant oils.
First-generation food-based feedstocks
are straight vegetable oils such as
soybean oil and animal fats such as
tallow, lard, yellow grease, chicken fat
and the by-products of the production
of Omega-3 fatty acids from fsh oil.
Soybean oil and rapeseeds oil are the
common source for biodiesel produc-
tion in the US and Europe in quanti-
ties that can produce enough biodie-
sel to be used in a commercial market
with currently applicable
PTQ Q1 2014 3
Editor
Chris Cunningham
editor@petroleumtechnology.com
Production Editor
Rachel Storry
production@petroleumtechnology.com
Graphics Editor
Rob Fris
graphics@petroleumtechnology.com
Editorial
tel +44 844 5888 773
fax +44 844 5888 667
Business Development Director
Paul Mason
sales@petroleumtechnology.com

Advertising Sales
Bob Aldridge
sales@petroleumtechnology.com
Advertising Sales Ofce
tel +44 844 5888 771
fax +44 844 5888 662
Publisher
Nic Allen
publisher@petroleumtechnology.com
Circulation
Jacki Watts
circulation@petroleumtechnology.com
Crambeth Allen Publishing Ltd
Hopesay, Craven Arms SY7 8HD, UK
tel +44 844 5888 776
fax +44 844 5888 667

PTQ (Petroleum Technology Quarterly) (ISSN
No: 1632-363X, USPS No: 014-781) is published
quarterly plus annual Catalysis edition by Crambeth
Allen Publishing Ltd and is distributed in the US
by SP/Asendia, 17B South Middlesex Avenue,
Monroe NJ 08831. Periodicals postage paid at New
Brunswick, NJ. Postmaster: send address changes to
PTQ (Petroleum Technology Quarterly), 17B South
Middlesex Avenue, Monroe NJ 08831.
Back numbers available from the Publisher
at $30 per copy inc postage.
Vol 19 No 1
Q1 (Jan, Feb, Mar) 2014
A window
half-open
A
s last year drew towards its close, two closely related but seemingly
contradictory items of news came through. First, Shell announced that
it would not go ahead with construction of a gas-to-liquids (GTL) plant
near Baton Rouge in Louisiana, this despite plentiful supplies of feedstock
natural gas at highly advantageous prices. Shell was thought to be reticent about
infated labour and construction costs in the US Gulf Coast region; the likely bill
for starting up the 140 000 b/d facility was already at the $20 billion mark.
Construction costs did not seem to put off South African Sasol when a few
days later it announced the award of engineering and design contracts to
Paris-based Technip for its proposed 100 000 b/d GTL plant to produce diesel
and petrochemicals at Lake Charles, for a proportionally comparable outlay.
So what tipped the balance for or against, depending on which company
was making the decision? It is no coincidence that the two have big stakes in
GTLs fagship production sites, Pearl and Oryx in Qatar. Both certainly went
ahead with those projects, Shell with Pearl and Sasol with Oryx.
Although Shell is known to have encountered severely infated start-up
costs in the Arabian/Persian Gulf too, what makes the difference in the US is
more likely to be the price differential between oil and gas.
Right now, gas is cheap. In Qatar, the states huge, offshore North Field
resource makes the Gulf state the worlds richest per capita. Qatar can export
its LNG product at whatever the market will allow for it to remain competi-
tive but at home the price is always right for GTL, whatever the market says.
In the US at the time of Shells Lake Charles decision, the gas price spread
topped out at a lowly $3.80/million BTU while West Texas Intermediate
crude was trading around $98/bbl, a price set to rise on lower stocks as
refneries came out of turnaround. If ever the time was right for GTL this
would seem to be it, but not for Shell.
As usual these days, the answer to the question lies somewhere in the
shale. Low gas prices appear set to stay, in North America at least and per-
haps across the globe as local technical and environmental issues surround-
ing shale gas reserves are resolved. For the time being, North America
appears a natural choice for expansion of the worlds GTL capacity.
Meanwhile, although the high price of oil remains relatively stable the bets
are not on it remaining that way.
On the face of it, making fuels and chemicals from natural gas rather than
petroleum makes sense. The white wax produced by Fischer-Tropsch synthe-
sis in a GTL plant is a convenient and unmessy starting point for a broad
range of products. Desulphurising natural gas is a lot cheaper than hydrogen
production and hydrodesulphurisation in a refnery, and fuels derived from
GTL are not only inherently low in sulphur but high in cetane number. You
could say: gas is too good to burn.
The problem for proponents of GTL technology is that the world is set up
to refne petroleum for its automotive fuels in particular. GTL is a window
of opportunity technology that requires the right price differentials without
established competition from oil refning on its doorstep. If a window opens
for GTL, you need to be certain that it will stay open.
CHRIS CUNNINGHAM
p
t
q
Y L R E T R A U Q Y G O L O N H C E T M U E L O R T E P
ed com copy 3.indd 2 12/12/2013 10:50
As the R & D team lead in oil re ning catalyst development for CRITERION, Ferry Winter creates
ground-breaking catalysts from conception through commercial phases for naphtha hydrocrackers.
Working with a team of zeolite experts, he helped create Z-NP10 and Z-NP20, two recent
hydrocracking catalysts that make use of our new, advanced zeolite Y technology. These catalysts
deliver performance through world leading activity and naphtha/kerosene selectivity, which can
improve pro tability. These two seeds grown by Ferry and his team are now bearing fruit for our
re nery customers around the world.
You put a thousand seeds in the ground
and only one might grow. We make
sure its a winner.
Meet Ferry Winter:
Growing Serious Science, Research Scientist.
Leading minds. Advanced technologies.
www.CRITERIONCatalysts.com
criterion.indd 1 09/12/2013 11:35
Maria van der Hoeven
Executive Director
International Energy Agency
T
he world is undergoing an
era of economic rebalancing
toward emerging markets,
and that is especially true within
the oil market, where shifting
patterns in consumption and
production further complicate
the broader trends. The global
energy map is not just shifting
east. Traditional producers like Saudi Arabia are seeing
booming demand, while heavily consuming markets
such as North America are experiencing production
revolutions thanks to new techniques and technologies.
New trading patterns and technological advances
also mean new scope for international market reform
creating a common incentive for market stability
and investment promotion. The next fve years will
consolidate the rise of the non-OECD in virtually
every aspect of the oil market, and most of the growth
(though by no means all of it) will come from East-of-
Suez economies. The idea that the emerging markets
and developing economies would eventually overtake
the OECD in oil demand is nothing new, but it is
happening faster than expected.
One of the consequences of those changes is a
transformed global product supply chain. New, non-
OECD mega-refneries are challenging OECD refning
economics, at least beyond the US. Their expanding
reach is accelerating the globalisation of the product
market, particularly in the case of refneries geared
toward export. With it come the beneft of greater
market fexibility in the dispatch of product supply,
but also longer supply chains, higher
reliance on stocks to meet demand,
diminished visibility in inventory
levels, increased disruption risks,
reduced market transparency and,
possibly, greater price variation
between key markets and also
between seasonal peak and troughs
in demand.
The IEAs Medium Term Oil Market
Report 2013, released in May, offers
projections to 2018. As refnery
capacity additions are forecast to
increase faster than world demand
over the medium term, the persistence of excess
refning capacity is expected to weigh on refning
margins, severely affecting the less effcient and older
refneries in the world. The overall upgrading ratio
gradually increases from 44% in 2012 to 47% in 2018 as
new refneries focus mainly on heavy crude oil
processing, mostly from the Middle East and Latin
America. The capacity expansions in Asia and the
Middle East, however, represent two different distinct
investment strategies. Whereas Asian capacity
development is meant to service rapidly rising
domestic demand, Middle East exporters (mostly in
joint venture with OECD refners or Chinese
companies, and mostly in Saudi Arabia) are climbing
the value chain and expanding into products and
petrochemicals.
This has put OECD refners in particular
European ones into a tight spot. Facing weak
demand, tighter regulations, and an aging capital
stock, these refners have little choice but to increase
their competitiveness through restructuring and
consolidation. The refning sector in OECD Americas is
undergoing a complete restructuring, with each
regional district trying to optimise its crude slate
between light domestic crude oil and heavy imported
oil. High sustained margins and the outlook for
continued growth pushes US refners to invest in deep
conversion and light oil processing. Recent
developments in Spain and Portugal show that
investment into deep conversion units can indeed
present a proftable alternative to closures. Yet despite
these efforts, the overcapacity which exists today is
bound to force a continuation of the capacity shut-ins
we have seen since 2008. Over that period, 15
European refneries have closed with a total capacity of
1.7 mb/d, and more are expected in the coming years.
Japanese refners will close a total of
around 800 kb/d by next year, in
line with government regulations
aimed at increasing conversion
yields.
Looking out to 2035, the World
Energy Outlook 2013 shows the
global refning sector set to
experience continued turbulence
amidst declining OECD demand and
Asian growth. Strains on the refning
system are amplifed by the
changing composition of feedstocks.
A growing share of oil supply
www.eptq.com PTQ Q1 2014 5
Outlook for 2014
What are the important trends affecting the downstream processing industry this year?
Executives and experts forecast challenges and prospects that could affect protability
Figure 1 Regional share of CDU expansions
Other Asia
14%
Latin
America
14%
Middle East
22%
Other
5%
China
45%
outlook copy 2.indd 1 11/12/2013 15:57
bypasses the refning system altogether, including most
natural gas liquids as well as oil products produced
directly from gas or coal. As a result, global demand
for refned products grows by only 10 mb/d to 2035
less than the growth in both overall liquids demand
(16.8 mb/d) and net refnery capacity additions (13
mb/d). Over-capacity in the sector is therefore a long
term phenomenon as well, but prospects for OECD
refners vary.
The outlook for the United States is helped by the
increasing availability of local crude, but the benefts to
refners vary by location thanks to new supply sources
and ongoing infrastructure constraints. Still, as the net
North American requirement for imported crude all
but disappears by 2035, the region becomes a large
exporter of oil products. On the other hand, medium
term European diffculties are compounded to 2035 by
declining local crude production (particularly from the
North Sea), product demand that is heavily skewed
towards diesel, and disappearing export markets for
gasoline.
For the non-OECD, the longer outlook period also
reveals new trends. Middle Eastern refning capacity
additions, initially focused on boosting product
exports and adding value to oil otherwise exported
as crude, turns to domestic needs as regional demand
rises to 2035. And in Asia, the region emerges as
the unrivalled centre of global trade, drawing in a
rising share of crude not only from the Middle East,
but also from Russia, Africa, Latin America,
and Canada. Massive additions to refning capacity
are still not suffcient to meet regional demand, and
both China and India are net oil product importers in
2035.
The changes occurring in the refning balance are
refective of both the traditional refning model in
which increasingly consumptive countries prefer to
import crude and refne it for domestic production
and also a new phenomenon in the rise of global
refning centres. The increase of product trade refects a
desire of some exporters to capture the added value of
exporting refned product, and also an effort to
diversify into highly specialised industries such as
petrochemicals. While some export oriented refning
capacity additions in the Middle East could be needed
to meet growing regional demand after 2025 or so, the
trend could catch on in North Africa, West Africa,
Brazil, or Russia with even greater implications for
the future of OECD refning as global product trade
increases more substantially. And all of this could also
have implications for global oil security a core
mission of the IEA as countries consider how best to
manage strategic oil stock balances and emergency
policy. Longer and more specialised product trade
routes can also render supply disruptions more acute.
Like with globalisation more generally, increasing
international oil trade will also mean the increased
need for international cooperation to secure
and manage that trade and thus the need for the
kind of international energy governance provided by
the IEA.
Daniel M McCarthy
President, CB&Is Technology
Operating Group
Member, Board of Directors, Chevron Lummus
Global LLC
T
he powerful impact of
technology changes in todays
energy markets has created
major paradigm shifts in barely 10
years. As a result of these technical
innovations, billions of dollars in
investments are now moving to
the sources of new energy supplies
with somewhat surprising results
and consequences.
Some may call these disruptive technologies, but
forward thinkers will realise they are opportunities to
improve productivity. As an example, the planning of
large, new olefn or liquefed natural gas capacity in
the United States is no longer considered so radical,
now that the economical natural gas feedstocks are
produced and available in the US.
This is good news for refning, petrochemical and
energy markets as we regularly see examples where a
technology innovation can dramatically shift products
or businesses. In some cases, it may take 20 or 30 years
for the technological changes to work through the
system and be fully implemented. Chevron
commercialised Isodewaxing catalysts to produce
higher quality Group II and III lube base oils 20 years
ago, but solvent dewaxed Group I lube oils are still
common today. For new technology deployment to be
rapid, its scale and beneft must be very large and
widespread.
The deployment in the US of new hydraulic fracturing
(fracking) technology for shale oil and gas deposits has
been both rapid and widespread, generating enormous
growth in new oil and gas supplies. We could assume
that other countries such as Russia, China and
Argentina, with similar or even greater shale assets,
would also be candidates for such rapid growth.
However, in these countries the investment capital and
technology deployment have been slow and limited, not
rapid and widespread. These three markets should
eventually catch up to the new shale technology
deployment, but probably not in this decade.
The discovery and production of oil and gas from
shale has generated a new dynamic and neither oil
producers, refners, nor petrochemical operators are
sheltered from its ripple effects. The rush to purchase
and produce attractive shale oil assets has been
predicated on US crude oil prices of approximately
$100/bbl for West Texas Intermediate (WTI) crude oil.
The current production rates have driven crude oil
prices below $100/bbl. Further rapid increases in supply
levels through new oil producing wells or oil associated
with gas produced for LNG could cause WTI to fall to
$80/bbl or less, resulting in a signifcant reduction in
new drilling and curtailment of supply growth.
Since 2003, the impact on refning is equally dramatic
6 PTQ Q1 2014 www.eptq.com
outlook copy 2.indd 2 13/12/2013 11:44
For more information about UOP adsorbents, visit www.uop.com/adsorbents
2013 Honeywell International, Inc. All rights reserved
too late
UOP adsorbents cut your risk of downtime and equipment failure
due to corrosion.
Minimize contamination with proven UOP adsorbents. With a vast portfolio of molecular
sieve and activated alumina adsorbents and as the designer of many plants in operation
today, UOP has been at the forefront of contaminant removal for more than 60 years.
UOP serves a variety of process types and industries through a wide range of adsorbent
solutions backed by the support and technical expertise of a comprehensive team of
experts. So dont put your equipment and processes at risk. Trust UOP for the exact
adsorbent solution you need before its too late.
uop too late.indd 1 09/12/2013 11:33
as light paraffnic crude oils from shale have displaced
more naphthenic crude oils in the refners diet.
Combine this crude oil supply quality change with
declining gasoline demand in the US and Europe and
the cost of gasoline octane has risen to historic levels.
As illustrated in Figure 1, the cost to raise one barrel of
gasoline by one octane point has almost doubled in the
past 10 years. We expect this to drive scientists and
engineers to develop new gasoline reforming
technology and catalysts.
The worldwide petrochemical industry accounts for
feedstock demand of about 6% of the total natural gas
and liquid fuels produced worldwide. The
petrochemical industry represents a value of roughly
$3 trillion in economic output. This sector also feels the
economic ripples from growing US shale oil and shale
gas production. We have already mentioned how
primary olefn demand is driving producers to secure
low cost supplies of ethane and LPG from shale gas.
For petrochemical producers securing feedstock from
refneries, the changes are equally dramatic.
US aromatic derivative petrochemical producers have
typically purchased high octane gasoline blending
streams containing benzene, toluene and xylene (BTX)
as feedstocks for aromatics extraction and manufacture
of aromatic polymers such as polystyrene, and chemical
derivatives such as paraxylene. As seasonal gasoline
demand would rise or fall, the price and supply of high
octane BTX would follow along to meet the octane
demand of the gasoline market. As seasonal gasoline
demand dropped off, excess BTX streams would be
available at lower prices for aromatic chemical
producers, until recently.
Figure 2 shows the trends of US mixed xylene and
gasoline prices. They generally track each other. From
2004 to 2010, the average price spread was
approximately $30/bbl but as octane demand and
value have increased since 2011, the spread has
increased to about $50/bbl. This change has made it
more expensive for aromatic petrochemical producers
to secure feedstocks. It has forced producers to secure
less expensive feedstocks elsewhere and in turn made
aromatic polymers less competitive with olefnic
8 PTQ Q1 2014 www.eptq.com
polymers. If this trend continues, new reforming and
aromatics extraction capacity will be constructed in
locations with cost competitive feedstocks.
There are many examples in our own research and
development activities where we expect and search for
breakthrough technology innovations that can
dramatically change fuels and petrochemical products
and markets. Some future breakthrough technologies
may include producing ethylene and propylene
directly from methane, direct coupling of methane
derived methyl groups into hydrocarbons, full
hydrocracking conversion of petroleum residues to
light clean products, biofuels innovations, and more
cost effcient polymers and specialty materials.
In the oil industry, the Hubbert peak theory is
named after American geophysicist M King Hubbert,
who in 1956 observed that the amount of oil under the
ground is fnite; therefore the rate of discovery which
initially increases quickly must reach a maximum and
decline. This theory has worked very well worldwide
describing declines in oil output until the fracking
revolution pushed new reserves into the equation and
pushed out the date of peak oil by many years. Such
breakthroughs are not common. However, when they
occur, they create many new opportunities and are a
welcome reward for those who drive change.
ISODEWAXING is a registered trademark of Chevron.
Rajeev Gautam
President & CEO
UOP LLC, A Honeywell Company
L
ooking back on 2013, it
was an impressive and
eventful year for the refning,
petrochemicals, gas processing,
and renewables industries.
Global production of heavy
crudes and diesel demand
continued to grow, as did demand
for polyester, plastics and detergents. And this has
been the case especially in developing economies
2.2
2.4
2.0
1.8
1.6
1.4
1.2
$
/
o
c
t
a
n
e

b
b
l
1.0
J
u
l

1
9
9
8
A
p
r

2
0
0
1
J
a
n

2
0
0
4
O
c
t

2
0
0
6
J
u
l
y

2
0
0
9
A
p
r

2
0
1
2
D
e
c

2
0
1
4
Figure 1 US retail Octane Index Barrel (OIB) price
150
200
175
125
100
75
50
25
$
/
b
b
l
0
S
e
p

2
0
0
2
J
a
n

2
0
0
4
M
a
y

2
0
0
5
O
c
t

2
0
0
6
F
e
b

2
0
0
8
J
u
l

2
0
0
9
N
o
v

2
0
1
0
A
p
r

2
0
1
2
A
u
g

2
0
1
3
D
e
c

2
0
1
4
USGC mixed xylenes contract
NY reg. gasoline spot
Figure 2 Mixed xylene pricing vs gasoline
outlook copy 2.indd 3 11/12/2013 15:58
relentless
UOP has been serving catalyst customers for
decades, and constant innovation ensures well
be there for decades to come.
Pilot plants. A true commitment to fast, thorough research and
development. A vast team of experienced, dedicated engineers
and scientists. The reasons to choose UOP catalysts are many,
but perhaps the biggest of all is that UOP never quits innovating. When you use UOP
catalysts, youll know youre using products specically designed to help optimize
your operation. UOP invented the rening and petrochemical technology used in most
operations today, and that knowledge and expertise comes through in every catalyst we
develop. Youll always get the high stability and global support your business needs, and
youll be working with a company that will keep you on the leading edge year after year.
For more information about UOP catalysts, visit www.uop.com/catalysts.
2013 Honeywell International, Inc. All rights reserved
SPM-UOP-35 Adsorbents Ad_v6.indd 1 2/27/13 2:11 PM
uop restless.indd 1 09/12/2013 11:32
where refners look for methods to increase production
of petrochemicals. Technological advancements in
natural gas have made the economical development of
shale and offshore gas resources possible to a greater
extent than ever before. Meanwhile, regulation,
technology and economics are creating incentives for
producers to ramp up production of renewable fuels.
But 2013 was not immune to challenges. While some
of these will continue into 2014, increased population
growth and middle class expansion will create higher
demand for energy sources such as natural gas, coal
and oil.
Government regulations will mean even tighter fuel
specifcations and higher renewable content. And the
market will continue to demand higher yields at lower
costs. While these challenges will vary by region, the
whole world will feel the impact.
In nearly every case, process technologies provide the
most economical, effcient and sustainable solutions to
directly meet these challenges.
Within refning, heavy crude production will
continue to grow and demand for low value residual
fuels will decline. There will be a growing need for
technologies such as UOPs Unifex process a slurry
hydrocracking technology that converts low quality
residue streams to high quality transportation fuels
while minimising byproducts, achieves a maximum
conversion of 90%, and produces an attractive
feedstock for subsequent high quality diesel
production.
In addition to maximising the yield from every barrel
of oil, refners are looking for ways to increase
petrochemicals production to meet global demand.
The global propylene gap will continue to grow as
traditional crackers shift to lighter, lower propylene
feedstocks such as ethane. There is strong demand for
propane hydrogenation technologies such as UOP
Olefex process, which produces on-purpose propylene
from propane at a very high yield with the
lowest cash cost of production and the highest return
on investment.
In areas with limited availability to conventional
olefn feedstocks, cheaper, more abundant feedstocks
such as coal will help producers address demand in
new ways. The UOP/Hydro Methanol-to-Olefns
(MTO) process converts methanol derived from non-
crude oil sources such as coal or natural gas to
ethylene and propylene at very high effciencies.
Natural gas not only will become a more common
ingredient for high value petrochemical feedstocks, but
it is on track to become one of the worlds top energy
sources.
Global unconventional gas resources are estimated to
yield 250 years of supply. Effcient contaminant
removal, purifcation and liquids recovery will be
essential steps to effciently monetise natural gas
resources.
The UOP Russell product line offers modular plants
that enable shale and conventional natural gas
producers to remove contaminants and recover high
value natural gas liquids used for petrochemicals and
fuels. UOP is continuously improving techniques for
gas treating and contaminant removal to ensure that
the gas can be cleaned to liquefed natural gas
specifcations.
In addition to natural gas, new, more effcient
alternative energy resources are on the horizon. These
renewable fuels and chemicals will support growing
global energy needs, while also addressing
environmental concerns.
UOPs frst Ecofning unit for renewable diesel
production is on-line, with more units in design and
construction. This technology is preferred by refners
due to the drop-in nature of the fuel product and the
ability to leverage existing infrastructure.
In 2014 and beyond, the oil and gas industry will be
driven by the development of global technologies that
are highly effcient, use cost advantaged resources and
provide compelling environmental benefts.
Eric Benazzi
Vice President Marketing & External
Communication
Axens
A
t the end of the year
2013, our future seems
uncertain. Contradicting
economic indicators are preventing
professionals from identifying a
clear trend for 2014.
Global GDP growth should reach
3.6% in 2014. This indicates that
the global economy continues to
recover post-2013, where growth should settle at 2.9%.
However, after a boost in capital infows, emerging
economies are facing a diffcult situation. This
monetary abundance led to the formation of credit
bubbles in Brazil, Russia, India, China and Turkey,
which increased infationary pressures.
The necessary resorption of fnancial imbalance in
emerging nations has led to a slowdown in their
domestic demand and activity, which was exacerbated
by a decrease in world trade due partially to Europes
poor economic health. To measure the decrease in
trade intensity it can be noted that exports as a share
of the world GDP, which rose steadily from the mid-
1980s to 2008, has remained stable since, slightly above
30%.
The growth of emerging nations import volumes
may therefore continue to decline, as it did in 2012,
holding back the driving force behind world trade if
this crisis was not curbed.
While writing these lines, GDP in the US is forecast
to grow 2.6% in 2014. Exploration and development of
unconventional gas and oil resources gave the US a
competitive advantage allowing them to stimulate
their petrochemical industry while having cheap
energy.
However, the latest episode in the US budget crisis
raises concerns about a possible recovery in 2014.
10 PTQ Q1 2014 www.eptq.com
outlook copy 2.indd 4 11/12/2013 15:58
cbi.indd 1 10/12/2013 09:54
limiting factors love
limitless possibilities
Overcome limiting factors affecting refnery capacity and
operating fexibility with BASF innovative FCC products,
services and solutions. Our products deliver value to
enhance sustainability and performance.
At BASF, we create chemistry for a sustainable future.

Catalysts

Co-Catalysts

Additives

Services and Solutions


www.catalysts.basf.com/refning
basf.indd 1 6/12/12 13:23:11
Indeed, Americas debt ceiling has only been raised for
a few months. Lack of a compromise in February
would lead to a destabilisation of the US economy that
would compare to the 2008 subprime mortgage crisis.
Such a scenario was deemed unthinkable given its
devastating consequences.
So, which indicators can we reasonably count on?
The global fuel and petrochemical demand should
experience growth, sustained by a growing world
population and its aspiration to a greater mobility and
lifestyle. Growth in the demand for refnery products
could increase by 1.1 million b/d, to reach a 2014
global oil demand of about 92 million b/d. On-road
diesel will continue to drive the demand, calling for
additional hydroprocessing capacities.
Product specifcations and environmental quality
standards will continue to tighten worldwide calling
for advanced technologies and innovative solutions.
Impulse, our latest hydroprocessing catalyst
technology and Symphony, our latest reforming
catalyst, are the result of a long term commitment to
clean fuels and our answer to market trends.
Feedstock and streams produced in refning,
petrochemical sites and natural gas felds are
continuously requiring solutions to remove
contaminants subject to stringent requirements and
emission limits. As a result, we deliver a complete
range of technologies and products and Claus catalysts
enabling the removal of all kind of contaminants such
as Cl, F, As, Si, Hg, H
2
S, RSH, and so on.
Demand in main petrochemical products should
keep growing at a rate of 4% to 5%. In regions where
demand is high or resources abundant, the
construction of mega scale complexes integrating
refning and petrochemical units will allow recovering
valuable products from low value heavy feeds and
reaching higher proftability. Regarding projects where
multiple technologies are involved, Axens Integrated
Process Solution is nowadays a must.
Those integrated mega scale complexes will continue
to require the latest technologies to convert the bottom
of the barrel and to maximise motor fuel yields,
notably diesel. VR ebullated bed hydrocracking plants
combined with solvent deasphalting or cokers units
and VGO hydrocrackers will play a key role as well as
scheme combining AR/VRDS and RFCC units when
naphtha/gasoline production is targeted.
In Europe, where refneries are facing strong
competitive pressure, energy costs represent more than
50% of operating expenses, so any improvement in this
feld will improve competiveness. Many new refning
entrants on the market have invested in the assets and
developed their ability to process a higher number of
crudes. In order to help EU refners on this path,
Axens revamping offer helps improve and upgrade
their existing assets and assists in improving process
energy effciency in order to increase competitiveness
at a minor risk.
Innovation will continue to be the best tools to
respond to market changes and to improve
proftability.
Roberto Penno
Chief Executive Ofcer
Global Engineering & Construction Group
Foster Wheeler
W
e are certainly seeing
robust activity in all of
the hydrocarbon related
business sectors in which Foster
Wheelers Global Engineering
and Construction Group operates:
onshore and offshore upstream oil
and gas, midstream/LNG, refning,
chemicals, and power, as well
as seeing continued investment in pharmaceuticals,
power, environmental and minerals and metals, where
we also operate.
This year there has been a very strong focus in the
industry on North America, with more projects
moving forward in both petrochemicals and chemicals
and in LNG liquefaction and more emerging too. Asia
and the Middle East also continue to see strong project
activity.
I would like to highlight a number of key themes we
are seeing which also underpin our companys E&C
strategy going forward.
Local service delivery has always been important to
us and is becoming increasingly important to our
clients. We have continued to develop operations in
key local markets, both organically, in joint venture,
and through M&A. Key markets in which we have
recently strengthened our ability to deliver the global
Foster Wheeler brand and quality locally include Saudi
Arabia, Mexico, and China, to name just three. I will
add to this the US, where we are already strong and
well known but have continued to enhance our
presence and capabilities. Our Houston operation has
had a great year with major wins, especially in
chemicals, and we acquired construction specialists
Yonkers Industries at the end of 2012.
Many clients with upstream and downstream
businesses are allocating more and more of their
planned capital spend to the upstream sector. This
sector, both onshore and offshore, remains a
strategically important market for us and one in which
we have continued to develop our skills, service
portfolio and geographic presence, with acquisitions in
the last 12-15 months in additional early consultancy
capabilities, in SAGD with the acquisition of Three
Streams Engineering in Canada, and in on-going asset
support.
We see more and more clients looking to develop
www.eptq.com PTQ Q1 2014 13
Product specications and
environmental quality
standards will continue
to tighten worldwide
outlook copy 2.indd 5 11/12/2013 15:58
long term strategic relationships with a limited
number of E&C contractors to ensure availability of
high quality resources. For a number of years we have
had a strong focus on developing long term
relationships with clients, working with them
wherever in the world they invest. We aim to match
our skills and geographic footprint development to
align with our clients long term investment plans, so
that we are well placed to meet their needs right
through the project lifecycle, across their entire project
portfolio and wherever in the world they plan to
invest. We have seen a signifcant number of key wins
in 2013 coming from both long standing relationships
and also strategic relationships that are more recently
established.
Projects also keep on growing in size and complexity,
strengthening the requirement for contractors like
Foster Wheeler who have the global horsepower and
track record, technical expertise, and regional and local
knowledge to design and execute these projects safely
and successfully.
Competition remains strong. We focus on
opportunities where we believe we have
differentiators, such as our technologies; our know-
how; our client relationships; our global presence and
our ability to work with clients from the very earliest
phases of projects to help them shape their
investments; and our proven track record of safely
delivering technically complex and very large projects,
often in challenging locations.
Charles T Drevna
President
American Fuel & Petrochemical Manufacturers
F
rom my vantage point in
Washington, DC, I cannot help
but see the year ahead as one
flled with many policy challenges
for the refning and petrochemical
industry.
A year ago, I was cautiously
optimistic that US President Barack
Obama would work to advance the
all of the above energy strategy that he so often and
adamantly talked about during his run for re-election.
But during his February State of the Union address,
and then with the release of his climate change plan
in late June 2013, it became clear that the anti-fossil
fuel sentiments of this Administration will continue
into 2014, and throughout the remainder of his
presidency.
Among the many expected battles ahead is fnding a
palatable solution to the sour Renewable Fuel Standard
(RFS). Progress was made in late 2013 when EPA
recognised the potential adverse effects on consumers of
increasing the ethanol volume in the fuel supply and
reduced the 2014 mandatory volumes. Even with the
cut, greater reductions in the biofuel mandate are
necessary if consumers are to avoid all the detrimental
impacts of the statute. Opposing sides in this battle will
undoubtedly continue to draw lines in the sand until
Congress repeals this broken law or fnds another
solution that considers the impact of this pernicious
policy on consumers, today and into the future.
But, our challenges do not end there. The industry
will continue to face a growing number of climate
regulations that confict with one another. Research has
shown that the RFS increases ozone and greenhouse
gas (GHG) emissions and is in confict with CAFE
Standards, as well as other Clean Air Act regulations;
while Tier III regulations result in greater greenhouse
gases, thus contradicting EPAs GHG regulations.
These conficts could leave many facilities across the
country questioning how they will meet some
requirements without violating others.
Finally, one of the largest challenges may come with
the Obama Administrations recent Executive Order
(EO) that adjusted the metrics by which all Federal
agencies assess the costs of carbon emissions, known
as the social cost of carbon (SCC) calculation.
This recent EO is comprised of subjective metrics
that were developed to try and quantify the alleged
societal benefts of reducing only greenhouse gas
emissions.
Unfortunately, the Social Cost of Carbon scheme will
only add additional burdens on the consumer because
it infates the potential benefts of regulatory actions in
a manner that hides the true economic and societal
costs of new rules. Applying these calculations to
assess the costs and benefts of regulations could
erroneously be used to justify discontinuing the use of
traditional, affordable energy sources through overly
stringent regulations in future rulemakings.
Although the challenges ahead are great, the future
holds enormous opportunity for the refning and
petrochemical industry and the United States. We are
currently witnessing the beginnings of a far reaching
manufacturing renaissance brought on by the recent
increase in domestic oil and natural gas production.
Today more than $100 billion in new manufacturing
infrastructure has been announced and is expected to
be developed throughout the United States during the
next decade.
We are the most fortunate manufacturing region in
the world today with advantaged raw materials,
advanced infrastructure and expertise along the entire
supply chain. The petrochemical industry is on the
verge of an exciting expansion, but it can only
materialise if the federal government stays out of the
way and allows the market to work. My industry will
work tirelessly to advance this development
throughout 2014 and well into the future.
14 PTQ Q1 2014 www.eptq.com
We are currently witnessing
the beginnings of a far reaching
manufacturing renaissance
outlook copy 2.indd 6 11/12/2013 15:58
www.exxonmobil.com/tsl
ExxonMobil
Champion Technology
for Lube Basestocks
MWI
TM
, MSDW
TM
, MAXSAT
TM
Sweep the Competition
Providing the World with Lube Basestock Technologies and Catalysts
exxon.indd 1 13/12/2013 10:56
pcs myth.indd 1 8/3/12 09:55:39
Q
We are nding a build-up of organic fouling material
in a fractionator after the coker. It does not appear to be
asphaltene based. Is this some form of polymer build-up and
how can we deal with it?
A
Celso Pajaro, Manager of Renery Applications, Celso.
Pajaro@sulzer.com, Sulzer Chemtech USA
For high severity cokers, this is not an uncommon
problem; polymer type material is found below the
LCGO draw. The key to minimise its occurrence is to
reduce liquid residence time within the column.
If trays are used, they should be equipped with anti-
fouling features such as fxed valves, sloped
downcomers and fow enhancers to reduce stagnation
and areas of recirculation on the tray deck.
Grid type packings can be used in high temperature
pumparound areas to minimise liquid hold-up and
residence time. Also, spray nozzles should be used for
the feed device. Strainers for spray nozzles should be
well oversided to capture any coke type material that
accumulates in the pipe.
Any pumparound piping operating at temperatures
higher than 500F (260C) should be as short as possi-
ble and should be designed for high velocities in order
to minimise liquid residence time.
Q
We have problems containing the catalyst in our FCC unit
and this is leading to contamination of our slurry oil. Is there
a simple way of narrowing down the probable cause(s) of our
catalyst losses?
A
Paul A Diddams, Senior VP FCC Additives, Paul.Diddams@
matthey.com; Bart de Graaf, FCC R&D Director, Bart.deGraaf@
matthey.com; and Charles Radcliffe, Technical Consultant,
FCC and Rening, Charles.Radcliffe@matthey.com, Johnson
Matthey
Troubleshooting FCC catalyst losses is fairly routine for
FCC catalyst suppliers involve your catalyst
supplier in resolving the issue.
Important considerations:
1. Are chemical analyses showing catalyst losses (not
coke or iron scale)?
2. How did the losses start? Following a TAR, unit
upset, feed change or other unusual operational
situation?
3. Was there a step change in losses or did the problem
gradually get worse?
4. What is the magnitude of losses? Consistent with
primary or secondary cyclone issues or attrition
source?
5. What is the PSD of the slurry fnes very fne
indicates attrition, coarse indicates hole/cyclone issues?
6. Do you have historical slurry PSD from a period
where losses were normal?
7. Are the losses only on the reactor side or have
regen losses increased too?
8. What does the whole catalyst mass balance look like:
losses on RX-side, RG-side, CAR, e-cat withdrawal,
e-cat fushing, additives?
9. Is the stripper behaving normally? What is the H-in-
coke? Has the regenerator T increased with the losses?
10. What are the steam rates going into the FCC any
valves vibrated open (for instance, emergency riser
steam)?
11. What are the bed levels and superfcial velocities in
the riser, regen, and so on?
12. Are cyclone inlet/outlet velocities within design
limits?
13. What is the superfcial velocity at the feed injector
tips? Air grid nozzles? Too high can cause attrition.
14. Fresh catalyst: is attrition index in the correct
range? Is particle density too low? Are there excessive
fnes (0-20%)?
15. Check for substantial changes in slurry yield (feed/
conversion/distillation).
A
Stuart Foskett, Regional Technical Service Manager for
Asia, stuart.foskett@basf.com, BASF Catalysts
Catalyst loss problems can be the result of a variety of
different issues. In this example the losses are specif-
cally noted as occurring from the reactor side. The frst
troubleshooting step is to seek a particle size distribu-
tion (PSD) of the catalyst losses by collecting a
representative slurry oil sample, ashing the slurry and
analysing the collected solids. The PSD may have one
or more peaks. A peak at around 1-2 indicates fnes
generated by attrition, a peak at about 20-30
represents typical cyclone losses and a peak at around
50-80 indicates loss of full range particles. It is good
practice to obtain a PSD for slurry oil ash and regener-
ator third stage separator fnes at least once per
quarter so there is historical data to compare against in
a troubleshooting situation. BASF offers a slurry and
fnes analysis service at no cost for FCC catalyst
customers.
If the majority of losses are attrition fnes smaller
than 5, then all potential attrition sources in the riser
and reactor should be investigated. Common problem
areas include feed nozzles (plugged nozzles or erosion)
and stripping steam distributors (erosion, holes,
cracks), and hardware damage from wet steam. In
www.eptq.com PTQ Q1 2014 17
ptq&a
Additional Q&A can be found at www.eptq.com/QandA
Q&A.indd 1 13/12/2013 11:47
some cases, it is possible to identify the problem area
by reducing gas fow and observing a change in losses,
for example by reducing stripping steam. Excessive
attrition occurring in the riser or stripper will often
result in higher losses from the regenerator as well.
A change in attrition properties of the catalyst itself
is sometimes a contributing factor towards attrition
losses. In general, poor attrition resistance of the cata-
lyst can be ruled out as a primary cause if there is a
sudden step increase in attrition losses. If the increase
in losses is more gradual, check whether the timing
corresponds to a change in catalyst grade, and also
review the fresh catalyst quality analyses to confrm
the catalyst properties remain in the normal range.
A large peak around 50-80 can indicate loss of full
range particles through a holed or cracked plenum, or
by a cyclone failure. Potential cyclone problems can
include diplegs plugged by coke deposits, often associ-
ated with incomplete feed vaporisation, failure of a
fapper valve or trickle valve, a holed cyclone and
cyclone dipleg fooding. A change in the cyclone pres-
sure drop relative to the expected value is one possible
signal of cyclone issues. Occasionally it is possible to
identify a missing or stuck-open fapper valve by rais-
ing the reactor bed level to submerge and seal the
cyclone dipleg.
Other information that can be helpful in identifying
loss problems includes a regular catalyst mass balance
around the reactor and regenerator (both daily and
monthly), historical slurry alumina and BS&W analy-
sis, calculation and monitoring of cyclone velocity,
cyclone pressure drop, steam and air distributor pres-
sure drop, and feed nozzle pressure surveys. As with
all troubleshooting, simply looking for changes in
operating conditions that coincide with the timing of
increased losses may provide further clues of a poten-
tial cause.
A
George Yaluris, FCC Technical Service Advisor, george.
yaluris@albemarle.com, Albemarle
Troubleshooting catalyst losses either from the reactor
or the regenerator side of the unit is not a simple
undertaking. A systematic approach must be applied
to identify the symptoms, quantify the losses, conduct
investigative work to troubleshoot the causes, and
identify the best possible course of action. The good
news is that the source of catalyst losses can be
narrowed down to a short list of possible causes. A
solution without having to shut down the unit may be
possible depending on the nature of the problem.
On the reactor side, catalyst losses manifest them-
selves as increased solids in the slurry. They can lead
to increased exchanger fouling, erosion, slurry pump
failure and the need to dispose of oily fnes.
Troubleshooting these losses must start before the
problem begins by routinely collecting data to estab-
lish the units normal operation baseline. These data
should include accurate accounting of the amount of
catalyst lost under normal operation, catalyst data
(including fresh, equilibrium, and fnes particle size
distribution [PSD], fresh catalyst attrition, ABD, and so
on), collection system and unit design details, and
operating details (for instance, temperatures, pressure
surveys, steam rates).
Once the problem is identifed and quantifed,
pinpointing the cause involves a detailed review of
possible causes to determine if they can be excluded.
There are three common types of causes: operation
changes, catalyst attrition and collection system hard-
ware failure. The unit operations should be carefully
reviewed for any changes which may be causing or
contributing to the increase in slurry solids. Examples
are increased feed and catalyst addition rates, changes
in the reactor/stripper level, and excessive vapour/
solids fow rates leading to cyclone fooding. The fresh
catalyst properties should also be reviewed to ensure
the attrition resistance and other relevant physical
properties of the catalyst have not changed.
The slurry fnes should be carefully analysed for
changes in the PSD and morphology (for instance, by
SEM imaging). This analysis can provide important
clues. An increase in the proportion and absolute
amount of fragmented particles (typically in the 20 to
40 range) may indicate the existence of an attrition
source. On the other hand, a fattening of the fnes PSD
distribution and an increase in the amount of larger
particles being lost may indicate a decrease in cyclone
effciency for example due to cracks and small holes.
The presence of equilibrium size catalyst can indicate
larger holes in the catalyst separation hardware, or
dipleg plugging, fooding or breaking, including
dipleg valves not opening. If the unit includes a device
for separation of solids from the slurry, the operation
of this hardware should also be reviewed.
The data should then be evaluated in an attempt to
locate the area of the problem. In addition to examin-
ing the catalyst and fnes data as well as the operating
data (such as changes in the unit pressure and
temperature profle), step tests can be conducted to
further troubleshoot the problem, identify the location
and potentially provide relief from the losses. Gamma
ray scans may also be done to identify changes in the
catalyst fow and to help locate the cause.
Troubleshooting a catalyst loss problem requires
detailed familiarity with the unit design, current confg-
uration, hardware conditions and operations, as well as
experience in conducting this kind of work. It also
requires working closely with the catalyst vendor who
will typically provide much of the analytical work
required. If the problem is related to operating or cata-
lyst changes, a solution can often be devised to address
it. If the problem is related to mechanical failure, a
remedy may be possible while the unit is running.
18 PTQ Q1 2014 www.eptq.com
Excessive attrition occurring in
the riser or stripper will often
result in higher losses from the
regenerator as well
Q&A.indd 2 11/12/2013 15:54
pcs 1.indd 1 8/3/12 09:50:20
20 PTQ Q1 2014 www.eptq.com
in PNA content from 0.48 to 0.55. Notice also that the
conversion decreases despite the fact that the FCC feed
nitrogen content is at the lowest level suggesting the
PNA feed content has the larger impact on FCC
conversion.
Clearly, driving the hydrotreater to remove nitrogen
and PNAs improves FCC product value, but this needs
to be balanced against the increased costs of higher
hydrogen consumption and shorter cycle length that
result from this mode of operation.
The shorter run lengths from operating in PNA
mode can be extended with the use of a sulphur reduc-
tion catalyst or additive in the downstream FCC unit.
FCC gasoline sulphur reduction of 20% or higher from
a product such as D-PriSM or GSR-5 additives, or
SuRCA catalyst would enable the pretreater to run
beyond the point when product sulphur levels exceed
target levels, while FCC gasoline product sulphur
remains within the limits needed for gasoline pool
sulphur compliance. Extending operation in PNA
mode closes the gap in run length versus a hydro-
treater in the HDS mode of operation, while
continuing to provide FCC feed of higher quality than
HDS mode. Use of OlefnsMax and OlefnsUltra addi-
tives during maximum LCO operation can assist in
maintaining liquid yields and gasoline octane.
In addition to optimising the hydrotreater catalyst
system and the hydrotreater operation, the FCC cata-
lyst formulation can also be included as one of the
variables that can be adjusted to maximise proftability.
Adjusting FCC catalyst formulation as the hydrotreater
moves from start of run to end of run can also signif-
cantly offset a decline in product quality and preserve
product value. In this case, using FCC products such
as Midas for maximum LCO selectivity with improved
bottoms cracking or materials from the Genesis and
Achieve Catalyst technology platforms provide fexibil-
ity in catalyst formulations for conventional as well as
unconventional feeds.
All of the combinations for optimising the FCC
pretreater/FCC complex demonstrate the need for
refners to follow an integrated approach to managing
the catalysts and operation of the FCC pretreater and
FCC unit. Grace Catalyst Technologies is well suited to
assist with this endeavour through its Advanced
Refning Technologies (ART) and Refning
Technologies businesses. The complexity of
combinations of catalyst design and operat-
ing conditions for both the FCC feed
hydrotreater and the FCC unit presents a
signifcant optimisation opportunity for
refners to drive the combined operation to
maximum product value.
A
George Anderson, Global Hydroprocessing
Specialist, george.anderson@albemarle.com,
Albemarle
In general, it is advisable to hydrotreat
heavy vacuum gas oil (HVGO) before it is
fed to the FCC unit. HVGO typically
contains metals (Ni, V), contaminants (As,
Changes in the catalyst physical properties or the use of
a fuidisation aid such as Albemarles Smoothfow may
help the unit run until the next turnaround, provided
that the failure does not worsen. Of course more serious
hardware failures will eventually require a shutdown to
repair the damage. In this case, troubleshooting the
problem will help the operator better prepare for the
shutdown and locate the damage faster, thus quickening
repairs and decreasing downtime.
Q
Is it advisable to hydrotreat heavy vacuum gas oil before it
is fed to the FCC unit? If so, why?
A
Charles Olsen, DHT R&D and Technical Services Director,
Chuck.olsen@grace.com, Advanced Rening Technologies
FCC feed pretreating offers the potential to signif-
cantly improve refnery economics. Advanced Refning
Technologies has previously described the benefts of
hydrotreating FCC feed on the FCC product yields and
sulphur content in earlier publications (see for example
NPRA papers AM 02-39, AM 04-29, AM 07-55, AM
11-21). That work has demonstrated that both the
hydrotreating catalyst system and the operating strat-
egy for the pretreater are critical to providing the
highest quality feed for the FCC as well as the ability
to tailor the FCC product slate.
In general, NiMo based catalysts produce higher
quality FCC feed, with both lower nitrogen and PNA
content. The operating mode of the hydrotreater can
also be used to further improve the FCC feed. It is
generally accepted that removing PNAs and nitrogen
from the FCC feed will improve FCC performance and
increase conversion. As an example, Figure 1
summarises FCC pilot plant data for several hydro-
treated FCC feeds. The left y-axis represents the FCC
conversion and the right y-axis represents the FCC feed
quality, shown as a ratio of either product nitrogen
content to feed nitrogen content or product PNA
content to feed PNA content. Clearly, the PNA and
nitrogen content of the FCC feed have a strong effect
on FCC performance. At the highest pretreater severity
the FCC conversion actually decreases by about 1 wt%
relative to the previous severity in the chart. The
decrease in FCC conversion coincides with an increase
72
76
75
78
77
74
73
71
70
69
68
67
F
C
C

c
o
n
v
e
r
s
i
o
n
,

w
t
%
66
Increasing pretreater severity
0.6
0.8
0.7
0.5
0.4
0.3
0.2
0.1
N
i
t
r
o
g
e
n

o
r

P
N
A

c
o
n
t
e
n
t
,

p
r
o
d
u
c
t
/
f
e
e
d
0
Nitrogen
PNA
FCC
Figure 1 Pretreater PNA and nitrogen conversion effect on FCC conversion
Q&A.indd 3 11/12/2013 15:55
Innovative Catalytic Solutions
Grace FCC Catalysts
and Additives
n
Industrys Broadest Catalyst Portfolio
n
Flexible Functionality for Processing
Unconventional Feeds
n
Global Manufacturing Footprint
n
World-class R&D
n
Industry-leading Technical Service
grace.indd 1 10/12/2013 09:55
22 PTQ Q1 2014 www.eptq.com
and nitrogen contents leads to lower SOx and NOx
emissions from the FCC unit regenerator. This trans-
lates into lower NOx stack emissions for almost all
refneries, and either lower SOx stack emissions, or
lower treating costs to meet SOx emission regulations.
A
Paul A Diddams, Senior VP FCC Additives, Paul.Diddams@
matthey.com; Bart de Graaf, FCC R&D Director, Bart.deGraaf@
matthey.com; and Tom Ventham, Senior Technical Service
Engineer, Tom.Ventham@matthey.com, Johnson Matthey
Resid FCC units are designed to process HVGO with-
out necessarily hydrotreating it frst.
Advantages of pre-hydrotreating HVGO:
1. Contaminant levels are decreased: V, Ni, N, S
lower delta coke, lower CAR, lower stack emissions
(SOx, NOx)
2. Coke precursor levels are decreased: CCR lower
delta coke, relieves regen T and air constraints
3. Partial aromatics saturation increases feed crackabil-
ity lower delta coke, improves conversion, increases
gasoline selectivity.
Disadvantages of pre-hydrotreating HVGO:
1. Capex required to build a HVGO hydrotreater
2. Opex required to run a HVGO hydrotreated
3. Increased hydrogen consumption on the refnery
4. More H
2
S/NH
3
to the Claus plant.
Outside the US, the only justifcation is the increased
conversion (volume expansion if this word would have
been used abroad); inside the US, environmental consid-
erations will have a substantial input to the decision.
We have seen some refneries in Europe, particularly
Eastern Europe, switch to HT VGO FCC feed (or,
perhaps more precisely in many cases, MHC feed) as a
catch-all solution originally forced by a move to Euro
IV/V and the need to reduce gasoline sulphur. In
many of these cases, they then struggle to close the
heat balance. This could be seen as a signifcant disad-
vantage in addition to those given above.
So, it is great to have, but a full economic assessment
is required to see if the advantages outweigh the
disadvantages.
Q
We want to send coker naphtha for hydrogen production.
Can you recommend a hydrotreating scheme to facilitate this?
A
Charles Olsen, DHT R&D and Technical Services Director,
Chuck.olsen@grace.com; Brian Watkins, Manager, Plant &
Technical Service Engineer, Brian.watkins@grace.com; and
Brian Slemp, Technical Service Manager, Brian.slemp@grace.
com, Advanced Rening Technologies
A properly designed coker naphtha treating system for
the preparation of hydrogen plant feed requires special
consideration to mitigate the impacts of contaminants
and olefns. A traditional moderate pressure hydro-
treater with signifcant recycle gas fow and multiple
catalyst beds with suffcient quench to moderate the
exotherm from the large olefn content is preferred to
prepare coker naphtha for steam reformer feed.
Processing coker naphtha can have several undesir-
able effects on the performance of the hydrotreater if
Si), relatively high sulphur and nitrogen, and heavy
aromatics with 3 rings. The range of typical proper-
ties for HVGOs is shown in Table 1.
FCC units generally operate better on cleaner feeds.
Removal of Ni and V in an FCC pre-treat (FCC-PT)
unit helps the FCC catalyst (and e-cat) retain better
activity and reduces H
2
generation in the FCC unit.
Basic nitrogen titrates the FCC catalyst and inhibits
catalytic cracking activity. Thus, HDN in an FCC-PT
unit has signifcant FCC unit performance benefts.
Heavy aromatics do not crack in the FCC unit unless
they are at least partially hydrogenated. Many 2-, 3-
and 4-ring aromatics will pass on directly to the LCO,
HCO and slurry products. The remaining aromatics
will form coke on the FCC catalyst and deactivate it.
For those refneries that do not hydrotreat HVGO feed,
use of FCC catalyst with high accessibility and high
metals tolerance (such as Albemarles Upgrader and
Amber catalyst families) is highly recommended.
Reducing HVGO sulphur content in FCC-PT does
not have a major impact on FCC unit operating eff-
ciency, but it does have a major impact on FCC unit
product quality. FCC naphtha typically contains
sulphur levels that are ~5-7% of the sulphur levels for
hydrotreated FCC unit feed. On the other hand, FCC
naphtha contains sulphur levels that are ~10% of the
sulphur levels for unhydrotreated FCC unit feed. Thus,
the beneft of HDS in the FCC-PT unit is two-fold: 1)
the FCC unit feed has a lower sulphur level and 2) a
lower proportion of the FCC unit feed sulphur will
end up in the FCC naphtha product. As gasoline
sulphur limits are reduced (such as with implementa-
tion of Tier 3 gasoline regulations in the US beginning
in 2017), hydrotreating of HVGO feeds in FCC-PT
units becomes even more critical for many refneries.
FCC unit LCO product quality is also impacted by
FCC-PT operations. The beneft of FCC-PT is that the
LCO product from hydrotreated feed contains lower
sulphur than LCO generated from unhydrotreated feed
and generally has higher yield from hydrotreated feed.
This beneft may be partially offset by the fact that the
LCO from hydrotreated feed may be more aromatic.
This may result in higher H
2
consumption during
diesel hydrotreating and lower ULSD product cetane
number.
A fnal beneft from hydrotreating the FCC unit feed
is the benefcial effect on environmental emissions and
related clean-up costs. Reducing FCC unit feed sulphur
Feed property Typical values
Boiling range, F 650-1100
Sulphur, wt% 1.0-3.0
Nitrogen, wppm 1000-3500
Aromatics, wt% 35-55
CCR, wt% 0-4
Nickel + vanadium, wppm 1-10
Silicon, wppm 0-4
Typical properties of HVGO Feeds to FCC-PT units
Table 1
Q&A.indd 4 11/12/2013 15:55
www.jmprotech.com
UK Tel +44 (0)1642 553601 Fax +44 (0)1642 522542
At Johnson Matthey we dont just provide world
class hydrogen catalysts, we provide knowledge
and understanding of the hydrogen process
and catalytic interactions through a unique
range of services to enhance plant operation
and improve efciency and reliability.
service,
support,
solutions.
j mathey.indd 1 11/12/2013 09:33
24 PTQ Q1 2014 www.eptq.com
include: AT724G, which is used in applications requir-
ing moderate activity and high silicon capacity;
AT734G, which has high silicon capacity and high
arsenic capacity; and AT714G which has low activity
and ultra high silicon pick-up capacity.
The potential polymerisation of the cracked naphtha
can be addressed with a diolefn reactor and/or appro-
priate activity and size grading to ensure a suffcient
run life. A diolefn reactor is operated at a much lower
temperature in order to selectively catalyse the diolefn
saturation reaction and avoid any excess heat genera-
tion from sulphur and nitrogen removal. Liquid
product recycle may also be considered as an addi-
tional method to provide a heat sink to control reactor
temperature rise.
Coker naphtha is known to vary dramatically
during the various portions of the coking cycle.
Special consideration of the feedstock design basis
needs to refect the various portions of the coking
cycle. This will ensure the new hydrotreater design
can accommodate not only the average feed proper-
ties, but both the maximum and minimum naphtha
properties experienced during the coking cycle. Steam
reforming hydrogen production is an extremely high
temperature operation sensitive to carbon lay down
and coking. Naphtha end point control and carbon
speciation are important considerations to determine
if the hydrogen unit reformer is capable of handling
variable higher carbon number feeds. The use of the
pre-reformer may be required to facilitate the feed
change.
Q
We are nding particulates and water in our product streams.
Is it best to trace back for the problem or is there some means of
separating the contaminants without further processing?
A
Paul A Diddams, Senior VP FCC Additives, Paul.Diddams@
matthey.com and Charles Radcliffe, Technical Consultant,
FCC and Rening, Charles.Radcliffe@matthey.com, Johnson
Matthey
Diffcult to generalise without knowing which
product(s) but defnitively identify and fx. The follow-
ing are good general rules for troubleshooting:
a) Check levels in water boots of upstream vessels
b) Any aqueous treatment upstream check levels and
condition of demisters; check for foaming
c) Check upstream water coolers for leaks, with water
side the higher pressure
d) Analyse water for clues as to source, for instance
cooling/BFW/condensate water treatment chemicals
e) Analyse particulates for clues to source.
the catalyst system is not properly designed to deal
with the unique challenges of the feed. In general,
coker stocks have a higher level of olefns present from
the coking process. Once in the hydrotreater these
olefns will quickly get saturated (see Figure 1) result-
ing in high hydrogen consumption and generation of a
lot of heat. As a general rule of thumb, one mole of
hydrogen is required per mole of carbon-carbon double
bonds, or between 5-10 times the bromine number
reduction in standard cubic feet of hydrogen per barrel
(scf/b). This additional heat (130-160 Btu/scf of hydro-
gen consumed), if not managed properly, will initiate
additional reactions, quickly creating a very high
temperature rise. The high temperatures can accelerate
coking and can lead to olefn polymerisation, resulting
in a dramatic increase in pressure drop. This can set an
upper limit on how much coker naphtha can be
processed based on the need to limit the heat rise or
hydrogen consumption.
(C)
n
-C=C
+H
2

(C)
n
-C-C + Heat

Easy
Figure 1 Olen saturation
A system that is properly sized and activity graded
is extremely important when processing coker feeds.
ART recommends utilising a grading system to help
mitigate pressure drop build-up. A large inert hold
down ring (GSK-19) with a very high void fraction
used for trapping large particulates is placed at the top
of the catalyst bed. A smaller diameter macroporous
ring (GSK-9) that traps iron as well as other fne partic-
ulates is typically used as the next layer. After that,
two types of smaller rings are used as active grading.
These materials have a low level of active metals
which help to begin olefn saturation reactions as well
as provide additional external void space at the top of
the catalyst bed. Avoiding the use of any highly active
catalyst at this point is also recommended. Below the
grading system, it is recommended to use a layer of
larger size (1/10in or 1/12in) catalyst which provides
activity for olefn saturation and additional void space
for pressure drop mitigation. This layer is often a cata-
lyst which is suitable for trapping silicon (and arsenic)
which is another concern when processing coker
naphtha.
Coker naphtha contains a signifcant amount of
poisons that need to be removed prior to being used as
feed to a hydrogen plant. The reactor must be large
enough to trap the silicon and arsenic normally found
in coker products. ART has extensive experience in
coker naphtha hydrotreating and has several products
designed to combat the contaminants found in coker
products. ART also has a portfolio of high activity
naphtha catalysts designed to not only remove high
levels of sulphur and nitrogen but also remain tolerant
to silicon contamination. ART utilises several grades of
silicon and arsenic trapping materials depending on
the level of contaminants present in the feed and the
need to balance HDS/HDN activity. These grades
A system that is properly sized
and activity graded is extremely
important when processing
coker feeds
Q&A.indd 5 11/12/2013 12:04
For over 60 years, weve been there for our refning
and petrochemical customers. Linde Engineering
North America Inc. offers single source responsibility
for technology, engineering, procurement and
construction - T-EPC.
Selas Fluid refnery and petrochemical fred heaters
Oxidation and incineration technologies
Engineered revamps and rebuilds
Hydrogen and synthesis gas plants
Air separation plants
Were local. Were global.
And were proud to be both.
Headquarters: Five Sentry Parkway East, Blue Bell, PA 19422 USA Phone +610.834-0300
Texas Offce: 3700 West Sam Houston Pkwy. South, Suite 425, Houston, TX 77042 USA Phone +281.717-9090
Al-Khobar: Linde Arabian Contracting Phone +966.3.887-0133
www.lindeus-engineering.com
linde NA.indd 1 12/09/2013 11:06
26 PTQ Q1 2014 www.eptq.com
from 3.5 wt% to 0.5 wt% globally by 2020. The Marine
Environment Protection Committee (MEPC) also
agreed to use equivalence technology such as
on-board scrubbing for post-treatment instead of using
low sulphur bunker fuels.
In reality, it will likely be diffcult to implement
post-treatment scrubbing in the ECA by 2015. Thus,
clean maritime fuel is the more realistic solution. This
is expected to increase demand for marine distillate oil
(MDO) and marine gas oil (MGO) by approximately
500 000 b/d. The increased demand will likely be
supplied from distillate production using existing
hydrotreating capacity as well as additional capacity
coming online in the near future.
In order to meet the 2020 mandates (S <0.5 wt), the
industry will require approximately 2-3 million b/d of
clean bunker fuel to cover increased global consump-
tion. This may cause a signifcant shift in quality from
heavy, high sulphur fuels to lighter, lower-sulphur fuels
(MDO, MGO). For this to occur, there will need to be a
considerable increase in global capacity for bottoms
upgrading and secondary hydrotreating by 2020.
Consequently, bunker fuels will need to be sold at a
premium to justify the substantial refnery investment.
As an alternative, shipping industries have been seri-
ously examining implementation of scrubbing units
using currently available high sulphur fuel oil (HSFO).
The technology has been commercially proven to
reduce SOx emission by 97% or higher. It also prevents
increased demand for MDO and MGO supply.
A
Charles Radcliffe, Technical Consultant, FCC and Rening,
Charles.Radcliffe@matthey.com, Johnson Matthey
Diffcult to answer without more specifcs, but there
are two options for possibly meeting this:
a) More active catalyst, possibly dense load if not
already done so; more frequent change out; if using
mixed Co/Mo & Ni/Mo then perhaps shifting more
towards Co/Mo will help.
b) Remove/reduce the highest sulphur feed compo-
nent (LCO and coker gasoil).
Q
We suffer occasional corrosion problems with the
exchangers in our crude unit overhead. We assume these
events depend on variations in crude quality. Is there some
way we can monitor and treat ows on-line? Or is new
metallurgy a more economic option?
A
Berthold Otzisk, Consulting Engineer, Otzisk@kurita.de,
Kurita Europe GmbH
The main question is whether there is a permanent
corrosion attack, resulting in occasional heat exchanger
leakages.
Crude oil contains various chloride salts. The initial
corrosion control system is the electrical desalter,
where corrosive salts are removed from crude oil. The
crude oil quality can be a prime reason for operational
problems. Some crude oils tend to create very stable
emulsions, resulting in unwanted desalter upsets and
increased overhead corrosion potential. Sodium
A
Celso Pajaro, Manager of Renery Applications, Celso.
Pajaro@sulzer.com, Sulzer Chemtech USA
The presence of water is typically due to the use of
stripping steam. One way is to change strippers from
steam to reboiler, another option is to install either a
salt dryer or high effciency coalescer (like Sulzer
Dusec). These devices are able to remove submicron
water droplets. Both need the stream to be at storage
temperature to minimise water solubility to maximise
water removal.
Particles are typical corrosion products that can be
minimised by upgrading metallurgy in the points
where corrosion is higher. Another option is installing
particle removal flters, or to just allow a product stor-
age tank to act like a big settler, allowing the
accumulation of the particles which will be removed
during tank maintenance.
Q
A big drop in the allowable sulphur content of marine
gas oil is looming (from 1% down to 0.1% from 2015). Is it
possible that we could meet this using existing hydrotreating
capacity in the renery or will we need to invest in new plant?
A
Hiroshi Toshima, Global Application and Marketing
Specialist, HPC, hiroshi.toshima@albemarle.com, Albemarle
The International Marine Organisation (IMO) has
amended MARPOL Annex VI to decrease the sulphur
content of bunker fuel from 1.0 wt% to 0.1 wt% in the
designated Emission Control Areas (ECA) by 2015, and
challenges remain in process confguration, catalyst
development and materials development when
membranes are employed for optimisation of the fow
diagram. Oxygen supply (by defnition) is also a
requirement for a partial oxidation plant (both CPO
and POX).
In catalytic steam reforming, technology licensors
offer different commercially proven solutions that are
based primarily on furnace technology, and it remains
the most widely applied technology for producing
hydrogen. Steam reforming is particularly effective
when methane is available as an economical source of
hydrocarbon feed (for example, from shale gas).
However, in cases where heavier hydrocarbons are
used as feed, a pre-reformer is usually applied. This
not only increases costs, but also the susceptibility to
upsets. In these cases, partial oxidation becomes a
competitive option.
Although CPO of methane is still very much in the
development phase, it remains theoretically favourable
in terms of energy and CO
2
balance if one can manage
to selectively carry out the reaction of methane with
oxygen to CO. Some interesting concepts have been
presented for the debottlenecking of existing refnery
steam reformers by adding a CPO unit, which is
always a smaller investment in terms of size and capi-
tal when compared to steam reforming or adding a
POX unit.
A
Luigi Bressan, Director of Process and Technologies, Foster
Wheeler Italy
In the majority of cases, the feedstock dictates which
process is more attractive. Generally, where it is possi-
ble to use catalytic steam reforming, this technology is
preferred. Feedstocks ranging from methane down to
naphthas can be readily processed in a catalytic steam
reformer; for heavier feedstocks, partial oxidation is
the only viable alternative.
A
Lorena Mosca, Technology Project Coordinator, KT-
Kinetics Technology, l.mosca@kt-met.it
If we consider catalytic partial oxidation of natural gas
against steam reforming, it will compete economically
when CO
2
removal from syngas is mandatory (zero
CO
2
emission hydrogen production). In all other cases,
it depends on the cost of the oxygen available at the
industrial site. Refneries where O
2
is used also for
sulphur recovery units, for wastewater treatment or
gasifcation are candidates for low-cost oxygen
production.
Q
We are maximising propylene in the FCC and would also
like to maximise xylenes production from our FCC gasoline.
What additional process step(s) do we need?
A
Stuart Foskett, BASF Regional Technical Manager, Stuart.
Foskett@BASF.com
There is only limited fexibility when adjusting the
aromatics yield in FCC gasoline. Increasing the catalyst
rare earth content results in a small increase in aromat-
ics by hydrogen transfer reactions. Unfortunately, this
is opposite to the requirements for maximum propyl-
ene, where hydrogen transfer should be minimised to
preserve olefns from cracking by ZSM-5 to propylene.
Otherwise, simply operating the unit to maximise
conversion will enhance both the propylene yield and
aromatics in the light FCC gasoline fraction.
An increase in the benzene yield occurs when
increasing the propylene yield, as shown in Figure 1. A
possible explanation for this are the cyclisation reac-
tions involving light olefns, although less clear,
commercial data also suggest a possible increase in
total aromatics associated with a higher propylene
yield and/or ZSM-5.
Since aromatics do not crack, FCC gasoline aromatics
will also increase with a higher mono-aromatics
content in the FCC feed. For units with a feed hydro-
treater, increasing the hydrotreating severity or seeking
catalysts with greater activity for aromatics saturation
will increase mono-aromatics by saturating more
multi-ring aromatics, although hydrogen consumption
will also increase.
Some refners believe that higher aromatics produc-
tion can be achieved by processing FCC gasoline
through the reformer, typically a heart-cut (170-300F,
76-149C). FCC gasoline is a poor reformer feed,
with higher nitrogen and N+2A than straight-run
feed, and olefns that must be saturated in the
hydrotreater.
6 PTQ Q4 2013 www.eptq.com
www.eptq.com PTQ Q3 2013 121
India Nagi-Hanspal is Lead Renery Engineer:
Technical Services with Dorf Ketal Chemicals,
Mumbai, India, and is primarily responsible
for providing renery process improvement
strategies through unit monitoring. She
provides chemical solutions to Dorf Ketal site
teams globally and holds a MEng degree in
chemical engineering from Imperial College
London. Email: indianagi@dorfketal.com
Mahesh Subramaniyam is Director of Research
& Development with Dorf Ketal Chemicals and
leads the companys chemical developments.
He holds a number of patents for oil treatment
chemicals and engineers products to address
emerging issues in the hydrocarbon industry.
He holds a PhD in chemistry from Indian
Institute of Technology, Mumbai.
Email: drmaheshs@dorfketal.co.in
Parag Shah works in Global Renery Technical
Services with Dorf Ketal Chemicals, where he
is instrumental in software development for
desalter adequacy testing and monitoring
fouling in preheat exchanger trains. His
experience includes more than eight years in
the worlds largest grassroots renery and he
holds a BEng in chemical engineering from
Mumbai University.
Email: paragshah@dorfketal.com
corrosion also declined substan-
tially, illustrating the importance of
metallurgy selection for corrosion
control.
Conclusion
With proper planning, high-acid
crudes can be processed success-
fully if appropriate strategies are
implemented before adding high-
acid crudes to the feed diet. Success
requires a thorough understanding
of corrosion processes, identifca-
tion of susceptible systems and
consistent monitoring to adapt
treatment as system dynamics and
crude diets change.
CEION is a mark of Teledyne Cormon and
Hydrosteel is a mark of Ion Science.
References
1 Houlton G, Crude demand to increase, feed-
quality changes in store, Oil and Gas Journal,
2010.
2 Zhang L, Shortage of crudes, not products, to
drive Asian rening market, Oil and Gas Journal,
2011.
corrosion occurs. Once sensors are
installed, the system collects data
and reports trends over time.
HTCI case study
One of the worlds largest refneries
processed blends with an average
TAN of 1.31 mgKOH/g and a
sulphur content of 1.28 wt%. The
heavy vacuum gas oil loop was
monitored closely for corrosion.
TAN in the stream was approxi-
mately 1.79 mgKOH/g with 1.19
wt% sulphur. The key performance
indicator for the trial was corrosion
control at a maximum of 8 mpy.
Dorf Ketal HTCI treatment results
are shown in Figure 6.
Testing by an independent third-
party laboratory confrmed the
effectiveness of the formulation. As
can be seen in Figure 7, HTCI
reduced corrosion rates by more
than 80% in a laboratory blend
with a TAN of 5.0 mgKOH/g.
Although the reduction in corro-
sion on carbon steel is clearly the
most dramatic, 410 stainless steel
82 PTQ Q2 2013 www.eptq.com
content of 5-8 wt% largely present in the unstripped
products.
The levels of sulphur and nitrogen in coke are much
higher than in the feed. In the regenerator, carbona-
ceous coke and unstripped hydrocarbons are
combusted to CO
2
, CO and H
2
O. Sulphur is oxidised to
SO
3
, SO
2
, COS and H
2
S. Nitrogen behaves differently;
as oxygen reacts with the coke matrix to form CO
2
, CO
and H
2
O, much of the nitrogen may initially form
HCN.
This is similar to the chemistry observed in coal
combustion. HCN is a thermodynamically unstable
species under FCC regenerator conditions and, given
suffcient time and temperature, HCN will be fully
converted and no trace of HCN will be found in the
regenerator. Under FCC regenerator conditions, N
2
is
the most stable nitrogen species, followed by NO,
which at thermodynamic equilibrium should reach a
stable concentration of about 10 ppm. Evidently, nitro-
gen chemistry is not at thermodynamic equilibrium,
but is under kinetic control.
Investigation of coke combustion
Temperature programmed oxidation (TPO) may be
used to study coke combustion. In the TPO experiment,
a sample of coked catalyst (spent catalyst) is heated in
an inert atmosphere up to a chosen initial temperature
(say, 150C). The gas supply is then switched to the
combustion gas-containing oxygen (and optionally other
selected gases). The temperature is ramped (at, say,
15C/min) to 730C, then held isothermally.
Combustion fue gases are continuously monitored by
infrared and mass spectroscopy.
The frst nitrogen species evolved (at the lowest
temperature) is HCN (see Figure 1). HCN generation
starts at very low temperatures, already around 450C
(780F). At this low temperature, the rate of conversion
of HCN to thermodynamically more stable NO or N
2
is
relatively slow. NO is not observed until the fue gas
temperature reaches about 600C (1100F). As the
temperature is increased further, the HCN concentra-
tion declines as NO increases. The HCN is no longer
stable enough to survive so it and subsequent N in
coke are converted into NO or N
2
. Despite being ther-
modynamically unstable under FCC regenerator
conditions, N
2
O is also observed at similar tempera-
tures to NO. This indicates that N
2
O may also be a
reactive intermediate in the formation of NO.
Increasing the oxygen content of the combustion gas
from 3 vol% to 4 vol% increases the rate of HCN
conversion substantially. The frst traces of NO in fue
gas form at a lower temperature of about 550C
(1000F).
These simple experiments show that HCN is readily
formed during coke combustion at low temperatures,
especially if oxygen concentrations are high at these
low temperatures. Once the temperature is high
enough, the HCN becomes unstable and is readily
converted to NOx or N
2
.
The implication of this is that the points of greatest
HCN formation in the FCC regenerator are likely to be
Product Nitrogen, wt%FF
Fuel gas NH
3
/HCN 5-15
Gasoline 1-5
LCO/diesel 10-20
HCO/bottoms 25-35
Coke 35-60
Typical nitrogen balance: wt% feed nitrogen to FCC products
Table 1
compounds present in FCC feedstocks; these are meas-
ured and reported as total nitrogen and basic nitrogen.
Typically, about 30-50% of the feed nitrogen is basic.
These nitrogen species strongly adsorb on acid sites on
the catalyst and are thereby transported with the cata-
lyst into the regenerator, where they are combusted
together with the coke. As a rule of thumb, about half
of the feed total nitrogen ends up being combusted in
the FCC regenerator. Table 1 shows a typical FCC
nitrogen balance.
Coke composition varies, depending on feed prop-
erties and stripper effciency. Coke consists of
carbon-rich polycyclic aromatic structures containing
heteroatoms and contaminant metals as well as
unstripped hydrocarbon products (for example,
10-30% of coke may be gasoline, diesel and fuel oil
range products that could not be stripped from the
pores of the catalyst). Typically, coke has a hydrogen
j matthey.indd 2 11/03/2013 13:02
8 PTQ Q3 2012 www.eptq.com
before etherifcation. A TAME unit is an opportunity to
reach a higher RON pool. Oligomerisation of C
5
into
distillate (PolyFuel process) is an opportunity to switch
gasoline into diesel, particularly in Europe, where
refners face surplus gasoline production.
Q
What arrangement of trays and packing designs will
give me the best effciency of fractionation in my crude and
vacuum distillation units?
A
Celso Pajaro, Manager of Refnery Applications, Sulzer
Chemtech USA, Celso.Pajaro@sulzer.com
There is no specifc arrangement of packing and trays
for getting the best effciency of fractionation. The choice
in column internals is a function of the type of crude
being processed and the column diameter available.
Structured packing has been used in every section of
crude columns, often to substantially increase effciency.
In many cases, the effciency within a given section can
be increased by two to three theoretical stages each.
However, its use in the top section can be a challenge
with particularly corrosive crudes. Monel and AL6XN
materials have been used successfully in these applica-
tions, but care needs to be taken to properly match the
correct material with the level and type of corrosivity.
In such corrosive sections, trays have an innate
advantage due to their higher thickness. Some heavy,
high-sulphur crudes also have issues with chloride salt
deposition, which can plug packing and trays, so
specially designed fouling-resistant trays such as
Sulzers VG AF have been used to increase column run
length and maintain tray performance.
If corrosion is not an issue, entrainment-resistant
trays, such as Sulzers Umbrella Floating Mini (UFM)
valve trays will give the highest effciency over the
widest operating range. The downward vapour fow
from these valves improves mixing at the tray deck
level and minimises entrainment, maintaining the high-
est possible effciency from turndown to full capacity.
For the kero through diesel fractionation sections, the
UFM valve trays are the best economical option. Also,
structured packing can be used to improve diesel-
AGO fractionation.
The wash oil section is characterised by low liquid
rates coupled with high vapour rates. This is ideal for
structured packings. With its high effciency, a well-
designed structured packing wash section can produce
high yields of high-quality AGO under a wide variety
of conditions. Sulzer MellapakPlus high-performance
structured packing provides a high combination of eff-
ciency and capacity.
The stripping section is typically designed with trays
with a variable open area to maximise their effciency.
Structured packing has been used in this section with
some success; its low pressure drop allows for better
stripping. However, plugging has been experienced
with some diffcult crude types.
For vacuum columns, structured packing is the clear
choice due to its lower pressure drop that allows mini-
mum pressure at the fash zone and maximum feed
vapourisation. For columns with a steam stripping
section, trays are the typical solution, although struc-
tured packing has also been used due to the lower
pressure drop that maximises stripping. There are
several vacuum columns successfully operating with
structured packing in their stripping section. As is the
case with crude columns, structured packing can be
subject to fouling when processing diffcult crudes.
Q
Is there a case for hydrogen production via coke
gasifcation over steam methane reforming?
A
Girish K Chitnis, Licensing Director, Technology Sales
and Licensing, ExxonMobil Research and Engineering
Company, girish.k.chitnis@exxonmobil.com
Hydrogen demand in modern refneries continues to
increase, with additional hydrotreating and hydro-
cracking units being deployed to meet the various
clean fuels mandates and product demands with
heavier crudes and additional conversion capacity.
Historically in low-conversion refneries, byproduct
hydrogen from catalytic reforming of naphtha for motor
gasoline production was principally used to meet hydro-
gen demand. Although catalytic reforming technology
has advanced over the years, resulting in increased
hydrogen yield due to lower operating
pressures,
1,2
it does not produce enough hydrogen to
want to keep the FCC at full feed rate. An option would be to
add partial conversion hydrocracking before the FCC. The
feed rate to the hydrocracker (HC) can be fixed so that the
FCC can be kept full, but there is substantial additional diesel
produced. Of course, some additional VGO is required, either
purchased or through additional crude runs.
The pressure of the partial-conversion HC needs to be
optimised. A high-pressure unit will be able to make ultra-
low-sulphur diesel (ULSD) directly. However, capital cost is
high. A less capital-intensive option is to use a lower-pressure
HC, with some additional post-treat of the diesel integrated
with the HC. If there are other difficult diesel stocks that need
to be treated in the refinery, such as FCC light cycle oil (LCO),
it may also be possible to economically utilise the post-treat
unit for them.
With regard to FCC LCO, this stream is an inferior
potential diesel component with its low cetane value, even
after hydrotreating. Partial hydrocracking of LCO (at
moderate pressure) can be considered to upgrade the
material. While some diesel molecules will be lost to lighter
material, the diesel retained is upgraded to a higher-quality
diesel blending stock.

What are some of the most attractive isomerisation


configurations and catalysts available to meet the
growing demand for light paraffin isomerisation? What
can be done to lower the equipment cost, such as the
recycle hydrogen compressor?

Bruno Domergue, product line manager, Axens, bruno.


domergue@axens.net: Gasoline specifications are getting
tighter when it comes to octane, sulphur, olefins and
aromatics, which is why interest in isomerisation will grow
accordingly. With its octane number up to 25 numbers higher
than the C
5
/C
6
cut and the total absence of sulphur, benzene
and olefins, isomerate is the ideal gasoline blendstock. From
a processing scheme standpoint, several isomerisation
options are available, which include the once-through (the
most simple) and the deisohexaniser recycle (the current
favourite), ranging up to complete iso-normal paraffins
separation and recycle (a must in the near future for some
geographic areas). Of the various types of catalyst available,
chlorinated alumina is the most profitable because it gives
the highest possible RON, being by far the most active
catalyst. It also has the lowest operating and investment
costs, owing to low operating temperatures, low recycle flows
(for recycle units) and no need for a recycle compressor.
Margaret Stine, marketing manager, process technology &
equipment refining market, UOP LLC, margaret.stine@
uop.com: The UOP Penex and Par-Isom processes are the
most common processes chosen to meet the worldwide needs
for light paraffin isomerisation. Penex uses the aluminum
chlorided catalysts I-8 Plus and I-82. The UOP catalysts
process all feedstocks, including those with high C
7
hydrocarbons or benzene content. Penex is frequently
coupled with deisopentaising and/or deisohexanising to
reach high isomerate product octane. The process is very
efficient and does not require a recycle compressor.
Par-Isom has been widely accepted as the best means of
converting an idle process unit, such as a fixed-bed reforming
unit, to light paraffin isomerisation. Little capital investment
is required, because the idle reforming unit contains all the
major equipment required for Par-Isom operation. No
additional feed treatment equipment is necessary beyond
typical hydrotreating. The newest Par-Isom catalyst, PI-242, is
the highest-activity non-chlorided catalyst available on the
market today. It is contaminant-tolerant and can recover
from water or sulphur upsets.
Stephan Zuijdendorp, technical sales manager,
isomerisation, Albemarle Catalyst Company BV, stephan.
zuijdendorp@albemarle.com: With the latest generation of
chlorinated alumina isomerisation catalysts, refiners save on
platinum investments and gain from long lifecycles and
high conversion. Albemarle delivers such catalysts with the
ATIS range. In units of IFP-Axens and ABB Lummus Global
design, and others, ATIS-1L and ATIS-2L bring such
improvements, which finally result in a significant cost
reduction. Depending on the conditions per unit, a technical
and economic evaluation will show if revamping from
hydrogen recycle to hydrogen once-through will be of
interest. The lower conversion in the liquid phase might be
compensated by more active catalysts and reduced yield loss
by cracking.

With the growing demand for isobutane as an


octane-enhancing refinery feedstock, as well as a
feedstock for the petrochemical route to propylene oxide,
what opportunities are available for increasing the
capacity and efficiency of existing deisobutanisers
similar to Figure 1?

QUESTI ONS & ANSWERS


PTQ Q2 2006
18
Kidextractor
Quarter page
After you have used
it the first time, the
Kid Extractor will
become your trusty
hydraulic-tube-
bundle-extractor
KIDExtractor Ltd.
P.O. Box 11, Zebbug
MALTA
Tel. 00356-21-462891
Fax. 00356-21-462755
MOBILE: 00356-94-20596
E-mail: idrojet@videobank.it
Website: www.kidextractor.com
Of the various types of catalyst available,
chlorinated alumina is the most profitable
because it gives the highest possible RON,
being by far the most active catalyst
ENQUIRY NO 214
www.ptqenquiry.com
PTQ&A copy 1 8/3/06 21:12 Page 3
Q&A copy 11.indd 3 1/6/12 12:56:32
dorfketal.indd 6 10/06/2013 16:24
Q&A copy 15.indd 2 16/09/2013 13:55
Q&A.indd 6 13/12/2013 11:48
www.eptq.com PTQ Q1 2014 27
temperature and pH, and tends to be localised in the
dew point zone. High TAN crudes and high sulphur
crudes are generally worst. Talk to chemical treatment
companies about corrosion monitoring and institute a
control programme, typically a combination of lming
amine to protect steel and neutralising amine to raise
pH.
Q
What is the best approach to cut NOx emissions from an
ethylene cracking furnace?
A
Paul A Diddams, Senior VP FCC Additives, Paul.Diddams@
matthey.com; Charles Radcliffe, Technical Consultant, FCC and
Rening, Charles.Radcliffe@matthey.com; Wayne Armstrong,
Business Manager Europe, FSU & Africa, Wayne.Armstrong@
matthey.com; and Joseph McCarney, Business Development
Manager, Joseph.McCarney@matthey.com, Johnson Matthey
Growing concern over air quality and local pollution
has led to an increased interest in abatement of NOx
emissions. NOx may be a problem in a number of renery
processes including steam methane reforming, uid
catalytic cracking as well as renery heaters and boilers.
NOx emissions may be substantially reduced (over 90%)
via selective catalytic reduction (SCR). Ammonia (NH
3
)
is introduced into the exhaust gas where it mixes with
the NOx and as the homogeneous NOx/NH
3
mix ows
through the channels of the honeycomb SCR catalyst the
NOx is neutralised at the surface producing nitrogen
and water (see Figure 1). SCR requires an exhaust gas
temperature in the range 200-480C. The process can be
optimised for minimum pressure drop.
In general, burner NOx depends on ame tempera-
ture. LoNOx burners are designed for staged
combustion to lower peak temperature. If not already
tted then this is the rst step; if tted, check opera-
tion for even burn and try adjusting primary/
secondary air ratios.
a) Eliminate nitrogen sources in the feed
b) Use low temp processes such as low NOx burners
c) Selective catalytic reduction needs an operating
temperature in the range 200-480C and needs reduc-
ing agent ammonia.
chloride is thermodynamically stable and will leave
the fractionator with the bottom product. Magnesium
and calcium chlorides thermally decompose at rela-
tively low temperatures. About 90% of the MgC
l2
and
about 15% of the CaC
l2
are converted to metal oxides
and hydrochloric acid (HCl) in the presence of water
vapour (hydrolysis). The HCl boils overhead in the
distillation column and dissolves in the overhead accu-
mulator drum. Severe corrosion of HCl occurs at the
dew point and the rst drops of condensate have a
very low pH of 12. Corrosion protection can be
achieved by neutralising the acidic species with suit-
able neutralising amines to a pH of 5.5-6.5.
Ammonia is not a suitable amine because of its high
volatility. The disadvantage of many neutralising
amines is the formation of corrosive amine salts. They
have a low pH and require a good water washing
system for salt removal. Kuritas patented ACF tech-
nology is a perfect corrosion inhibitor alternative,
based on a very strong organic amine. The formed
ACF salts are liquid salts with a neutral pH and very
high water solubility. Additionally, a lming amine can
be injected to provide a good corrosion protection of
the metal surface.
A perfect corrosion control programme combines
corrosion inhibitor injection with the execution of
analysis methods. The installation of electrical resis-
tance probes at several locations in the overhead
system may help to monitor corrosion rates continu-
ously. Operations, which mature daily, are the
analysis of the overhead receiver water draw. A
continuous online pH monitoring should be done
there, because dew point pH measurement is not
feasible. The iron concentration in the accumulator
drum should not exceed 1 mg/l in the dened pH
range of 5.5-6.5. When the salt content of the desalted
crude oil is 3.5 mg/l, the chloride content of the over-
head receiver water can be expected to be 15-20 mg/l.
Corrosion and fouling of the overhead system may
occur when the chloride content exceeds 20 mg/l. A
50 ppm concentration of hydrochloric acid is quite
corrosive to carbon steel.
The use of corrosion inhibitor programmes is a
cost-effective, attractive option. The
installation of dosing equipment can be
realised easily. In case of very severe
corrosion the selection of corrosion
resistant alloys (CRA) should be the
rst choice. They can resist corrosion
more effectively than carbon steel mate-
rials. CRAs are far more expensive but
longer lasting materials. These alloys
are mixtures of various metals such as
iron, molybdenum, nickel, chromium,
copper, cobalt, tungsten and/or
titanium.
A
Charles Radcliffe, Technical Consultant,
FCC and Rening, Charles.Radcliffe@
matthey.com, Johnson Matthey
Crude overhead corrosion is related to Figure 1 Reactions in selective catalytic reduction
NO
X
NH
3
N
2
H
2
O
Basic reactions
4 NO + 4 NH
3
+ O
2
4 N
2
+ 6 H
2
O
6 NO
2
+ 8 NH
3
7 N
2
+ 12 H
2
O
Undesired parallel reactions
SO
2
+ O
2
SO
3
NH
3
+ SO
3
+ H
2
O NH
4
HSO
4
Q&A.indd 7 11/12/2013 15:55
For more information, contact us at processplants@mustangeng.com,
call 713-215-8000 or visit www.mustangeng.com/process.
People Oriented...Project Driven

Were Big
without the burdens
As a Top 10-ranked rm in reneries and petrochemical plants by
Engineering News-Record, we have the expertise and international
presence to execute and deliver your project anywhere in the world.
And with a exibility and responsiveness that belies our size
thats why were big without the burdens.
mustang.indd 1 09/12/2013 11:29
I
n todays world of increasing costs and challenged
proft margins, achieving high availability (the most
useful measure of reliability) is rapidly becoming even
more essential for many downstream operators. In order
to achieve high availability in the long term, consistent
application of sound reliability principles is required.
The contributions to reliability of all those involved in
the design, operation and maintenance of a facility are
discussed in this article.
The article will look at answering the question Whose
responsibility is reliability?. In the authors experience
(at petrochemical and refnery facilities) the meaning of
reliability is frequently misunderstood. Using examples
from everyday life, the article will start with a defnition
of reliability. The application of reliability theory in
practice, with examples from the authors experience in
operating facilities, will be used to explain who can
affect the reliability of an industrial facility.
What is reliability?
For manufacturing facilities such as refneries, with
ever increasing pressure on margins, ensuring assets
remain operational has more signifcance now than
ever before. It is frequently the case that those facilities
that are reliable are also successful fnancially, environ-
mentally and in terms of safety.
Reliability is often seen as a function of the mainte-
nance or reliability departments and is typically
thought to describe equipment that does not fail.
When a piece of equipment does fail in some way, it is
frequently viewed as a reliability failure and a failure
of the maintenance department.
So what is reliability? Here is a useful defnition
from Lees Loss Prevention in the Process Industries: reli-
ability is The ability of an item to perform a required
function under stated conditions for a stated period of
time.
There are some key words here that are worth
discussing further: ability of an item to perform a required
function.
Reliability needs to be seen as provision of a function
and not that a component or piece of equipment will
never fail or require maintenance work. This important
distinction needs to be understood, in particular when
the challenges of designing a piece of equipment that
will not require some maintenance work or need
replacement due to wear and tear are considered.
Also of importance in the defnition is the principle
that reliability needs to be considered under stated
conditions in other words, within the design limits
of the equipment and normally for a specifed time
period. Examples are tubes in fred heaters that are
designed with a 100 000-hour life, and bearings used
in rotating equipment with a 10 000-hour design life.
To further explain these defnitions, consider what
reliability can mean for an item with which we are
probably familiar, the family car. A possible defnition
of reliability for a family car would be that it can carry
a family, with luggage, between two locations when
required.
Other functions might also include:
The capability to travel at a required speed and with
acceptable acceleration
To protect occupants in the event of a crash, typically
against industry standards such as the European New
Car Assessment Programme - Euro NCAP
1
Have climate control to keep occupants cool in warm
weather and warm in cold weather.
Now consider some stated conditions that might
apply:
Drive within speed limits
Drive on normal roads
Drive in normal/design temperatures
Fill with fuel before tank is empty
Get the car serviced at quoted intervals and get
defects corrected
Complete daily/weekly checks (by driver) and
get defects corrected.
The reliability principle that time limits need to be
considered is now becoming apparent:
The car needs to be serviced at specifed intervals
Fuel needs to be added based on fuel used and tank
level
Typically cars will not be expected to last more than
a certain total mileage (>200 000 miles would be
considered a good design life).
Consider the car example further. If the tyre on your
car only lasts for 20 000 miles, how can this be consid-
ered reliable?
A tyre has its own stated conditions (load, grip,
speed and wear ratings), and the function of getting
from A to B can still be reliably accomplished as long
as the tyre is checked regularly and replaced when
worn or damaged.
Consider the scenario of a puncture. Whilst the tyre
itself may have failed, the function of the car getting
from A to B can still be accomplished by the use of a
spare wheel/tyre.
This principle of having a spare piece of equipment
for those items that need regular servicing/mainte-
nance or that can suffer a random failure in service is
very common in process industries.
www.eptq.com PTQ Q1 2014 29
Processing Trends
Whose responsibility is reliability?
ROBIN PEATEY Foster Wheeler
proc trends copy 11.indd 1 11/12/2013 15:52
If a large number of identical components or systems
is considered, a number of early or infant mortality
failures could be expected, followed by a period with
low random failures and then a signicant increase in
failure rate as the equipment wears out.
Using the family car again as an example, in the
1970s-1980s, a brand new car was expected to suffer a
number of problems (failures) in the rst six months.
Then there would be a number of years during which
the car was seen to be reliable with only a few random
failures. As the car reached 10 years or so, many parts
started to wear out and most cars were deemed to be
at end of their life.
On the surface, this seems to be a sensible theory.
However a 1978 report for the aviation industry
5

showed that failures in real life seldom follow this
pattern. They identied that failures could be catego-
rised by six failure curves, with 11% of failures
aligning with a time based failure mechanism (see
Figure 2).
This is important to understand as there are still
many facilities and companies that
base their maintenance strategies
on the belief that equipment mainly
fails based on time or time in
service.
So if 89% of components do not
fail based on time in service, how
do they fail? Figure 3 shows the
percentage of component failures
that align with a number of varia-
tions of possible random failure
patterns.
It may be thought surprising that
components that exhibit a high fail-
ure rate during early service life
and then settle to a low random
failure rate for the rest of their
service life account for almost 70%
of components.
Reliability strategies clearly need
to focus signicant attention on
reducing the incidence of these
infant mortalities. The reduction in
early failures for modern cars is a
good indication that this is possi-
ble. However, strategies also need
to be in place to manage both the
random failures and the relatively
small number of failures that can
be predicted based on service life.

Reducing infant mortality
If a high percentage of equipment
failures are a result of infant mortal-
ities, then clearly signicant effort
needs to go into reducing or
preventing their occurrence.
The causes of this type of failure
are so numerous that a detailed
discussion is outside the scope of
30 PTQ Q1 2014 www.eptq.com
Another example: the function of moving a liquid is
typically achieved using a lead pump and installed
spare pump. The spare pump is used when the lead
pump has failed. This highlights the importance of
considering the function that is required, rather than
whether a component can wear out or fail. As long as
the spare can be used until the lead component can be
repaired or replaced, then the required function can in
principle still be satised. A cost/benet analysis is, of
course, a key element in developing the sparing
philosophy.
Reliability theory
Now that we have established that reliability is more
about maintaining a required function than on whether
an individual component fails, let us consider what we
can do to protect the reliability function when an indi-
vidual component within a system fails.
Classical reliability thinking believed that most
components followed the bath tub failure pattern
described in Figure 1.
3
Infant
mortality
F
a
i
l
u
r
e

r
a
t
e
Falling
apart
Optimum point
to replace
Useful working life
(random failures)
Time
Figure 1 Bath tub curve
F
a
i
l
u
r
e

r
a
t
e
5%
Time
F
a
i
l
u
r
e

r
a
t
e
2%
F
a
i
l
u
r
e

r
a
t
e
4%
Figure 2 Time-based failure curves
F
a
i
l
u
r
e

r
a
t
e
F
a
i
l
u
r
e

r
a
t
e
F
a
i
l
u
r
e

r
a
t
e
68%
14%
7%
Time
Figure 3 Random failure curves
proc trends copy 11.indd 2 11/12/2013 15:52
7130 South Lewis, Suite 335, Tulsa, OK 74136, call 1-918-496-7599 or visit our website www.callidus.com
2013 Honeywell International Inc. All rights reserved.
next generation combustion
UOP Callidus is an established world leader in process burner,
flare and thermal oxidizer technology and continues to break
new ground with innovative new products and solutions for the
most difficult combustion challenges.
UOP Callidus next generation process burners were the first to break the
10ppmv NO
x
barrier, our flares set the standard for smokeless capacity
and reliability, and our thermal oxidizers reach near 100% destruction
efficiency on various fuel streams. Our commitment to technical excellence is driven by our focus to
provide a custom designed solution for every project based on each customers specific requirements.
Our technology is tested and validated using the most advanced CFD capabilities combined with
exceptional design experience.
Our combustion products are in operation in refining and petrochemical facilities around the world.
When you need next generation combustion equipment and solutions for new construction or retrofit
operations, choose UOP Callidus.
uop next gen.indd 1 09/12/2013 11:34
heurtey.indd 1 10/12/2013 13:46
www.eptq.com PTQ Q1 2014 33
this article. However, in the authors experience, to
successfully reduce infant mortality failures an organi-
sation needs to have clear strategies and procedures in
place in a number of areas.
Design
Inadequate design typically accounts for only around
4% of failures. However, failures from design defects
can often be very severe, with high profle examples:
The Mercedes A-Class car made headlines in 1997
after rolling on its roof during a road test that included
a violent lane change manoeuvre known infa-
mously as the elk test
2

Explosion at the Nypro (UK) Caprolactam Plant at
Flixborough due to failure following a modifcation.
3
Let us consider in more detail an example of the
poor reliability of a pair of refnery pumps. Two
pumps were installed: a lead pump and an installed
spare pump. The spare pump was run once a month
for a few hours to ensure that it would operate
correctly in the event that the lead pump failed. Every
month it was found to run with high vibration and the
bearings were found to be damaged.
Investigation of the problem identifed brinelling of
the bearings. Brinelling is the surface deformation
caused by stress that exceeds material limits. It usually
occurs in situations where there is a load with a high
force distributed over a relatively small surface area.
Brinelling typically results from a heavy or repeated
impact load.
Investigation revealed that inadequate design of the
concrete foundations was allowing vibration (from the
operating pump) to be transmitted to the stationary
spare pump resulting in brinelling damage to the
stationary pump bearings.
Due to the prohibitive cost of correcting the design
fault, mitigation of the problem was to run each of the
pumps for a week at a time. Whilst not eliminating the
problem, this reduced the brinelling damage to a low
level and allowed for an increase in bearing life.
Providing design assurance is a signifcant topic in
its own right. Many tools and processes are required to
provide such assurance and can include hazard and
operability studies (HAZOPs), layers of protection anal-
ysis (LOPA) and qualitative risk assessment (QRAs).
In addition to the engineering designers themselves,
many different disciplines, including (but not limited
to) process engineers, maintenance engineers, commis-
sioning engineers, quality control engineers, inspection
engineers and operations specialists, are required to
contribute to ensuring design assurance.
Purchasing
If inappropriate or poor quality materials/equipment
are purchased, failures as a result of unexpected corro-
sion and/or many other failure mechanisms or
non-compliance can occur. Such failures are often
unexpected, can present signifcant safety risks and
often result in long system outages whilst the correct
equipment is obtained. Robust processes for the
specifcation, purchasing and then subsequent quality
control of purchased equipment is essential in order to
reduce such failures.
A worrying example was the alert issued by a
company that had found a quality issue with a valve
purchased from a low cost supplier. On inspection
prior to installation, the company discovered that the
bolted bonnet was held together by two short studs
tack welded together (rather than studs of the correct
length). If this valve had been put into service, a cata-
strophic failure is likely to have occurred.
Strategies for purchasing, including checks on the
accuracy/quality of purchase requisitions, policies to
purchase only from approved supplier lists, rigorous
bid evaluation processes (where technical compliance,
price and life cycle cost are assessed) and high quality
inspection are all essential to avoid material related
reliability failures.
Stores/warehousing
If equipment is not adequately stored to prevent corro-
sion and/or mechanical damage, it can either cause
delays due to the part needing to be replaced because its
poor condition has been identifed before installation, or
result in early failure if the item is put into service.
A frequent example of incorrect storage is for rotating
machinery (pumps, compressors, turbines). If not in
service (or for spare shafts that are stored horizontally),
the shafts can easily become bowed and require
replacement. This can be prevented by either regularly
turning the rotating elements or, for large compressor
shafts, by storing them vertically.
Also consider the storage of consumable items like
lubricating oil.
Figure 4 is a familiar sight, but what is the problem
with storing sealed oil drums outside?
Rain water can be drawn along the threads of the
sealing plug due to a partial vacuum being generated
in the drum due to changes in temperature (see
Figure 5)
Water contamination of the oil results
The presence of water in lubricating oils can
shorten bearing life down to 1% of the expected
The execution of functional
safety assessment and validation
must take place on every safety
project and must be carried out by
functional safety experts
Figure 4 Oil drum store
proc trends copy 11.indd 3 11/12/2013 15:52
bearing life or less, depending on the quantity present
Oil goes cloudy when contaminated with water
but bearings can lose 75% of life due to water before
oil becomes cloudy
Solutions to this problem are varied, and include
storing the drums horizontally, putting them in a
covered warehouse or protection of the drum with a
plastic cover (similar to a dustbin lid).
Installation
The following are some examples of where installation
(human error) can contribute to early failures:
Contaminated lubricating oil
Pump to motor shaft misalignment
Debris in pipes left over from construction
Filter elements not installed
Vent/drain plugs and caps not installed
Valves installed in the wrong direction
Lagging damaged (or not installed)
Field welds not painted
Spring hangers not unlocked
Wrong material used
Corrosion before entering service (preservation
required).
It is clear from the above that almost everyone
involved in the many aspects of a facility can play a
role in preventing these infant mortality failures, from
the project team responsible for design and construc-
tion to the operations team responsible for
commissioning and operating the facility.
Rigorous and comprehensive management systems,
processes and procedures, such as the ones in place at
Foster Wheeler, are required to be in place to minimise
the likelihood of infant mortality failures.
Managing random failures
Reducing the incidence of early or infant mortalities,
whilst essential, does not address management of
random failures that will occur during the life of a
plant once the infant mortality period has been
successfully passed. It should also be remembered that
maintenance work can reset the reliability clock, rein-
troducing the potential for infant mortality failures. It
is therefore important to make sure that maintenance
tasks are performed to add value and do not increase
the risk of reintroducing infant mortality (see Figure 6).
Typically, infant mortality accounts for 10-20% of
failures, with 80-90% of failures occurring in the
constant/random part of the curve.
34 PTQ Q1 2014 www.eptq.com
John Moubray
4
states that for the failure modes expe-
rienced in modern industrial systems, the following
strategies need to be applied:
33% require detective maintenance (checks for
hidden failures)
25% require predictive maintenance (check if it is
failing)
5% require preventive maintenance (time based
restoration)
33% can be allowed to run to failure
4% require a redesign.
Predictive maintenance
An understanding of the potential-failure (P-F) Curve is
important in understanding predictive maintenance (see
Figure 7).
Predictive maintenance can only be used where the
condition of a component deteriorates over a signi-
cant length of time, and where the condition of the
equipment can be monitored. The P-F Curve describes
the interval between the potential and functional fail-
ure of an item and supports the identication of
suitable condition monitoring tasks and the most sensi-
ble frequency for these tasks.
An example is the condition of bearings in rotating
equipment such as a pump. Whilst bearings typically
have a design life (maybe 10 000 hours), inuences
such as the condition and contamination of lubricant,
external stresses and speed mean that a bearing will
normally fail in the random part of the failure graph
without achieving its design life. This partly explains
why failures frequently do not reach the wear-out
portion of the bathtub curve, as Figure 1 shows.
Of use to us is that, as the condition of a bearing
deteriorates, its condition can be monitored by a vari-
ety of what are called condition based tasks including
vibration analysis, temperature surveys and, in some
cases, by analysis of oil samples.
Consider a strategy that uses vibration analysis to
monitor the condition of the bearing. Typically it is
possible to detect the early signs of the bearing starting
to fail about six weeks prior to the bearing actually
failing. The main benets of this are:
1) Repair can be planned for a time when the function
is not required or when the function can be maintained
by switching to a spare pump
2) The high cost of repairing the pump damage caused
by a collapsed bearing can be avoided.
Back to our car example: a failing wheel bearing on a
Rainwater
Figure 5 Water ingress into oil drum
Initial period of
infant mortality
Unnecessary
maintenance
Infant mortality
restarted
F
a
i
l
u
r
e

r
a
t
e
Time
Figure 6 Infant mortality reintroduced by unnecessary maintenance
proc trends copy 11.indd 4 11/12/2013 15:52
Reliable SwiSS Quality
API 618
Rod load up to 1'500 kN/335'000 Ibs
Power up to 31'000 kW/42'100 HP
Full Range:
You get moRe than Just a
PRocess gas comPRessoR
Lubricated up to 1'000 bar, non-
lubricated up to 300 bar
For highest availability: We recom-
mend our own designed, in-house
engineered compressor valves and
key compressor components
Designed for easy maintenance
We are the competent partner
with the full range of services
worldwide
www.recip.com/api618
YouR BeneFit:
lowest liFe cYcle costs
burckhardt.indd 1 7/3/12 21:39:05
36 PTQ Q1 2014 www.eptq.com
part of achieving high reliability and avoid-
ing expensive equipment repairs.
Note that for components other than bear-
ings the P-F Curve and the types of
condition monitoring tasks are likely to be
different.
Detective maintenance
Where failure of a component happens
quickly or its condition cannot be moni-
tored, a different approach is needed.
Examples of equipment that fail in this way
are electrical components such as light bulbs
or circuit boards. One minute they are work-
ing, the next they have failed. In reality they
still follow a P-F Curve. For a light bulb, the
time from seeing it start to ficker to it actu-
ally failing is so short (maybe only a few
seconds) that it is not helpful in preventing the failure.
To understand detective (or failure fnding) mainte-
nance, consider another car related example, failure of
a car brake light (see Figure 8).
Consider the hypothetical situation where a brake
light has failed after two months. In case A, where the
driver checks the brake light after 12 months, fnds it
failed and then replaces it, the availability in that year
would be only 17%.
If the failure fnding task of checking the brake light is
reduced from 12 to six months, as in case B, the avail-
ability of the brake light will have increased to 67%.
Cases C, D and E show how by increasing the failure
fnding task to monthly, then weekly, then daily, the
availability can be increased to 99.7%.
In the case of brake lights, most cars now have three
brake lights. This may explain why car manufacturers
recommend checking brakes on a fairly infrequent
basis (say on a monthly basis), as the loss of one light
can be tolerated for a short period due to the redun-
dancy provided by the other two brake lights.
Consider now a motorbike with only one brake light;
bikers checking the single brake light every time they
ride is a must.
This type of failure fnding inspection task is typi-
cally used for equipment whose function is not
continuous, and where its failure is hidden. A good
example would be a high level trip on a process plant.
The trip is only required to operate in the event that
the level reaches an unusually high level. In normal
operation if the trip is not working then it would not
be noticed until it fails to prevent a high level
condition.
So, whilst most disciplines are needed for the reduc-
tion or prevention of infant mortalities, managing the
failures occurring on the random part of the failure
curve is principally the responsibility of the operations
and maintenance groups. They need to identify and
implement the most appropriate type of task (time
based replacement, condition monitoring or fault
fnding) in order to either prevent loss of the system
function or to avoid an unnecessary repair cost.
car can be replaced at a time when use of the car is not
required. If the bearing is allowed to collapse, the
likely consequence is loss of control and a potential car
crash.
Vibration analysis can detect the onset of a bearing
failure about six weeks before it fails, explaining why
pump vibration analysis is normally conducted on a
monthly basis. If this is done less frequently the bear-
ing failure may not be discovered until too late and
functional failure is experienced. If the frequency of
analysis is increased, the higher cost will deliver little
added beneft.
Whilst use of vibration analysis is well proven, it is
worth mentioning that there are failure mechanisms
that can occur more quickly and result in the vibration
analysis missing the failure. In order to prevent the
actual catastrophic failure of bearings experiencing
such rapid failures it is often necessary to use the
complementary strategy of operator checks during a
shift. An operator using normal senses (although the
author would not recommend using taste!), which is
signifcantly less sensitive than vibration analysis,
would be able to detect a bearing problem maybe two
to three days before an actual failure. Hence, the rela-
tively low cost of an operator checking all pumps on
the plant whilst on a walkabout during a shift is a vital
0%
Time
C
o
n
d
i
t
i
o
n
100%
Failure starts
here (P)
Operator
touch detects
problem
Vibration analysis
detects problem
(Typically) problem
detected 6 weeks
before failure
(Typically) problem
detected 2-3 days
before failure
Equipment
fails here (F)
Figure 7 P-F Curve for a bearing failure
a
b
c
d
e
0 1 2 3 4 5 6 7 8 9 10 11 12
Time, months
12-month check / PM = 17% availability
Assuming brake light
fails after 2 months
6-month check / PM = 67% availability
3-month check / PM = 92% availability
Weekly check / PM = 98% availability
Daily check / PM = 99.7% availability
Figure 8 Effect of detective maintenance frequency on availability
proc trends copy 11.indd 5 11/12/2013 15:52
www.eptq.com PTQ Q1 2014 37
and RCM will identify the appropriate detective,
predictive and preventive maintenance tasks.
Predominantly, these tasks are performed by either the
operators of the facility or by the maintenance and
inspection departments.
References
1 European New Car Assessment Programme, November 2013, www.
euroncap.com
2 Kotabe M, Helsen K, Global Marketing Management, 2010, Wiley, 5th
edition.
3 Mannan S, Lees Loss Prevention in the Process Industries, 2012,
Elsevier, 4rd edition.
4 Moubray J, 1997, Reliability Centred Maintenance, Industrial Press Inc.
5 Nowlan F S, Heap H F, Reliability-Centered Maintenance, 1978, report
No A066-579 by United Airlines under the sponsorship of the Ofce of
Assistant Secretary of Defense, Washington, US.
Further reading
Centre for Chemical Process Safety, 2007, Guidelines for Risk Based
Process Safety, American Institute of Chemical Engineers.
Robin Peatey is Chief Engineer Maintenance and Commissioning
with Foster Wheeler, Reading, UK. He has over 25 years experience
at various petrochemical and renery facilities, holding a variety of
roles as Reliability Engineer, Area Maintenance Engineer and Design
Engineering Manager. He graduated from Loughborough University and
is a Member of the IMechE. Email: solutions@fwc.com
Whilst identifying suitable maintenance tasks can be
based on operating experience, processes such as failure
modes and effects analysis (FMEA) and reliability
centred maintenance (RCM) can provide a robust meth-
odology for the selection of suitable maintenance tasks.
It is worth being aware that sometimes the cost of
maintenance can exceed the beneft, and, where this is
the case, it can be acceptable to allow a component to
run until failure and then repair or replace it, assuming
the other failure consequences are acceptable.
Conclusion
From the discussion in this article it can be seen that,
in fact, reliability is everyones responsibility.
Robust strategies and procedures during the life-
cycle of the asset need to be in place to reduce or
prevent the occurrence of infant mortality failures. To
be effective, these will involve almost all disciplines,
including in the following areas:
Design for lowest life-cycle cost
Buy for total cost of ownership
Store equipment to retain its reliability (this includes
preservation)
Install/start-up with discipline and precision
Operate with care and within process limits
Maintain the plant/equipment with precision for
reliability.
In order to manage the unavoidable incidents of
random failures a good understanding of reliability
theory and application of analysis tools such as FMEA
www.eptq.com PTQ Q4 2013 77
backwash section. Omitting the
backwash section is therefore a
means to tolerate higher glycol
content in the feed gas. However, it
will also lead to increased amine
losses (higher opex) and is there-
fore not the most preferred
mitigation measure. Also, this
glycol mitigation measure is limited
to a certain glycol feed gas concen-
tration range only.
Option 2
Instead of increasing the capex,
there is also an option to go ahead
with a standard design of the
AGRU by accepting increased oper-
ational expenditures (opex). Not
installing a water-backwash section
in the absorber top will slightly
increase the vapour pressure losses
of glycol exiting the absorber top
compared to a plant with a
Options for mitigation
There are several options to reduce
the impact of glycols coming into
the AGRU with the feed gas. To
decide which option is most appli-
cable, knowledge of the use of any
glycol upstream of the AGRU is
very important: which type of glycol
was applied at which temperature,
what is the resulting glycol vapour
pressure in the feed gas, and so on.
Not considering glycol in the feed
gas can lead to malperformance of
the AGRU. Even though this point
may only be reached after several
years of operation, mitigation meas-
ures will usually be required sooner
or later. In the following, several
options are presented:
Option 1
Consider the glycol content in the
feed gas during the design phase;
additional margins in the solvent
circulation rate, heat exchanger
duties and absorber height can
compensate for glycol ingress into
the AGRU within a certain range.
As a consequence, mainly capital
expenditures (capex) will increase.
50 400
Temperature, F
V
a
p
o
u
r

p
r
e
s
s
u
r
e
,

p
s
i
a
Water
Activator
MDEA
MEG
TEG
DEG
Figure 16 Vapour pressure curves for thermal reclamation
and the lower temperature would
not allow complete vaporisation of
the main feedstock, leading to
unnecessary coke formation and
deposition in the feed zone.
Another option is to use a sepa-
rate riser, where the temperature
and catalyst-to-oil ratio can be opti-
mised and where the processing of
the LCN does not interfere with the
cracking of the main feedstock. In
this case, the increase in propylene
yield is still relatively small (about
2 wt% on top of the one achieved
with the main feedstock), with low
conversion per pass, a high gas
yield and a great portion of the
olens produced being further
converted due to secondary reac-
tions into aromatics. Another
option to dispose of the LCN is to
send it along with the straight-run
light naphtha to a steam cracker.
Table 2 compares the yields that
can be obtained when cracking
naphtha on a steam cracker versus
in a separate riser on a FCC unit.
The trends are similar in both
cases. As the severity increases, the
14 PTQ Q3 2013 www.eptq.com
ethylene yield increases; the
propylene yield in the steam
cracker decreases with increased
severity, whereas it increases in the
FCC unit. C
4
yields remain about
the same, but the naphtha yield
decreases with increased severity,
hence the once-through conversion
of naphtha on the steam cracker is
higher than on the FCC unit.
From a yields perspective, dispos-
ing of the LCN through the steam
cracker versus a separate riser in
the FCC could yield better
economic returns, but this needs to
be further investigated for each
specic case.
Processing C
4
s
There are several commercially
proven ways to upgrade C
4
s
produced from the FCC unit and
steam crackers.
6
Figure 1 illustrates
some of the options.
Although the butadiene content
of C
4
s from the steam cracker is
signicantly higher than that of the
FCC unit, the potential to recover
butadiene is lost when the light
recovery section of the steam
cracker and FCC unit are
integrated, because of the dilution
effect that the C
4
s from the FCC
unit have on the combined C
4
s
stream.
The main options for the produc-
tion of on-purpose propylene are:
Direct recycle of C
4
s to a separate
FCC riser
Via metathesis
Oligomerisation combined with
recycle of the oligomerate to be
processed in a separate riser of the
FCC unit.
The direct recycle of the C
4
s cut
involves a secondary high-severity
riser parallel to the main riser
designed to upgrade the C
4
s into
ethylene, propylene and catalytic
naphtha. Some fuel gas is also
produced. The products from both
risers merge at the reactor outlet
and travel as a common stream to
the main fractionator.
An advantage of this process is
that selective hydrotreatment for
the diene conversion of the C
4
s
stream as well as the removal of
fw.indd 3 10/06/2013 12:33
q4 basf.indd 8 12/09/2013 16:18
proc trends copy 11.indd 6 11/12/2013 15:53
Your Single Source for New or Retrot Internals
Why wait 10, 15 or 20 weeks for your delivery! Ask AMACS about our Fast
Track delivery for trays and a complete range of tower internals. With complete
in-house engineering and fabrication, we can use your existing drawings or
modify them to improve your process.
Trays (numerous options)
Sieve or perforated
Bubble cap trays
Cartridge trays
Dual ow
Bafe
Valve
www.amacs.com 24hr EMERGENCY SERVICE (281) 716-1179
Visit our new website at
www.amacs.com
Mesh-numerous alloys-plastics
Vane (Chevron)
Multi Pocket Vane Insertion Mistx

MIST ELIMINATORS RANDOM PACKING


Catalyst bed supports
Basket strainers
Nozzles Outlet/Inlet baskets
Distributors
Hub and header laterals
WEDGE WIRE SCREEN
Wide variety of random packing
types, sizes and materials in stock!
Ask us how our patented SuperBlend
2-Pac can increase your capacity
and efciency.
DISTRIBUTORS
& SUPPORTS
Manufactured to customer
specications or engineered to
meet performance requirements.
COALESCERS
Oil water separations Haze removal
from fuels Removal of tower wet reux
Caustic treater applications
STRUCTURED PACKING
Woven, sheet metal, and knitted
structured packing. Built to spec
or performance requirement.
amacs.indd 1 09/09/2013 17:14
HS-FCC for propylene: concept to
commercial operation
T
he fuid catalytic cracking
(FCC) process has undergone
a long evolution of hardware
and catalyst changes, from bed
cracking with amorphous catalyst
to short contact time riser cracking
with sophisticated zeolite catalyst
systems. Improvements to the
process have provided a wide
degree of fexibility to selectively
target the production of distillates
or gasoline, or propylene from
VGO and residue feeds, thereby
making FCC the most widely used
conversion process.
More generally, the objective of
the process is to produce high
valued products, and increasingly
this includes fuels and petrochemi-
cals, such as light olefns and
aromatics. At present, over 30% of
the worldwide propylene supply
comes from FCC-related processes
(FCC, RFCC, DCC). Fluctuating
product demand and price have
caused most new project develop-
ers to demand product fexibility
for long-term proftability and
process integration with petro-
chemical facilities for added
synergy and cost savings.
In order to respond to these
market demands, a new high sever-
ity down fow FCC (HS-FCC)
process has been developed by an
alliance of Saudi Aramco, JX
Nippon Oil & Energy (JX) and King
Fahd University of Petroleum and
Minerals (KFUPM), culminating in
a 3000 b/d semi-commercial unit in
operation since 2011 in Japan (see
Figure 1). The process provides a
high light olefn yield from a wide
variety of feedstocks utilising high
severity reaction conditions, a novel
down fow reaction system and
A FCC process provides a high light olen yield from a wide variety of feedstocks
utilising high severity reaction conditions and a novel down ow reaction system
NICOLAS LAMBERT Axens IWAO OGASAWARA JX Nippon Oil & Energy
IBRAHIM ABBA Saudi Aramco HALIM REDHWI King Fahd University of Petroleum & Minerals
CHRIS SANTNER Technip Stone & Webster Process Technology
proprietary catalyst. HS-FCC is
now available for licence from a
Global Alliance by Axens and
Technip Stone & Webster Process
Technology.
Features of HS-FCC
FCC utilises acidic zeolite catalysts
to crack heavy hydrocarbons into
lighter fuels such as gasoline and
distillate and, under more severe
conditions, into lighter olefns such
as propylene and butylene (and, to
a lesser extent, ethylene). Complex
secondary reactions that can
degrade the primary products to
less valuable components should be
limited to retain product selectivity
and refnery proftability. For
HS-FCC, the objective is to not only
improve the selectivity for normal
fuels production, but also to
maximise the potential of light
olefn and petrochemical produc-
tion at high severity. HS-FCC
provides a total system to maximise
product selectivity and, in particu-
lar, propylene yield. Three key
elements are required to attain this
objective:
www.eptq.com PTQ Q1 2014 39
Figure 1 HS-FCC semi-commercial unit
axens.indd 1 12/12/2013 10:58
40 PTQ Q1 2014 www.eptq.com
products and parallel bi-molecular
hydrogen transfer reactions leading
to paraffn formation and aromisa-
tion of naphthenes. Managing the
acid site density of the catalyst can
suppress hydrogen transfer and
isomerisation reactions to maximise
olefns production. When coupled
with ZSM-5 pentasil cracking cata-
lyst additives, the increased olefns
in the gasoline cut can be selec-
tively cracked to further increase
the propylene yield.
The HS-FCC catalyst uses a high
USY zeolite content system with
very low acid site density, formu-
lated to minimise hydrogen transfer
reactions for high olefn selectivity,
and low coke and gas selectivity.
This catalyst has been shown to be
more effective for propylene
production when coupled with
ZSM-5 additives (see Figure 2).
Commercial catalysts and HS-FCC
catalyst exhibited a similar trend in
gasoline and propylene yield as a
function of conversion (severity),
but the customised HS-FCC catalyst
was much more effective in feed-
ing the ZSM-5 additive with more
olefns, and more accessible linear
olefns, to produce more propylene.
1

Optimised reaction conditions
When targeting maximum petro-
chemicals production, HS-FCC
operates under more severe condi-
tions than conventional FCC. The
main reaction conditions applied
and the advantages and challenges
presented are shown in Table 1.
High reaction temperature
coupled with short contact time
increases the primary reactions
towards olefns, while limiting the
unwanted secondary reactions of
hydrogen transfer and thermal
degradation. A consequence of the
increased severity and short time is
the need for higher catalyst circula-
tion (catalyst-to-oil mass ratio, or
Highly selective catalyst and
additive system
Optimised reaction conditions
Down fow, short contact time
reaction system with rapid catalyst
separation.
The balance of these elements
and realisation at commercial scale
is the key to success.
Catalyst system
The catalytic cracking reaction
pathways are complex, with the
primary formation of olefnic
75 80 85 90
Conversion, mass%
75 80 85 90
Conversion, mass%
15
25
20
10
5
P
r
o
p
y
l
e
n
e

y
i
e
l
d
,

m
a
s
s
%
0
Commercial FCC catalyst (octane catalyst)
Commercial FCC catalyst (activity catalyst)
+10% ZSM-5 additive
+10% ZSM-5 additive
+10% ZSM-5 additive
HS-FCC catalyst
40
60
50
30
20
G
a
s
o
l
i
n
e

y
i
e
l
d
,

m
a
s
s
%
10
A
B
Figure 2 Proprietary catalyst boosts ZSM-5 effectiveness for more propylene
Advantages Challenges
High temperature High conversion and olens selectivity Increased thermal cracking, product
degradation
Short contact time Reduced secondary reactions and Reduced conversion, rapid mixing
thermal cracking and separation required
High catalyst/oil Increased catalytic cracking Very high catalyst circulation,
uniform ow, mixing and separation
Reaction conditions and advantages of HS-FCC in petrochemicals production
Table 1
FCC HS-FCC
Reaction T,C 500-550 550650
Contact time, s 25 0.51.0
Catalyst/oil, wt/wt 5-8 2040
Reactor ow Up ow Down ow
Typical operating conditions for FCC
and HS-FCC
Table 2
axens.indd 2 11/12/2013 12:17
www.eptq.com PTQ Q1 2014 41
C/O) to provide the required heat
to the reactor and suffcient catalyst
activity to achieve high conversion
at short contact time. The range of
operating conditions for a conven-
tional FCC and HS-FCC are
summarised in Table 2.
Down ow reaction (DFR) system
The specifc reaction conditions
with very high C/O result in
certain challenges in a conventional
up fow FCC riser reactor system,
where the catalyst required for the
reaction is lifted up the reactor pipe
or riser by the vaporised and
cracked hydrocarbon feed. In up
fow fuid-solid systems, the solids
or catalyst are conveyed upwards
against the force of gravity by drag
forces from the rising gases (hydro-
carbons). As a result, all riser
reactor systems have varying
degrees of catalyst back-mixing and
FCC
up flow riser
Reactor residence time
Low
conversion
HS-FCC
down flow
Over
cracking
Back
mixing
Feed + catalyst
Plug
flow
Feed + catalyst
Figure 3 Up ow vs down ow residence time proles
70 75 80 85 90
Conversion, wt%
60
50
40
G
a
s
o
l
i
n
e
,

m
a
s
s
%
30
A
0 10 20 30 40
C
2
-C
4
olefins, mass%
60
40
20
50
30
10
G
a
s
o
l
i
n
e
,

m
a
s
s
%
0
B
Riser
Down flow
Riser
Down flow
Figure 4 Selectivity benets of a down ow reaction system
4
Table 2
scan points representing one
as-built pipe. Evidently, that is only
one clash, but some systems will
register it as thousands of clashes,
making it difcult to identify the
real clash and to resolve it
effectively.
Ensuring that an accurately
designed pipe spool actually ts
correctly on site requires accurate
fabrication. Leading 3D design
solutions such as our PDMS or
AVEVA Everything3D can not only
generate fully detailed fabrication
drawings automatically, they can
also perform manufacturability
checks at the design stage to help
maximise fabrication quality. It is
also now possible to scan a
completed fabrication and compare
it against the design model, to
quickly verify its accuracy and
resolve any errors early.
Owner-operators considering
placing revamp projects should
review a contractors capabilities in
these areas. The most capable typi-
cally achieve less than 1%
design-related rework costs, even
on complex projects. Our vision of
plant design for lean construction
goes further than this; our goal is to
use laser scanning, among other
tools, to completely eliminate
rework in construction. This vision
is discussed further in a business
paper.
2
Plan the installation
Scheduling the revamp installation
involves similar considerations to
planning the original survey. EPCs
and owner-operators must collabo-
rate closely to achieve a
well-executed installation. Here, the
power of 3D design can add
considerable value.
Reverse engineering objects that
are to be removed enables the crea-
tion of accurate demolition
drawings and the determination of
weight and centre of gravity, which
informs the correct use of handling
equipment. Model animation
enables planners to evaluate
proposed task sequences and to
check that, for example, items can
be moved safely in the available
space constraints. Design review is
arguably even more important for
revamps than for new-build
www.eptq.com Revamps 2013 31
projects; revamps take place on
plants that may contain hazardous
chemicals, temperatures or
pressures. Evidently, these consid-
erations imply the need for
well-specied, specialist plant
design software.
Business process
The introduction to this article
referred to the need for efcient
business processes. That is hardly a
great insight, but, in the engineer-
ing industry, business processes are
inextricably linked with the engi-
neering and design technologies
that generate project information.
Best practice is therefore to select
solutions that can share informa-
tion efciently and reliably. An
efcient revamp project workow
can thus be achieved using 3D
design technology that integrates
both with laser scan data sources
and with engineering data sources,
so that engineering and design
information can be kept synchro-
nised as the project progresses.
From this, new design can automat-
ically generate accurate materials
requirements that feed into an
enterprise resource management
(ERM) system.
Such engineering, design and
information management technolo-
gies now exist and are in use on a
wide variety of new-build projects.
Their ability to support efcient
business processes becomes even
more important on revamp projects
with their need for on-time, right-
rst-time, low-risk installation.
References
1 Lighting the Way, www.aveva.com/
publications
2 Plant Design for Lean Construction:
Innovation for a new era in plant design, www.
aveva.com/publications
Gary Farrow is Vice President 3D Data Capture
with AVEVA in Cambridge, UK. He works with
customers in the use of 3D data capture
technology to increase productivity and to
advance the performance of AVEVAs LFM
software. A mechanical engineer, he has been
involved in 3D laser scanning from its inception
in the late 1990s, initially undertaking projects
delivering data and 3D models, including a huge
scanning project for Fluor/TCO in Kazakhstan.
I think Ive got liquid carryover.
what can I do about it?
Read more on this topic at
www.amacs.com
It happens in petrochemical
plants, refineries, and anywhere
else that the gas approaching
a compressor is wet. Traces of
aqueous or organic liquid escape
the inlet knockout drum, often
intermittently, and silently damag-
ing the compressor. Telltale signs
include pitting corrosion, salt
deposits, and diluted lubricants.
Phone:+1-713-434-0934 Fax: +1-713-433-6201
amacs@amacs.com
AMACS
Compressor suction drums:
Knockout
drums
Stage 2 Stage 1
Cooler
Typical 1980s mist eliminator technology
Instead of trying to repair
symptoms, look for the root cause,
which usually involves the mist
eliminator in the knockout drum.
Problems may include improper
mist eliminator specifications,
overloading, uneven velocity, waxy
deposits, liquid slugs, foaming,
incorrect installation, and several
other possibilities. New, high-
capacity, high-efficiency mist
eliminator technologies pay off the
first time you avoid shutdown.
rev aveva.indd 3 10/09/2013 11:20
axens.indd 3 11/12/2013 20:06
42 PTQ Q1 2014 www.eptq.com
maximum gasoline yield is about 5
wt% higher in the down fow
system. When olefns are of inter-
est, the more selective down fow
reaction environment can produce
substantially more light olefns
(C
2
-C
4
) at the same gasoline yield
compared to a conventional up
fow system (see Figure 4b).
Although the idea of a controlled
high severity, short contact time
down fow reaction has been
considered for some time, achieving
this successfully on a commercial
scale has been elusive. Extensive
plug fow reaction conditions, as
summarised by Cheng.
2
When plug
fow conditions are achieved, more
selective primary cracking results
in greater selectivity. FCC pilot
work demonstrating the effects of
short contact time and down fow
have been reported by Del Poso
3

and Abul-Hamayel
4
(see Figures 4a
and 4b). The general trend is that of
greater gasoline selectivity at short
contact time down fow, with a
maximum yield achieved at a
higher conversion level. This effect
is seen in Figure 4a, where the
refux along the walls, particularly
in the feed injection or catalyst
pick-up zone at the bottom of the
riser reactor. At very high C/O,
signifcant back-mixing is unavoid-
able. This problem is overcome in a
down fow reactor (DFR), where
both the catalyst and feed fow
downwards together (see Figure 3).
Down fow fuid-solid reaction
systems have been of increasing
interest in recent years to achieve
Figure 5 HS-FCC demonstration unit
60
100
90
80
70
50
40
30
20
10
Y
i
e
l
d
,

m
a
s
s
%
0
Dry gas
LPG
Gasoline
LCO+
Coke
Bench scale (0.1 bpd) Demo plant (30 bpd)
Propylene
10.5%
Propylene
10.6%
RON
98
RON
99
Figure 6 Bench scale vs demonstration scale results on low sulphur VGO at high severity
without ZSM-5
Lift air
Main air
Air outlet
Catalyst circuit
Air outlet
Catalytic
circulation
hopper
Catalytic
circulation
hopper
Injector
Separator
Total height: 35m (115ft)
Catalyst inventory: 20T
Max. catalyst circulation: 1.0T/min
Figure 7 500 b/d equivalent cold ow testing to scale up and optimise the reaction system
axens.indd 4 11/12/2013 12:17
www.eptq.com PTQ Q1 2014 43
obtained along with a very high
octane gasoline.
Work immediately began on
scale-up to a commercial unit.
Important lessons were learned
concerning equipment design, and
larger-scale cold ow work was
undertaken by JX in Japan at the
500 b/d equivalent scale to opti-
mise the feed injection zone and
separator design (see Figure 7).
This work was coupled with CFD
simulations to assist in larger-scale
equipment design.
6
Semi-commercial unit
With the successful demonstration
of HS-FCC technology at the 30
catalyst circulation loop and
reactor-separator equipment to
validate the design of the demon-
stration unit.
The demonstration unit (see
Figure 5) was operated from 2003-
2004 at the Aramco Ras Tanura
renery. Results from the demon-
stration unit validated the HS-FCC
concept, with good agreement
between 0.1 b/d pilot results and 30
b/d demonstration (see Figure 6).
5, 6

A low sulphur VGO was cracked
at high severity in both the pilot
and demonstration units using only
the new HS-FCC catalyst without
ZSM-5 additive. A very high
propylene yield, over 10%, was
pilot work at the 0.1 b/d scale
demonstrated the principle, catalyst
system and operating conditions,
but did not address how rapid
mixing, reaction and efcient cata-
lyst/gas separation can be achieved
at a large scale with a target resi-
dence time on the order of 0.5
seconds. On a commercial scale,
equipment design for very short
contact time with the mechanical
integrity to withstand high-velocity
catalyst circulation in a coking envi-
ronment requires extensive research,
development and demonstration.
R&D history
The challenges of developing
this new technology required a
systematic research program under-
taken by JX, KFUPM & Saudi
Aramco with the support of Japan
Cooperation Center, Petroleum
(JCCP). Early pilot work by both JX
and KFUPM in 1996-2000 demon-
strated the benets of high severity
operation at controlled short
contact time in down ow mode.
Aramco became an active partici-
pant in the scale-up effort to design
a 30 b/d demonstration unit. JX
conducted large-scale, 30 b/d
equivalent, cold ow testing of the
VGO + HDT VGO + VGO+
HC Btm VGO DAO AR
Feed SG 0.845 0.879 0.891 0.915
Reactor T, C 575 595 580 600
Conv, w% 93.2 83.7 83.0 82.4
Light olens, wt% 39 34 31 31
C
2
= 4 4 3 3
C
3
= 19 17 15 15
C
4
= 16 13 13 12
C
5
-220 gasoline, wt% 35 34 34 34
RON 98.5 98.1 98.1 98.4
Semi-commercial unit performance
Table 3
Feed Feed
Product
Quench
Catalyst
Catalyst
Injectors
Separator
CFD
simulation
Combined kinetic and
hydrodynamic model
27 lump kinetic model
DFR
(down flow
reactor)
Assembly of a large
number of small reactors
Feed Catalyst
Product
15
25
30
20
10
5
P
r
o
p
y
l
e
n
e

e
s
t
i
m
a
t
i
o
n
,

m
a
s
s
%
0
0 5 10 15 20 25 30
Experimental data,
mass%
Feed B
Feed A
Feed C
Feed D
Figure 8 Combined kinetic and hydrodynamic modelling assists design and scale-up
axens.indd 5 11/12/2013 12:17
44 PTQ Q1 2014 www.eptq.com
produced becomes a signicant
boost to the economics. The gaso-
line also has value beyond fuels,
with an octane of 98-99, olen
content of 25-40 wt% and 35-50
wt% aromatics.
The testing programme will
continue, with 100% residue crack-
ing trials to begin soon. With a
controlled short contact time, high
C/O and p lug ow reaction
system, HS-FCC is well adapted to
be highly selective for both light and
residue feed conversion to
petrochemicals.
Throughout the programme,
equipment evaluation, inspection
and reliability data continue to be
gathered to guide further develop-
ment and scale-up to a fully
commercial scale of at least 30 000
b/d. In parallel to this work, CFD
simulation of the DFR and separator
hydrodynamics are being combined
with a kinetic model to analyse the
results, validate the kinetic models,
and enable accurate predictions at
commercial scale for future feeds
and reactor congurations.

HS-FCC in the family of catalytic
cracking processes
The HS-FCC process expands the
operating window of catalytic crack-
ing to encompass heavier feeds and
greater propylene potential.
Commercial processes for high
propylene production from light
distillate feeds and residue feeds
include DCC,
8
high-propylene FCC
(HP FCC) and resid to propylene
(R2P). More severe conditions for
residue feeds to attain a higher
propylene yield have proven chal-
lenging in the past due to undesired
secondary reactions. High severity,
combined with an optimised cata-
lyst system and a controlled short
contact time DFR reaction system,
allows the new HS-FCC technology
to provide selective conversion with
lower fuel gas production and a
greater olen and petrochemicals
yield even with heavy residue feeds.
Indeed, the selectivity of the system
presents opportunities to crack a
wide range of conventional and
unconventional feedstocks.
The technology mapping by
severity and feedstock is shown in
Figure 9.
evaluate yields and product proper-
ties for widely different feeds and
to demonstrate equipment reliabil-
ity. Preliminary results showing
yields for several blends of VGO,
hydrocracker (HC) bottoms, DAO
and atmospheric residue are shown
in Table 3. Combined light olen
(C
2
-C
4
) yields of 30-40 wt% have
been demonstrated with 15-19 wt%
propylene and 4 wt% ethylene. The
yield of butenes is similar to propyl-
ene and offers opportunities for
greater petrochemical integration,
including oligomerisation and the
FlexEne conguration for even
higher propylene production.
7

These results are without the use of
post-separator quench injection,
which will improve olen selectiv-
ity further. The catalyst system
continues to be optimised for the
various feeds.
When viewed from a petrochemi-
cals perspective, the ethylene
b/d scale completed, it was time to
look forward to scaling up to a full-
sized commercial unit and to plan
for future licensing of the technol-
ogy. Several FCC licensors were
interviewed and evaluated before
Axens and Technip Stone &
Webster Process Technology were
selected to assist in the design of a
3000 b/d semi-commercial unit,
plan for a larger commercial unit,
and serve as exclusive licensor for
the HS-FCC technology, relying on
its extensive knowledge in FCC
and RFCC design.
A complete 3000 b/d HS-FCC
unit with main fractionator, gas
plant and ue gas treatment was
designed for the JX Mizushima
renery. Chiyoda Engineering
performed the detailed engineering
and construction of the plant (see
Figure 1), which was put on stream
in early 2011.
Performance trials are on-going to
21
25
23
19
17
15
13
11
P
r
o
p
y
l
e
n
e

y
i
e
l
d
,

w
t
%
9
Heavy residuals Light residuals VGO H
1
-H
2
VGO
DCC
HS-FCC
R2P
HP-FCC
Figure 9 Family of high-propylene catalytic cracking processes
HS-FCC
Oligomerisation
Polynaphtha
PRU
Steam
cracker
VGO
resid.
Aromatics
complex
Paramax
C
2

Mixed C
3
HCN
Mixed C
3
LCN
Oligomers
recycle
Paraffinic
raffinate
Polymer grade
ethylene
Polymer grade
propylene
Bz + PX + OX
Fuels
HS-FCC unit Petrochemicals
Figure 10 Integrated renery-petrochemical complex
axens.indd 6 11/12/2013 12:18
www.eptq.com PTQ Q1 2014 45
8 Dharia D, Increase light olens production,
Hydrocarbon Processing, April 2004.
9 Roux R, Upgrading of heavy cuts into max
olens through HS-FCC, JPI Petroleum Rening
Conference, Tokyo, 2012 and www.axens.net.
Nicolas Lambert is Technologist in Axens
Middle Distillates & Conversion Business Line,
focusing on FCC technology. He is a graduate
of Arts & Mtiers ParisTech.
Iwao Ogasawara is Facility Planner of Technical
& Engineering Service Department, Rening
Technology & Engineering Division, JX Nippon
Oil & Energy Corporation. He holds BS and MS
degrees in chemical engineering.
Ibrahim A Abba is Chief Technologist of the
Chemicals Research Division of Saudi Aramco
Research & Development Center. He holds a
PhD from the University of British Columbia.
Halim Redhwi is the CEO (A) of Dhahran
Techno-Valley Company and a Professor in
the Chemical Engineering Department, King
Fahd University of Petroleum & Minerals, Saudi
Arabia. He holds BS, MS, and PhD degrees in
chemical engineering.
Chris Santner is Senior Director of Catalytic
Cracking Rening Technology with Technip
Stone & Webster Process Technology. He holds
BS and MS degrees in chemical engineering
from the University of Houston.
Technip Stone & Webster Process
Technology are now offering
HS-FCC technology on behalf of
the HS-FCC Global Alliance team.
FlexEne is a mark of Axens.
References
1 Maghrabi A, HS-FCC process for maximized
propylene production, 10th Annual Saudi-
Japanese Symposium on Catalysis in Petroleum
Rening and Petrochemicals, Dhahran, 2000.
2 Cheng Y, Downer reactor: from fundamental
study to industrial application, Powder
Technology, 183, 2008.
3 Del Poso M, Development of ultra selective
cracking technology, 2nd IFP and S&W FCC
Forum, The Woodlands, Texas, 1996.
4 Abul-Hamayel M A, Comparison of downer
and riser based uid catalytic cracking process
at high severity conditions: a pilot plant study,
Petroleum Science Technology, 22, 2004.
5 Redhwi H, Meeting olens demand in a
novel FCC technology, 18th World Petroleum
Congress, South Africa, 2005.
6 Okazaki H, High severity Fluidized Catalytic
Cracking (HS-FCC) go for propylene!, 20th
World Petroleum Congress, Doha, 2011.
7 Ross J, (R)FCC product exibility with FlexEne,
WRA Downstream Asia, Singapore, 2011 and
www.axens.net.
With the option to operate at
conventional severity or high sever-
ity, the refner will have the ability
to select an operating mode and
feedstock best suited to the prevail-
ing economic conditions. A high
severity product slate rich in olefns
and aromatics also makes integra-
tion with petrochemicals plants
more attractive so that the natural
synergy of shared intermediate
products and recovery schemes can
be realised.
9
An example of HS-FCC
integration with a petrochemical
complex is shown in Figure 10.
Global Alliance for commercialisation
The HS-FCC technology is the
product of systematic process
research, catalyst development,
pilot work, 30 b/d demonstration
unit testing, and ongoing semi-
commercial operation and testing at
the 3000 b/d scale. These successful
results and the modelling tools
developed for further scale-up
make the technology ready for
commercialisation. Axens and
www.eptq.com PTQ Q4 2013 97
technology, as the worldwide
industrial standard for clean syngas
production, it provides a clean
hydrogen product and enables
economic carbon capture.
Lurgi Rectisol, MPG, OxyClaus and Purisol are
marks of Air Liquide Global E&C Solutions
(Lurgi GmbH). Selexol is a mark of UOP, a
Honeywell company.
Max-Michael Weiss is Director Innovation,
Clean Conversion, with Air Liquide Global E&C
Solutions/Lurgi GmbH. He graduated as Diplom
Chemie Ingenieur (chemical engineering) from
the Technical University of Karlsruhe, Germany.
Helmut Heurich is Director for Renery
Applications in the HyCO Product Line in
Global Engineering & Construction Solutions
of Air Liquide. He studied process technology
at the Technical University of Braunschweig.
Delphine Roma is the Global Marketing
Manager in charge of the rening industry
within Air Liquide Global E&C Solutions. She
holds an MSc and engineering degree from the
cole des Ponts et Chausses in Paris.
Stefan Walter is Head of Department,
Gasication Technologies, with Air Liquide
Global E&C Solutions/Lurgi GmbH. He
graduated as Diplom Verfahrensingenieur
(process engineering) from the Technical
University of Aachen, Germany.
also integrated solutions for steam/
energy generation and CO
2

handling. Air Liquide can also
provide an (over-the-fence) supply
of air gases, power and steam,
CO
2
compression/liquefaction and
transportation.
Conclusion
In the context of ever more stringent
environmental regulations for refn-
ers, there is a trend to increase
residue conversion with hydrocrack-
ers. Depending on the conversion
rate of the units, the heavy bottom
yield will range from 10-20%. Using
Air Liquide Global E&C Solutions
Lurgi MPG technology-based
hydrogen production allows a
refner to transform these residues
into an amount of hydrogen that
balances the refnerys needs.
MPG is a proven and reliable
technology for securing the hydro-
gen supply to a refnery. It further
avoids the production of petroleum
coke, and helps to consume much
less natural gas and water. In
combination with the Lurgi Rectisol
of most refnery off-gases, reduc-
tion in natural gas and water
consumption, and the recovery of
CO
2
for EOR or sequestration.
With the confguration shown in
Figure 7, the feedstock to the MPG
unit is normally reduced to that
amount needed to satisfy the
demand of the complex. If syngas
production exceeds the amount
needed to produce the required
hydrogen (depending frst of all on
the crude quality/origin), the
surplus syngas can be used for
power generation with gas turbines
or the production of chemicals/
fuels. Air Liquide Global E&C
Solutions can provide the applica-
ble technologies (see Figure 7).
Besides MPG technology, including
air separation, CO shift, syngas
cleaning, Lurgi Rectisol, PSA and
methanation, the company also has
technologies for sulphur recovery
(preferably OxyClaus, since oxygen
is available), technologies for chem-
icals (such as methanol and
propylene) and fuels production
(for instance, Fischer-Tropsch), but
www.eptq.com PTQ Q3 2013 99
cracking reactions over the strictly
thermal coking reactions that occur
in the traditional delayed coker
operation. During development, it
has been observed that OptiFuel
Technology shifts the delayed coker
yields towards more valuable
products, with reduced amounts of
dry gas and coke.
Pilot plant verication
Albemarle and OFTG have
conducted a series of pilot plant
runs at Penn State University (PSU).
The scope of these studies has been
to quantify the benefts of the tech-
nology as a function of additive
composition, feed properties and
operating conditions.
vessels needed for additive mixing
and storage.
The yield improvements seen with
this technology are hypothesised to
be the result of reactions in both the
liquid and vapour phases, which are
directly infuenced by the additive.
The active sites of the additive are
intended to preferentially catalyse
Albemarle has designed an addi-
tion system to ensure proper mixing
of the solid and liquid portions and
to avoid solids settling. The supply
consists of a liquid carrier and the
additive supplied to the refnery in
bags or bulk shipments for mixing
on-site. The additive injection
system design minimises the size of
6
10
8
4
2
0
-2
4
A
b
s
o
l
u
t
e

d
e
l
t
a
s
,

w
t
%

o
f

f
e
e
d
6
A1
4.32
8.83
4.51
4.01
8.77
4.76
3.38
6.52
3.14
4.43
3.00
1.42
1.44
0.44
0.99
1.57
0.67
2.25
2.28
0.43
1.85
1.49
0.30
1.79
Coke, %
C
3
+ liquid, %
Fuel gas, %
A2 A3 A4 B1 B2 B3 B4
Feed and formulation
Figure 3 Changes in pilot plant yields with application of OptiFuel Technology: pilot plant
runs at Penn State University
Base Predicted
Coke 32.2% 28.5%
Dry gas 5.6% 4.4%
C
3
+ liquid yields 62.2% 67.1%
Projected yields for commercial
application of OptiFuel Technology
Table 4
www.eptq.com PTQ Q2 2013 43
Very Heavy Crude Upgrading Long Term R&D
Opportunities, 1994.
2 Yui S, Chung K H, Syncrude upgrader revamp
improves product quality, Oil Gas J, 2007, Vol.
105, 46, 52.
3 Chrones J, Germain R R, Bitumen and heavy
oil upgrading in Canada, Fuel Sci Tech Int, 1989,
7, 783.
4 Rana M S, Samano V, Ancheyta J, Diaz J
A I, A review of recent advances on process
technologies for upgrading of heavy oils and
residua, Fuel, 2007, 86, 1216.
5 Speight J G, The Chemistry and Technology
of Petroleum, 2007, 4
th
ed, CRC Press/Taylor &
Francis, Boca Raton, FL.
6 Sayles S, Romero S, Understand differences
between thermal and hydrocracking,
Hydrocarbon Process, 2011, Sept, 37.
7 Martinez J, Sanchez J L, Ancheyta J, Ruiz R
S, A review of process aspects and modeling
of ebullated bed reactors for hydrocracking of
heavy oils, Catal Rev Sci Eng, 2010, 52, 60.
8 Yui S, Producing quality synthetic crude oil
from Canadian oil sands bitumen, J Jpn Petrol
Inst, 2008, 51, 1.
9 Yui S, Athabasca oil sands produce quality
diesel and jet fuels, Oil Gas J, 2000, Vol. 98, 47,
58.
10 Yui S, Chung K H, Processing oil sands
bitumen is syncrudes R&D focus, Oil Gas J,
2001, Vol. 99, 17, 46.
Variante 4a
A Member of Unitherm-Cemcon Group
CS Combustion Solutions GmbH
Simmeringer Hauptstrasse 24
A-1110 Vienna
Austria
T+43 1 740 40 2812
F +43 1 740 40 2815
sales@comb-sol.com
www.comb-sol.com
your experts
n ACID/TAIL GAS
n LIQUID SULFUR
n SPENT ACID
n CHEMICAL/PETROCHEMICAL RESIDUES
The team of CS Combustion
Solutions is experienced in
engineering, supply and
commissioning of vertical
and hori zontal burners
and combustors for
more than 20 years.
11 Wadsworth D, LC-Fining options for heavy
oil upgrading, Proceedings of the NPRA Annual
Meeting, San Diego, CA, 9-11 March 2008.
12 Ordorica-Garcia G, Croiset E, Douglas P,
Elkamel A, Gupta M, Modeling the energy
demands and greenhouse gas emissions of the
Canadian oil sands industry, Energy Fuels, 2007,
21, 2098.
13 Morawski I, Mosio-Mosiewski J, Effects of
parameters in Ni-Mo catalysed hydrocracking
of vacuum residue on composition and quality
of obtained products, Fuel Process Technol,
2006, 87, 659.
14 Danial-Fortain P, Gauthier T, Merdrignac
I, Budzinski H, Reactivity study of Athabasca
vacuum residue in hydroconversion conditions,
Catal Today, 2010, 150, 255.
15 Ding F, Ng S H, Xu C, Yui S, Reduction of
light cycle oil in catalytic cracking of bitumen-
derived crude HGOs through catalyst selection,
Fuel Process Technol, 2007, 88, 833.
16 Botchwey C, Dalai A K, Adjaye J, Kinetics
of bitumen-derived gas oil upgrading using a
commercial NiMo/Al
2
O
3
catalyst, Can J Chem
Eng, 2004, 82, 478.
17 Yui S, Sanford E, Kinetics of aromatics
hydrogenation of bitumen-derived gas oils, Can
J Chem Eng, 1991, 69, 1087.
18 Yui S, Sanford E, Mild hydrocracking of
bitumen-derived coker and hydrocracker heavy
gas oils: kinetics, product yields, and product
properties, Ind Eng Chem Res, 1989, 28, 1278.
19 Yui S, Removing diolens from coker
naphtha necessary before hydrotreating, Oil
Gas J, 1999, 97, 36.
20 Chang A-F, Liu Y A, Predictive modeling of
large-scale integrated renery reaction and
fractionation systems from plant data. Part 1:
hydrocracking processes, Energy Fuels, 2011,
25, 5264.
Anton Alvarez-Majmutov is an NSERC Visiting
Fellow at CanmetENERGY working on bitumen
upgrading process modelling and simulation.
He holds a PhD from Mexican Institute of
Petroleum (IMP).
Jinwen Chen is a Senior Research Scientist and
Group Leader at CanmetENERGY. He holds
a PhD in chemical engineering from Tianjin
University.
Mugurel Munteanu is a Lead Process
Engineer at CoSyn Technology, a division of
WorleyParsons, in Edmonton, Canada. He holds
a PhD in chemical engineering from Laval
University, Canada.
canmet.indd 6 08/03/2013 13:04
albemarle.indd 4 07/06/2013 20:04
air liquide.indd 6 12/09/2013 16:47
axens.indd 7 12/12/2013 10:58
sandvik.indd 1 09/12/2013 11:31
Optimising distillation column
product quality
I
n an attempt to ensure that prod-
uct specifcations are always met
in distillation unit operations,
refnery operations will often run a
process unit at higher reboiler
duties than required, resulting in
excessive energy use and sub-
optimal product yields. Process
units are run in this less than eff-
cient fashion to provide enough of
a cushion for process variability
resulting from changes in process
conditions, such as changing feed
compositions, feed rate and feed
temperature. Therefore, accounting
for process variability can be crucial
when optimising distillation unit
operations. This article shows how
to determine the optimum operat-
ing targets when product
specifcation constraints need to be
met in the presence of process
variability.
We present a methodology that
uses a process simulator along
with Monte Carlo simulation to
account for changing process
conditions. By using a process
simulator, the performance of the
unit can be evaluated outside the
current operating range. To illus-
trate the application of the
methodology used, the optimisa-
tion of the operations of a
propylene-propane splitter distilla-
tion unit was chosen.
1
Figure 1
provides a schematic of a typical
propylene-propane splitter distilla-
tion unit. A mixture of
approximately 75% propylene and
25% propane enters the unit.
Reboiler steam provides the
energy required to separate the
feed into a side draw propylene
product and a bottoms propane
product. Chemical grade propylene
Process simulation and random sampling were used to optimise product targets
for a propylene splitter unit
JOSE BIRD and DARRYL SEILLIER
Valero Energy Corporation
specifcations were assumed to be
a minimum 93% propylene purity
and a maximum 15 ppm methyl
acetylene (MA) concentration. A
concern that arises when operating
the propylene splitter close to the
propylene purity specifcation is
the corresponding increase in the
MA concentration of the propylene
product, which might result in
off-spec product. If this happens,
the product is downgraded to
refnery grade propylene.
Accurately predicting the perfor-
mance of the distillation unit at
lower propylene purities needs to
consider changes in feed composi-
tion, feed temperature and feed
fow rate as operations is facing
this challenge. The use of a process
simulator with cases generated
with Monte Carlo random
sampling allowed for proper
modelling of the distillation
process mechanisms in the pres-
ence of process variability.
The following steps summarise
the methodology used in this
study:
Process simulation model cali-
brated to current unit operations
Probability distributions gener-
ated for input process variables
Monte Carlo simulation used to
generate process simulation cases
Process simulations conducted
for all cases generated.
Steps 2-4 are repeated for the
different propylene purity scenarios
considered. The process simulation
results are then used to generate
joint probability density functions
of propylene purity percentage
and MA concentration. The result-
ing joint probability distributions
are used to assess the ability of
www.eptq.com PTQ Q1 2014 47
Feed
155
81
1
10
Reflux
Cooling
water
Splitter
vent
Side draw:
propylene
Bottoms:
propane
Steam
Figure 1 Propylene splitter schematic
valero.indd 1 11/12/2013 12:21
48 PTQ Q1 2014 www.eptq.com
were considered to be primary
sources of process variability: feed
temperature, feed rate, feed propyl-
ene composition, feed MA
concentration and propylene prod-
uct percent purity. These inputs
were assumed to follow independ-
ent normal distributions. Historical
data with the APC system turned
the process to meet product
specifcations at different operating
targets and to determine the associ-
ated optimum unit operating
targets.
A detailed description of the
analysis methodology is provided
below, followed by results of analy-
sis and conclusions.
Analysis methodology
To examine the effect of process
variability on the ability of
the process to meet product
specifcations, Monte Carlo simula-
tion was used to generate a set of
process simulation cases that
captured existing process variabil-
ity. The following model inputs
30%
50%
40%
20%
10%
0%
95.4 95.7 96.0 96.3 96.6 96.9 97.2 97.5 97.8
Propene product
= 96.92
= 0.38
15%
25%
20%
10%
5%
0%
83.8 88.8 93.8 99.8 103.8
Feed temperature
= 93.25
= 4.90
15%
25%
20%
10%
5%
0%
4.5 7.5 10.5 13.5 16.5 19.5 22.5 25.5 28.5
Feed MA
= 17.87
= 5.34
15%
25%
20%
10%
5%
0%
6625 6875 7125 7375 7625 7875 8125 8375 8625
Feed rate
= 7536.72
= 450.12
15%
25%
20%
10%
5%
0%
67.5 69.0 70.5 72.0 73.5 75.0 76.5 78.0 79.5
Feed propene
= 73.07
= 2.59
Distribution of product propene Distribution of feed temperature
Distribution of feed rate Distribution of feed MA
Distribution of feed propene
Figure 2 Distribution of input variables
valero.indd 2 11/12/2013 20:22
on were used to build the probabil-
ity distributions of these inputs.
Figure 2 shows histograms and
the calculated normal density func-
tions for each of the model inputs.
The mean and standard deviation
for each input random variable
were calculated and used to model
the corresponding distributions.
The propylene purity value
generated for each of the cases was
then used to estimate the propane
concentration in the propylene
product. The propane in the
propylene product was set as one
of the process specications in the
simulation. The concentration of
the remaining trace components
was assumed to be a constant when
computing the propane concentra-
tion in the propylene product.
Figure 3 highlights the analysis
methodology. First, 200 independ-
ent random samples for each of the
model inputs is generated for the
propylene purity target scenario
considered. The propylene purity
target was assumed to be equal to
www.eptq.com PTQ Q1 2014 49
the mean of the propylene purity
distribution. To generate the differ-
ent propylene purity target
scenarios, the mean of the propyl-
ene purity was shifted while the
standard deviation was kept the
same. Two hundred cases for each
propylene purity target scenario
were generated, with each case
representing a different combina-
tion of the model inputs based on
the 200 independent random
samples generated. The SAS
Analytics procedure PROC
SIMNORMAL
2
was used to gener-
ate the random samples. The SAS
Analytics scalable software envi-
ronment was selected for this
project due to its extensive statisti-
cal and charting capabilities.
The 200 cases were then evalu-
ated using a process simulator. The
Petro-SIM 4.1 process simulation
software was used, and the model
inputs entered using an Excel
spreadsheet interface within Petro-
SIM software. The results from the
simulations were then used to esti-
mate the incremental prot for each
propylene purity scenario against
the base case. The calculated incre-
mental prot gures were then
used to generate a prot response
surface.
3,4,5
The prot response
surface was generated using two
SAS Analytics procedures: Proc
G3GRID and PROC G3D.
2
The
simulation results were also used to
generate contour maps representing
the joint probability distribution of
propylene purity and MA concen-
tration in the propylene product.
The contour maps of the joint prob-
ability distributions were generated
using the SAS Analytics procedure
PROC KDE
2
, which uses kernel
density estimation.
The process simulator was cong-
ured to use three process
specications: condenser tempera-
ture, propane weight percent in the
propylene product, and propylene
weight percent in the bottoms
propane product. The propane
weight percent in the propylene
product was calculated from the
30
25
20
15
10
5
M
A
0
90 92 94 96 98
Purity, %
Propylene product percent purity
and methyl acetylene distribution
94.9% target
0.13
0.12
0.11
0.10
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
Feed temperature
Feed C
3
=
Feed MA
Process
simulator
Feed rate
Propylene purity
PP splitter optimisation
P
r
o
p
e
n
e

in

p
r
o
p
a
n
e
,

%
6
P
r
o
f
i
t
,

$
M
M
3
0
3
6
4
.0
0
3
.2
5
2
.5
0
1
.7
5
1
.0
0
9
4
.
0
0
9
4
.
7
5
9
5
.
5
0
9
6
.
2
5
9
7
.
0
0
P
u
r
it
y
, %
Figure 3 Analysis methodology schematic
valero.indd 3 11/12/2013 20:22
50 PTQ Q1 2014 www.eptq.com
Excel spreadsheet interface and
capability to run multiple cases.
To model the propylene-propane
splitter, a preconfgured distillation
column unit consisting of a reboiler
and a condenser was used. Figure 5
is a simplifed process fow diagram
of the distillation unit used.
The output streams consisted of
an overhead vent stream, a propyl-
ene side draw stream and a
bottoms propane product stream.
The feed stream fow rate, tempera-
ture and pressure were specifed as
inputs. Compositions of each
component were specifed accord-
ing to the reconstituted feed stream
calculated from the output streams.
Table 1 provides feed fow rate,
pressure and temperature, and
Table 2 the calculated feed
compositions.
The distillation column was
confgured as shown in Figure 6.
Two product streams exit the
tower: a bottoms propane product
and a propylene side draw product.
A vapour stream leaves the column
as an overhead vent. Column pres-
sure ranged from 265-282.6 psia
from the top to the bottom of the
column. The variability of the over-
head temperature was minor, so it
was kept constant at 100F (38C)
for all simulations. The process
simulator was confgured to calcu-
late reboiler duty as well as fow
rates and compositions for the three
output streams. Table 2 provides
Base model description
To build the propylene-propane
splitter model, data historian infor-
mation including fow rates,
compositions, pressures and
temperatures were gathered for
feed and product streams.
Distillation column data including
number of trays as well as pressures
and temperatures for condenser and
reboiler were obtained. Component
material balances using feed and
product fow rates and composi-
tions were performed to test the
validity of the data. Propylene and
propane product compositions and
fow rates were used to reconstitute
the feed stream.
Petro-SIM 4.1 process simulation
software was selected as the process
simulator due to its user-friendly
difference after subtracting the
propylene purity weight percent
and the concentration of trace
components from 100%. The weight
percent propylene in the propane
product was varied from 1-4% and
was assumed to be a constant for
each propylene purity scenario. To
reduce the number of cases required
to estimate the incremental proft
response surface, the experimental
central composite design confgura-
tion shown in Figure 4 was used.
3,4

This confguration allows for
consideration of second order terms
in the construction of the proft
response surface. The different
scenarios were defned by the
combination of average propylene
purity and the propylene concentra-
tion in the propane.
[95.5, 1.0]
200 cases
[95.5, 2.5]
200 cases
[97.0, 1.4]
200 cases
[94.0, 1.4]
200 cases
[97.0, 3.6]
200 cases
[94.0, 3.6]
200 cases
[94.0, 2.5]
200 cases
[97.0, 2.5]
200 cases
[95.5, 4.0]
200 cases
Figure 4 Central composite design scheme
Splitter
feed
Splitter
vent
Propene
product
QC
Propane
QR
Refinery
plant to
Crude-100
T-100
Figure 5 PetroSIM simulation model of the
distillation unit
Conditions Splitter feed Propane bottoms Propylene product
Temperature, F 76 137 118
Pressure, psig 272 283 271
Flow rate, bbl/day 7094 1432 5561
Stream conditions
Table 1
Composition Splitter feed Propane bottoms Propylene product
Propane, wt% 21.27 95.48 2.92
Propylene, wt% 77.86 2.50 97.01
M-acetylene, wt% 0.00 0.01 0.00
Ethane, wt% 0.47 0.00 0.07
i-Butane, wt% 0.31 1.56 0.00
1-Butene, wt% 0.02 0.10 0.00
i-Butene, wt% 0.03 0.15 0.00
n-Hexane, wt% 0.02 0.10 0.00
Cyclohexane, wt% 0.02 0.10 0.00
Stream compositions
Table 2
valero.indd 4 11/12/2013 20:22
www.eptq.com PTQ Q1 2014 51
prot response surface constructed
using the results of the nine
scenarios. To generate the response
surface, the SAS Analytics proce-
dure PROC G3GRID was rst used
to generate a ner grid covering the
propane, and 94.4% purity with
3.6% in the propane. The 94.4%
purity with 3.6% propylene in the
propane is preferred, since it
resulted in 0% off-spec product.
Figure 7 shows the incremental
the calculated compositions for the
propylene and propane products
for the base case.
The process simulations used
three model specications:
Condenser temperature
Propylene weight percent in the
bottoms propane product
Propane weight percent in the
propylene product.
The condenser temperature was
xed at 86F (30C), which repre-
sents the average condenser
temperature based on actual opera-
tions. The variability of this
temperature was minor, so it was
kept constant for all simulations.
The propylene weight percent in
the propane product was varied
between 1-4% and kept constant for
any given set of 200 cases corre-
sponding to a given propylene
purity scenario. The propane weight
percent in the propylene product
was computed for the given propyl-
ene purity, assuming a xed value
for other trace components. To
match the actual propylene weight
percent purity and the propane
weight percent in the bottoms, the
distillation column tray efciencies
were adjusted.
Results and analysis
The analysis assumed a hypotheti-
cal base case with an average 97%
propylene purity and 2.5% propyl-
ene in the propane product.
Incremental prot gures versus
the base case were calculated for
each of the propylene purity
scenarios considered. US Gulf
Coast prices were assumed for
chemical grade propylene, renery-
grade propylene and propane.
Table 3 summarises the incremental
prot and the calculated percentage
of off-spec results for the nine
scenarios considered. The incre-
mental prot gures shown
represent the average of the 200
cases of the corresponding propyl-
ene purity scenario. Note that cases
with a combination of low propyl-
ene in the propane product and
lower propylene purity resulted in
a high percentage of off-spec and a
corresponding negative incremental
prot. Based on these results, the
best two cases were 94% purity
with 2.5% propylene in the
Figure 6 Distillation column conguration
PP splitter optimisation
P
r
o
p
e
n
e

i
n

p
r
o
p
a
n
e
,

%
6
P
r
o
f
i
t
,

$
M
M
3
0
3
6
4
.0
0
3
.2
5
2
.5
0
1
.7
5
1
.0
0
9
4
.
0
0
9
4
.
7
5
9
5
.
5
0
9
6
.
2
5
9
7
.
0
0
P
u
r
i
t
y
,

%
Figure 7 Prot response surface
Purity, % Propylene in propane, % Incremental benet , $MM/yr Off-spec, %
94.0 2.5 4.6 4.0
94.4 1.4 (4.1) 37.5
3.6 4.4 0.0
95.5 1.0 (5.8) 37.0
95.5 2.5 2.9 0.0
95.5 4.0 2.2 0.0
96.6 1.4 1.1 0.5
3.6 0.3 0.0
97.0 2.5 0 0.0
Incremental benets and off-spec results
Table 3
valero.indd 5 11/12/2013 20:22
52 PTQ Q1 2014 www.eptq.com
optimum operating targets. The
path was designed with three incre-
mental steps so that the
performance of the process can be
monitored along the path and
corrective action taken if actual unit
performance does not follow the
model results. For instance, if the
nal and most aggressive step
results in unacceptable MA levels,
the propylene in the propane could
be increased to 4% to reduce the
MA concentration in the propylene
product and meet the chemical
grade propylene specications.
Figure 9 shows the joint density
contour maps generated for the
nine propylene purity target
scenarios considered. The joint
probability densities of propylene
purity and MA concentration were
represented using contour maps,
with reference lines added to high-
light the product specication
prot response surface shows that
the optimum falls in the top left
corner of the operating range high-
lighted in red, with propylene
purity ranging between 94% and
95% and propylene weight percent
in the propane product between
2.5% and 4%. Figure 8 illustrates a
suggested path to move the process
from initial operating targets to the
range of 94-97% propylene purity
and 1-4% propylene in the propane
product. The SAS Analytics
procedure PROC G3D
2
was then
used to generate the incremental
prot response surface using the
ner grid values generated with the
SAS Analytics procedure PROC
G3GRID.
2

Examination of the incremental
25
20
15
10
5
M
A
0
92 93 94 95 96 97 98
Purity, %
0.20
0.18
0.16
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00
25
20
15
10
5
M
A
0
92 93 94 95 96 97 98
Purity, %
0.55
0.50
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
25
20
15
10
5
M
A
0
92 93 94 95 96 97 98
Purity, %
0.55
0.50
0.65
0.60
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
0.095
0.000
25
20
15
10
5
M
A
0
92 93 94 95 96 97 98
Purity, %
0.24
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
0.22
25
20
15
10
5
M
A
0
92 93 94 95 96 97 98
Purity, %
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
25
20
15
10
5
M
A
0
92 93 94 95 96 97 98
Purity, %
25
20
15
10
5
M
A
0
92 93 94 95 96 97 98
Purity, %
0.095
0.000
25
20
15
10
5
M
A
0
92 93 94 95 96 97 98
Purity, %
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00
25
20
15
10
5
M
A
0
92 93 94 95 96 97 98
Purity, %
0.14
0.12
0.10
0.08
0.06
0.22
0.20
0.18
0.16
0.04
0.02
0.00
96.6% target and
1.4% propene
97.0% target and
2.5% propene
96.6% target and
3.6% propene
95.5% target and
1.0% propene
95.5% target and
2.5% propene
95.5% target and
4.0% propene
94.4% target and
1.4% propene
94.0% target and
2.5% propene
94.4% target and
3.6% propene
Figure 9 Joint probability density graphs
PP splitter optimisation
P
ro
p
e
n
e
in
p
ro
p
an
e
, %
5.0
P
r
o
f
i
t
,

$
M
M
2.5
0
2.5
5.0
3
.
2
5
2
.
5
0
1
.
7
5
1
.
0
0
9
4
.0
0
9
4
.7
5
9
5
.5
0
9
6
.2
5
9
7
.0
0
P
u
r
i
t
y
,

%
4
.
0
0
94.4, 3.6
97.0, 2.5
96.1, 2.9
95.3, 3.2
Figure 8 Suggested process optimisation path
valero.indd 6 11/12/2013 20:22
www.eptq.com
constraints. The contour maps were generated using
the SAS Analytics procedure PROC KDE
2
, which uses
Kernel density estimation. The 3 x 3 panel chart was
constructed using the SAS Analytics SAS procedures
PROC GSLIDE and PROC GREPLAY. Note that the
joint density distribution is very compact at the
higher purities as long as the propylene concentration
in the propane product is above 2.5%. Note also that
when the propylene in the propane product was
either 1% or 1.4%, the variability of the MA increased
signifcantly, as refected by the wide span of the
density contour map. This effect is more noticeable
when propylene purity was also reduced, as is
evident from the (95.5%, 1%) and (94.4%, 1.4%)
scenarios.
Conclusions
This article illustrates the application of a methodol-
ogy developed to optimise the propylene splitter unit
product quality targets in the presence of varying
process conditions. The methodology uses a process
simulator along with Monte Carlo random sampling
to optimise product quality outside the current oper-
ating range. Two product specifcation constraints
were considered in this study: 1) minimum propylene
purity and 2) maximum MA concentration. The meth-
odology can also be applied to situations with a
higher number of product specifcations. The example
presented here assumed that input process variables
followed independent normal distributions, but other
probability distributions such as a multivariate
normal with correlation structure or independent
gamma distributions can be considered for model
inputs. The same methodology can be used to opti-
mise other refnery and petrochemical units, where
fnal product specs need to be met in the presence of
varying process conditions. The results of the analysis
were used to determine optimum operating targets
for the unit APC system.
PROC SIMNORMAL and PROC G3GRID are marks of SAS Analytics Inc.
Petro-SIM is a mark of KBC Advanced Technologies. Excel is a mark of
Microsoft Corporation.
References
1 Gary J H, Handwerk G E, Kaiser M J, Petroleum Rening: Technology
and Economics, 5th ed, CRC Press, 2007.
2 SAS Customer Support, 2005, http://support.sas.com.
3 Castillo E, Process Optimization: A Statistical Approach, Springer,
2007.
4 Montgomery D C, Myers R H, Response Surface Methodology:
Process and Product Optimization Using Designed Experiments, Wiley,
1995.
5 Montgomery D C, Peck E A, Vining G G, Introduction to Linear
Regression Analysis, 5th ed, Wiley, 2012.
Jose Bird is Director Energy System and Metrics with Valero Energy
Corporation. He is responsible for implementing statistical solutions in
the areas of energy efciency, process optimisation, process monitoring
and ethanol manufacturing operations.
Darryl Seillier is Technology Advisor with Valero Energy Corporation.
He is responsible for company-wide process improvement in the areas
of energy efciency, hydrogen systems and process optimisation.
www.eptq.com
forming inside the transfer line. The conventional analy-
sis showed plug fow and slug fow in horizontal and
vertical pipes respectively for Case 3.
Figure 10b shows the volume fraction contours for
Case 4, which corresponds to the start-up case with
higher vapour fow. The fow regime in Case 4 is similar
to slug fow. Even though the gas volume fraction has
increased to 0.933 from 0.8 in Case 1, the slug fow
regime still prevailed in the transfer line. Conventional
analysis showed slug fow in both horizontal and verti-
cal pipes for Case 4.
To understand the effect of variation of vapour
volume fraction in the start-up case on pressure exerted
on the wall, pressure fuctuations in the transfer line
over time were studied (see Figure 11). It was found
that pressure fuctuations for Case 3 are low compared
to Case 4. This is due to liquid-dominated operation in
the transfer line. The pressure fuctuations for Case 4 are
very similar to Case 1. This is because the slug fow
regime prevailed in the transfer line for Case 4.
Conclusions and recommendations
CFD analysis of a two-phase vapour-liquid mixture fow
in the transfer line of a delayed coker unit was carried
out at different modes of operation. The fow during
start-up operation, wherein vibrations were observed,
was compared with normal operation where vibrations
were absent. While simulation showed large slugs during
the start-up operation, mist fow prevailed during normal
operation. Large slugs of liquid lead to large pressure
fuctuations, ultimately resulting in vibrations during
start-up. Further, CFD simulations were carried out to
determine the conditions which would avoid vibrations
in the transfer line. Lowering the steam fow rate was
found to be the only solution available for reducing trans-
fer line vibrations. Even at higher steam fow rates, a slug
fow regime prevailed. Good agreement was found
between fow regimes obtained from CFD and conven-
tional analysis. The recommendation was to reduce the
steam fow to avoid vibrations during start-up operation.
References
1 Green D W, Perry H B, Perrys Chemical Engineering Handbook, 8th Edition.
2 De Schepper S C K, Heynderickx G J, Marin G B, CFD modeling of all
gasliquid and vapourliquid ow regimes predicted by the Baker chart,
Chemical Engineering Journal, 138, 349357, 2008.
3 Fluent 6.3 Documentation, User Guide, Ansys Inc, USA.
Rahul C Patil is Senior Manager of the modelling and simulation group
of rening R&D at Reliance Industries Ltd, Jamnagar, Gujarat, India. He
holds a Masters in chemical engineering from IIT Kanpur and a degree in
chemical engineering from ICT Mumbai.
Dr Randheer Yadav specialises in the application of computational uid
dynamics to systems involving multi-phase ows. He holds a PhD in
chemical engineering from ICT Mumbai.
Dr Ajay Gupta is Assistant Vice President and leads the modelling and
simulation group of rening R&D at Reliance Industries Ltd, Jamnagar,
Gujarat, India. He has Bachelors, Masters and Doctoral degrees in chemical
Engineering from IIT, Delhi, India.
Dr Asit K Das is Vice President and heads the Rening R&D at Reliance
Industries Ltd, Jamnagar. He holds a degree in chemical engineering from
Jadhavpur University, West Bengal, a Masters in chemical engineering from
IIT, Kanpur, and a doctoral degree in chemical engineering from Ghent
University, Belgium.
who are not desk-bound and can beneft from access to
real-time data, trends and alerts anytime, anywhere.
Plant managers and production engineers can use
mobile BI anytime, anywhere in order to frst under-
stand an issue and then propose solutions to problems
in a time-frame not previously experienced with tradi-
tional desktop solutions. The prevalence of mobile
devices is transforming the process industries. Mobile
solutions empower decision-makers to have immediate
access to important data, enabling them to make
informed and quick decisions to improve proftability.
Easy, digestible analysis of plant information, even in
remote locations, helps industry to react to adverse
changes and keep the operation performing to targets.
The ability to access and analyse real-time plant data
has enormous benefts. In the past, users needed to be
in the control room or in front of a monitor to track
and manage manufacturing performance. Mobile BI
has proven to be more effective when users are
provided with visualisation tools (charts, graphs,
portals and so on).
For todays engineers, the message is simple: mobile
intelligence provides the platform to achieve greater
proftability. State-of-the-art mobile software, such as
Aspen InfoPlus.21 Mobile, enables faster decision-mak-
ing and troubleshooting and displays critical,
up-to-date information. The software functionality
improves employee effciencies by simplifying routine
engineering analysis tasks, such as examining and
comparing process data, reducing root cause analysis
time and easily fnding KPI data that will enable the
engineer to respond to changing process conditions.
Conclusion
Over the past 30 years, MES technology has evolved to
help refners survive in highly competitive markets.
Real-time data and decision support tools provide
access to plant information to allow quick and timely
responses to production issues that negatively infu-
ence effciency, quality and regulatory compliance.
MES is essentially the nucleus of the operation, which
links all capabilities of the business. It is an integrated
set of production activity and support software
designed to harmonise and optimise the plant.
The bottom line is that effective production drives
operational excellence, enabling better and faster deci-
sions. Software technology helps refners achieve
consistent performance across all assets. It also defnes
the importance of real-time business performance
management: plan, execute, monitor and respond to
change immediately on all time horizons. History has
shown that manufacturing execution systems have laid
the foundations to help refners across the globe
strengthen their competitiveness and build upon the
pillars of proftability.
Marty Moran is the MES Manager at AspenTech. During 30 years in the
process industries, he has worked as a consultant in over 65 reneries,
chemical, gas plants and other manufacturing environments. He
holds a US patent for multivariable control and a degree in chemical
engineering from the University of Illinois.
www.eptq.com
www.eptq.com
cause sediment formation in the heavy fuel oil. The
fouling tendency can be shown to increase exponen-
tially with conversion increase (see Figure 1). Therefore,
the value beneft of the conversion increase can be lost
to fouling and sediments. However, this fouling-to-con-
version relationship can also show that there is no major
advantage in decreasing severity whenever fouling rates
are acceptable or controllable. To do so would simply
result in lost conversion without a corresponding value
improvement in terms of sediment deposition control
and run lengths. This relationship brings about the
concept of optimising conversion as a function of the
rate of fouling and fuel oil sediment formation.

Fouling problems, their monitoring and control
Problem areas
The sections most prone to fouling are the atmospheric
column, vacuum column and preheat exchangers. At
very high conversion, the reactor and separator may
also suffer from high coke generation. Extensive foul-
ing of the separators and columns can lead to
unplanned shutdowns, downtime and lost production.
The same trends for conversion severity are valid for
sediment generation in the fuel oil. Below certain limits
of conversion the fuel will be stable, while above
certain limits the tendency to generate sediment with
time cannot be controlled.
From the above considerations, it is clear how setting
the proper operating conditions is important, as this
enables the best trade-off between conversion maximi-
sation and production of stable fuel and acceptable
rates of fouling. Optimal plant management requires
continuous control for resid product stability. The
stability is related to the tendency to produce fouling
deposits and generate sediment.
Optimal severity depends on the properties of the
feed being processed. The feed changes whenever the
refnery feed quality, residual feed make-up or the
plant feed rate changes.
1
Feed composition-related
factors that may infuence severity/conversion are the
stability reserve of asphaltenes in the vacuum resid
(often reported as p-value), the content of asphaltenes
and the intrinsic solubility of these asphaltenes. Low
stability reserve, high asphaltene content and poor
solubility will all contribute to an increased tendency
to generate coke and unstable residuum product and
fouling deposits.
The amount of metallic contaminants (especially
sodium and iron) is another factor that can impact
process performance by affecting catalyst performance
and, in some cases, by increasing coking tendency,
favouring dehydrogenation and conversion of coke
precursors those less soluble thermally cracked
asphaltenes into coke.
To avoid deposit generation, the LC Finer catalyst
plays a major role.
3
The type of catalyst utilised in the
process can have a great effect. In ebullated-bed reac-
tors, the catalyst is changed continuously to maintain
catalyst activity and to remove metals from residue oil.
The effect of the catalyst operations on the process and
fouling can be monitored easily with sophisticated
effciency trays must be used to reduce refux ratio
and lower the overhead cooling load. Flue gas
turbines, energy recovery hydraulic turbines, energy-
saving motors, frequency conversion motors and air
fow regulation systems of compressors should be
employed to recover pressure energy and reduce elec-
tric energy consumption. High-effciency intensifed
burners need to be used in furnaces to improve eff-
ciency. The exhaust temperature of furnaces should be
reduced to improve thermal effciency by 2-3%. New
insulation material needs to be employed to reduce
the heat loss of equipment and pipelines.
Low-temperature heat should be utilised by apply-
ing low-pressure steam generation technology,
low-temperature Organic Rankine Cycle (ORC)
systems and Kalina Cycle systems.
Energy optimisation of the area should be carried
out. Heat integration between the refnery and local
co-generation power plant needs to be realised. A
large quantity of low-temperature heat in a refnery
cannot be recovered because a heat sink is not availa-
ble, and this portion of low-temperature heat may be
used as a heat source for demineralised water and
boiler feed water in a cogeneration power plant. Also,
integration with the local chemical plant is important
to achieve material exchange and optimisation, as well
as optimum energy use. Integrated energy optimisa-
tion of the area should be carried out rather than
energy optimisation of a single refnery.
Conclusions
The priority of refning technology development and
the confguration of refning units in China have
distinct characteristics. During a long period in the
future, the FCC unit will still be the main secondary
conversion unit for gasoline and diesel production.
Therefore, improving the quality of FCC gasoline
and diesel is very important for oil product quality
upgrading technology in the future.
With the trend towards poor-quality crude oils,
future refneries should further optimise process fows
and change unit confguration so as to improve the
comprehensive utilisation rate of resources, and to
meet the requirements of energy savings and emis-
sions reduction.
More detailed classifcation of crude oil fractions
and the consideration of various narrow-cut process-
ing technologies will actively promote the low carbon
emission of refneries. Narrow-cut processing will
become the development trend in the overall process
fow optimisation of refneries in the 21st century.
Sun Lili is Vice President of Sinopec Engineering Incorporation in
charge of processing solution studies, engineering, construction and
startup of grassroots refneries and refnery modifcation projects. She
has over 20 years experience in engineering design and startup of
hydroprocessing units, has won several Prizes of National Scientifc
and Technological Progress, and holds a BS in petroleum refning from
China University of Petroleum.
Email: sunlili.sei@sinopec.com
www.eptq.com
www.dpharp.com
High Performance
Unmatched precision and real world
accuracy
Unconditional long term stability with lowest
Total Cost of Ownership (TCOj
Multi-sensing functionality
Safety
lnherently fail-safe sensor with no
undiscovered failure modes.
Certifed to lEC61508 as standard for single
use SlL2 and dual use SlL3
Robustness
Rugged, yet compact construction
Multiple material options for harsh industrial
environments
ln-built overpressure mechanism for
increased reliability
A New Standard
in Pressure
Measurement
EJA-E-270x86.indd 1 07/09/12 13:50
sinopec.indd 8 10/12/12 13:02:15
baker hughes.indd 2 08/03/2013 12:10
aspentech.indd 3 07/06/2013 19:12
ril.indd 6 13/09/2013 11:42 valero.indd 7 11/12/2013 12:22
Treating Solutions
Proven
Results
Merichem Company provides a diversifed portfolio of products and services to a wide range of
operations, including the refning and petrochemicals industries. With more than 1,000 licensed
process units worldwide, Merichem is the leader in licensing patented process technologies and
supplying proprietary equipment to provide hydrocarbon treating solutions and by-product
management services that remove sulfur and other impurities from hydrocarbon liquids and gases.
SCAN THE QR CODE WITH YOUR
MOBILE DEVICE TO SEE OUR
COMPLETE PORTFOLIO OF
TREATING SOLUTIONS.
C
M
Y
CM
MY
CY
CMY
K
Merichem Company.pdf 1 5/22/2013 8:09:34 AM
merichem.indd 1 06/06/2013 10:23
Extending the treatment of highly sour
gases: cryogenic distillation
C
ompanies willing to produce
large gas felds with very
high amounts of CO
2
have to
face a constraint related to the
essential need to reduce the atmos-
pheric emissions of greenhouse
gases.
The economics may also be
improved by the growing accept-
ance of the re-injection of separated
H
2
S and CO
2
, for reutilisation to
enhance oil recovery. Separated
acid gas re-injection into a depleted
reservoir or an aquifer, as a feasible
alternative to costly sulphur recov-
ery to a diminishing sulphur
market, or to limit atmospheric
emissions of greenhouse gases,
increases the number of highly sour
gas felds that can be reconsidered
as exploitable to produce much
needed natural gas.
1,2
These new constraints lead to the
development of more energy eff-
cient technologies for acid gas
separation, adapted to these new
production schemes. With this
objective, in addition to the
AdvAmine series of gas sweetening
processes using amine based
solvents and discussed in a preced-
ing article (PTQ, Q4 2013), Total,
IFP Energies nouvelles and
Prosernat have developed the Sprex
and SprexCO2 processes for the
production of highly sour gas
reserves with acid gas re-injection.
This second article discusses the
benefts of the Sprex and SprexCO2
processes.
Cryogenic distillation for the
sweetening of super sour gases
Some gas felds contain very high
amounts of H
2
S (more than 30
vol%) or CO
2
(up to 70 vol%) in
Cryogenic bulk removal of H
2
S or CO
2
offers an economic advantage when the
separated acid gases from super sour gas are re-injected
FRANOIS LALLEMAND, GAUTHIER PERDU and LAURENT NORMAND Prosernat
CLAIRE WEISS Total
JULIA MAGNE-DRISCH and SEBASTIEN GONNARD IFP Energies nouvelles
natural or associated gases. Even
though amine processes can be
optimised to treat very sour gases,
the high cost associated with sweet-
ening may make the production of
these super sour gases uneconomic
under certain gas price conditions.
Permeation membranes have
been used commercially since the
1980s for the bulk removal of CO
2
from gases with a very high CO
2
content down to levered CO
2
content generally between 5% and
10%. With the current status of this
technology, permeation membranes
can only be used to treat gases with
very limited H
2
S content, as they
are very sensitive to this chemical.
Membrane units are relatively
simple and use very little energy,
however permeation membranes
are not selective and one must
accept large methane losses with
the separated acid gases, even with
dual-stage membrane units with
inter-stage recompression. Provided
that adequate preconditioning of
the gas is performed, membrane
life is considered as acceptable to
limit the cost of replacement.
However, the preconditioning unit
is generally fairly large, which
substantially increases the cost of
the sweetening. The process further
needs post-treatment with an amine
solvent to reach tight CO
2
specifcations.
Cryogenic distillation processes
offer many advantages when the
separated acid gases need to be
re-injected, to limit undesired
sulphur production or to minimise
greenhouse gases emissions to the
atmosphere. They are very selective
towards light hydrocarbons, and
the separated acid gases (H
2
S and/
or CO
2
) are recovered in the liquid
state under pressure. Producing the
acid gases as a high pressure liquid
saves expensive and energy
consuming compression require-
ments, because the pumping duty
is much lower.
It is possible to produce
pipeline quality gas with cryogenic
distillation. Such technologies
require dehydration of the sour gas
prior to entering cryogenic separa-
tion. Separating CO
2
or H
2
S down
to a commercial specifcation
requires a very low temperature in
the refux drum, corresponding to a
very signifcant refrigeration
requirement. Furthermore, CO
2
separation down to commercial
levels suffers from limitation due to
CO
2
freezing conditions in the top
section of the cryogenic distillation
column. Some processes address
this limitation by adding, for exam-
ple, a suitable hydrocarbon solvent
to the top section of the demetha-
niser to stay outside the CO
2
freezing conditions, or by using a
column with a specifc frozen CO
2
remelting zone.
The Sprex and SprexCO2
processes, jointly developed by
Total and IFP Energies nouvelles/
www.eptq.com PTQ Q1 2014 55
Separating CO
2
or H
2
S
down to a commercial
specication
requires a very low
temperature in the
reux drum
prosernat.indd 1 11/12/2013 20:20
56 PTQ Q1 2014 www.eptq.com
gas by producing a partially sweet-
ened gas, which can then be
processed by a smaller, conven-
tional amine sweetening unit,
capable of meeting the most severe
H
2
S and CO
2
gas specifcations
(pipeline gas or feed for a LNG
plant)
It produces a high pressure liquid
H
2
S soup (50 to 65 bar) that can be
easily re-injected into a geological
reservoir.
In this basic version of the
process, which has been demon-
strated in an industrial context at
the Lacq plant,
4
in the southwest of
France (see Figure 3), the H
2
S
content in the gas leaving the Sprex
unit is about 10-12%. The refux,
consisting of cold, dry H
2
S, dehy-
drates to some extent the incoming
gas in the upper zone of the Sprex
column. The minimum temperature
reached in the units low tempera-
ture refux drum is limited to about
-30C, so as to remain outside the
hydrate zone at all points of the
unit. This rules out the necessity of
installing a dehydration unit
upstream of the Sprex column.
The H
2
S separated out in the
process is produced in a high
pressure liquid phase, requiring
considerable less energy for re-
injection into an underground
reservoir, as the large, multistage
acid gas compressors can be
replaced by injection pumps.
Prosernat, are bulk fractionation
processes. The Sprex process for
bulk H
2
S removal does not require
upstream dehydration of the gas.
The refrigeration requirements of
Sprex and SprexCO2 are limited
compared to those of other cryo-
genic processes and, as the
temperature in the SprexCO2 is not
as low, the operating conditions are
far away from the CO
2
freezing
region, therefore avoiding the need
for an additive fractionation and
recycle or for a remelting zone.
When pipeline or LNG specifca-
tion is required, the Sprex or
SprexCO2 unit is easily and
economically combined with a
solvent based acid gas removal
unit.
Bulk H
2
S removal with the
Sprex process
The Sprex process was jointly
developed to improve the
economics of the production of
ultra sour gas with high H
2
S
content when the separated H
2
S is
re-injected. The process was
improved and several other patents
were fled in the 2000s.
3
Figures 1
and 2 show the principles and
process fow diagram of the staged
acid gas separation from a very
sour natural or associated gas using
Sprex followed by a conventional
solvent sweetening plant.
The process has the following
advantages:
2

It reduces the H
2
S content in the
Very sour
natural gas
Acid gas liquid
high pressure
Sales gas
Remaining
acid gas
Pre-ex
AGRU
Acid gas
removal unit
Figure 1 Staged acid gas separation with
the Sprex process
TEG
energised
MDEA
Sour gas
30-35C
50-65 bar
H
2
S soup
60-70C
Refrigerant
30C
Pipeline
gas
Fuel
gas
Acid
gas
Sprex
Re-injection well

Figure 2 Sprex simplied process ow diagram
Figure 3 The Sprex demonstration plant at Lacq in France
prosernat.indd 2 13/12/2013 12:02
Now that the amine sweetening
unit located downstream has less
H
2
S to separate, the amine solvent
circulation rate can be substantially
decreased, as can the size of the
equipment. This drastically reduces
both the investment cost and
energy consumption.
To illustrate the benets of using
the technology, the two schemes
shown in Figure 4 have been
compared: a reference case using
an amine unit, a base case using
Sprex for bulk removal and an
amine unit as a nishing unit.
The comparison has been made
for a sour gas containing 33 mol%
H
2
S and 10 mol% CO
2
at 70 bar.
The sales gas specication has been
considered at 4 ppm mol of H
2
S
and less than 2 mol% CO
2
. The acid
gas containing H
2
S and CO
2
is
re-injected at 200 bar.
Comparison of the utilities
consumption of the two schemes
shows that the compression power
is much lower for the base case
than for the reference case (see
Figure 5).
As far as steam consumption is
concerned, the decrease linked to
the use of the technology is more
than 30% compared to the full
amine scheme (see Figure 6). This
leads to a saving in energy
consumption of 23% for the Sprex +
amine scheme compared to the
amine only scheme.
Moreover, the comparison of the
two schemes has shown that over-
all capex is decreased using Sprex.
Depending on the overall ow rate
www.eptq.com PTQ Q1 2014 57
to be treated, Sprex can also lead to
a reduction in the number of trains,
further reducing the capex and the
overall plant footprint, and thus
lowering the toxic risk.
When considering a possible H
2
S
increase with time, due to H
2
S
re-injection in the producing reser-
voir, the Sprex design is less
sensitive than the amine unit
design. This reduces the cost of the
necessary oversizing of treatment
capacity to handle such an increase
in H
2
S concentration.
When the sour gas contains
mercaptans, the Sprex condenses
those sulphur species and enables
the operator to meet their sales gas
specication, which greatly simpli-
es downstream treatment.
Finally, in the case of distant
nishing treatment (a downstream
amine unit located in a different
location), the pretreated gas stream
from Sprex (bulk removal) is dry
and monophasic, which allows easy
design and operation of a transfer
line.
Brief review of development
and validation
From the rst patent led in 1994
to formal qualication of the
process for the sweetening of super
sour gases with acid gas re-injec-
tion, a long process had to be
cleared to meet the challenges
imposed by the innovative concept
of Sprex. The main steps of this
work have been:
Develop and validate a liquid/
liquid/vapour thermodynamic
model within the range of composi-
tions and operating conditions of
the process: the Gasweet model was
calibrated from literature and
Amine unit
(MDEA)
Sprex
(H
2
S cryogenic
separation)
Acid gas
re-injection
Amine unit
(MDEA)
Acid gas
re-injection
Reference case: full amine
Base case: Sprex + amine
Figure 4 Sprex economic evaluation:
considered cases
60
100
90
80
70
50
40
30
20
10
C
o
m
p
r
e
s
s
i
o
n

p
o
w
e
r
,

%
0
Sprex +
amine
Amine
only
60%
100%
Figure 5 Sprex economic evaluation: acid
gas compression power
60
100
90
80
70
50
40
30
20
10
R
e
l
a
t
i
v
e

s
t
e
a
m

c
o
n
s
u
m
p
t
i
o
n
,

%
0
Sprex +
amine
Amine
only
69%
100%
Figure 6 Sprex economic evaluation: steam
consumption
60
100
90
120
110
80
70
50
40
30
20
10
P
r
e
s
s
u
r
e
,

b
a
r
0
0 25 50 75 100 125 150 175 200 225 250
Temperature, C
Liquid-liquid
Liquid-vapour
Liquid-liquid-vapour
bubble curve
Liquid-vapour
dew curve
Liquid-liquid-vapour
Liquid-liquid-vapour
dew curve
Vapour
Figure 7 Phase diagram of H
2
O/H
2
S/ CH
4
/CO
2
system. Comparison between experimental
data and the Gasweet model
prosernat.indd 3 11/12/2013 12:33
58 PTQ Q1 2014 www.eptq.com
Figure 7 shows that the devel-
oped thermodynamic model
accurately predicts the vapour
liquid equilibrium and the
liquid-liquid-vapour equilibrium of
such complex systems
Select and validate a liquid/
vapour/solid model able to predict
hydrate formation conditions with
very low water contents. As very
limited data were found in the
literature, hydrate dissociation
temperatures were measured at
high pressure for hydrocarbon/
acid gases mixtures containing up
to 80% H
2
S and with low water
contents. The model that was even-
tually selected as giving the best
results was an in-house model (see
Figure 8)
Develop and optimise the Sprex
process scheme. The original
scheme, patented in 1994, has
evolved and matured to the current
scheme, including a distillation
column where the sour gas is dehy-
drated with cold liquid H
2
S in the
upper zone, and light hydrocarbons
and CO
2
are stripped from the H
2
S
soup in the lower section
Run the necessary testing to
determine the corrosivity of super
sour gas mixtures towards carbon
and stainless steel under the oper-
ating conditions of the Sprex unit,
and acid gas injection pipelines and
wells, to allow the safe selection of
materials of construction
Develop adequate risk manage-
ment studies and mitigation
measures to manage the risk
involved in handling dense phase
H
2
S rich mixtures in large process-
ing units
Select and qualify specifc pieces
of equipment, such as liquid H
2
S
pumps.
The decision was made to build
and operate an industrial demon-
stration plant, to validate the Sprex
concept, unit design parameters,
metallurgical recommendations,
process control system, normal and
emergency operating procedures,
and to verify the robustness of the
process.
The pilot unit was installed and
integrated in a large commercial
sour gas plant located in southwest
France in a populated area (see
Figure 3), with all of its constraints,
tans, COS, CS
2
and methanol, in a
wide range of temperature/pres-
sure conditions, from -85C to
150C and up to 150 bar.
specifcally acquired isotherms, for
fuid compositions including light
and heavy hydrocarbons, H
2
S, CO
2
,
nitrogen, water, aromatics, mercap-
30
35
25
20
15
10
5
I
n
-
h
o
u
s
e

m
o
d
e
l
,

C
0
0 5 10 15 20 25 30 35 40
Lab data, C
Hydrate prediction temperature
Lab data in-house model
CH
4
/CO
2
/H
2
S system
Figure 8 Hydrate prediction temperature with very low water content: in-house model
8
0
2
4
6
10
12
14
16
18
T
r
a
y

n
u
m
b
e
r
20
0 10 20 30 40 50 60 70 80
H
2
S content, %
Cas 30% yH
2
S exp
Cas 30% yH
2
S cal
Figure 9 H
2
S concentration prole: pilot results vs Gasweet model
30
50
45
40
35
25
20
15
10
5
T
e
m
p
e
r
a
t
u
r
e
,

C
0
0 250 500 750 1000 1250 1500 1750 2000 2250 2500
Time, seconds
Start of hydrate
formation
Temperatures
inversion
Hydrate
melting
MeOH
injection
TR11
TR13
TR9
TR15
Figure 10 Forming and melting hydrate in a Sprex pilot plant: temperatures on selected trays
prosernat.indd 4 11/12/2013 12:34
www.eptq.com PTQ Q1 2014 59
specifcation achievable in the
distillation outlet around 20%
CO
2
mol at 30/40 bar.
The SprexCO2 process is compet-
itive in comparison with
conventional CO
2
separation
processes (amines) or other hybrid
solutions (membranes + amines)
when CO
2
concentrations are
higher than 40%, and CO
2
re-
injection is required for enhanced
oil recovery and/or carbon seques-
tration. Carbon dioxide is recovered
gases containing a very high CO
2
content (40% or so), possibly
together with high concentrations
of H
2
S.
The SprexCO2 process is based
on an optimised CO
2
distillation,
which is performed at conditions
outside the CO
2
crystallisation
domain. The parameters and limits
to avoid crystallisation are illus-
trated in the distillation chart based
on the main binary mixture (CO
2
/
CH
4
). Figure 11 shows a typical gas
such as variations in raw gas fow,
conditions and quality, production
upsets, and so on.
3
The unit
processed sour gas of variable H
2
S
content (from 18% to 40%), oper-
ated by conventional plant
personnel, from the beginning of
2005 to 2006.
Operation has demonstrated
excellent agreement between the
thermodynamic models (L/V equi-
libriums [see Figure 9] and
hydrates formation conditions);
measured the actual effciency of
the trays used in the column; vali-
dated the process and safety
engineering studies; and validated
the adequacy of the choice of mate-
rials of construction for large scale
units using corrosion coupons
installed in selected places of the
pilot unit.
Operating procedures have been
optimised and validated.
4
The
stability and robustness of the
process was demonstrated through
long, continuous operation runs.
The behaviour of the plant was also
tested under startup and shutdown
conditions, including emergency
shutdown procedures. Hydrate
formation and melting in the unit
were tested, and the column has
proven to easily recover stable
operation by quickly melting the
hydrates through adequate adjust-
ments in the operating conditions
and methanol injection (see Figure
10).
The process was eventually
formally qualifed by Total for
implementation in the development
of large super sour gas felds, using
the companys formal internal qual-
ifcation procedure for innovative
processes.
Cryogenic bulk removal of CO
2

With a dehydration unit upstream
of the cryogenic fractionation
column, the gases can be cooled to
well below the -30C of the basic
version of the Sprex process, which
improves the effciency of the H
2
S
separation process and allows the
separation of CO
2
from the gas.
However, the limit then becomes
the temperature at which CO
2
crys-
tallises. The enhanced version of
the Sprex process, called SprexCO2,
allows bulk removal of CO
2
from
40
0
10
20
10
20
30
50
60
70
80
90
T
e
m
p
e
r
a
t
u
r
e
,

C
100
0 10 20 30 40 50 60 70 80 90 100
CO
2
, mol%
Gas
Liquid
30 bara
CO
2
crystallisation limit
50 bara
40 bara
Figure 11 CO
2
/ C
1
distillation chart
Pre-treatment
From
pre-treatment
FC
Refrigeration
loop
Cold box
Sprex
column
To
re-injection
To amine
unit
CO
2
soup
Refrigerant
Precooling
Thermal
integration
Figure 12 SprexCO2 simplied process scheme
prosernat.indd 5 11/12/2013 12:34
60 PTQ Q1 2014 www.eptq.com
fexibility to a certain extent by
means of bulk and fnishing separa-
tion units in series, in order to cope
with fuctuations in feed gas
composition or temporary offset
conditions.
The level of complexity of the
proposed CO
2
separation scheme
remains moderate: no prototype
equipment is required, minimum
acid gas compression system is
required (only for the acid gas sepa-
rated in the amine unit) and spare
capacity considered is adequate to
achieve availability levels typical of
acid gas removal units.
In terms of capital investment,
the estimation made in the case
study shows an overall reduction in
the range 17% to 20% by using
SprexCO2 bulk separation instead
of membranes.
Conclusions
Technologies to treat sour natural
or associated gas are numerous, but
none has shown to be as versatile
and economic as those using amine
based solvents.
In the preceding article (PTQ, Q4
2013), it was demonstrated that
amine processes can be optimised
to effciently and economically treat
any type of highly sour gas.
To be truly competitive, the
removal of acid gas components
H
2
S and CO
2
from highly or super
sour gases requires the optimum
choice of process. Within the
AdvAmine series of processes, the
energizedMDEA process can be
amine combination has been evalu-
ated as an alternative, using
commercial membrane vendors
design data.
Removing CO
2
from 70% to 20%
(our reference case) represents
more than 90% removal of the CO
2
from the feed gas. Subsequent
amine units, using the AdvAmine
technology, remove all the remain-
ing CO
2
down to the gas
specifcation required.
The better performances of the
SprexCO2 option, compared to
those of the membrane alternative,
are shown in Figures 13 and 14.
SprexCO2 maximises the gas
valorisation by halving the overall
auto-consumption needs and
signifcantly reducing overall
hydrocarbon losses with the sepa-
rated CO
2
.
As Figure 14 shows, there is
roughly an order of magnitude
difference in power required for
CO
2
recompression. From an ener-
getic point of view, one of the
interests of SprexCO2 is, indeed,
that the CO
2
is recovered in a liquid
phase, so that a compression step is
not necessary.
The case study also demonstrates
that the two processes, SprexCO2
and AdvAmine, are energetically
paired, with benefts in terms of
waste heat recovery optimisation.
One of the advantages of the
proposed CO
2
separation scheme is
the possibility to increase process
in a liquid soup phase, so injection
is realised by pumps instead of
multistage compressors, with
considerable energy savings.
In order to match fnal gas export
specifcations (pipeline or LNG
plant feedstock), the SprexCO2 unit
can be associated with a fnishing
AdvAmine unit.
A simplifed SprexCO2 scheme,
including upstream pre-condition-
ing units and downstream CO
2
boosting and transfer/injection
pumps, is shown in Figure 12.
The scheme in Figure 12, for the
purposes of illustration, is typical:
some alternative process schemes
are achievable, depending on the
raw gas conditions and fnal
specifcations required for export
gas (pipeline, LNG feed or other
uses).
An example of the use of the
SprexCO2 technology is presented
below. The case study compares a
SprexCO2 + amine combination
and a membranes + amine combi-
nation. The range of methane and
CO
2
in the raw gas composition is
shown in Table 1.
The gas contains a few per cent
of C
2
+ and inerts as well.
The base case for natural gas
sweetening includes the SprexCO2
technology for bulk removal of CO
2

and a fnishing gas treatment based
on AdvAmine technology (ener-
gizedMDEA). The membranes +
60
100
90
80
70
50
40
30
20
10
M
a
t
e
r
i
a
l

b
a
l
a
n
c
e
,

v
o
l
%
0
Sprex CO
2
Membranes
HC losses
Valorised gas
Auto-consumption
60
100
90
80
70
50
40
30
20
10
P
o
w
e
r

c
o
n
s
u
m
p
t
i
o
n
,

%
0
Sprex CO
2
Membranes
10%
85%
Figure 13 Overall performances
comparison: SprexCO2 + amine vs
membranes + amine
Figure 14 Power consumption dedicated
to CO
2
recompression: SprexCO2 + amine
vs membranes + amine
Mol%
Methane 24-27%
CO
2
68-71%
Fluid composition
Table 1
The acid gas loading
of amine solutions
needs to be
maximised to reduce
the required amine
solution ow rate
prosernat.indd 6 11/12/2013 12:34
www.eptq.com PTQ Q1 2014 61
sour gas processing: H
2
S removal with the Sprex
process, SPE paper IPTC 10581-PP, International
Petroleum Technology Conference held in
Doha, Qatar, 21-23 Nov 2005.
4 Huygue R, Lemaire E, Lecomte F, Deixonne
M, Lajeunesse H, Operational validation of the
Sprex process for bulk H
2
S removal, Sulphur
Conference, Vienna, Austria, Oct 2006.
Francois Lallemand is Senior Advisor with
Prosernat.
Gauthier Perdu is Deputy Process Director
with Prosernat and Head of the Amine Process
Technology Group within Prosernats Process
department. He holds a Dipl. Engineer and Post
Graduate certicate in chemical engineering
from ENSIGC Toulouse.
Laurent Normand is Business Development
Manager with Prosernat. He holds a PhD in
material science and graduated as an engineer
in physics from the cole Nationale Suprieure
des Mines de Paris.
Claire Weiss works for Total EP in the process
department within the Acid Gas group.
Julia Magne-Drisch is Gas Treatment Project
Manager at IFP Energies nouvelles. She holds a
PhD in hydrotreatment.
Sebastien Gonnard is Process Engineer
at IFP Energies nouvelles, in charge of
process development and techno-economic
evaluations.
separated acid gases from super
sour gas are re-injected. The Sprex
and SprexCO2 processes, jointly
developed by Total, IFPEn and
Prosernat, can be combined with an
amine unit to reduce the cost of the
production of gases containing
more than 25-30% H
2
S or over
40-45% CO
2
. They provide an
economic and easy to operate alter-
native to amines, to membranes
and to other more complex cryo-
genic fractionation processes for the
treatment of such highly sour
gases. The Sprex series of processes
was demonstrated in a unit
installed at the Lacq sour gas plant,
and the technology is now ready to
be implemented at industrial scale.
References
1 Lallemand F, Aimard N, Rocher A, Sour gas
production: moving from conventional to
advanced environmentally friendly schemes,
Sogat Conference, Abu Dhabi, 29 Apr3 May
2007.
2 Lallemand,F, Minkkinen A, Processes combine
to assist acid gas handling, reinjection, Oil and
Gas Journal, 21 Jan 2002.
3 Lallemand F, Lecomte F, Streicher C, Highly
optimised to reduce energy require-
ments, taking advantage of the
ability of MDEA to release a
substantial part of the absorbed acid
gases in a low pressure fash to
produce the semi-lean stream in a
split-fow process confguration.
The acid gas loading of amine solu-
tions needs to be maximised to
reduce the required amine solution
fow rate. Operating amine plants at
high loadings has been proven by
Total through extensive operating
experience, and a liquid side cooler
scheme on the acid gas absorber
was developed to further increase
this loading and reduce the cost of
highly sour gas sweetening.
When the mercaptans content of
highly sour gas becomes an issue to
meet a stringent total sulphur
commercial gas specifcation, the
HySweet process from Total may
be considered.
Even though amine based sweet-
ening has proven to be a very
effcient technology, cryogenic bulk
removal of H
2
S or CO
2
offers an
economic advantage when the
www.eptq.com PTQ Q4 2013 109
personnel. This starts with the initial
design of the power system by the
electrical engineer, where all factors
such as steady-state load require-
ments, ambient conditions, and
quantied fault conditions are thor-
oughly considered. Management
and engineers must also work with
the vendors to be certain that all the
that are further discussed in the
report.
Six Sigma
Strategies for process improvement
such as Six Sigma rely on identify-
ing and removing defects within a
process and carrying out proper
operational practices by qualied
and, of course, bad media coverage
and negative publicity. Sudden,
unplanned shutdowns are never
good for the reliability of a ren-
erys operating units, especially the
larger ones. Motors and rotary
equipment are damaged when
suddenly taken from full speed to a
dead stop, and consume a great
deal of energy while restarting. But
if steps are taken to minimise
unplanned failures, all of these
consequences can be reduced.
Finding the most reliable electrical
equipment
Reneries need to work with elec-
trical technology vendors to design
and construct the most reliable grid
for each unit. This cooperation can
help reneries improve their energy
security and efciency. Table 1
shows a list of selected major
companies from those that were
surveyed by Hydrocarbon
Publishing Company in preparation
of its report. The survey covers 120
companies from around the world
that provide various technologies
Energy
efficiency
Reliability
Financial
performance
Plant safety
and liability
Negative
publicity
Environmental
concern and
liability
Asset
management
Power failure
impacts
Figure 2 Impacts of renery power failures
To keep pace with the demanding
quality requirements of modern
fuels, advanced, precise and easy to
use analytical technology is required.
With a complete range of XRF and
ICP spectrometers, SPECTROs
unique solutions for at-line and
laboratory elemental analysis
are capable of meeting the most
demanding product specification
testing requirements.
Elemental Analysis
of Fuels and Oils
Discover more exciting details at
www.spectro.com/fuels
Determination of Sulfur and other
elements at-line and in the laboratory
hpc.indd 5 13/09/2013 11:51 prosernat.indd 7 13/12/2013 12:03
www.energyexchange.co.uk/europeanfuels
marketing@theenergyexchange.co.uk
+44 (0)207 384 8012
Organised by: Supporting Partner: Media Partners:
Europes number 1 strategic
event for the fuels industry
Book before the 20
th
of
December and save 400
Use VIP code DF10
18
th
March 2014 // Rome, Italy
An unparalleled line up of
reners and industry experts:
Giacomo Rispoli
Executive Vice President
RSP/R&M - Ricerca, Sviluppo
Tecnologico e Progetti -
Rening & Marketing Division
ENI
Philippe Marchand,
Directeur of Biotechnology -
New Energies Division,
Total
Hugh Tucker,
Technical Director,
UKPIA
Alessandro Bartelloni,
Policy Director,
Europia
00000 EFC A4 Advert_Layout 1 13/12/2013 11:55 Page 1
euro fuels.indd 1 13/12/2013 12:41
Optimising hydrogen sulphide scavenging
S
cavenging hydrogen sulphide
(H
2
S) from petroleum products
may be easy, but doing it eff-
ciently is much more challenging.
Too often some important aspects
are overlooked, causing extra costs
in the best case and off-specs in the
worst ones. Through the example of
a case history we want here to show
the impact that a proper approach
in treatment management can have
on dosage rates and therefore costs.
H
2
S risk
Risk is known to be the product of
the probability of an event and its
consequences, and H
2
S is a poison-
ous gas, lethal at certain
concentrations, that can be present
in petroleum products. This is why
it represents a high risk for the
petroleum industry. But there is
more to this issue.
H
2
Ss specifc features make it
worthy of special attention. Being a
colourless gas, its presence can be
identifed only by smell which can
be deceptive. H
2
S can be detected by
a characteristic odour of rotten eggs
when present in very low concentra-
tions (1.4 g/m
3
or 0.0047 mg/kg),
but the senses are soon saturated
and deactivated by an increase in
concentration, meaning that when
the concentration exceeds 100 mg/
kg humans are no longer able to
smell it. This is perhaps one of the
most dangerous aspects of this gas;
its presence ceases to be perceptible
by the human senses as its concen-
tration becomes more dangerous.
And exposure at a concentration
equal to a few hundred mg/kg may
already be lethal.
H
2
S is present, in extremely varia-
ble amounts, in crude oils.
Selecting the right hydrogen sulphide scavenger is important, but appropriate
treatment setup and tailored monitoring procedures are key to avoiding off-specs
GIACINTO GRANDE and ALESSANDRA BERRA
Chimec
Advanced refning processes gener-
ally allow for high, if not complete,
removal of H
2
S from light and
middle distillates, with rare but
possible exceptions.
The situation is different when it
comes to heavy fuel oils. Generally
speaking, H
2
S content in residues is
the result of the sum of what
remains trapped during the distilla-
tion processes, what is formed as a
result of cracking phenomena, and
what is removed by desulphurisa-
tion and/or stripping processes.
The nature of the sulphur
compounds present in the crude,
and the severity of cracking applied,
infuence the amount of H
2
S that
can be produced as a result of these
phenomena. Conditions suffcient
for the formation of H
2
S, for exam-
ple, can be reached by the bottom of
the vacuum distillation columns.
The nature of residue refning
processes is more oriented towards
the formation of H
2
S (cracking) than
its removal (desulphurisation) to the
point that its content in the liquid
phase can reach values even higher
than 50 mg/kg even though it is
mostly in the range 5-20 mg/kg.
The values mentioned above may
seem lower than the safety thresh-
olds for human health but there is
an important difference: the phase
to which these concentrations refer.
When we speak about petroleum
products, H
2
S content is reported as
concentration in the liquid phase.
From this liquid, H
2
S is then
released into the vapours above the
oil, where it becomes dangerous to
humans. The amount of H
2
S devel-
oped in the vapour phase depends
on several factors, including the
characteristics of the oil, conditions
of temperature and pressure, and
obviously the H
2
S concentration in
the liquid.
For residues and heavy fuel oils,
the levels of H
2
S in the vapours are
normally 50-100 times those in the
liquid. It is therefore clear why even
a few parts per million of H
2
S in the
liquid phase can be so dangerous.
H
2
S specication
On 1 July 2012, a H
2
S specifcation
was implemented in a revision of
ISO 8217. A limit of 2 mg/kg in the
liquid phase was established for all
marine diesel and residual grades.
This specifcation set some new
operating requirements and new
challenges for all companies
producing marine fuels.
Refneries have means of
controlling H
2
S content in the resi-
dues and fuel oils they produce,
although it is important that this
does not have negative effects on
other aspects of production. But
failing to optimise operating condi-
tions can lead to a H
2
S content as
much as an order of magnitude
higher than the minimum required.
Achievement of the H
2
S specifca-
tion in ISO 8217 therefore starts
with the proper management of
fuel oil production processes. When
this is not possible, or the H
2
S level
is not suffciently low, the specifca-
tion has to be met through the
implementation of other solutions.
The main and most common
among these is the use of specifc
additives able to remove H
2
S by a
chemical reaction, so-called H
2
S
scavengers. The alternative is the
installation of stripping columns
for the residue, but this solution is
seldom applied due to high instal-
www.eptq.com PTQ Q1 2014 63
chimec.indd 1 11/12/2013 12:39
64 PTQ Q1 2014 www.eptq.com
care about time. It is taken for
granted that the sample must be
representative of the fuel produced
and that the container must be well
sealed to avoid any H
2
S loss. Care
must be given to the contact time
allowed for the scavenger to
execute its action, especially when
the contact time available in the
feld is limited.
In the case study described here,
the treated samples were taken at a
point in the fuel rundown some 30
minutes from the scavenger injec-
tion point, but the time between
sampling and testing varied, differ-
ing by up to a few hours (see
Figure 1). That was causing some-
times misleading results because,
oddly enough, when the refnery
was more effcient and the fuel was
tested a very short time after
sampling, high H
2
S content could
be found. This was wrongly
ascribed to poor effciency or insuf-
fcient dosage rates.
The consequences of this errone-
ous evaluation were magnifed by
the corrective measures applied in
case of off-specs: a doubling of the
scavenger dosage rate until the H
2
S
content returned below the 2 mg/
kg level.
As a consequence of the particu-
lar operating conditions, the needs
of the refnery, the applied proce-
dures and the absence of any
technical assistance from the H
2
S
scavenger supplier, the dosage
rates applied were high and outside
the standard.
When Chimec was asked to set
up an alternative treatment, we
performed, frst of all, a laboratory
study to select the most suitable
product and the theoretical dosage
rate necessary to fulfll the specif-
cation. The tests were performed by
simulating plant operating condi-
tions and carrying out a frst
standardisation of the sampling
and analytical procedures.
The results of the study showed
that it was possible to apply, with
the additive selected, a dosage rate
approximately 50% lower than the
one previously applied. Our choice
of scavenger was also triazine based.
The different dosage ratio could also
be the result of a higher concentra-
tion in comparison to the previous
lation and operating costs. In some
cases, there may be a competitive
solution through the use of
additives.
Even though some refneries were
already dealing with H
2
S scaveng-
ing before 2012, the number of
applications sharply increased after
the implementation of the new
specifcation, but experience in the
proper management of similar
processes did not spread with the
same speed.
Through the presentation of a
case study, this article will make a
general discussion of the relevant
aspects of the management of H
2
S
scavenger treatment and the impact
they can have on the overall results.
A deeper treatment of the issue
would require the use of confden-
tial information that cannot be
provided in these circumstances,
but Chimec is available to provide
further clarifcation of our experi-
ences and technologies.
Case study
Refnery scheme, operating condi-
tions and H
2
S content are among the
fundamental variables to be taken
into account for the proper setup of
an H
2
S scavenger application.
Monitoring, on the other hand, is a
key aspect of treatment manage-
ment and, to achieve real cost
optimisation, specifc programmes
should be set up based on the char-
acteristics of the site.
The best way to show this is
through real experience. It is not
the specifc case that this article
wants to focus on, as it may differ
from the situation in other refner-
ies, but the basic principles applied.
These principles have wider
applicability.
The refnery described in this case
study produces marine residual
fuels through the blending of a
vacuum residue and a mix of fux-
ants. The main source of H
2
S is the
vacuum residue while its content in
the lighter components is negligible.
When we approached the refnery
for the frst time, an H
2
S scavenger
application was already in place. A
triazine based scavenger, the most
common chemistry for this type of
application, was injected in the
rundown to the storage tanks.
Even though the refnery was
applying a good monitoring
programme, the dosage rates were
unusually high. Two very important
aspects were that the starting H
2
S
content in the heavy fuel oil was
quite variable, ranging between 10
mg/kg and 40 mg/kg, and that the
refnery had very fexible produc-
tion planning which made it
necessary to have the H
2
S scavenged
in a short time to make the fuel
always available when needed.
In particular the heavy fuel oil
needed to be on specifcation by the
time it reached its intermediate stor-
age tanks, before fnal blending, and
the minimum storage time in these
tanks was only a few hours, the
time therefore available for the H
2
S
scavenger to complete its action.
The high variability of the H
2
S
content was mainly due to the
changing operating conditions of
the vacuum distillation unit and the
crude slate, in particular in the
alternation of low and high sulphur
campaigns.
The refnery, aware of the impor-
tance of these factors, was
monitoring the blank H
2
S content
once a day and every time relevant
changes in the operating conditions
were implemented. At the same
time, treated heavy fuel oil in the
intermediate tanks was sampled
and analysed three times a day to
verify continuous fulfllment of the
H
2
S spec.
Monitoring scavenger treatment
There are two main targets in
monitoring H
2
S scavenger treat-
ment: reliable measurement of the
H
2
S content and fast response to its
variations. Once these two aspects
are guaranteed, over-dosage and
off-specs can be minimised. A deep
knowledge of the system is neces-
sary to identify the critical variables
and to defne the best procedures.
Continuous assistance and collabo-
ration during the process make
these procedures effective.
With regard to reliable measure-
ments, every technician knows that
data can only be as good as the
analysis that produces them, and
the analysis can be only as good as
the sample. In H
2
S scavenging
applications, this means taking due
chimec.indd 2 11/12/2013 12:39
A wide range of vessel internals in down fow processes to
retain the often costly media and to provide a collection area for
the process fow across the entire vessel diameter or length
All assemblies required in a radial fow reactor vessel, custom
built to satisfy every installations needs
Scallops to meet the various conditions of your process and to
ensure best fow distribution, optimum mechanical resistance
and long service life
Expertise and global feld service from initial product design
through proprietary fabrication, inspection, testing, installation
and start-up
JOHNSON SCREENS BRAND OF PRODUCTS
AND SOLUTIONS FOR THE HYDROCARBON
PROCESSING INDUSTRY

As the manufacturer of the Johnson Screens

brand, Bilfnger Water Technologies provides:

BILFINGER WATER TECHNOLOGIES


www.water.bilfnger.com
Australia - Asia Pacifc
Phone +61 7 3867 5555
Fax +61 7 3265 2768
asiapacifc.water@bilfnger.com
Japan
Phone +81.45.661.3575
Fax +81.45.661.1921
japan.water@bilfnger.com
France
Phone +33 5 4902 1600
Fax +33 5 49021616
france.water@bilfnger.com
North & South America
Phone +1 651 636 3900
Fax +1 651 638 3171
usa.water@bilfnger.com
PTQ_Issue Q1.indd 1 11/12/2013 15:22:43 biflinger.indd 1 11/12/2013 14:57
scavenger, but we could not verify
this. The following evidence, in any
case, showed us that most probably
this was not the case.
The next step was to verify the
laboratory results in the feld
through an industrial test.
Management of the scavenger
dosage rates appeared from the
beginning more diffcult than
expected. The high variability of the
H
2
S content and the excessive
corrective measures the refnery
used to apply did not permit keep-
ing the dosage rate at the intended
levels. In particular, dosage rate
management was unbalanced by a
fear of producing off-specs,
regarded as a bigger threat by
comparison to the extra costs arising
from overdosage. Initially, it was
possible to apply an average dosage
ratio only 15% lower than the one
applied by the previous supplier.
Our approach was to put more
66 PTQ Q1 2014 www.eptq.com
focus and effort into standardisation
of the monitoring procedures and in
particular into the contact time
experienced by the fuel samples
before testing. More care was also
given to the time between a varia-
tion in H
2
S level and the follow-up
adjustment in dosage rate. Each
delay had a potential impact on
chemical consumption due to the
consequent overdosages and
off-specs that could result from it.
The better results obtained helped
also to increase the confdence of
refnery personnel in dealing with
H
2
S levels near the specifcation.
This allowed us to reduce the
corrective measures and therefore
provided further beneft through a
reduction in chemical consumption.
At the end of the frst optimisa-
tion period, the average dosage rate
had been reduced by 35%
compared to the level applied by
the previous supplier. Still, we had
not reached the target identifed in
the laboratory study, and in our
opinion the response time to the
variations in H
2
S content, still too
high, was one of the main reasons.
In order to reduce this response
time, we needed to anticipate a
correction to the dosage rate before
the result of an H
2
S test was availa-
ble. In other words, it was
necessary to develop a tool able to
predict with a certain accuracy the
vacuum residue H
2
S content from
the plant operating conditions.
First, we identifed the process
variables that could have a major
impact on H
2
S content and, among
these variables, those that could be
more easily monitored. Then we
developed a model able to correlate
these variables with H
2
S concentra-
tion and optimised it using data
available from the plant.
It appeared from the beginning
that the prediction system was able
to give a good indication of the H
2
S
trend (see Figure 2). The model
enabled preliminary optimisation
of the dosage rate at the same time
as a variation in operating condi-
tions (and hence H
2
S content)
occurred, so that only fne tuning
was necessary after identifcation of
the new H
2
S level through labora-
tory analysis.
The prediction system could not
be used for a full replacement of
laboratory analysis; this was not
the scope for which it was devel-
oped. An improvement in its
precision was later obtained
through the implementation of new
models and more variables, but still
its intrinsic error suggested that it
could only be used in combination
with laboratory analysis.
The prediction system was then
implemented in the form of a
simple mask, available in the
control room PCs, so that operators
and supervisors could be constantly
more involved. Implementation of
the prediction system enabled
further improvement in the average
dosage rate, reducing it to approxi-
mately 50% of the level applied
under the previous regime.
Figure 3 shows the trend of
dosage rate reported as parts per
million of additive for each part per
million of H
2
S. The explicit numbers
St. mixer
Fluxant
HFO
VDU
H
2
S
scavenger
~4h
30 minutes
4-6 Barg
130-140C
= sampling point
Figure 1 Heavy fuel oil rundown line
30
50
45
40
35
25
20
15
10
5
H
2
S

c
o
n
t
e
n
t
,

m
g
/
k
g
0
10 11 12 01 02 03 04 05
Month
Predicted (moving averages)
Measured (moving averages)
Figure 2 Prediction system results
chimec.indd 3 11/12/2013 12:39
www.eptq.com PTQ Q1 2014 67
With broad experience in the management
of treatments applied to bunker fuel, his
main areas of activity include ow improvers
for middle distillates and residues. He holds
a masters degree in chemistry from Rome
University. Email: ggrande@chimec.it
Alessandra Berra is Technical Development
Manager with Chimec Fuel Additives
Technological Units. She has over 15 years of
experience in fuel additives development and
marketing in the downstream and renewable
fuel areas and holds a masters degree in
chemistry from Rome University.
Email: aberra@chimec.it
key to avoiding unwanted off-specs
and unnecessary extra consumption
of chemicals.
To achieve these targets, collabo-
ration between customer and
supplier is fundamental to identify
critical variables, available tools
and options and areas for improve-
ment. And it is only through such
collaboration that new ideas and
tailored solutions can come alive.
Giacinto Grande is Product Manager with
Chimec Fuel Additives Technological Units.
are confdential information but the
trend may be easily seen. The aver-
age dosage rate during our
management was reduced by 37.5%
as a result of implementation of the
predictive software.
As a result of the good results
achieved, the industrial test turned
into a continuous application to the
full satisfaction of the customer and
the supplier. The model algorithm
and parameters are controlled and
updated when necessary by Chimec
personnel, to maintain the effective-
ness of the software.
Conclusions
The case study presented here
shows how an H
2
S scavenger appli-
cation may need particular care to
provide real cost optimisation. Not
all refneries have conditions simi-
lar to those discussed and the
necessary solutions may differ but
it is the approach that must be the
same in all the cases.
Good additive selection is impor-
tant, but generally not suffcient. A
proper treatment setup and tailored
monitoring procedures provide the
R
a
t
i
o

a
d
d
i
t
i
v
e
:

H
2
S
Time
Treatment start-up Further optimisation
achieved after software
implementation
Average dosage rate:
37.5% relative to our
current start-up
Figure 3 Dosage trend
www.eptq.com PTQ Q4 2013 27
potentially to the revamp of exist-
ing facilities. The zero gasoline
production refnery is a practical
and viable reality.
References
1 Gillis D, Unique opportunities with proven
technologies to maximise residue conversion
& renery margins, Asia Technology forum,
Bangkok, 10-11 Oct 2012.
2 Edwards S M, et al, Relative economics of
mixed C
4
s processing routes, PTQ, Q1 1998, 1.
Blasis Stamateris is Downstream Business
Consultant in the Business Solutions Group
of Foster Wheeler, UK, participating in renery
conguration studies, feasibility studies and
FEED work for grassroots ren eries, CTL
facilities and renery upgrades. A graduate in
chemical engineering, he has over 25 years
experience in oil rening and upgrading.
Email: blasis_stamateris@fwuk.fwc.com
Dan Gillis is Director, Rening for Foster
Wheeler, US. He provides technical direction
and applications development for Foster
Wheelers proprietary heavy oil technology. He
is an engineering graduate of the University of
Saskatchewan, Canada.
Email: dan_gillis@fwhou.fwc.com
and competitive confguration.
Certainly, if the objective is to
maximise revenues with proven
technology, in a reliable environ-
ment, the confguration that has
been developed by Foster Wheeler
utilising its SDA + DCU combina-
tion offers a competitive advantage.
How to make the right decision
The integration of refning and
petrochemical facilities offers
several opportunities for adding
value to refnery streams by
exploiting synergies between differ-
ent process units while rationalising
investments that can positively
impact project economics.
Different refnery confgurations
can be tailor-made to suit a range
of processing objectives, including
variations to the propylene-to-
ethylene ratio, and the production
of middle distillates and aromatics.
Capacities of the process units can
be adjusted to accommodate differ-
ent crude oils.
The concepts outlined are appli-
cable to grassroots projects and
of magnitude total installed cost
(TIC) estimates that have been
prepared, based on Foster
Wheelers in-house cost estimating
database.
As can be seen, the investments
for the upgrading section of the
refnery confgurations are of the
same order of magnitude. With
similar product slates and invest-
ments, the differences in economic
indicators between the refnery
confgurations will not be signif-
cant enough to clearly favour one
confguration over another.
Therefore, when it comes to the
decision of which refnery confgu-
ration to choose, other factors
besides the product slate and
investments, such as operating
complexity and associated reliabil-
ity of the units, should be taken
into consideration.
Ultimately, the selected option is
based on an individual refners
objectives. The SDA + DCU options
represent well-proven technologies.
If there is a suitable outlet for the
coke produced, this is a viable
www.eptq.com PTQ Q3 2013 81
customer conveyed the need to
process more oil sands-derived
coker gas oil. Consequently, the
rst initiative was to plan a transi-
tion to a more metals-tolerant
catalyst system, and Ascent
DN-3551 catalyst was selected for
Cycle 2. This catalyst has been
shown to have superior feed poison
tolerance and was considered to be
better suited to accommodate the
expected poisons that would be
associated with processing incre-
mentally higher percentage of oil
sands-derived coker gas oils.
The second initiative included
optimisation of the demet catalyst
package to enhance catalyst stabil-
ity. Prior to Cycle 1s completion
and the DN-3551 loading, we
advised the customer that the opera-
tion was being subjected to a higher
feed endpoint and the catalyst
system was at risk of accelerated
aging due to feed resid and metals
entrainment. Subsequent detailed
feed component analysis at
Criterions Research facility
conrmed elevated feed poisons
by DN-3551. The points enclosed in
the highlighted areas labelled
DN-3651 represent samples tested
as part of scale-up and manufactur-
ing optimisation. Use of high
throughput experimentation
allowed rapid development of a
reliable manufacturing process
yielding the highest practical cata-
lyst performance.
Applying technology improve-
ments such as DN-3651 requires a
holistic joint approach to ensure
that the full benets of the technol-
ogy can be taken advantage of by
the operator, taking into account
the site-specic constraints and
objectives.
In this example, Criterion utilised
a collaborative approach with its
customer to improve catalyst
performance and facilitate more
difcult oil sands-derived feed
processing capabilities. Three initia-
tives were implemented.
The original cycle 1 utilised
Criterions Centinel DN-3100 cata-
lyst as the primary conversion
system. During the cycle, the
tubular ow reactors with auto-
mated process control and sampling.
The use of this multi-tube reactor
system allows signicant accelera-
tion of catalyst development relative
to conventional testing techniques.
Leads generated with the high
throughput equipment were
conrmed by conventional-scale
pilot plant testing.
The target performance level for
DN-3651 was an improvement in
HDN activity of at least 10F
(5.5C) relative to DN-3551 (~20
RVA) with equivalent or better
HDS activity with a challenging
design feed containing bitumen-
derived vacuum gas oil and heavy
coker gas oil. Figure 14 illustrates
the high throughput experimenta-
tion HDN and HDS activity data
obtained during the prototype
development phase as well as
during the scale-up and commer-
cialisation phase. The activity as
presented in these plots is the
decrease in temperature required to
achieve the target S or N level in
the product relative to that required
Kurita Europe GmbH
.
Industriering 43
.
41751 Viersen
.
Germany
Phone:+49 2162 9580110
.
www.kurita.de
We keep a close eye on your plant effciency.
The trouble-free operation of your plant is the precondition for maximum economic effciency. Kuritas patented ACF technology inhibits the formation
of chloride and ammonium salts, thereby preventing fouling and corrosion. This is only one example of how Kurita protects your production unit against
unscheduled shutdowns, extends its runtime and lowers process costs while at the same time increasing plant and operational safety. Particularly
for refneries and petrochemical plants, Kurita offers innovative and sustainable process technologies as well as a broad portfolio of products for water
management. From consultation through to implementation and maintenance, we support you with a comprehensive range of services.
As you can see, we keep a close eye on the effcient operation of your plant.
Process technologies for refneries and petrochemical plants
Water treatment technologies for cooling water, boiler water, waste water and reverse osmosis
RZ_Anzeige_Quer_24052013_MQ.indd 1 24.05.13 15:24
cri.indd 8 10/06/2013 16:38
f wheeler.indd 8 12/09/2013 13:41
chimec.indd 4 11/12/2013 12:39
Onis lnc. | One Riverway | Suite 1700 | Houston | Texas 77056 | USA
1 (713) 840 6377 | sales@onislineblind.com | www.onislineblind.com
www.onislineblind.com
onis.indd 1 10/06/2013 13:10
Flare Netware H
2
S Service
Representation in : Europe (Germany, Great Britain, Finland, Netherland, Spain)
Latin America (Argentina, Brazil, Mexico, Venezuela)
Asia and Middle East (China, India, Kuwait, Russia, Saudia Arabia,
Singapore, South Korea, Taiwan)
onis.indd 1 11/12/2013 10:35
Overcoming high conductivity in process
condensate
I
n 2007 the Bahrain Petroleum
Companys (Bapco) Low Sulphur
Diesel Production (LSDP)
Complex came onstream. The
complex centres on a 60 000 b/d
heavy vacuum gas oil (HVGO)
hydrocracking unit, which is
supplied with hydrogen from the
No.2 Hydrogen Plant (2H2P). 2H2P
is a conventional unit, licensed
from Technip USA, consisting of
zinc oxide (ZnO) purifcation, a
top-fred Steam Methane Reformer
(SMR), High Temperature Shift
(HTS) conversion and a pressure
swing adsorption (PSA) unit. The
ZnO, SMR and HTS catalysts were
all supplied by Johnson Matthey.
The feed to 2H2P is sweet natural
gas and the plant design capacity is
100 million scfd of 99.9% purity
hydrogen. 2H2P is a large exporter
of 600 psig superheated steam to
the rest of the refnery.
After start-up, Bapco found that
the water in the refnery condensate
systems had high conductivity and
had to be dumped to sewer to
prevent it from entering the boilers
in the Bapco operated power plant.
The cause was ultimately attributed
to the presence of ammonia (NH
3
)
and methanol (MeOH) in the 2H2P
process condensate, which was
recycled within 2H2P for the gener-
ation of steam that was then
exported, condensed and recycled
to the power plant. The high
conductivity masked the presence
of any leaks from the seawater
cooling system into the refnery
condensate systems and was also a
threat to the integrity of these
systems due to corrosion. The 2H2P
process condensate, equivalent to
200 US gal/min, therefore had to
Analysis of design and operations enabled a rener to eliminate contaminants
from hydrogen plant process condensate by installing a high pressure stripper
TAPAS KANTI SAMADDAR and GREGORY M LILBURNE The Bahrain Petroleum Company BSC(c)
CALUM MCINTOSH and MILETA BABOVIC Johnson Matthey
be routed to the oily water sewer
(OWS). Consequently, the produc-
tion of make-up water to the
refnery, the fash evaporated distil-
late (FED) from the desalination
units, had to be increased and a
signifcant fnancial penalty of the
order $100 000 per month was
incurred. Hence a permanent solu-
tion was required.
Bapco undertook a detailed
review and evaluation of different
options for reducing NH
3
and
MeOH levels in the process
condensate. This also involved
discussions with several experts
including the catalyst vendor,
process licensor, other hydrogen
plant operators and water treat-
ment consultants. Ultimately the
company decided that the most
cost effective approach was to
install a high pressure (HP) Steam
Stripper.
The new stripper was designed
by Technip USA. Johnson Matthey
was again closely involved to
provide catalyst performance
predictions for the revamped
process fow scheme. The stripper
was successfully commissioned in
Q1 2011.
The process engineering and
catalytic issues associated with the
new HP Steam Stripper, as well as
the concerns, problems and limita-
tions that arose during project
development and unit start-up, and
how they were overcome, are
presented in this article.
Process condensate ow scheme
(original design)
The original process fow scheme
before the installation of the new
HP Steam Stripper is shown in
Figure 1. About 200 US gal/min of
process condensate is produced in
2H2P during normal operation. The
condensate system design that
Bapco selected was the same as that
used in the older No.1 H
2
Plant
(1H2P), which had never had any
problems.
In the 2H2P fow scheme, the
desulphurised natural gas is frst
mixed with high pressure super-
heated steam and this combined
stream is further heated in the feed-
steam superheat coil in the SMR
and fows into the top of the
reformer tubes. The reformer effu-
ent is cooled to about 650F (343C)
in the process gas boiler and fows
into the HTS reactor. The HTS reac-
tor effuent is then cooled to 110F
(43C), which also causes the steam
in the effuent to condense. The
process condensate is separated
from the hydrogen-rich process gas
in the PSA feed knock out (KO)
drum. The gas is routed to the PSA
unit for fnal purifcation.
In the original as-built unit
confguration, the process conden-
sate from the KO drum was mixed
with FED make-up water and
routed to the deaerator for degas-
sing, using low pressure (LP) 25
www.eptq.com PTQ Q1 2014 69
High conductivity
in the renery
condensate return
systems was caused
by NH
3
in the 2H2P
process condensate
bapco.indd 1 11/12/2013 14:55
70 PTQ Q1 2014 www.eptq.com
Bapco established that the high
conductivity in the refnery conden-
sate return systems was caused by
the presence of NH
3
in the 2H2P
process condensate, at typical levels
of 500 wppm, which had ended up
in the superheated steam system,
via the steam drum. In addition to
NH
3
, the condensate also contained
other dissolved gases like carbon
dioxide (CO
2
), carbon monoxide
(CO) and other by-products includ-
ing MeOH and formic acid.
Bapco had previously experi-
enced elevated concentrations of
NH
3
in the process condensate in
the No.1 Hydrogen Plant (1H2P)
but this was a short term phenome-
non attributed to high activity fresh
catalysts and it disappeared within
four to six weeks after start-up.
However on 2H2P, the problem
persisted and even after six months
of operation the NH
3
had only
decreased to 300 wppm. The MeOH
concentration was 200 wppm.
During this period, Bapco
worked very closely with Johnson
Matthey, the 2H2P catalyst vendor,
to identify and understand the
reaction mechanisms that were
occurring.
Johnson Matthey confrmed that
at the 2H2P operating conditions,
the concentrations of NH
3
and
MeOH would persist throughout
the four-year operating cycle.
Ammonia and methanol formation
The NH
3
and MeOH (and other
by-products) that ended up in the
process condensate are formed,
respectively, in the SMR and HTS
reactor.
Ammonia formation
The key reactions related to NH
3
formation in the SMR are:
N
2
+ 3H
2
2NH
3
H -92 kJ/kg-mol N
2
(1)
CH
4
+ H
2
O CO + 3H
2
H +206 kJ/kg-mol CH
4

(2)
The overall NH
3
production is a
trade-off between several compet-
ing factors, as follows:
Higher feed N
2
content increases
the NH
3
(reaction #1). Bapcos
natural gas contains 11 mol% N
2
,
psig steam. The deaerator operating
pressure is 15 psig. The conden-
sate/FED from the deaerator was
heated and recycled to the steam
drum for the generation of 600 psig
steam. The steam is superheated
and part of this stream is routed to
the SMR as feed process steam,
with the remainder being exported
to all parts of the refnery.
Condensate from the refnery
steam users is collected in the High
Grade (HG) and Low Grade (LG)
condensate systems and returned to
the power plant, where it is mixed
with FED and fows into the boilers
as boiler feed water (BFW).
Conductivity meters are installed
on the condensate return headers to
detect leaks from the seawater cool-
ing systems into the steam/
condensate systems. When the
conductivity goes high, above
25 S, all the condensate is auto-
matically dumped to the OWS to
ensure that any hardness salts do
not enter the boilers. They would
cause scale formation, corrosion
and hot spots. Dumping of conden-
sate is a very costly emergency
procedure because the FED produc-
tion rate from refnery desalination
units must be substantially
increased to meet the refnerys
FED and steam demand.
The problem: impurities in process
condensate
During commissioning of 2H2P, the
process condensate was initially
routed to sewer. However when
2H2P was brought fully onstream,
the condensate was routed to the
deaerator and then to the steam
drum, with some of the 2H2P super-
heated steam being exported into the
refnery steam system. After about 30
minutes, the refnery condensate
conductivity shot up from a typical
value of < 10 S to 140 S. As a
result, the condensate recovery
systems high conductivity alarms
were activated and all the condensate
was automatically dumped. The situ-
ation returned to normal when the
process condensate was again routed
to OWS in 2H2P.
Through a series of plant trials,
Steam
Methane
Reformer
Zinc
Oxide
Deaerator
Steam
Drum
HTS
PSA Feed
KO Drum
Process Gas Boiler
Fuel
gas
Natural
gas
Export
Superheated
Steam
Flash
Evaporated
Distillate
FI
FV
H
2
to
PSA
Oily water sewer
LP
steam
Vent
Figure 1 Original process ow scheme
bapco.indd 2 11/12/2013 14:55
which is unusually high and so
drives up the NH
3
make compared
to most other natural gas fed H
2

plants
Increasing the steam to carbon
(S/C) ratio reduces the N
2
partial
pressure, which decreases NH
3
levels
Higher operating pressure drives
reaction #1 equilibrium to the right,
hence more NH
3
is produced
Operating temperature also has
an impact, but the situation is
complex, with competition between
equilibrium and kinetics. Higher
temperature and pressure increase
any reaction rate, but whether a
higher reformer outlet temperature
gives more or less ammonia is also
affected by a balance between reac-
tion #1 equilibrium and the steam
reforming equilibrium (#2). A
higher temperature produces more
H
2
from the steam reforming reac-
tion, which pushes the NH
3
equilibrium to the right, but it also
pushes reaction #1 equilibrium to
the left because it is exothermic.
Each case needs to be assessed
individually.
The difference in the NH
3
levels
in the process condensate from
Bapcos 1H2P and 2H2P is due to
the relative operating conditions
prevailing at the outlet of the
reformer in each case. The impact
of N
2
content and the SMR operat-
ing conditions is summarised in
Table 1.
The initial high activity of fresh
SMR catalyst means that NH
3

production activity is also high.
Experience shows that, in practice,
the actual start-of-run (SOR) NH
3

levels in the condensate are
around 30% of predicted equilib-
rium values, while at end-of-run
(EOR) they are about 10% of the
equilibrium predictions due to
lower catalyst activity. For 2H2P,
the SOR and EOR estimated NH
3

levels were, respectively, 420
wppm and 140 wppm. The predic-
tions are consistent with Bapcos
measured SOR concentrations of
about 500 wppm. For 1H2P, the
predicted NH
3
levels at SOR and
EOR were, respectively, 140 and 47
wppm. The corresponding feld
measurements were 45 and 8
wppm.
www.eptq.com PTQ Q1 2014 71
Since equilibrium levels of
ammonia are not attained in either
1H2P or 2H2P, the limiting factor
on NH
3
formation is kinetics.
2H2P also operates with half the
S/C ratio of 1H2P, leading to
higher inlet N
2
concentration, and
at a higher operating pressure and
tube exit temperature. All three
factors drive the rate of
NH
3
formation, resulting in higher
NH
3
levels in 2H2P compared to
1H2P.
The short term phenomenon of
elevated concentrations of NH
3
in
the process condensate at SOR in
1H2P, and the subsequent drop
within four to six weeks after SOR,
is likely a result of the high S/C
and steam promoted passivation
of the catalyst towards the NH
3
formation reaction over that
period. The high S/C ratio used in
1H2P reduces the partial pressures
of nitrogen and hydrogen, and
therefore reduces the potential for
ammonia formation. It should also
be noted that the two reformer
units are of different design, as
well as operating at different
conditions, which results in differ-
ent temperature and gas
composition profles.
The NH
3
dissociation reaction,
which is the reverse of the NH
3
formation reaction, also infuences
overall NH
3
production:
2NH
3
N
2
+ 3H
2
H +46 kJ/kg-mol NH
3
(3)
This reaction has been studied at
various temperatures, pressures
and with various catalyst types.
1
As well as being the catalyst of
choice for reforming, nickel is also
effective in the cracking of NH
3
,
which is favoured by high tempera-
tures and low pressures. In
this respect the temperature is
favourable in 2H2P, but the pres-
sure is not. The reverse is the case
for 1H2P.
Formation of methanol and other
by-products
2
Methanol is produced in the HTS
converter, and especially in the Low
Temperature Shift (LTS), which is
installed in 1H2P. It is formed
through copper promoted side reac-
tions of H
2
with carbon monoxide
(CO) and carbon dioxide (CO
2
). (In
addition to iron oxide and chro-
mium oxide, the HTS catalyst
contains a small amount of copper
oxide.
2
The LTS catalyst is copper
based.) The relevant reactions are:
CO + 2H
2
CH
3
OH H -90 kJ/mol (4)
CO
2
+ 3H
2
CH
3
OH + H
2
O H -50 kJ/mol (5)
The high operating temperature
in the HTS converter kinetically
favours MeOH formation.
However, the high temperature
also drives the equilibrium for reac-
tions #4 and #5 to the left i.e. lower
MeOH formation. The NH
3
from
1H2P 2H2P
Inlet S/C ratio 6 3
Reformer inlet N
2
, mol% wet 1.9 3.2
Tube outlet temperature, F (C) 1480 (804) 1580 (860)
Tube outlet pressure, psig 384 436
Predicted equilibrium NH
3
levels in process condensate, wppm 470 1400
Effect of N
2
content and SMR operating conditions on NH
3
formation
Table 1
1H2P 2H2P
S/dry gas ratio 1.062 0.513
Inlet concentrations CO/CO
2
/H
2
, mol% wet 3.9/5.9/35.3 8.9/5.7/45.3
Temperature, F (C) 630 (330) 630 (330)
Pressure, psig 380 430
Predicted equilibrium MeOH levels in process condensate, wppm 66 447
HTS: effect of gas composition and operating conditions on MeOH formation
Table 2
bapco.indd 3 11/12/2013 14:55
72 PTQ Q1 2014 www.eptq.com
stripping and absorption processes,
as well as HP steam stripping,
although the experience of at least
one other rener showed that LP
stripping alone was not a viable
option.
Through discussions with other
hydrogen plant operators, technol-
ogy licensors and water treatment
consultants, Bapco established that,
on large scale hydrogen plants, HP
steam stripping was typically used
for NH
3
and MeOH removal. This
was a proven approach, which
would ensure that only clean
condensate would be recycled to
the steam drum for steam genera-
tion, and also allow the stripping
medium to be returned to the
process and minimise the loss of
water.
The process condensate is a
complex equilibrium system involv-
ing the ionic species such as NH
4
+
,
CO
3
2+
and OH
-
ions, as well as the
actual gas molecules. Fairly severe
operating conditions are needed to
maximise the concentrations of the
molecular form of the impurities,
for instance, NH
3
and CO
2
, which
can then be stripped out of the
condensate.
4
High temperatures
also reduce the solubility of these
gases and facilitate stripping.
The use of high pressure satu-
rated stripping steam ensures high
enough operating temperatures to
achieve the required stripping
performance. In Bapcos case, the
operating pressure was selected so
that the overhead steam containing
the impurities could be routed
back to the process. The MeOH
would be destroyed in the
reformer, and the NH
3
would be
recycled and reach an equilibrium
concentration.
A sketch of the new HP Steam
Stripper with the key operating
conditions is shown in Figure 2.
The column has two beds of
random packing and was installed
during the shutdown for turna-
round and inspection in Q1 2011.
The details are as follows:
Diameter: 4ft 6in (1.4m)
Height: 72ft (22m) (tan-tan)
Materials of construction: carbon
steel with 3mm type 304L SS
cladding
Two packed beds: IMTP #40 type
The actual MeOH concentrations
in the condensate that were meas-
ured in Bapcos 2H2P were about
200 wppm, as compared to the
predicted equilibrium value of 447
wppm.
Technology selection for ammonia
and methanol removal
The main objective was to select a
process for the removal of the
ammonia, methanol and other
impurities from the process
condensate, to ensure that they did
not end up in the 2H2P export
steam to the rest of the renery.
Only proven technologies, which
could be practically and cost-
effectively integrated into the exist-
ing plant, with minimum downtime
and minimum environmental
impact, were considered.
Several technologies for removing
NH
3
and MeOH from process
condensate are available, ranging
from steam and air stripping to
biological processes.
4
Each has its
own advantages and limitations,
and their efcacy depends on the
system being treated and the
process requirements. There is no
common design solution. Bapco
also reviewed low pressure
the SMR passes through the HTS
catalyst bed effectively unchanged.
There is also potential for MeOH
to react with NH
3
to form amines
exhibiting the well known sh-like
odours. However, due to the
prevailing conditions in the bed,
monomethylamine formation does
not occur to a signicant extent.
Ethanol can form at about 10% of
the methanol concentration from
condensation-type reactions of
methanol. Formic acid and acetic
acid can also be formed over the
HTS bed, with typical levels of tens
of wppm in the process
condensate.
3
The effect of feed composition
and operating conditions on MeOH
formation in the HTS converter is
shown in Table 2.
MeOH formation across the HTS
catalyst bed does not decrease
signicantly over the life of the cata-
lyst. The dominant factors are the
operating conditions: temperature,
pressure and reactant concentra-
tions. Although the model assumes
that an equilibrium level of MeOH
is generated across the converter,
some of the available evidence
shows MeOH levels in the conden-
sate are less than equilibrium levels.
Process condensate
Flow: 255 Std US gpm
Temperature: 445F
Estimated composition
H
2
O: 99.04 wt%
CO
2
: 0.80 wt%
NH
3
: 1260 ppmw
MeOH: 360 ppmw
Stripper overhead
Flow: 40.4 Mlb/hr
Temperature: 475F
Pressure: 525 psig
Estimated composition
H
2
O: 96.97 wt%
CO
2
: 2.53 wt%
NH
3
: 0.39 wt%
MeOH: 0.11 wt%
HP steam saturated
Flow: 45 Mlb/hr
Temperature: 490F
Pressure: 605 psig
Neutralising amine
Stripped condensate
Flow: 263 Std USgpm
Temperature: 475F
Pressure: 525 psig
Estimated composition
H
2
O: 100.0 wt%
CO
2
: 0.0 wt%
NH
3
: <0.50 ppmw
MeOH: <10.0 ppmw
Figure 2 HP Steam Stripper details
bapco.indd 4 11/12/2013 14:55
GE Works to redene pump efciencies.
In processing industries, GEs SPS pumping
systems work to increase efficiency, reduce
environmental impact and provide versatile
pumping solutions. For example, our multi-stage
centrifugal SPS pumps are a widely-used, efcient
solution for circulation, transfer boosting and
disposal applications. Our SPS systems provide
enhanced pumping efciencies by lowering noise
and vibration levels, decreasing construction
lead times and delivering improved reliability and
extended runtimes. With GE, you have a worldwide
support system ready to provide engineering,
eld service or sales support.
Take a look at our results, and see how
GE can work for you.
Call +1 713 260 6731 or e-mail sps.inquiry
@
ge.com
www.geoilandgas.com/sps
GE SPS Refining Ad_PTQ_1-18-13.indd 1 1/22/13 9:50 AM
ge.indd 1 07/03/2013 13:37
quest .indd 1 11/12/2013 09:36
www.eptq.com PTQ Q1 2014 75
stripper would be designed to
achieve <0.5 wppm NH
3
in the
stripped condensate, although the
process guarantee would be for
<3 wppm. This was acceptable to
Bapco, since it meant that after
dilution with the FED make-up
in the deaerator, the specifcation
of <1.0 wppm for the export
steam would be met. The
actual performance in 2011/2012
was consistently less than 1.0
wppm NH
3
in the stripped
condensate.
The major concern with the meth-
anol was that it could form organic
acids, which would require higher
dosage rates of sodium phosphate
for neutralisation. As a result, the
foaming tendency in the steam
drum would increase, causing
carryover of sodium salts to
the steam superheat coils and
steam turbines, which was not
acceptable.
Based on feedback from water
treatment specialists, Bapco
concluded that MeOH levels of <10
wppm in the stripped condensate
would not cause signifcant prob-
lems. The licensor advised that this
limit was achievable in practice and
that it was generally used in their
designs. However, they would not
provide a guarantee of perfor-
mance. Bapcos water treatment
consultant also confrmed that <10
wppm was typical of actual indus-
try performance, and Bapco
specifed this as the basis for
design.
Lower MeOH levels could also
have been achieved, but only if a
low pressure stripper were to be
installed downstream of the HP
Steam Stripper. However, in this
case, there was no operational
proprietary software. The predicted
equilibrium concentration of NH
3

in the condensate was 1260 wppm,
which was signifcantly higher
than the measured concentration in
the original fow scheme.
The methanol concentration of
360 wppm was also estimated by
the licensor using their thermody-
namic equilibrium model. The
actual maximum concentration
measured during normal operation
was 335 wppm.
Stripped condensate product
The concentrations of ammonia and
methanol in the stripped conden-
sate were equally important.
However, there is no generally
accepted specifcation for either
component in FED (BFW) or steam.
Bapco canvassed several water
treatment specialists and other
refners, and decided on a specifca-
tion of <1 wppm NH
3
for the 2H2P
export steam. This limit was set to
avoid yellow-metal corrosion in the
steam and steam condensate
circuits.
The licensor would not accept
this specifcation of <1.0 wppm
NH
3
for either the export steam or
the process condensate. However, it
was eventually agreed that the
304 SS random packing (0.25mm
thickness)
Bed height: 24ft (7.3m) each
Distributors: Inlet Model 116
Intalox Deck Gravity Distributor
and Redistributor Model 117.
The revamped process fow
scheme is shown in Figure 3. The
process condensate from the PSA
feed KO drum is now pumped
through the stripper feed-effuent
plate heat exchanger (PHE) into the
top of the column. As it fows
down through the packed beds, it
contacts saturated 600 psig steam,
which strips out the impurities. The
clean stripped condensate heats the
stripper feed in the feed-effuent
PHE, is combined with make-up
FED, and fows through the deaera-
tor and then into the steam drum
for steam generation. The column
overhead steam, containing NH
3
,
MeOH, CO
2
, CO and other impuri-
ties, is routed to the reformer
process steam feed system, and the
combined stream ties into the sweet
natural gas feed line. The gas-steam
feed fows through the superheat
coil and into the reformer tubes.

Design issues
The key design issues were related
to the feed impurities levels, the
stripping performance effciency,
the specifcation of the suction
conditions for the new process
condensate pump, and the new
stripper feed-effuent PHE.
Impurities and performance
specications
Process condensate feed
It was vital to specify the correct
ammonia concentration in the
process condensate feed to the
stripper. Measurement of NH
3
levels in the condensate being
routed to sewer was easy.
However, the stripper overhead
steam would be routed to the SMR
as feed and, although MeOH is
decomposed in the reformer, the
NH
3
is essentially recycled, eventu-
ally reaching an equilibrium
concentration in the process
condensate. Modelling of this
system was therefore done by the
process licensor using the model
that had been set up originally for
the design of 2H2P and their
Component
1
Process condensate feed Stripped condensate product
Design basis Typical Design basis Actual
Ammonia, wppm 1260 1000 to 1400 0.5 <1.0
<3.0 guarantee
[Methanol + ethanol], wppm 360 100 to 300 <10 Not detectable

Note 1: Bapcos water treatment consultant advised that the CO
2
and O
2
should be, respectively, less than
0.5 wppm and 0.007 wppm at the deaerator outlet. Bapco adopted a conservative approach and used
these specications for the stripped process condensate ex the HP stripper; that is, the inlet to the deaerator.
In normal operation, these parameters are not measured.
Stripper design and actual impurity concentrations
Table 3
The concentrations
of ammonia and
methanol in the
stripped condensate
were equally
important
bapco.indd 5 11/12/2013 14:55
76 PTQ Q1 2014 www.eptq.com
reliability and experience of others,
there was no justifcation for a spare
exchanger. The equipment operated
without any problems for almost 18
months, but then leaks developed.
Unfortunately, the PHE could not
be inspected or repaired locally and
had to be returned to the overseas
manufacturer. Consequently, this
forced a long term shutdown of the
stripper and the costly routing of
the process condensate to OWS
again. The reason for the leaks is
not yet known.
Reformer steam ow measurement
Accurate and reliable measurement
of the total steam fow and the feed
gas fow to the reformer is vital for
control of the S/C ratio and to
ensure that it does not fall below
3.0, which is the minimum safe
operating limit for 2H2P.
As Figure 1 shows, in the original
confguration the fow meter (FI) on
the reformer process steam feed
line was upstream of the fow
control valve (FV). Hence, the pres-
sure at the fow orifce was constant
and did not need pressure or
temperature compensation.
However, as Figure 3 shows, in
the new fow scheme there would
be no fow or pressure control on
the overhead steam from the strip-
per, and the column pressure
would foat on the reformer system
back pressure. Since the stripper
overhead steam line would tie into
the fresh process steam feed
system, the total combined steam
fow had to be measured and
controlled. Therefore, the locations
of the existing steam FV and FI
were swapped. The stripper over-
head product steam now enters the
system between the FV and the FI,
and the total steam fow is essen-
tially controlled by the FV, which
regulates the fow of the fresh
steam feed.
As a result of this control scheme,
the operating conditions at the
steam FI varied, and pressure and
temperature compensation of the
fow measurement was required to
obtain an accurate total steam fow
rate and steam to carbon ratio.
SMR furnace duty and temperatures
As was described above, the
the process condensate feed and the
stripped condensate effuent were
very low, of the order of 30F (17C).
Rigorous in-house simulations
showed that if conventional shell
and tube heat exchangers (SHE)
were to be used, 10 single pass
shells or four F-type shells would be
required. This approach would have
been very costly, resulted in high
pressure drops, and required a more
complex piping arrangement and a
signifcantly larger equipment foot-
print area. This conclusion was also
confrmed by the licensor. The solu-
tion was to use a PHE.
Although SHEs are widely used
in petroleum refneries, PHEs are
less popular. Bapco is no exception,
although a few PHEs are used in
low pressure, non-hydrocarbon
applications. The refners main
concerns were reliability, mainte-
nance and plate sealing issues at
high pressure. However, the licen-
sor recommended the use of a
single PHE with an all-welded
plate construction, which would
address these concerns. Such equip-
ment had been successfully used in
similar services on other hydrogen
plants, was considered to be relia-
ble, and would have an operating
cycle of several years. This
approach would minimise invest-
ment and maintenance costs, and
simplify the piping confguration.
Moreover, the footprint area would
be 85 to 95% less than the SHE
arrangements. The use of a PHE
would therefore facilitate installa-
tion and integration into the
existing fow scheme.
Bapco accepted the licensors
recommendation and installed a
single PHE. Given the advice on
justifcation for the additional
investment.
Since start-up, little or no metha-
nol has been detected in the
stripped condensate. The stripper
design impurities levels and actual
performance are summarised in
Table 3.
Condensate pump net positive
suction head available (NPSHA)
Specifcation of the NPSHA for the
new process condensate pump was
a serious challenge. The condensate
in the PSA feed KO drum contains
0.5 mol% dissolved gases. A deci-
sion had to be made on how much
free vapour, if any, would be
allowed at the inlet to the centrifu-
gal pump. Too much vapour would
lead to cavitation and mechanical
damage or loss of pumping capa-
bility. Too little or no vapour could
lead to a very low NPSHA specif-
cation and a costly pump, with
special design features to obtain the
extremely low NPSH required like
lower speed, double suction inlet,
larger impeller inlet (eye) area and
a fow inducer.
5,6

For example, zero vapour would
require the specifcation of an
NPSHA of just 3.7ft (1.13m). In
the unrealistic case of pure water
with no dissolved gas, the liquid
vapour pressure would be the
KO drum operating pressure and
the NPSHA would be 903ft (275m).
After considerable review and
evaluation, the effective vapour
pressure was selected to allow
2 vol% vapour at the pump suction,
which would have no adverse
effects on the pump i.e. no vapour
lock or cavitation.
5,6
The corre-
sponding NPSHA that was
specifed was 402ft (122.5m), and
the actual NPSHR of the new pump
was 7ft (2.13m). The pump has now
been in continuous trouble-free
service for 18 months.
Stripper feed-efuent plate heat
exchanger (PHE)
To improve effciency and minimise
cost and energy consumption, a
stripper feed-effuent heat exchanger
was used in the overall system
design. However, this presented
another unique challenge because
the approach temperatures between
Accurate and reliable
measurement of the
total steam ow and
the feed gas ow to
the reformer is vital
for control of the
S/C ratio
bapco.indd 6 11/12/2013 14:55
www.eptq.com PTQ Q1 2014 77
point on the process line to each
instrument. After these modifca-
tions, the problems of ammonium
bicarbonate salt deposition
disappeared.
Use of neutralising amine
In the original plant confguration,
neutralising amine was used in the
deaerator. Amine injection was also
required for the new HP Steam
Stripper to achieve a pH of 8.5 to 9
in the stripped process condensate
and prevent corrosion in the down-
stream carbon steel piping.
Initially, it was intended to use
the same amine for both services.
However, the amine used in
the deaerator needs to be more
volatile than that required for
the stripper, to enable it to partition
more towards the vapour phase to
maintain the pH of the condensed
steam. The stripper amine must be
less volatile to stay in the liquid
phase to maintain the pH of the
stripper process condensate.
Therefore, a new amine storage
stripper overhead system was a
serious and unexpected develop-
ment, which caused operational
and control problems.
The ammonium bicarbonate salts
formed from the reaction of the
NH
3
and CO
2
in the overhead
steam at temperatures below 170F
(77C). This meant that they
precipitated out in the instrument
impulse lines, causing blockages
and erroneous readings. The salts
also blocked impulse lines on the
emergency shutdown (ESD) fow
transmitters on the reformer inlet
process steam feed fow line,
resulting in unreliable readings
and a high risk of a plant
shutdown.
There were several unsuccessful
attempts to prevent salt formation
by insulating the instrument
impulse lines and the associated
instrument condensate pots. The
problem was eventually solved by
removing the condensate pots and
installing electrical heat tracing on
the impulse lines from the take-off
stripper overhead steam ties into
the fresh process feed steam
system. As a result, the combined
process steam temperature
decreased after start-up, which
reduced the temperature at the
inlet to the gas feed-steam super-
heat coil and therefore increased
the fring duty. However, the steam
superheat coil load also decreased
because a portion of the HP satu-
rated steam leaving the steam
drum is diverted to the stripper as
stripping steam, thereby reducing
the total steam for superheating.
Overall, there was a marginal
increase in reformer absorbed
heat duty and reformer fring,
the main effect being an increase in
the cross-over temperature, which
is a critical process variable trip
setting.
Start-up operating issues
Ammonium bicarbonate
salt formation
The formation and deposition of
ammonium bicarbonate salts in the
Steam
Methane
Reformer
Deaerator
Steam
drum
PSA Feed
KO Drum
Plate Heat
Exchanger
Fuel
gas
Natural
gas
H
2
to
PSA
New equipment
Stripping steam
LP
steam
Vent
High Pressure
Stripper
Process Gas Boiler
Zinc
Oxide
Export
Superheated
Steam
Flash
Evaporated
Distillate
HTS
FI
FV
Figure 3 Revamped process ow scheme
bapco.indd 7 11/12/2013 14:56
78 PTQ Q1 2014 www.eptq.com
References
1 Hacker V, Kordesch K, Ammonia Crackers,
Handbook of Fuel Cells Fundamentals,
Technology and Applications, 2003, ISBN
0-417-49926-9, Vol 3, Pt 2, 121-127.
2 Pach J (Johnson Matthey Catalysts), Causes
and consequences of by-product formation in
ammonia plants, 54th Symposium on Safety in
Ammonia Plants and Related Facilities, AIChE,
2009, Vol 50, 263-274.
3 Madsen J (Haldor Topse), Process
condensate purication in ammonia plants,
Symposium on Safety in Ammonia Plants and
Related Facilities, AIChE, 1991, Vol 31, 227-240.
4 Odenbrand I (Lund University) and Anderson
K H (Haldor Topse), Investigation of process
condensate treatment of a methane steam
reforming plant, Department of Chemical
Engineering, Lund University, 29 May 2009,
www.chemeng.lth.se/ket050/Finalreport2009/
Topso.pdf
5 Tsai M J, Accounting for dissolved gases in
pump design, Chemical Engineering, Vol 89, No
7, 26 Jul 1982, 65-69.
6 Chen C C, Optimal system design requires
the right vapour pressure heres how to
calculate it, Chemical Engineering, Vol 100, No.
10, 1 Oct 1993, 106-112.
Tapas Kanti Samaddar is the Process Specialist
Lubes in Technical Services with Bapco,
Bahrain. He has 20 years of experience, mainly
in technical support for petroleum rening,
specialising in hydrocracking, lube base oil and
uid catalytic cracking, and holds a Bachelor
of Technology degree in electrochemical
engineering from Madurai Kamaraj University,
India.
Gregory M Lilburne is the Manager, Technical
Services with Bapco, Bahrain. He has 33 years
of experience, mainly in technical support for
petroleum rening, holds BE and PhD degrees
in chemical engineering from the University
of Melbourne, Australia, and is a Fellow of the
IChemE and a Chartered Engineer (UK).
Calum McIntosh is Technical Sales Manager,
Middle East Region with Johnson Matthey. He
joined Johnson Matthey in 2002 on acquisition
of the ICI Catalyst Business, holds BSc and
MPhil degrees in chemistry from the University
of Heriot Watt in Edinburgh, Scotland, and is
co-author of several patents and academic
papers in the eld of catalysis.
Mileta Babovic is a Senior Process Engineer
with Johnson Matthey. He joined Johnson
Matthey in 2007 after several years in
academia and related R&D industry, focusing
on heterogeneous catalysis and process
intensication. He holds a MSc degree in
chemical engineering from the University of
Belgrade and a PhD from Newcastle University,
UK, and is co-author of several patents that
have seen successful industrial application.
including the original design basis
and the catalytic processes, as well
as the operation, control and
constraints, and ensured that
Bapcos requirements were prop-
erly met
It was vital to involve the right
people in the evaluation and
decision making process. Diagnosis
of the problem and establishing the
product purity specifcations
required input from Johnson
Matthey, the catalyst vendor,
water treatment specialists, process
licensors and hydrogen plant
operators
Throughout the project, it was
important to question and analyse
everything, even though this some-
times frustrated the licensor/
contractor. Bapco had to be
convinced, and able to minimise
the uncertainty and risk. Good
licensors/contractors understand,
respect and even welcome this
approach
Bapco made time for in-depth
PFD and P&ID reviews with the
licensor and the EPC contractor,
which paid dividends in highlight-
ing problems such as increasing the
stripper T/T length to ensure suff-
cient boot space for the packing
hold-up liquid, and installing a trip
on the stripping steam inlet to
prevent liquid carryover in the
event of a reformer trip
Despite Bapcos attention to
detail, there were some surprises
that needed to be addressed
quickly and effciently, including
the need for fow compensation on
the reformer steam fow and the
ammonium bicarbonate salt
formation.
As a result of these actions, the
HP Steam Stripper was successfully
commissioned and all specifcations
are being met.
tank and two high pressure
injection pumps were installed. The
amine is injected into the stripper
bottom head for better mixing
rather than into the bottom outlet
line. Directionally, this may also
elevate the pH of the steam-water
system inside the column, which
would shift the equilibrium in
favour of NH
3
, which can then be
stripped out.
The amine injection system works
very well and there have been
no pH control or corrosion
problems.
Lessons learned and concluding
remarks
Bapcos HP Steam Stripper has
been very successful in operation.
The problems of high conductivity
condensate have been eradicated,
ensuring that clean HP steam fows
to the superheat coils and steam
turbines, and saving 200 USgal/
min of condensate from going to
drain. The fnancial beneft is about
$100 000 per month. Several process
design and operating challenges
had to be overcome to achieve this
success, out of which came the
following key learning points
related to the original hydrogen
plant design and the stripper
project itself:
In the design of a new H
2
plant
or any revamp facilities, do not
forget the utilities. Always verify
the quality of the process conden-
sate and evaluate the potential
impact on the utilities systems. The
onus is on the client to decide on
the need or otherwise for process
condensate purifcation, but the
question must also be asked of the
licensor
Work closely with the catalyst
vendor to identify and determine
the nature of the side reactions and
the concentrations of impurities,
which may form and accumulate in
the process condensate and other
systems
Having a technically strong and
experienced in-house process
engineering, mechanical and metal-
lurgical engineering capability was
extremely benefcial, especially
given the uncertainties that arose.
These subject matter experts had an
intimate knowledge of 2H2P,
In the design of a
new H
2
plant or any
revamp facilities,
do not forget the
utilities
bapco.indd 8 11/12/2013 14:56
EDV

Technology provides refiners with an economic and reliable means for controlling
emissions of pollutants (PM, PM
10
, SO
X
, HAPs, NO
X
) from FCCUs and other stationary sources.
Write, call or fax Belco Technologies Corporation.
TM Belco Technologies Corporation
Reducing refinery emissions has never been so easy.
9 Entin Road, Parsippany, New Jersey 07054, U.S.A. Phone: 1-973-884-4700 Fax: 1-973-884-4775 www.belcotech.com
Copyright 2010 DuPont. The DuPont Oval Logo, The miracles of science

, and BELCO

are registered trademarks or trademarks of E.I. du Pont de Nemours and Company or its affiliates. All rights reserved.
Belco_Ad_7_2010_210x297mm.indd 1 9/29/10 4:58:27 PM
belco.indd 1 06/06/2013 10:16
GET A RETURN ON INVESTMENT
WITHIN 24 HOURS
Using ITW Online Cleaning on a proactive way will help improve
TURNAROUND OPERATIONS and OPERATIONAL EXCELLENCE
An entire Process Unit can be ITW Online Cleaned in as low as 24 hours on an oil-to-oil basis. Our cleaning technology will
clean all of the Units equipment quickly, safely and environmental friendly.
To recover a Units performance, ITW Online Cleaning can be applied in a Turnaround during shutdown operations, to
reduce/eliminate the need for mechanical cleaning, or during the run.
Furthermore, use of our cleaning technology eliminates the need to open or enter hazardous process equipment.
ITW Online Cleaning is the only technology which has an immediate Return On Investment (ROI).
By using ITW Online Cleaning during the run, to recover Units performance, at Unit start-up (just after 24h, ITW
Technology has been applied), all the benefits are immediately monetized in terms of:
- improved throughput
- improved energy efficiency
- improved product quality
- reduced emission taxes
Regular application of ITW Online Cleaning will target an increased run length under clean conditions, with a far higher
ROI.
For Turnaround applications, a further ROI increase can be added by applying ITW Improved Degassing/Decontamination
to achieve quick and effective safe entry conditions.
EVALUATE ITW ONLINE CLEANING TODAY TO IMPROVE YOUR PLANTS PROFITABILITY !
For more informations contact : ITW S.r.l.- C.da S.Cusumano - 96011 Augusta - Italy
Tel. +39 (0931) 766011
E-mail: info@itwtechnologies.com
www.itwtechnologies.com
Join ITW Team worldwide and send your Curriculum Vitae to : jobs@itwtechnologies.com
ITW
Innovative
Technologies
Worldwide
Now hiring
Professionals
Worldwide
itw.indd 1 11/9/12 11:45:38
Mercury treatment options for natural
gas plants
M
ercury in natural gas and
natural gas liquids is most
likely to be in the elemental
state. Although mercury has a high
boiling point (357C), it also has a
high vapour pressure, which makes
it very mobile. Mobility presents a
challenge for todays gas processors
as mercury can disperse through-
out gas plant assets, making it
diffcult to determine how and
where it should be removed. Left
unchecked, mercury will deposit on
surfaces including those common
to pipelines and plant assets.
Mercury can then desorb back into
gas streams, passing through
contaminated pipelines. Because of
this, long periods of time can elapse
between the installation of an
upstream mercury removal unit
(MRU) and the complete purging
of a pipeline. How best to remove
mercury has consumed a lot of
thought on behalf of processors and
treatment companies alike. Industry
experts suggests that removing
mercury as close to the front end of
a natural gas processing or gas
transmission system as is practica-
ble is the best choice.

Mercury removal process options
The market has a number of
approaches to treatment for
mercury removal. These options
can be categorised as regenerative
adsorbent and non-regenerative
adsorbent solutions for mercury
contaminant removal.
Non-regenerative adsorbent for
mercury removal
A non-regenerative MRU is sulphur
impregnated on carbon or metal
sulphide beds. The common and
Removing mercury as close as is practicable to the front end of a natural gas
processing system is the best choice
SATYAM MISHRA
UOP, a Honeywell Company
traditional approach to mercury
removal has been through the use
of sulphur impregnated carbon
beds. Existing sulphur impregnated
activated carbon options tend to be
less effective at positions upstream
of molecular sieve drying systems
or glycol injection due to the risk of
capillary condensation of water and
heavier hydrocarbons in the micro-
pores of the carbon sub-structure.
Sulphur impregnated carbon prod-
ucts are thus not a product of
choice, particularly where MRU
locations have been in the up-front
position, where raw gas is often at
or close to its dew point and
entrained liquids are common. This
is where fxed bed, metallic based
MRU products fnd greatest success
in their ability to treat wet gas
streams in up-front positions.
UOPs non-regenerative metal
sulphides (UOP GB series of prod-
ucts) are a set of versatile
non-regenerative mercury removal
adsorbents. They are operationally
fexible and can be used to process
gas that is at or close to its dew
point in a variety of process
locations.
Gas and liquid streams contain-
ing thousands of micrograms or
parts per billion levels of mercury
can be treated to extremely low
effuent levels using GB adsorbents.
These high capacity mercury adsor-
bents are engineered using a
copper based active component
fnely dispersed across an alumina
substrate. This high capacity leads
www.eptq.com PTQ Q1 2014 81
50
60
40
30
20
10
0
L
o
g

K

e
q
u
i
l
i
b
r
i
u
m
10
0 10 20 30 40 50 60 70 80 90 100
Temperature, C
2CuO + 2H
2
S + Hg = HgS + Cu
2
S + 2H
2
O
Cu
2
(OH)
2
CO
3
+ 2H
2
S + Hg = HgS + Cu
2
S + 2H
2
O + CO
2
2CuS + Hg = HgS + Cu
2
S
Cu
2
S + Hg = HgS + 2Cu
Figure 1 Thermodynamic driving force favours cupric sulphide over cuprous sulphide for
reaction with mercury
The common and
traditional approach
to mercury removal
has been the use of
sulphur impregnated
carbon beds
uop.indd 1 11/12/2013 12:41
82 PTQ Q1 2014 www.eptq.com
mercury desorption profle from
the HgSIV adsorbent is similar to a
typical water regeneration profle,
except that mercury is completely
removed from the HgSIV adsorbent
well before the full regeneration
temperature is reached (for water
removal).
Mercury exits the bed during the
regeneration step along with the
spent regeneration gas. Plants have
the option to consider installation
of a smaller, non-regenerative
guard bed to treat the mercury
laden, spent regeneration gas. This
ensures removal of mercury from
the gas processing section of the
plant.
Physically, UOP HgSIV adsor-
bents have a similar appearance to
conventional molecular sieves.
These HgSIV adsorbents are loaded
into an adsorption vessel in the
same way as are conventional
molecular sieves. There is no need
for special care such as the use of
nitrogen blanketing during the
installation. For unloading, only the
same precautions need to be taken
as when unloading conventional
molecular sieves.
Regenerative adsorbents for
mercury removal, such as HgSIV
adsorbents, can be used to treat
mercury laden gas in plants which
were not originally designed for
mercury from feed gas. These
adsorbents can be installed in the
existing dehydration beds without
capital expenditure of installing
non-regenerative MRUs. This also
helps users avoid additional pres-
sure drop upstream of cryogenic
separation.
HgSIV adsorbents also help users
treat their gas in situations where
their non-regenerative guard beds
are either out of service or are
under performing.
Mercury removal schemes
Non-regenerative bed for upstream
mercury removal
UOPs GB range of non-regenera-
tive metal sulphide adsorbents can
remove mercury from the raw gas,
upstream of the amine unit and the
dehydration vessels (see Figure 3).
Using larger MRU vessels protects
the brazed aluminum heat
exchanger and signifcantly reduces
to infrequent change-outs and a
longer lifespan, reducing the cost of
mercury removal over time. GB
adsorbents can be supplied in their
oxidised form or in their sulphided
form, which offers fexibility, and
they can be pre-sulphided or
sulphided in situ.
Mercury is reactively adsorbed
from the gas/liquid hydrocarbon
stream by reaction with cupric
sulphide. As Figure 1 shows, cupric
sulphide is the product of choice
for mercury removal over cuprous
sulphide.
Careful selection of the carrier
substrate is essential for maximis-
ing mercury adsorption as these
beds are designed to operate effec-
tively at short contact times and are
exposed to hydrocarbons close to
their dew point. Therefore it is criti-
cal to specifcally engineer the pore
distribution of the adsorbent. As
Figure 2 shows, an engineered mix
of micro- and mesoporous
structures ensures rapid mercury
adsorption with the ability to with-
stand liquid condensation and
minimise mass transfer limitations.
Regenerative adsorbent for mercury
removal
UOP HgSIV adsorbents are regen-
erative molecular sieve products
that contain silver on the outside
surface of the molecular sieve pellet
or bead. Mercury from the process
fuid (either gas or liquid) amal-
gamates with the silver and a
mercury-free dry process fuid is
obtained at the bed outlet. An addi-
tional layer of HgSIV adsorbent, to
an existing dryer confguration,
results in removal of design water
load and the mercury from feed
gas. Mercury and water are both
regenerated from the HgSIV adsor-
bents using conventional gas dryer
techniques where the mercury-
silver amalgam on the molecular
sieve is regenerated thermally. The
G
B

a
d
s
o
r
b
e
n
t
10 100 1000 10000 100000
Pore width,
Mesopores enable rapid
mercury scavenging
Macropores are preferable for
handling liquid condensation and
minimising mass transfer limitations
Figure 2 Pore features of UOP GB products
Feed gas
separator
UOP GB MRU
Dryers CO
2
removal
Raw
natural gas
UOP MOLSIV
adsorbents
Figure 3 MRU upstream of the acid gas removal section and dehydrators
uop.indd 2 11/12/2013 12:41
C O M P R E S S O R S n T U R B I N E S n G L O B A L S E R V I C E
www.elliott-turbo.com
The world turns to Elliott.
n Customer:
Ethylene plant, Saudi Arabia.
n Challenge:
Overhaul and repair four critical steam turbines in 12 days.
n Result:
Elliott coordinated six months of pre-planning
and completed the project four days early.
They turned to Elliott
for planning and execution.
The customer turned to Elliott because they understood that planning is everything
and that few companies can match Elliott in turnaround planning and execution.
Who will you turn to?
elliott.indd 1 10/12/2013 09:53
mercury contamination in and
around the process plant. This
option has become increasingly
popular since it minimises the total
mercury present before there is any
opportunity for mercury to migrate
to various locations within a gas
processing plant and avoid the risk
of subsequent partitioning into
processed natural gas and conden-
sate streams. This option also
avoids subsequent adsorption onto
any pipeline asset or piece of
equipment downstream.
While the recommendation to
position the MRU upstream of the
acid gas removal units remains the
ideal solution for natural gas
plants, there are many operating
facilities with MRU vessels down-
stream of the dryers (see Figure 4).
Placement of the MRU down-
stream of the dehydrator might not
appear to be a logical choice for a
84 PTQ Q1 2014 www.eptq.com
user. However, when plant opera-
tions decides to use sulphur
impregnated carbon as the adsor-
bent for mercury removal from gas,
the MRU location becomes critical.
Sulphur impregnated carbon based
mercury adsorbents are highly
sensitive to moisture and thus need
dry gas to effectively remove
mercury. This forces the location of
the MRU to be downstream of the
dryers.
However, this is not the preferred
MRU location for most natural gas
plants. Because the mercury
removal occurs just upstream of the
cold section, there will be mercury
in the acid gas removal system as
well as in the molecular sieve
regeneration gas. Additionally, this
option requires separate vessels
and adds to pressure drop
upstream of the cryogenic section.
Regenerative adsorbents for
mercury removal
For a regenerative mercury removal
option, HgSIV adsorbent can be
loaded with the dehydration molec-
ular sieve. This helps protect the
brazed aluminum cold box in the
cryogenic section. HgSIV adsor-
bents can be used in natural gas
plants to meet the processing
requirements of a mercury content
of less than 10ng/Nm
3
for the
treated gas.
Handling mercury contamination of
spent regeneration gas
UOP receives frequent enquiries
regarding use of HgSIV adsorbent
and handling of spent regeneration
gas. There are two practical
approaches to handling spent
regeneration gas from a dryer
installed with HgSIV adsorbent.
Some plants choose to route the
mercury laden spent regeneration
gas to the fuel gas header of the
plant or into a larger volume of
pipe gas where the contribution
from the spent regeneration gas is
relatively minor, thereby reducing
any concerns about reliability in the
downstream hardware.
The second option is to treat the
spent regeneration gas with a GB
non-regenerative adsorbent.
Spent regeneration gas treatment
to enable recycle to feed gas
UOP offers design and know-how
support for mercury removal from
spent regeneration gas. The design
incorporates the use of a small bed
of GB adsorbent (non-regenerative
mercury removal adsorbent). One
such scheme is shown in Figure 6.
The spent regeneration gas, after
Feed gas
separator
UOP MOLSIV
adsorbents
UOP GB MRU
Dryers CO
2
removal
Raw
natural gas
Figure 4 MRU downstream of the acid gas removal section and dehydrators
Feed gas
separator
UOP MOLSIV
adsorbent
UOP HgSIV
adsorbent
Dryers CO
2
removal
Raw
natural gas
Figure 5 Regenerative mercury removal along with dehydration
For a regenerative
mercury removal
option, HgSIV
adsorbent can
be loaded with
the dehydration
molecular sieve
uop.indd 3 11/12/2013 12:41
www.eptq.com PTQ Q1 2014 85
Satyam Mishra is Global Strategic Marketing
Manager for UOPs Adsorbents product line
in Honeywell UOPs Dubai ofce. In this
role, Satyam is responsible for coordinating
technology development and promotion
and implementation of the UOP Adsorbents
portfolio to the natural gas and petrochemical
industry. Satyam holds a bachelors degree in
chemical engineering from Indias Maharaja
Sayajirao University of Baroda and has been
involved with the oil and gas industry for the
last 15 years.
mercury treatment options and
arrangements. The decision for
using one or more of these schemes
would generally depend on the
operational and capital require-
ments of the user. However, for
addressing health, safety and envi-
ronmental concerns, the scheme
involving upstream mercury
removal offers maximum
insurance.
being cooled and passed through
a separator, is sent through a
small GB adsorbent bed. Since
regeneration gas fow is relatively
small compared to the feed gas,
the GB adsorbent requirement is
relatively small when compared to
the requirement for feed gas
adsorbent.
Natural gas liquids treatment
C
4
, C
4
+ and in some instances even
C
3
streams from the cryogenic
section of a natural gas plant
exhibit mercury contamination.
This mercury contamination can be
high in situations where upstream
mercury treatment is either ineffec-
tive or not present. Mercury
treatment from these liquid streams
is done principally in the same way
as for feed gas treatment. Options
for mercury removal are the
non-regenerative or regenerative
adsorbent beds designed for liquid
treatment.
Summary
UOP offers various reliable
Feed gas
separator
UOP MOLSIV
adsorbents
UOP HgSIV
adsorbent
Dryers CO
2
removal
Raw
natural gas
UOP GB MRU
Figure 6 Conguration showing regeneration gas mercury removal
www.eptq.com PTQ Q4 2013 65
than predicted corrosion rates, the
maximum allowable predicted
corrosion rates are then reduced
appropriately based on the uncer-
tainty in the corrosion model.
The maximum allowable
predicted corrosion rates from
Table 5 can then be used to esti-
mate a maximum percentage of
sulphur for each stream similar to
what was done in the frst example.
In this case, the 9Cr vacuum resid
piping would not make it through
a second turnaround cycle even
with the current feedstock. By plan-
ning to replace it in the next
turnaround, and the heavy diesel
and atmospheric resid piping at
the following turnaround, the
planned maintenance shutdown. In
this example, the next two turna-
rounds are being considered. Using
this approach, the maximum
acceptable corrosion rate is that
which would put the pipe at its
retirement thickness just as the
planned turnaround is starting. If
the piping only has to make it to
the next turnaround, a higher
corrosion rate would be more
acceptable than if it cannot be
replaced until some turnaround
further down the road. In this case,
it has been assumed that predicted
corrosion rates using the modifed
McConomy curves are +/- 50%. To
decrease the likelihood of an
unplanned failure due to higher
shows a simplifed example of how
the equipment condition-based
approach might be applied to set
limits based on piping condition. In
this example, the last inspection
was done in March 2011; the new
crude slate is expected to start in
2013; and there are planned turna-
rounds in 2016 and 2021.
The frst step using this approach
is to estimate the condition of the
pipe at the time when processing
the new crude is expected to
commence. The inspection depart-
ment should be able to provide the
measured pipe thickness from the
most recent inspection as well as
historical corrosion rates. Based on
the last known condition of the
pipe, the time and the historic
corrosion rate, it is easy to estimate
the thickness of the pipe at the time
when processing the new crude
commences. The inspection depart-
ment can also supply the retirement
thickness for the pipe, allowing the
remaining corrosion allowance to
be calculated.
Normally, piping replacements
would be done during a regularly
Max allowable predicted CR Maximum cut S
Line Material Op temp, To next T/A, To 2nd T/A, To next T/A, To 2nd T/A,
F mpy mpy wt% wt%
6 Heavy diesel 5Cr 625 29 11 >5 2.1
12 Atmospheric resid 9Cr 690 33 12 >8 4.4
8 Vacuum resid 9Cr 675 12 4 6.1 -
Maximum cut S for piping condition-based approach
Table 5
www.eptq.com PTQ Q3 2013 33
and DHP units. He holds a degree in chemical
engineering from Middle East Technical
University, Turkey, and is a certied Energy
Supervisor for industrial plants.
Email: osman.karan@tupras.com.tr
Mehmet Asim Ay is CCR/NHT/ISOM Units
Process Superintendent with Tpras Kirikkale
renery. He holds a degree in chemical
engineering from Middle East Technical
University, Turkey.
Email: MehmetAsim.Ay@tupras.com.tr
Koray Kahraman is CCR/NHT/ISOM Units
Process Chief Engineer with Tpras Kirikkale
renery. His six years of renery experience
includes the process side of hydrocracker and
hydrogen production plants, sulphur recovery,
NHT, ISOM, CCR and DHP units. He holds a
degree in chemical engineering from Middle
East Technical University, Turkey.
Email: koray.kahraman@tupras.com.tr
Arnaud Selmen is Axens Technology Manager
for Naphtha Hydrotreatment and Reforming
Technologies. He has worked mainly with
bottom-of the-barrel technologies, specialising
in heavy crude oil upgrading. He has also been
involved in the process design of aromatics
complexes and NHT, as well as reforming
units startup and troubleshooting. He holds
an engineering degree from the ENSGTI
engineering school and a DEA in renery
process modelling from IFP School.
Email: Arnaud.SELMEN@axens.net
year without a major interruption,
indicating that the vibration prob-
lem was correctly identifed. This
case shows that a good licensor and
refnery relationship is essential for
solving problems that require both
technological and operational
experience.
Osman Kubilay Karan is the Hydroprocessing
Units Process Superintendent with Tpras
Kirikkale renery. His 25 years of renery
experience includes the operational and
process sides of crude, vacuum units,
hydrocracker, hydrogen production plants, CCR
measurements are taken at the
outlet stream of this adsorber and
the results are 0 ppm HCl, whereas
inlet concentration averages
30 ppm HCl.
Conclusion
Currently, the H
2
-rich gas compres-
sors are running smoothly without
any problem. The compressors
were recently opened by the
mechanical maintenance group and
no green oil formation was found,
although they ran for almost one
Reduction
chamber
Reaction
section
Dedicated
chlorine trap
Separator
drum
H
2
to
reduction
H
2
from
reduction
To booster
compressor
Recycle gas
Catalyst stream
Process stream
Figure 3 PFD after adsorber drum
www.eptq.com PTQ Q2 2013 21
rener and particularly the ability
to anticipate market needs in differ-
ent regions as constraints evolve.
PolyFuel and PolyNaphtha are trademarks of
Axens.
Marielle Gagnire is Technology Manager
for hydroprocessing and olens technologies
downstream FCC, especially oligomerisation
and etherication technologies, in Axens
Marketing, Technology and Technical
Assistance Department. She is an engineering
graduate from the Ecole Nationale Suprieure
de Chimie de Paris, and holds a post-graduate
engineering degree from the IFP School.
Annick Pucci is Deputy Product Line Manager
in the eld of light ends hydrotreatment and
a specialist in rening olens processing,
particularly for FCC efuent upgrading.
She holds a bachelors degree in chemical
engineering from Ecole Nationale Suprieure
des Industries Chimiques de Nancy, France.
Emilie Rousseau is a Strategic Marketing
Engineer in Axens Marketing Department. She
holds a chemical engineering degree from the
Ecole Nationale Suprieure des Ingnieurs en
Arts Chimiques et Technologiques de Toulouse,
a masters in chemical engineering from
Imperial College in London and a masters in
energy economics and corporate management
from IFP School.
It is in Europe where the
difference between renery yield
structure and market demand is
critical, especially since conven-
tional rening tools do not have the
exibility to reduce excess gasoline
production and to increase the
amount of middle distillates.
Moreover, with European reneries
facing increasing difculty in nd-
ing export markets for their excess
gasoline and given the tensions in
middle distillate supply, PolyFuel
should full a primordial role in
adjusting the gasoline-distillate
production to better t market
demand.
In other regions, new tendencies
such as shale oil and shale gas are
revolutionising the US market,
providing additional light products
and consequently inuencing
market balance and prices. Today
in the US, as a result of the impact
of shale gas on the cost of LPG,
PolyFuel is already protable for a
mixed feed of LPG and C
5
/C
6
cut.
The exibility of the new process
offers many advantages to the
The new technology is protable
taking into account todays US
market prices even if the middle
distillate price is lower than the
gasoline price. Indeed, as a result of
shale gas production, LPG prices are
low. Adding LPG (C
3
and/or C
4
) cut
in a PolyFuel unit lowers feedstock
costs and contributes to increased
protability, while maximising
middle distillates production in the
renery.
To reach 15% IRR for PolyFuel
with prices based in 2012 in the US
Gulf Coast, the middle distillate
price can be $96/t lower than gaso-
line. If the middle distillate price
were equal to the gasoline price
($1129/t) and the LPG price kept at
$636/t, the IRR would reach 28%.
Conclusion
With the world market for middle
distillates growing and a reduced
demand for gasoline in certain
regions, the new process for olenic
gasoline oligomerisation allows the
renery scheme to be adapted to a
maximum distillate mode.
Taking the Guesswork Out of CO
2
Removal
ProTreat Simulation Software
Get Real
Dont Guess!
SYNGAS, LNG AND CARBON CAPTURE
ProTreat accurately predicts column performance.
100% Mass-Transfer-Rate based no ideal stages.
Grassroots? Revamp? Designing for Packing?
Using MDEA with piperazine? HotPot?
Build a virtual plant with the industrys
most advanced gas treating simulator.
www.ogtrt.com +1 281 970 2700 info@ogtrt.com
Learn how state-of-the-art simulation science can benet you.
OGT-0117 PTQ Half Horizontal Ad.indd 1 11/13/12 10:40 PM
axens.indd 7 08/03/2013 11:49
tupras.indd 4 07/06/2013 17:58
corrmat.indd 5 12/09/2013 14:33
uop.indd 4 11/12/2013 20:13
IDTC & BBTC 2014
LISBON, 6-9 MAY
Targeting Operational Excellence, Energy Efficiency & Profitability
European Refinery Competitiveness & Challenges Ahead
Impact of Legislative & Demand-Related Changes on EU Refining Industry
How Galp Energia Integrates & Optimizes its Refining System
First-hand Experiences from MOL, Sarlux & Kuwait Petroleum
Focus on Refinery-Petrochemical Integration
Energy Efficiency, Sulphur Management & Emissions Reduction
Latest Technology & Catalyst Developments
Profitability Strategies & Optimisation Tools
The Industry Standard for Excellence
Overview of Global Heavy Crude Production; Oil Product Demand Outlook
Bottom of the Barrel Approaches & Solutions in Central European Refineries
Exclusive: Galp Energias Sines & Matosinhos Refineries Upgrading Project
First-hand Experiences from NIS, OMV & Petrochina
Comparison of Delayed Coking, SDA, FCC & Hydrocracking Solutions
Thermal & Gasification Processes
Advances in FCC Catalysts & Technologies
New Developments in Hydroprocessing
Sponsored by:
Speakers include:
Register now at www.europetro.com
E: conferences@europetro.com T: +44 20 7357 8394
AIR LIQUIDE
ART
AXENS
BASF
BECHTEL
CLG
CONCAWE
DAILY THERMETRICS
DUPONT
EURECAT
EXXONMOBIL R&E
FLOWSERVE
FOSTER WHEELER
GALP ENERGIA
GTC TECHNOLOGY
hte
HTI HEADWATERS
KBR
KUWAIT PETROLEUM
MOL
NIS
OMV R&M
PETROCHINA
PROMETHEUS
SARLUX / SARAS
SFA PACIFIC
SHELL GLOBAL SOLUTIONS
TECHNIP
UOP
WOOD MACKENZIE
Co-hosted by:
epc.indd 1 12/12/2013 16:29
Role of FCC catalyst in renery
protability
T
he FCC unit is the main
contributor (51-52%) to the
refnery gasoline pool of the
Lukoil Neftohim Burgas refnery in
Bulgaria, with the next largest
contributor being the reformer
(about 27%). The gasoline sensitiv-
ity (research octane number minus
motor octane number [RON
MON]) from the FCC unit and
reformer is about 12 and 11
numbers, respectively, resulting in
an overall MON shortage in the
refnery gasoline pool. Considering
MTBE has a MON of 94-97, alkylate
a MON of 93 (both of which are
produced from C
4
olefns) and FCC
gasoline a MON of about 82,
increasing the production of
C
4
olefns at the expense of gasoline
in the FCC unit will increase
the MON of the overall gasoline
pool.
A catalyst selection was
performed with the objective of
increasing MON. Grace proposed
the REsolution catalyst, which was
trialled in the FCC unit and subse-
quently led to an increase in
gasoline MON by 0.5 numbers.
This allowed the refnery to
increase the production of automo-
tive gasoline by 1.3%, and to
increase the share of premium
automotive gasoline from 68% to
73%. This resulted in an annual
six-fgure improvement in refnery
economics.
The Reid vapour pressure (RVP)
of gasoline from the FCC unit
correlates to the content of C
4

hydrocarbons in the gasoline, with
lower RVP values obtained by
producing more C
4
at the expense
of gasoline. Optimising the FCC
gasoline RVP during the winter
By switching to a rare earth-free FCC catalyst, a rener raised its output of
premium grade gasoline and increased its operating margin
IVAN CHAVDAROV, DICHO STRATIEV, IVELINA SHISHKOVA and ROSEN DINKOV Lukoil Neftohim Burgas JSC
VLADIMIR JEGOROV Grace Catalysts Technologies
PETKO PETKOV University Prof Dr Assen Zlatarov Burgas
season (RVP was reduced from 60
to 50 kPa) and increasing the C
4
olefns yield (leading to a higher
MON in alkylate production)
resulted in an additional improve-
ment in refnery economics by a
fve-fgure number (US $/y).
Lukoil Neftohim Burgass FCC unit
Commercial investigations were
carried out on the FCC unit, which
consists of a feed hydrotreater
section, the FCC reactor and regen-
erator, as well as the main
fractionator and vapour recovery
sections. The FCC reactor is
equipped with the modern UOP
VSS riser termination device and
the UOP Optimix feed injection
system. Typical feed for the FCC
unit is hydrotreated vacuum gas oil
distilled from Urals crude, of which
the physical and chemical proper-
ties are shown in Table 1.
www.eptq.com PTQ Q1 2014 87
Density at 20C, g/cm
3
0.895
Sulphur content, wt% 0.3
Total nitrogen, wt ppm 800
Content of Ni and V, wt ppm max 1.0
Conradson carbon, wt% max 0.1
Refractive Index at 20C 1.4994
ASTM D 1160 distillation
5 vol% 357
10 vol% 376
50 vol% 438
90 vol% 509
95 vol% 532
K-Factor 12.13
Molecular weight (Goosens) 369
Hydrogen content (ConocoPhillips) 12.5
Aromatic carbon content
(ConocoPhillips) 17.02
Gasoline precursors (LNB) 79.5
Physical and chemical properties of
hydrotreated vacuum gas oil feedstock
for Lukoil Neftohim Burgas FCC unit
Table 1
Optimisation of FCC catalyst
technology
In response to rapidly infating
rare earth metal prices in 2011,
Grace developed the REpLaCeR
series of rare earth-free FCC cata-
lysts. The REsolution catalyst
belongs to this series and is based
on the rare earth-free Z-21 zeolite.
Within each family of REsolution
catalysts, the ability to inde-
pendently adjust the activity and
selectivities of zeolite and matrix,
as well as the ratio of zeolite/
matrix activity, enables maximum
formulation fexibility. For low
metal applications, REsolution
catalysts are proven to match, and
even improve on, the performance
of traditional rare earth based cata-
lysts. To date, there have been
more than 15 successful applica-
tions of the REsolution catalyst in
the EMEA region.
Based on laboratory, pilot plant
and commercial data using
Catalyst X from Supplier 1, Lukoil
Neftohim Burgas switched to the
Grace REsolution catalyst in 2011,
with the objective of increasing the
FCC gasoline MON. As Table 2
shows, the REsolution catalyst
differs signifcantly from the previ-
ous catalyst, primarily in terms of
rare earth content, which is 16
times lower than in the catalyst
from Supplier 1. Even though
REsolution is a rare earth-free cata-
lyst, the e-cat still contains a small
amount of rare earth, as the inven-
tory was not 100% changed out.
Typical FCC unit operating condi-
tions from periods using Catalyst
X and the Grace catalyst are shown
in Table 3.
Table 4 shows the yield structure
grace.indd 1 12/12/2013 12:15
88 PTQ Q1 2014 www.eptq.com
higher hydrogen transfer activity,
resulting in a lower olefn content
of FCC gasoline, as well as the C
4

and C
3
fractions. This is due to the
Z-21 technology utilised in this
catalyst.
Figure 1 shows how the propor-
tion of different grades of
automotive gasoline within the
gasoline pool changed when
switching from Catalyst X to the
REsolution catalyst. The REsolution
catalyst increased the proportion of
premium gasoline A-95 from 68%
to 73% at the expense of the regular
grade.
Figure 2 highlights the structure
of the gasoline pool using Catalyst
X and the REsolution catalyst.
Catalyst X produced FCC gasoline
with a RON of 94 and a MON of
81.7, resulting in a gasoline pool
with a RON of 95 and a MON of
84.2, which complies with the
for FCC products obtained using
Catalyst X and the REsolution cata-
lyst. The slightly higher conversion
obtained with Catalyst X is due to
the higher catalyst to oil ratio
resulting from the higher outlet
temperature and lower feed
temperature used in this period.
The catalyst consumption rate was
the same for both catalysts, at 0.350
kg catalyst addition per ton of feed-
stock. Generally, both catalysts
displayed very similar product
selectivities.
Table 5 shows the FCC product
properties from Lukoil Neftohim
Burgas. Compared to Catalyst X,
the REsolution catalyst provided
FCC gasoline with an increased
MON of 0.5 numbers. This can be
explained with the increased
aromatics in gasoline. It is well
known that increased hydrogen
transfer activity results in the
increased conversion of olefns and
naphthenes to paraffns and
aromatics. Conventional under-
standing of typical FCC catalysts
would explain that a lower rare
earth content and smaller unit cell
size (as found in the REsolution
catalyst) should result in decreased
hydrogen transfer activity.
However, the data shown in
Table 5 demonstrate that the
REsolution catalyst displayed
E-cat properties Supplier 1 Grace
Catalyst X REsolution
Apparent bulk density (ABD), g/cm
3
0.85
Total surface area, m
2
/g 159 147
Matrix surface area, m
2
/g 55 51
Zeolite surface area, m
2
/g 104 96
Particle size distribution, %
0-20 m 0.1 2
0-40 m 1.8 5
0-80 m 46 45
0-149 m 94
Average particle size, mm 83 84
Unit cell size, 24.29 24.27
Al
2
O
3
, wt% 40.5 44.2
RE
2
O
3
, wt% 1.35 0.38
Na
2
O, wt% 1.15 0.28
Fe, wt% 0.61 0.42
V, wt ppm 206 208
Ni, wt ppm 25 63
Cu, wt ppm 25 8
Physical and chemical properties of e-cats used in the study
Table 2
Operating conditions Supplier 1 Grace
Catalyst X REsolution
Flow rates
Hydrotreated feed, t/h 220 229
Unhydrotreated feed, t/h 11 7
Recycle, t/h 2 5
Dispersion steam, kg/h 3000 3435
Riser steam, kg/h 2000 2000
Stripping steam, kg/h 5881 6200
Air ow rate, kNm
3
/h 134 118
Fresh catalyst ow, t/d 2 2
Catalyst to oil ratio 8.1 7.9
Temperatures, C
Combined feed temperature 317 331
Riser temperature 535 533
Regenerator dense phase temperature 668 668
Regenerator dilute phase temperature 678 677
Air temperature 179 184
Steam temperature 262 247
Operating conditions in Lukoil Neftohim Burgass FCC unit
Table 3
Product, wt% Supplier 1 Grace
Catalyst X REsolution
Dry gas (C
2
-) 3.8 3.5
Total C
3
s 8.0 8.2
Total C
4
s 14.0 13.8
Gasoline (C
5
-195C) 51.1 51.2
LCO (195-296C) 9.8 10.3
HCO (296-360C) & slurry 9.0 8.9
Coke 4.3 4.1
Conversion 81.2 80.8
Yield structure in Lukoil Neftohim
Burgass FCC unit
Table 4
Product, wt% Supplier 1 Grace
Catalyst X REsolution
Gasoline
MON 81.7 82.2
RON 94.0 94.0
RVP, kPa 55.1 57.9
FIA hydrocarbon
Composition, vol%
Saturates 37.8 41.7
Olens 39.2 32.5
Aromatics 23.0 25.8
C
3
s, wt%:
Propylene 80.8 78.8
Propane 17.9 20.9
C
4
s, wt%:
i-butane 33.5 39.5
i-butylene 17.3 16.0
n-butane 6.7 7.6
Total olens 56.3 53.3
Product properties in Lukoil Neftohim
Burgass FCC unit
Table 5
grace.indd 2 11/12/2013 12:48
O
ryx GTL slurry phase distillate
(SPD) raw wax catalyst slurry is
the system where Everlasting Valves are
installed. This unique valve was designed
nearly 100 years ago for blowing down steam
locomotive boiler solids. The U.S. Department
of Energy in development of the clean coal
technologies selected this valve as state of the art
valve for dry solids and slurries.
A major licensor of rening processes now
species the Everlasting rotating disc metal seated
valve for Fluid Catalyst Crackers handling fresh
catalyst and for hot CAT withdrawal. It is also listed in
the Best Practices of a world class petroleum rener
for continuous catalyst reforming. A preeminent
catalyst additive supplier has standardized on this
valve concept for all their chemical injection units.
Spent catalyst
recyclers use
them on their
vacuum truck load
out systems. The
flat rotating disc
renews the sealing
surfaces each time
the valve is cycled. No
other valve is similar, with vacuum through 10,000
psig (689b) and temperatures to 1500F (815C). See
more specications on the reverse side.
Whether its fresh, spent, dry or slurry use
the Everlasting Rotating Disc Valve and see your
CAT valve problems disappear. There are many
ways of skinning the CAT, but one best valve for
catalyst slurries.
Call us today at 1-908-769-0700 or visit our website at (www.everlastingvalveusa.com)
American Boiler Manufacturers Association
Theres more
than one
way to skin the
CAT!

VALVE COMPANY, INC.


Everlasting

Valve Manufacturers Since 1906


Boiler Blowoff Valves
Bulk Material Valves
Process Valves
everlasting.indd 1 09/12/2013 11:27
requirements of standard
EN228:2012 for RON but not MON.
Switching to the REsolution catalyst
90 PTQ Q1 2014 www.eptq.com
increased the MON of FCC gaso-
line from 81.7 to 82.2, resulting in a
nal MON of 84.5 for the gasoline
pool with the RON unchanged.
Since FCC gasoline accounts for the
majority of the gasoline pool (over
50 vol%), the FCC catalyst clearly
has a signicant impact on the
octane number of the total gasoline
pool.
Based on operational data from
the FCC unit, the economic ef-
ciency of using both catalysts from
different suppliers was determined
by applying a renery model,
which uses software for linear
programming by Honeywell
(RPMS). The results from RPMS
demonstrate that replacing Catalyst
X with the Grace catalyst results in
an increased prot of 4% for the
overall oil rening operations of
Lukoil Neftohim Burgas. This
demonstrates the important role
that the FCC catalyst plays in ren-
ing operations protability.
Optimisation of FCC gasoline Reid
vapour pressure
A signicant parameter of FCC
gasoline is Reid Vapor Pressure
(RVP),
1
which increases linearly
with an increasing amount of C
4
components. Lukoil Neftohim
Burgas performed a commercial
test operating with two different
RVP values of FCC gasoline. Table
6 shows how the yield of FCC
products changed when switching
from an RVP of 60 kPa to 50 kPa,
with the yield of C
4
s increasing by
0.6 wt% at the equal expense of
gasoline.
The effect of operating with two
different RVP values on the ren-
ery gasoline grades produced is
shown in Figure 3. The total
amount of automotive gasoline
produced by the renery decreased
by 0.18 wt%, while overall LPG
production increased by 3.7 wt%.
This resulted in an increase of
Premium A-95 production from
71% to 73% at the expense of the
Regular A-92 grade. Reducing the
RVP from 60 kPa to 50 kPa
improved renery protability by
about 2.5%.
Summary
It has been demonstrated that the
FCC catalyst is an excellent tool for
improving overall renery prota-
bility. Optimising FCC technology
by switching to the Grace
REsolution catalyst led to an
increase in the FCC gasoline MON
by 0.5 points, which allowed the
renery to produce 5% more
premium grade gasoline.
The overall renery margin was
improved further by optimisation
of the FCC gasoline RVP. A reduc-
tion in the RVP from 60 to 50 kPa
resulted in an increase in LPG
production by 3.7% at the expense
of gasoline pool reduction. The
Regular A-92
Gasoline A-93
Premium A-95
Super A-98
Catalyst X
2%
1%
29%
68%
REsolution
2%
25%
73%
Figure 1 Renery gasoline grades produced
by Lukoil Neftohim Burgas using Catalyst
X and the REsolution catalyst
FCC gasoline
Prime-G
unit
Prime-G
unit
Reformate
Alkylate
Straight-run
naphtha
Methyl tertiary
butyl ether
Bioethanol
Gasoline pool
RON 95.0
MON 84.2
Gasoline pool
RON 95.0
MON 84.5
1.9 vol%; RON 94.0; MON 81.7 7.0 vol%; RON 94.0; MON 82.2
51.7 vol%
RON 93.1
MON 81.7
53.6 vol%
RON 94.0
MON 81.7
26.7 vol%; RON 99.9; MON 89.0
11.1 vol%; RON 97.2; MON 93.2
4.2 vol%; RON 63.2; MON 60.2
4.4 vol%; RON 112.0; MON 93.0
0.01 vol%; RON 108.0; MON 93.0
51.7 vol%
RON 94.0
MON 82.2
44.7 vol%
RON 93.1
MON 82.2
27.7 vol%; RON 99.9; MON 89.0
10.9 vol%; RON 96.8; MON 93.1
5.1 vol%; RON 63.2; MON 60.2
4.2 vol%; RON 118.0; MON 94.0
0.1 vol%; RON 108.0; MON 93.0
Catalyst X REsolution
Figure 2 Structure of gasoline pool at Lukoil Neftohim Burgas using Catalyst X and the
REsolution catalyst
Product, wt% RVP
60 kPa 50 kPa
Dry gas 5.32 5.30
C
3
s 7.10 7.15
C
4
s 12.65 13.25
Gasoline 50.61 50.01
LCO 11.01 11.04
HCO 5.10 4.92
Slurry 4.00 4.10
Coke 4.21 4.23
Conversion 79.89 79.94
FCC yields at different gasoline
RVP values
Table 6
grace.indd 3 11/12/2013 12:48
www.eptq.com PTQ Q1 2014 91
Ivelina Shishkova is R&D Department
Manager with Lukoil Neftohim Burgas. She
holds a MS in organic chemistry engineering
and a PhD in petroleum rening from Soa
Chemical and Technological and Metallurgical
University, and has authored more than 20
technical papers.
Email: Shishkova.Ivelina.K@neftochim.bg
Rosen Dinkov is the Quality Manager in the
Process Engineering department of Lukoil
Neftohim Burgas. His research interests include
crude oil characterisation, bio/conventional
fuels blends characterisation and modelling of
renery distillation processes. He holds a MS
in organic chemistry engineering from Burgas
University and a PhD in the technology of
fossil and synthetic fuels from the University
of Chemical Technology and Metallurgy, Soa.
Email: Dinkov.Rosen.K@neftochim.bg
Vladimir Jegorov is the Sales Development
Manager for Grace in the CIS region. Prior to
joining Grace, he was an FCC process engineer
at the Mazheikiai renery in Lithuania.
Petko Petkov is a full professor and rector
of the Burgas University Assen Zlatarov. He
teaches in the social science department in
the eld of oil rening and lubricants, and has
authored more than 180 scientic papers and
ve books.
Email: PST_Petkov@abv.bg
higher FCC C
4
yield led to higher
production of alkylate, which
resulted in the production of 2%
more premium grade gasoline.
References
1 Watanbe K, Nagai K, Aratani N, Saka Y,
Chiyoda N, Mizutani H, Techniques for octane
enhancement in FCC gasoline, 20th Annual
Saudi-Japan Symposium, Dhahran, December
2010. 17. Montgomery J A, Guide to Fluid
Catalytic Cracking, Part 1, 1993.
Ivan Chavdarov is a Chemical Engineer in
the Process Engineering department of Lukoil
Neftohim Burgas, Bulgaria. His activities are
focused on guiding the operation of the units
of the FCC complex, troubleshooting support
and optimisation of the performance of the
FCC complex.
Email: Chavdarov.Ivan.S@neftochim.bg
Dicho Stratiev is Chief Process Engineer with
Lukoil Neftohim Burgas. He holds a MS in
organic chemistry engineering, and a PhD and
a DSc in petroleum rening from the Burgas
University Assen Zlatarov. He has authored
more than 130 papers.
Email: Stratiev.Dicho@neftochim.bg
Regular A-92
Premium A-95
Super A-98
RVP = 60kPa
1%
28%
71%
RVP = 50kPa
1%
26%
73%
Figure 3 Effect of changing the RVP on
renery gasoline grades produced during
the Resolution catalyst period
MINIDIS ADXpert
Grabner Instruments Messtechnik GmbH | A-1220 Vienna, Austria | Dr. Otto-Neurath-Gasse 1
Phone +43-1-282 16 27-0 | Fax +43-1-280 73 34 | info.grabner-instruments@ametek.at
QUICK, ACCURATE, PORTABLE
DISTILLATION
Portable, True Atmospheric D7344 Distillation
Excellent D86, ISO 3405, IP 123 Correlation
Fully Automated Measuring and Cleaning
6 mL of Sample
No Glassware, no Open Flame
Distillation of Gasoline, Jet Fuel, Diesel, Solvents,
Unknown Samples, Raffinates and Biodiesel
Specified for Use by US Pipelines
ADXpert half page 1.indd 1 06.12.2013 11:46:37
grace.indd 4 11/12/2013 12:49
blow...
Our heater doesnt
...the ames downward, against the natural buoyancy of the
ue gases. Our steam methane reformer design res upward,
reducing fan power consumption by 40 - 50%, to save
operating costs.
Choose the Foster Wheeler Terrace Wall Steam
Reformer for your hydrogen plant.
Learn more at www.fwc.com/terracewall
foster wheeler.indd 1 08/03/2013 10:49
Field experience with a Claus furnace
checker wall
D
epending on the licensor,
there are a number of differ-
ent architectures that may be
placed in the Claus furnace to infu-
ence performance. Each has its pros
and cons. Regardless of design,
mechanical stability has tradition-
ally been problematic.
Blasch was tasked to develop a
checker wall that would be easy to
install, mechanically stable under
all conditions, allow for design
with varying degrees of open area,
and could incorporate the use of an
integral manway, if desired.
Concerns at this point were strictly
mechanical; there were no requests
for mixing or fow management.
The goal was simply to put in a
checker wall that could survive a
campaign intact.
The result was the Blasch
HexWall checker wall. The blocks
are designed to be stacked dry and
are mechanically engaged through a
series of tabs and slots, so the wall
is quite stable, even when several
metres in diameter. No mortar is
used between the blocks, allowing
the assembly to accommodate the
thermally driven expansion and
contraction that comes with reaction
chamber operation. The blocks may
also be reused.
The initial checker wall blocks
were designed to be 9in wide, the
same as a row of bricks in the
hotface. As larger furnaces were
encountered, additional designs
utilising longer blocks were devel-
oped so as to retain a reasonable
aspect ratio of wall width to height.
The largest walls (>5m) utilise 18in
wide blocks.
The walls may either be erected
into a slot in the existing lining, or
A new design of Claus furnace checker wall delivers improved mechanical
performance and process ow characteristics
JEFFREY BOLEBRUCH Blasch Precision Ceramics
MOSSAED Y AL-AWWAD Saudi Aramco
MENG-HUNG CHEN CPC Corporation
installed against the hot face brick
with a course of brick on either side
to lock it in. The interlocking nature
of the blocks makes provisions for
expansion management straightfor-
ward and effective. The openings in
the blocks are round, so that any
portion of the block not directly
supported by other blocks is in the
form of an arch, and, as such, has
greater mechanical stability than a
fat span.
Over the next several years, as
the stability and durability of the
checker walls was proven out,
some operators, who previously
spent their time rebuilding walls,
began to question what they were
actually designed to do other
than remain standing for an entire
campaign.
After considerable consultation
with process licensors, engineering
companies, end users, and miscella-
neous stakeholders, it became
apparent that almost everyone
agreed that some combination of
time, temperature and turbulence
were the variables that played the
largest roles on the process constit-
uents. That makes sense, as
the three Ts of reaction kinetics,
and their effects on process
effciency, are well documented.
Blasch was determined to fnd a
way to use the HexWall checker
wall to infuence these parameters.
We selected turbulence, thinking
that if we could improve mixing in
the furnace, we could improve eff-
ciency, but soon learned that they
are all inter related, and we ended
up learning a lot more about resi-
dence time, residence time
distribution, and plug fow vs
stirred tank design along the way.
A vectoring hood was developed
that would ft into the outlet ends
of the existing blocks, and could be
oriented and secured in the feld in
such a way as to redirect, or
vector, the downstream fow
allowing each individual block in
the assembly to contribute to the
creation of the overall desired fow.
The vortex was initially selected
because it was felt that this confgu-
ration made the best use of the
furnace volume and created a fow
pattern that yielded a long path
length with a very tight residence
time distribution. This variation of
the HexWall checker wall was
christened the VectorWall.
There are now nearly two dozen
VectorWalls installed worldwide,
as well as approximately 100 of the
earlier HexWalls. Highlighted
below is a pair of VectorWall instal-
lations where it was possible to
contrast stability related details
from before and after the
installation
Field experience with the VectorWall
checker wall: mechanical
One of the larger diameter
VectorWalls done to date is
installed in Saudi Aramcos Berri
www.eptq.com PTQ Q1 2014 93
Time, temperature
and turbulence were
the variables that
played the largest
roles among the
process constituents
blasch.indd 1 13/12/2013 12:05
94 PTQ Q1 2014 www.eptq.com
and other effects of process upsets.
The exiting conventional checker
walls installed in Saudi Aramco
reaction furnaces either have a
cylindrical shape that resists high
stress due to the geometry, or a
matrix box shaped brick system
jointed with mortar that normally
failed due to thermal shock cycling,
mortar failure and vibration.
Modied Claus process
All Saudi Aramco SRUs use the
Modifed Claus process shown in
Figure 1, where acid gas is initially
treated in an acid gas scrubber and
knock out drum to remove water,
hydrocarbon liquids and diglycola-
mine. Then the acid gas is heated to
500F (260C) in a pre-heater and
introduced to a reaction furnace
where a third of the H
2
S is burned
in the presence of heated air to
produce SO
2
. The generated SO
2
reacts with the remaining H
2
S in
three in-series catalytic stages to
produce sulphur. Each stage
contains an auxiliary burner
(reheater), converter and a
condenser. Liquid sulphur is
collected from all condensers and
stored in a sulphur pit where it is
degassed and then exported. The
Gas Plant, in its SRU200. This
portion of the article is excerpted
from a paper written by Mossaed
Al-Awwad, Heat Transfer Engineer
for the Aramco Consulting Services
Department (CSD). Blasch grate-
fully acknowledges Saudi Aramco
for the following data.
Saudi Aramcos Berri Gas Plant
The current refractory system used
on the internal reaction furnace
lining in the Saudi Aramco sulphur
recovery unit (SRU) is a dual layer
of fre bricks as a hot face layer and
a castable layer as back-up with
stainless steel anchors to support
the castable refractory. Part of this
refractory systems internal features
is a checker wall that must be
incorporated into the refractory hot
face lining design. The existing
checker wall system that was used
for over 20 years frequently failed
due to improper design and instal-
lation techniques. This article
addresses the experience of a
checker wall refractory system, the
modifed system and its impact on
the reaction furnace operation.
1
Refractory lining description
The normal furnace operating
temperature at Saudi Aramco SRU
units is 1800-2000F (9801090C);
however, the hot face material must
be capable of withstanding temper-
atures of 2500-3000F (13701650C)
which can occur during start-up
fring of natural gas. All SRU reac-
tion furnace linings should utilise
at least 90% alumina hot face
material.
Brick linings at the hot face layer
are more durable than castable or
plastic ram materials. The initial
installation of brick may be slightly
more expensive and require a
highly skilled installer, but brick
normally provides a longer lining
life and requires less maintenance.
Part of this refractory systems
internal lining is a checker wall that
is made of a high density alumina
brick system that has good high
temperature strength to sustain the
harsh environment.
Checker wall system main function
The main function of the checker
wall is to have proper gas mixing
due to the residence time increasing
as a result of improved destruction
of H
2
S. In addition to that, the
checker wall works as a protector to
the tube sheet from radiant heat
Acid
gas
Air
blower
Blowdown
system
Reaction
furnace
Knockout
drum
Steam
drum
Steam
separator
Auxiliary
burner
T
Catalytic
converter
No. 1
condenser
Auxiliary
burner
Catalytic
converter
No. 2
condenser
Auxiliary
burner
Catalytic
converter
Thermal
oxidiser
Stack
No. 3
condenser
Sulphur
pit
No. 4
condenser
Liquid
sulphur
storage
MP/HP
steam
Clean gas to
atmosphere
LP
steam
Liquid sulphur
Tail gas
SO
2
Figure 1 Modied Claus process employed by Saudi Aramco
blasch.indd 2 11/12/2013 12:52
unreacted H
2
S and other sulphur
components, in the tail gas leaving
the last condenser, are oxidised to
SO
2
in a thermal oxidiser.
Modifed Claus reactions:
3 H
2
S + 3/2O
2
H
2
O + SO
2
+ 2H
2
S (1)
2 H
2
S + SO
2
2 H
2
O + 3/x S
x
(2)
Where x is = 1, 2, 3...
In a split fow confguration of
the Modifed Claus process, a third
of the acid gas passes through the
reaction furnace and waste heat
exchanger and the remaining frac-
tion is mixed with the waste heat
exchanger exit gases just before the
frst converter.
Hot process air comes from the
air preheater at 700F (370C) and
9 psig pressure to meet the acid gas
line at the burner nozzle. The air is
rationed to provide enough air to
burn a third of the total H
2
S in the
acid gas with no excess oxygen.
The exact amount of air needed to
maintain a H
2
S/SO
2
ratio of 2:1 is
controlled. The ratio of 2:1 H
2
S/SO
2

must be obtained in the combustion
chamber to progress all the way
through the train. If there is more
than 2:1 H
2
S/SO
2
in the reaction
furnace, the SO
2
content of the gas
will gradually lessen all the way
through the train leaving an excess
of H
2
S. If the ratio is less than 2:1
H
2
S/SO
2
in the reaction furnace,
the reverse happens with more SO
2

remaining than H
2
S in the tail gas.
On the initial start-up of the plant,
fuel gas is used to raise the tempera-
ture in the reaction furnace because
the burning characteristics of the
cold acid gas are poor. When the
main burner is lit on fuel gas, the
gas/air ratio is adjusted so that the
fuel gas is burned stoichiometrically.
Too much gas or substoichiometric
fuel gas fring will deposit carbon
on the catalyst and too much air or
above stoichiometric fuel gas fring
will cause a temperature runaway in
the converter bed if there is sulphur
present, due to the oxidation of the
sulphur.
Fuel gas addition is normally
used to warm up the train during
start-up prior to acid gas admission
and to clean up the converter beds
before a planned shutdown. The
www.eptq.com PTQ Q1 2014 95
hot process air and acid gas lines
are provided with 75 psig steam
injection. Steam is injected to main-
tain a stable fame and to keep the
fame temperature below 2400F
(1315C).
Investigative results
The exiting conventional checker
walls installed in Saudi Aramcos
reaction furnaces either have a
cylindrical shape that resists high
stress due to their geometry, or
matrix box shape bricks jointed
with mortar. Investigation revealed
that the existing checker wall has a
clear design defciency and should
be redesigned. The main cause of
the failure was due to thermal
cycling during the unit start-up and
shut-down. Liquid carry-over to the
combustion chamber loosens the
brick system and causes complete
failure. In addition to that, running
the unit above its design capacity
causes damage to the conventional
checker wall as it cannot sustain
high vibration due to over-design
fow rate.
Figures 2 and 3 show damage
caused to both checker wall
designs.
Additional investigation revealed
that, considering there were no
drastic changes in the internal
fring, gas composition and the
external ambient conditions, the
premature failure of the existing
refractory system was due to
several factors including the design,
bricks quality, installation, and so
on.
Investigative results and
recommendations
Saudi Aramco had bad experiences
with the current checker designs
due to: repeated failures, short
service life, prolonged repair shut-
down, and extensive repair costs.
It is recommended to use an
interlocking hexagonal ceramic
block system which is made of
ceramic (VectorWall, see Figure 4).
Because of its special design, the
modifed system which has hexago-
nal shape increases capacity and
lowers the operation and mainte-
nance costs of medium and large
diameter high temperature reaction
chambers such as those used in
Claus sulphur recovery units. This
is accomplished by altering the
reactant fow path creating signif-
cant mixing within the chamber
thereby optimising use of the avail-
able volume within the chamber.
Field experience with the
VectorWall checker wall: process
CPC Corporation - Taoyuan renery
CPC Corporation, Taiwan (CPC) is
Figure 2 Cylindrical shape conventional
checker wall
Figure 3 Matrix box shape checker wall
Figure 4 VectorWall at Berri Gas Plant
blasch.indd 3 11/12/2013 12:52
96 PTQ Q1 2014 www.eptq.com
Claus sulphur recovery process, the
F6301 reaction furnace is the heart
of No. 3 SRU at the Taoyuan ren-
ery. It was designed to process the
acid gas from an HDS (hydrodesul-
phurisation) plant in the renery.
The acid gas comes from two
sources, the amine regeneration
unit (ARU) clean sour gas stream
containing mostly hydrogen
sulphide and the sour water strip-
per (SWS) foul sour gas stream
with roughly equal amounts of
ammonia, hydrogen sulphide and
water vapour.
5
Figure 5 is a schematic of the
F6301 reaction furnace. It is a
refractory lined reactor with an
inner diameter of 2.1m and is about
9m long. A standard brick checker
wall formerly divided the furnace
into two zones. Acid gas from the
HDS was introduced both at the
burner and immediately down-
stream of the checker wall. The
hydrogen sulphide and ammonia
in the acid gas were converted to
elemental sulphur and SO
2
and N
2

respectively, in the furnace before
exiting the waste heat boiler. The
elemental sulphur condensed in the
cooling tubes was removed at the
liquid outlet. The leftover H
2
S and
the SO
2
formed in the furnace
reacted further downstream in a
catalytic reactor to form more
elemental sulphur.
The furnace was originally
equipped with a conventional 9in
brick checker wall. Prior to the
retrot, severe furnace vibration
and noise were observed, especially
at high furnace throughput. This
was believed to have been caused
by instability of the damaged
checker wall. Thus it was not possi-
ble to operate the furnace at high
temperatures, in particular in the
rst zone. In addition, deposits of
featuring a complex production
process and a complete range of
equipment. It has a capacity of
220 000 b/d of crude oil. The Dalin
renery, which split off from
Kaohsiung to become independent
in 1996, has four offshore mooring
buoys as well as both large and
small docks for the unloading and
loading of crude oil and petroleum
products. It has a topping capacity
of 300 000 b/d. The Taoyuan
Renery was established in 1976
and, following some debottleneck-
ing renovations and the addition of
a second distillation plant, currently
has a daily capacity of 200 000 bbl
of crude oil.
4
There are 15 sulphur recovery
units within CPC. As a part of the
the foremost energy enterprise in
the Republic of China with nearly
15 000 employees and a variety of
functional units around the island.
As a state run enterprise, the
company is responsible for the
development and supply of petro-
leum and natural gas, and is the
core of Taiwans petrochemical
industry.
2
CPC of Taiwan owns and oper-
ates three oil reneries, located at
Kaohsiung, Dalin and Taoyuan,
respectively. The three reneries
have a combined capacity of
720 000 b/d.
3
The Kaohsiung renery, which
has the longest history of the three,
is a large integrated oil rening and
petrochemical production facility
Zone (I) Zone (II)
H
2
S + NH
3

acid gas
Air
Fuel
gas
Boiler
inlet
Insulation
Start-up: open
Normal: close
H
2
S
acid gas
H
2
S
acid gas
Checker wall
Main burner
Figure 5 Schematic of Taoyuan renerys F6301 reaction furnace
1200
1400
1000
800
600
400
200
F
l
u
e

g
a
s

t
e
m
p
e
r
a
t
u
r
e
,

C
0
1300
1500
1100
900
700
500
300
100
2
3
/
1
1
/
2
0
1
1
1
2
/
1
/
2
0
1
2
2
/
3
/
2
0
1
2
2
1
/
4
/
2
0
1
2
1
0
/
6
/
2
0
1
2
3
0
/
7
/
2
0
1
2
1
8
/
9
/
2
0
1
2
1
7
/
1
1
/
2
0
1
2
2
7
/
1
2
/
2
0
1
2
1
5
/
2
/
2
0
1
3
6
/
4
/
2
0
1
3
2
6
/
5
/
2
0
1
3
1
5
/
7
/
2
0
1
3
3
/
9
/
2
0
1
3
Before revamp After revamp
Zone (II) flue gas
Zone (I) flue gas
Figure 6 A partially completed VectorWall
Figure 7 Flue gas temperatures before and after revamping
blasch.indd 4 11/12/2013 12:52
www.eptq.com PTQ Q1 2014 97
ammonium bisulphate/sulphate
were observed in the downstream
equipment due to insuffcient
destructions of the ammonia in the
furnace.
Retrot
In light of the above problems, CPC
decided to revamp the furnace in
May 2012. The revamp included
installing a new insulation refrac-
tory for the entire furnace and
replacing the checker wall with
a Blasch VectorWall. Figure 6
shows a picture of the VectorWall
during installation. It was built
against the insulating layer using
high alumina bricks as foundation
wedges backflled with tight
castable.
Results
Figure 7 compares the furnace
temperatures before and after the
revamp. It can be seen that after the
revamp, the Zone I temperature
increased to approximately 1400C.
However, the Zone II temperature
decreased about 200C to below
1000C.
The higher Zone I temperature
was benefcial not only to the
conversion of hydrogen sulphide
but also to the destruction of
ammonia in the acid gas. It is
believed that after the revamp more
ammonia was converted to N
2

in Zone I and less was carried
into Zone II for oxidation. Thus,
the Zone II temperature was
reduced.
Figure 8 compares the foul sour
gas processing rate. The amount of
the acid gas processed was
increased about 30% to 2000 m
3
/hr.
In addition, no more furnace vibra-
tions were observed.
Figure 9 compares the sulphur
yield. After the revamp, the aver-
age amount of sulphur produced
increased from approximately
105 t/d to about 120 t/d. This can
be attributed to the increased
amount of acid gas processed in the
furnace.
Cooperative agreement for
sulphuric acid applications
Following the frst presentation of
this technology, Blasch was
approached by DuPont Clean
2
3
/
1
1
/
2
0
1
1
1
2
/
1
/
2
0
1
2
2
/
3
/
2
0
1
2
2
1
/
4
/
2
0
1
2
1
0
/
6
/
2
0
1
2
3
0
/
7
/
2
0
1
2
1
8
/
9
/
2
0
1
2
1
7
/
1
1
/
2
0
1
2
2
7
/
1
2
/
2
0
1
2
1
5
/
2
/
2
0
1
3
6
/
4
/
2
0
1
3
2
6
/
5
/
2
0
1
3
1
5
/
7
/
2
0
1
3
3
/
9
/
2
0
1
3
Before revamp After revamp
1500
2500
2250
2000
1750
1250
1000
750
500
250
F
o
u
l

s
o
u
r

g
a
s

p
r
o
c
e
s
s
i
n
g

r
a
t
e
,

m
3
/
h
r
0
Figure 8 Comparison of acid gas processed
120
140
100
80
60
40
20
S
u
l
p
h
u
r

y
i
e
l
d
,

t
o
n
s
/
d
a
y
0
130
150
110
90
70
50
30
10
2
3
/
1
1
/
2
0
1
1
1
2
/
1
/
2
0
1
2
2
/
3
/
2
0
1
2
2
1
/
4
/
2
0
1
2
1
0
/
6
/
2
0
1
2
3
0
/
7
/
2
0
1
2
1
8
/
9
/
2
0
1
2
1
7
/
1
1
/
2
0
1
2
2
7
/
1
2
/
2
0
1
2
1
5
/
2
/
2
0
1
3
6
/
4
/
2
0
1
3
2
6
/
5
/
2
0
1
3
1
5
/
7
/
2
0
1
3
3
/
9
/
2
0
1
3
Before revamp After revamp
Figure 9 Comparison of sulphur yield

Figure 10 VectorWall after 18 months
service.
Technologies to explore application
of the VectorWall technology in
sulphuric acid applications with the
objective of developing a more eff-
cient, compact furnace design
utilising this technology.
DuPont was already using Blasch
HexWalls in two of its own plants.
The frst was a sulphur burning
plant in the Richmond, VA James
River plant. Two walls were
installed in 2007. The furnace shell
is 10ft in diameter. The walls were
both 9in thick.
The frst was constructed in a
choke ring confguration, and the
second as a standard checker wall
a confguration we would see
again in Claus plants later on. The
frst wall was located a mere 8ft 5in
from the burner, and the second
21ft 5in.
Figure 11 shows the walls during
an inspection in 2012, some fve
years after installation. They are
still in service.
Also, in 2007, a pair of HexWalls
blasch.indd 5 11/12/2013 12:52
98 PTQ Q1 2014 www.eptq.com
oped designs that are being
marketed.
Interestingly, the Borderland Plant
SAR was experiencing issues rela-
tive to NOx. A proprietary
VectorWall based solution was
modelled and installed in both
trains to work in conjunction with
the existing bafe walls in 2012, and
immediately resolved the concerns.
Furnace conguration and
VectorWall performance
So, what has been learned from all
this? Access to software and hard-
ware that can create a realistic,
accurate, multiphysic (heat, mass,
ow and kinetic) reactor simula-
tion, and someone that can run it
and analyse the results is hard to
come by.
Blasch has denitely seen some
trends from these installations, and
is continuing to use the relatively
simple cold ow modelling utilised
during the development of the
VectorWall, with a few tweaks.
The simplest furnace congura-
tion is the straight through
conguration (see Figure 13). All of
the reactants enter at one time and
in one place and react in similar
fashion. Maximum efciency in this
type of situation comes from
designing a ow conguration that
creates the tightest residence time
distribution, built around the opti-
mal residence time for that
particular process.
This conguration functions
effectively as a plug ow reactor,
and a consistent residence time
with a long spiral path length,
gives the most effective perfor-
mance, assuming the furnace in
question is the proper length for
the process in question.
The real world, unfortunately, is
not so tidy. Many times, in reality,
you have much more complicated
interactions, and the optimal solu-
tion is likely to be a compromise.
In Claus, often, one has to deal
with sour water stripper gas, which
contains ammonia, which if not
destroyed early and with great
enthusiasm, will create all sorts of
downstream issues.
In these cases, the creation of
multiple zones with unique charac-
teristics is desired. For example, in
Most recently, all three bafes in
the new Burnside, LA SAR furnace
came as Blasch HexWalls. This
furnace was the largest done at
DuPont to date, at just under 19.5ft.
The rst wall was 18in wide and
the remaining two were 13.5in.
At this point, Blasch began cast-
ing plugs for blocks it wished to
blind, rather than casting them
solid, which for very large walls
was a good idea from a structural
point of view.
At the time of this last installa-
tion, DuPont had not yet installed a
complete VectorWall in a furnace,
but after the performance of a
considerable amount of modelling
in Wilmington, DE, has now devel-
was installed in the El Paso, TX
Borderland plant (see Figure 12).
This was a spent acid regeneration
(SAR) plant, and each furnace was
designed with three brick bafes.
The rst (underow) walls in each
were replaced with the HexWalls
as this was before DuPont began to
explore the VectorWall and its
possibilities.
The mechanical stability of rst
(underow) wall in a decomposition
furnace is quite often problematic,
and these are generally designed to
be 18in thick when done in brick.
These furnaces were a modest 12ft
in diameter, and the HexWalls used
were only 9in wide and were
located 17ft from the burner.
Figure 11 Five years service at James River
Figure 12 New HexWall (left) and after ve years service (right) in the Borderland plant
Residence
time
0 0.5 1.0
Mixing
0% 100% 0%
Figure 13 Straight through furnace design
blasch.indd 6 11/12/2013 12:53
www.eptq.com PTQ Q1 2014 99
5 Meng-Hung Chen, Project Manager, CPC
Corporation, Taiwan.
6 Schmidt L D, The Engineering of Chemical
Reactions, 1998, New York: Oxford University
Press.
Jeffrey Bolebruch is Senior Market Manager,
Blasch Precision Ceramics, responsible for
sales to the chemical/petrochemical market.
He holds a bachelors degree in engineering
from the US Military Academy at West Point
and a MBA from Marist College, Poughkeepsie,
NY.
Mossaed Y Al-Awwad is a red equipment
specialist, Saudi Aramco Oil Company,
responsible for providing technical
consultations related to red equipment in
terms of design, technologies, and operation
support. He holds a bachelors degree
in mechanical engineering from KFUPM
University, Dhahran, Saudi Arabia.
Meng-Hung Chen is Manager of Technical
Service with the Rening & Manufacturing
Research Center of CPC Corporation. His
expertise is in the design and analysis of
process heaters and he holds a masters
degree in naval architecture engineering from
National Taiwan University.
Data from the eld continues to
point toward the directional
effectiveness of this simplied
modelling, and we continue to seek
cooperative efforts with licensors,
burner companies, and universities.
The goal is to learn enough about
how the VectorWall inuences the
combustion stage in Claus to help
plant operators make the very most
of the advantages the technology
can provide, as well as to identify
other combustion based processes
that can similarly benet from
VectorWall.
References
1 Mossaed Y Al-Awwad, Heater Engineer,
Saudi Aramco, Kingdom of Saudi Arabia.
2 www.cpc.com.tw/english/content/index.
asp?pno=24
3 www. cpc . com. t w/ engl i sh/ cont ent /
index01.asp?sno=715&pno=64
4 www. cpc . com. t w/ engl i sh/ cont ent /
index01.asp?sno=715&pno=64
the SWS case, higher temperatures
in zone 1 are desirable to promote
ammonia destruction, along with a
degree of back mixing to properly
blend it with the acid gas in that
zone. In zone 2, lower temperatures
are preferred, so as to discourage
NOx creation, and a return to a tight
residence time distribution built
around the optimal residence time.
In the case of a choke ring, as has
been shown, you are reecting back
some of the ow and creating a
degree of back mixing, but you are
still allowing a large ow to pass
straight through the centre of the
ring. Dumping the rest of the acid
gas into the ow at this point (even
closer to the tubesheet) creates a
real residence time issue. Many
people will also use a traditional
checker wall in front of the
tubesheet to protect it from
impingement by this strong jet.
The VectorWall, for reasons
described earlier, creates that back
mixing, but the ow through the
centre is then put into the swirling
motion that creates that strong
mixing and long path length, and
creates a plug ow conguration
and eliminates hot spots on the
tubesheet (see Figure 14).
The presence of the Vector Tiles
in the blocks also reects more
radiant heat back into zone 1, and
we can see the results in the CPC
data.
In terms of residence time, the
VectorWall and resultant rotation
put the ow into a plug ow
conguration more quickly than
the choke ring, which does seem to
allow a portion of the split ow
injected behind the choke ring out
of the furnace well before any
published minimum residence time
numbers are met (see Figure 15). It
is postulated that the dramatic
improvements seen in the eld are
due to this tighter residence time
distribution and the tendency of
the wall keep the amount of unre-
acted material from getting out of
the furnace to a minimum.
At this point Blasch is still
working with greatly simplied, cold
ow modelling, and not in any way
taking into account specic burner
congurations, the effects of temper-
ature, or any of the reaction kinetics.
160.0
148.8
137.1
125.7
114.3
102.9
91.4
80.0
68.6
57.1
45.7
34.3
22.9
11.4
0
Velocity
ft/s
Figure 14 Choke ring on top, VectorWall on bottom
Residence
time
7.1e
0.02
0
0.1
0.2
0.3
0.4
0.4
0.5
0.6
0.6
0.7
0.8
0.9
0.9
1.0
Figure 15 Choke ring on top, VectorWall on bottom
blasch.indd 7 11/12/2013 12:53
REFINING CHINA 2014
Rening China is the longest established international
conference in China dedicated to rening technology.
This longevity is the result of its recognised success
in bringing together international rening industry
companies and engineers and decision makers from
Chinese operating companies, a process greatly
assisted by our partnership with Sinopec.
The Sinopec alliance has, year after year, attracted
delegates not just from Sinopec itself but also from
Petrochina, CNOOC and independent producer
companies. Its success is also down to a strict
Zhao Jungui, Vice President and Secretary-General,
China Petroleum and Chemical Industry Federation
Opening Address
Cao Xianghong, President, Chemical Industry and
Engineering Society of China and Senior Consultant,
Sinopec Construction of safe and reliable, clean and
environmental protection in petrochemical enterprises
He Shengbao, Deputy director, Science and technology
management department, China National Petroleum
Corporation (CNPC) Green and clean production
technology of CNPC in rening enterprises
Hua Wei, Vice President and Chief Engineer, Sinopec
PRINCIPAL CHINESE SPEAKERS IN 2014
Regent Hotel, Beijing, 2-3 April
Join Us at the Top International Rening Conference in China
Sign up before 15/03/2014 and pay only 900
per delegate
Thereafter a late booking supplement of 200
per delegate applies
Group Discount
Enjoy 10% discount when you send 4 or more
delegates from the same company (applicable for
registrations received before 15/03/2014 only)
REGISTER EARLY TO ENJOY SAVINGS
Into its 8th year, REFINING CHINA was the rst rening technology conference in China and was hailed by many
past delegates as the best of its kind in China. The distinguished speakers that are conrmed thus far is an
afrmation that the conference will be yet another resounding success.
R
E
G
IS
T
E
R

N
O
W
Yanshan Petrochemical The green low-carbon path of
petroleum rening
Zhang Chunsheng, Vice President, Sinopec Jinling
Petrochemical Company Achieving clean and environment
friendly production by implementing green, low carbon
strategy
Huang Ziyou, Vice President, CNOOC Ningbo
Daxie Petrochemical Company The establishment and
perfection of rotating equipments monitoring and early
warning mechanism in petrochemical enterprises
Jin Chaokuan, Vice President, West Pacic Petrochemical
Company, Dalan, Safe use of petrochemical equipment
SPONSORS
For more information, visit www.reningchina.com
rc ad.indd 1 13/12/2013 12:51
Characterising and tracking contaminants
in opportunity crudes
C
rude oil characterisation and
tracking of specifc properties
and contaminants are critical
aspects in the design and operation
of the modern refnery. Given the
infux of new sources for oil
(synthetic crude from upgraders,
oil shale, etc), understanding the
effect of these crudes on the refn-
ery operation dictates how a facility
will incorporate these opportunity
crudes into their regular processing
mix.
The focus of this article will be to
share techniques and processes for
characterising and tracking critical
properties and contaminants,
particularly around the use of tools
and process simulations. The meth-
odologies will be examined both on
a theoretical and practical basis,
using real world applications and
case studies from design situations
and operating facility examples.
The availability of divergent
quality crude sources is having a
dramatic impact on the crude slate
of most refners. Those facilities
that are confgured for light/sweet
crude now have access to signif-
cant volumes of tight oil that match
nicely with the assets. Over the last
fve to 10 years, many refners have
gone through reconfgurations to be
able to handle heavy Canadian and
Venezuelan crudes.
As these assets come on-line, those
refners are beginning to process
these heavy crudes, but are also
mindful of the need to run tight or
conventional crudes to achieve
economic targets. Therefore, many
of these sites are faced with balanc-
ing their new asset capability
against changing availabilities of
both light and heavy crudes slates.
The specic properties of contaminants in opportunity crudes affect the selection
of crude slates for existing facilities and the design of revamps and new units
ROBERT OHMES
KBC Advanced Technologies
Ultimately, this dramatic shift in
crude availability is opening up the
crude slate that most refners
can process, and is infuencing a
facilitys need to track and under-
stand how different crudes will
and are impacting a facilitys
performance. KBC has been recom-
mending for some time that a
premium should increasingly be
placed on a refnerys optionality
in processing a wider, not
narrower, crude slate.
Critical stream properties
and contaminants
A full and rigorous discussion on
all the properties and contaminants
that impact a refnery is beyond the
scope of this article. However, a
high level summary is warranted to
provide context on the need for
refners to track and monitor these
properties and contaminants.
Tables 1 and 2 summarise key
properties and how they impact
refning operation.
www.eptq.com PTQ Q1 2014 101
Table 1
Property Comment Impact
API As the crude gravity Water/oil separation
approaches that of water,
diluent is needed to separate
out water from hydrocarbon
Sulphur High sulphur levels require H
2
Corrosion
and produce more H
2
S
Nitrogen High nitrogen levels require H
2
Corrosion
and produce more NH
3
PONA Drives gasoline/aromatic precursor Affects hydrogen addition
yields for clean products
Metals Ni/V/Fe High catalyst replacement cycle Catalyst deactivation
Metals Na/Ca/As/Ti Alkali metals require special Corrosion/catalyst
guard bed catalysts deactivation
Concarbon Requires carbon rejection Catalyst deactivation and
mechanism yields
Asphaltenes Increases potential for fouling that Fouling
requires shutdowns to resolve
Naphthenic acids High levels cause corrosion Corrosion/fouling
Compatibility Certain crude and blends are Affects allowable crude
incompatible blend and fouling
Chlorides Typically associated with alkali Corrosion
metals
Methanol Helps prevent hydrate formation Water/oil separation,
catalyst deactivation
Viscosity If too high to pump, requires High transportation costs
diluents or redesign
Key stream properties and impacts
Table 1
kbc.indd 1 11/12/2013 13:01
102 PTQ Q1 2014 www.eptq.com
Will resid processing capacity limit
refning crude rate and slate? Will
incompatibility impact the
onstream factor for the coker and
ebullated bed hydrocracker?
As indicated, these properties
essentially drive how a given refn-
ery will perform and how the
assets generate sustainable and reli-
able proftability. Therefore,
understanding these properties and
contaminants, not only in crudes
but also in intermediate streams
and fnal products, is critical to
selecting a crude slate and prepar-
ing and executing an operating
plan.
Property data sources
Several data sources exist to under-
stand and estimate the properties
of a given crude or refning stream.
Laboratory analysis
The best way to understand a
streams properties is to measure
them in a laboratory. While this
option may seem intuitively obvi-
ous, accurate measurement of a
given property is more diffcult
than it may initially seem.
First, multiple methods exist for
measuring a specifc property for a
given stream type. Most refners
utilise standard methods according
to internationally accepted measure-
ment techniques, such as those from
the American Society for Testing
and Materials (ASTM). However,
the refner must decide on the
method that is utilised in their facil-
ity or use methods that are dictated
by product sales requirements.
In addition to multiple methods,
any given method will have a level
of repeatability and reproducibility.
Table 3 provides some repeatability
and reproducibility values for vari-
ous properties and testing methods.
Therefore, when the laboratory
reports a given value for a prop-
erty, it really should be used within
the context of the values conf-
dence level rather than as an
absolute value. Also, the ability of
the laboratory to accurately meas-
ure a property is strongly
infuenced by the following factors:
Availability and accuracy of the
measurement equipment
Training of laboratory staff and
The following are typical exam-
ples of how these properties could
impact refning operations:
Crude/vacuum unit
Can the crude unit properly recover
and separate the amount of naphtha
and distillate in the crude? Does the
unit have suffcient metallurgy to
avoid naphthenic acid corrosion?
Are the crudes themselves compati-
ble with one another when mixed,
or will asphaltene precipitation
occur? Can the desalter suffciently
remove basic sediment and water
(BS&W) and salts to avoid down-
stream corrosion and plugging?
How are the crudes impacting
gasoil recovery and contaminant
levels in gasoil stream?
Light ends units
Are the unit off-gas compressors
suffciently sized to process the
native light ends in the crude? Does
the crude unit lose cutpoint due to
higher pressure operation? Can the
saturated gas plant maintain C
3+

recovery levels?
Naphtha units (hydrotreater
and reformer)
Can the hydrotreater and reformer
process all of the naphtha within
the crude? How does the N+2A
level impact reformer severity
requirements? Can hydrotreated
naphtha be routed to gasoline
blending to help manage pool
octane? What is the proper disposi-
tion for C
5
s and C
6
s when balancing
octane and RVP?
Distillate units (hydrotreaters)
Are trace unconventional metals
impacting catalyst deactivation?
Will poor cold properties from tight
oils adversely impact jet freeze and
diesel cloud and pour points?
Gasoil units (gasoil hydrotreaters,
uid catalytic crackers,
hydrocrackers)
How will the gasoil sulphur and
nitrogen impact these units? Can
the downstream amine, sour water,
and sulphur units manage the H
2
S
and NH
3
that are produced from
these units? How will metals,
Conradson carbon (concarbon), and
asphaltenes impact catalyst activity
and conversion? What are the yield
and capacity impacts of a given
gasoil stream? Is suffcient hydro-
gen available to meet contaminant
removal and/or conversion targets?
Resid units (coker, ebullated bed
hydrocracker, xed bed resid
hydrotreater)
How will the resid concarbon level
impact coker yields and utilisation?
Conventional metals Source Concerns
Nickel, vanadium In asphaltenes Natural organics Catalyst poison
Iron Iron oxides Corrosion products, Catalyst poison,
sulphides foulant
Silicon Polydimethlysiloxane Defoamer Catalyst poison
Arsenic As organics Natural organics Catalyst poison
Exotic metals Source Concerns
Phosphorous Pigging gel, Catalyst poison
acidising gel
Titanium With bitumen solids Naturally occurring Catalyst poison
Alkali metals Ca, Mg, Na Naturally occurring Catalyst poison, foulant

Calcium stearate Flow improver Crude fouling
Mercury Naturally occurring Catalyst poison
Selenium Naturally occurring Environmental
Crude metals and sources
Table 2
kbc.indd 2 11/12/2013 13:01
roMax process slmulauon soware by 8ryan 8esearch & Lnglneerlng, lnc.
oqloeetloq 5olouoos fot tbe Oll, Cos, kefoloq & cbemlcol loJosttles.
sales[bre.com www.bre.com 979 776-3220 uS 800 776-3220
Process
Comparing Physical
Solvents for Acid Gas
Removal
Insight:
Physical solvents such as DEPG, NMP, Methanol, and Propylene Carbonate
are often used to treat sour gas. These physical solvents differ from chemical
solvents such as ethanolamines and hot potassium carbonate in a number of
ways. The regeneration of chemical solvents is achieved by the application of
heat whereas physical solvents can often be stripped of impurities by simply
reducing the pressure. Physical solvents tend to be favored over chemical
solvents when the concentration of acid gases or other impurities is very high
and the operating pressure is high. Unlike chemical solvents, physical solvents
are non-corrosive, requiring only carbon steel construction. A physical solvents
capac|ly lor aosoro|rg ac|d gases |rcreases s|gr|lcarl|y as lre lerperalure
decreases, resulting in reduced circulation rate and associated operating costs.
Typical Physical Solvent Process
DEPG (Dimethyl Ether of Polyethylene Glycol)
DEPG is a mixture of dimethyl ethers of polyethylene glycol.
Solvents containing DEPG are marketed by several companies including
Coastal Chemical Company (as Coastal AGR

), Dow (Selexol), and UOP


(Selexol). DEPG can be used for selective H
2
3 rerova| ard car oe corlgured
to yield both a rich H
2
S feed to the Claus unit as well as bulk CO
2
removal.
DEPG is suitable for operation at temperatures up to 347F (175C). The
minimum operating temperature is usually 0F (-18C).
MeOH (Methanol)
The most common Methanol processes for acid gas removal are
the Rectisol process (by Lurgi AG) and Ifpexol

process (by Prosernat). The


ra|r app||cal|or lor lre Recl|so| process |s pur|lcal|or ol syrlres|s gases
der|ved lror lre gas|lcal|or ol reavy o|| ard coa| ralrer lrar ralura| gas
treating applications. The two-stage Ifpexol process can be used for natural gas
applications. Methanol has a relatively high vapor pressure at normal process
conditions, so deep refrigeration or special recovery methods are required to
prevent high solvent losses. The process usually operates between -40F and
-80F (-40C and -62C).
NMP (N-Methyl-2-Pyrrolidone)
The Purisol Process uses NMP

and is marketed by Lurgi AG.


Tre loW screres used lor lr|s so|verl are s|r||ar lo lrose lor 0EP0. Tre
process can be operated either at ambient temperature or with refrigeration
down to about 5F (-15C). The Purisol process is particularly well suited to the
pur|lcal|or ol r|gr-pressure, r|gr C0
2
synthesis gas for gas turbine integrated
gas|lcal|or coro|red cyc|e (l0CC) syslers oecause ol lre r|gr se|ecl|v|ly lor
H
2
S.
PC (Propylene Carbonate)
The Fluor Solvent process uses JEFFSOL

PC and is by Fluor
Daniel, Inc. The light hydrocarbons in natural gas and hydrogen in synthesis
gas are less soluble in PC than in the other solvents. PC cannot be used for
selective H
2
S treating because it is unstable at the high temperature required
to completely strip H
2
S from the rich solvent. The FLUOR Solvent process is
generally limited to treating feed gases containing less than 20 ppmv; however,
|rproved slr|pp|rg W|lr red|ur pressure lasr gas |r a vacuur slr|pper a||oWs
treatment to 4 ppmv for gases containing up to 200 ppmv H
2
S. The operating
temperature for PC is limited to a minimum of 0F (-18C) and a maximum of
149F (65C).
Gas Solubilities in Physical Solvents
All of these physical solvents are more selective for acid gas than
for the main constituent of the gas. Relative solubilities of some selected gases
in solvents relative to carbon dioxide are presented in the following table.
The solubility of hydrocarbons in physical solvents increases with
the molecular weight of the hydrocarbon. Since heavy hydrocarbons tend
to accumulate in the solvent, physical solvent processes are generally not
economical for the treatment of hydrocarbon streams that contain a substantial
amount of pentane-plus unless a stripping column with a reboiler is used.
Gas Component
DEPG
at 25C
PC
at 25C
NMP
at 25C
MeOH
at
-25C
H
2
0.013 0.0078 0.0064 0.0054
Methane 0.066 0.038 0.072 0.051
Ethane 0.42 0.17 0.38 0.42
CO
2
1.0 1.0 1.0 1.0
Propane 1.01 0.51 1.07 2.35
n-Butane 2.37 1.75 3.48 -
COS 2.30 1.88 2.72 3.92
H
2
S 8.82 3.29 10.2 7.06
n-Hexane 11.0 13.5 42.7 -
Methyl Mercaptan 22.4 27.2 34.0 -
Choosing the Best Alternative
A detailed analysis must be performed to determine the most economical
choice of solvent based on the product requirements. Feed gas composition, minor
components present, and limitations of the individual physical solvent processes are
all important factors in the selection process. Engineers can easily investigate the
ava||ao|e a|lerral|ves us|rg a ver|led process s|ru|alor sucr as ProVax

which has
oeer ver|led W|lr p|arl operal|rg dala.
For additional information about this topic, view the technical
article A Comparison of Physical Solvents for Acid Gas Removal at
http://www.bre.com/tabid/147/Default.aspx. For more information about ProMax,
contact Bryan Research & Engineering or visit www.bre.com.
bre.indd 1 7/3/12 10:45:54
adherence to a given procedure
Sampling, preparation, and
handling techniques of a given
stream sample.
Examples abound of refners that
were making critical decisions
based on inaccurate measurements
of properties due to problems in
these areas.
Repeatability
The difference between successive
test results obtained by the same
operator with the same apparatus
under constant operating condi-
tions on identical test materials
would, in the long run, in the
normal and correct operation of the
test method, exceed the value only
in one case in 20.
Reproducibility
The difference between two single
and independent results obtained
by different operators working in
different laboratories on identical
test material would, in the long
run, exceed the value only in one
case in 20.
Finally, measuring all streams
and properties is not practical nor
cost effective. In addition, being
able to measure a streams proper-
ties means that a given crude or
stream has already entered the
facility and been processed.
Therefore, the facility will be run
on a reactive rather than a predic-
tive basis. Hence, refners need
other methods to understand a
given crude or streams properties
before the stream is processed, let
alone purchased.
Paper assays
A crude assay provides a summary
of its properties for the various
104 PTQ Q1 2014 www.eptq.com
boiling point cuts that exist in the
crude, in addition to the quantity of
each boiling point cut. The assay
information is summarised in a
table or paper form.
These assays are generated by
specialty laboratories that take
samples of whole crude streams
and complete a series of separa-
tions and tests to measure the
amount of each cut and properties
of each respective cut. The crudes
are labelled by the source of the
crude, such as the region or
production feld. Assay databases
can be purchased, generated from
in-house data, or pulled from open
literature sources. The same testing
methods and challenges discussed
previously are utilised for and
impact generation of crude assays.
Additional issues can impact the
accuracy of a crude paper assay.
The properties and cutpoint content
of a given crude are strongly infu-
enced by the age of the production
feld. In general, as felds age the
amount of heavy material increases,
thereby concentrating contaminant
levels.
Content of trace metals and
contaminants is infuenced by the
chemicals and technologies used to
recover, separate, and transport the
crude.
Properties and cutpoint content
are impacted by the various wells
that are mixed together to produce
a given crude; or, in the case of
synthetic crudes, how the upgrader
facilities are performing and how
much raw or partially processed
bitumen is blended with processed
bitumen.
Since many tight oil sources are
being delivered by rail and truck
from various wells within a given
feld, the properties and boiling
point content of various crude ship-
ments can vary substantially.
Crude assays can be powerful
information in deciding a crude
slate for a given facility. The infor-
mation can be used by tools (which
will be discussed later) to estimate
the blended crude properties that
could be run in a given facility.
However, refners and crude trad-
ers can be lulled into a false sense
of security and knowledge as they
assume that a given crude will
have a fxed set of properties and
yields continuously.
In reality, the assay is a snap-
shot in time. Validations of crude
assays and backcasting of actual
versus predicted crude properties
and volumes are critical compo-
nents of effective crude selection
and management processes.
Plant data
Most refners regularly sample
plant streams and complete labora-
tory analyses, either as needed for
monitoring or dictating unit perfor-
mance or as required for selling
intermediate or fnal products. Due
to the infux of new crude sources,
a few specifc areas should be
reviewed and assessed within the
facility.
First, validation of crude assays is
becoming more critical. As an
example, refners processing tight
oils are observing swings in crude
API gravity of fve to 10 numbers,
or more, for a given shipment or
batch. Since API gravity is an indi-
cator of the amount of light and
heavy material, these levels of
swings can dramatically impact
how the refnery manages and
responds to these quality changes.
Therefore, the following recom-
mendations and considerations are
offered to improve crude validation
and monitoring:
Include crude assay validation as
part of the production planning
departments monthly backcasting
Compare properties of composite
crude blend against properties
calculated from crude assays
Ensure crude receipt tracking can
properly handle shipments via
truck or rail car
Validate that crude custody
Property Method Repeatability Reproducibility
Sulphur -~2 wt% ASTM D4294 0.06 0.25
Sulphur -~500 ppmw ASTM D4294 63 128
Concarbon -~ 2 wt% ASTM D4530 0.12 0.39
Concarbon -~ 0.5 wt% ASTM D4530 0.05 0.15
Nitrogen -~3000 ppmw ASTM D5762 261 798
Nitrogen -~30 ppmw ASTM D4629 1.1 4.7
Vanadium -~100 ppmw ASTM 5708 4 19
Vanadium -~100 ppmw ASTM 5863 11 21
Repeatability and reproducibility examples
Table 3
kbc.indd 3 11/12/2013 13:01
worldwide expertise on demand
Coldest. Hottest. Deepest. Furthest.
Thats where Zeeco people and products work today.
Our dedicated team of worldwide experts understand the
complexities of combustion across the globe. We have the
real-world experience and on-the-ground resources to
design and deliver proven combustion solutions for the global
refinery and petrochemical industries. Trust ZEECO

burners,
flares and thermal oxidisers. Trust our Rapid Response Team
for aftermarket sales, service, and engineered solutions.
Zeeco Global experience. Local expertise.
Zeeco, Inc. 22151 E. 91st St., Broken Arrow, OK 74014 USA
+1-918-258-8551 sales@zeeco.com zeeco.com
Zeeco, Inc. 2014
burners flares thermal oxidisers aftermarket parts, service & solutions
Zeeco_PTQ_Dec.indd 1 12/13/13 8:42 AM
zeeco .indd 1 13/12/2013 15:44
106 PTQ Q1 2014 www.eptq.com
crudes and conventional crude
sources. Laboratory and on-line
analysers can provide guidance on
how various mixtures will respond.
KBC Advanced Technologies, via
the acquisition of Infochem
Computer Services Ltd, is expand-
ing the usage of Infochems
Multifash technology in predicting
phase equilibria, including hydrates,
waxes, and asphaltenes. This tech-
nology is currently used extensively
in the upstream production side for
fow assurance modelling The
intention is to apply this same tech-
nology to downstream operations to
supplement and enhance KBCs
existing capabilities in predicting
incompatibility.
Methods and tools for tracking
contaminants
Refners need ways to predict
stream qualities throughout a facil-
ity to make decisions on crude
purchasing and operating strategy.
Ultimately, refners need to be able
to convert the raw data from crude
assays and plant stream data into
useful information. The following
will discuss the options available,
as well as the advantages and
disadvantages of each one.
Crude assay management system
Most refners rely on some crude
assay management system (CAMS)
to convert paper assay data into
usable information for making deci-
sions. Essentially, these software
programs allow the user to mix
multiple crudes together and gener-
ate properties and rates for streams
that would be produced in a crude
and vacuum unit.
The tools are relatively easy to
use and can quickly generate
stream properties that can be used
directly for crude selection or
within other simulation tools. These
systems can read in from crude
assay databases or generate assays
from measured or paper data. In
addition, the tools can rapidly
generate properties for new crudes
or for cutpoint adjustments.
The primary disadvantages of
these tools are that they do not
utilise rigorous fractionation and
thermodynamic packages to gener-
ate the cut properties. Instead,
The bottom line is that refners
need to be more proactive in stream
quality testing and monitoring to
manage new crudes and a diverse
crude slate.
Advanced techniques
The technology and science around
property measurement continues to
evolve. Though many techniques
are new and novel, the bulk of the
testing methods and the properties
that refners monitor are mature
and have not changed substantially
in decades.
However, there are advances in
technology that can improve stream
monitoring. On-line analysers are
being installed in production and
gathering stations, as well as refn-
eries, to monitor and track key
proprieties. Of key interest currently
are measurements for vapour pres-
sure and H
2
S for environmental
compliance and personnel protec-
tion purposes, particularly given
the usage of trucks and rail cars for
transporting tight oils.
Production entities, pipelines,
and refners are jointly developing
standard specifcations for crude oil
qualities. As an example, a specif-
cation standard for West Texas
Intermediate crude has been devel-
oped by the partner companies in
the Crude Oil Quality Association
(COQA) and is now under review
and consideration by NYMEX for
inclusion in setting qualities for
trading that crude.
Advances continue in the devel-
opment of testing methods and
on-line analysers for conventional
and trace metals, such as chlorine,
silicon, phosphorus, and others.
Groups such as COQA and
Canadian Crude Oil Technology
Association (CCOTA) are promot-
ing fundamental research projects
on crude oil quality and testing
methods.
Websites, such as www.crude-
monitor.ca, can provide data on
property variability of select crude
types and agglomerated crude
sources.
Crude compatibility has drawn
renewed interest in the last fve
years, given the desire by many
refners to process highly paraffnic
tight oils with heavy Canadian
transfer meters are properly and
regularly calibrated and account for
changes in crude specifc gravity
Calculate crude blend on a daily
basis, based on crude receipts, tank
gauges, oil movements, and tank
balances
Install proper crude sampling
stations, per ISO 3171 and ASTM
D4057 and D4177
Utilise the proper water in crude
and BS&W testing methods
Use composite crude sampling
stations at the crude unit to provide
a representative sample of
processed crude
Develop internal specifcations
for crude quality and understand
the incremental economic and
performance impact for relaxing
those specifcations
Essentially, treat crude quality
impacts as one would any other
standard operating limit (SOL)
Consider using advanced control
systems to improve how crude and
vacuum units respond to quality
changes.
Second, for streams beyond
crude, the following suggestions
are offered to improve stream
property tracking and
management:
Develop a list of approved prop-
erty testing methods for the
laboratory, and ensure the labora-
tory staff are properly trained and
have the proper equipment
Consider periodic round-robin or
independent laboratory testing to
validate and certify that the inter-
nal laboratory is providing accurate
and repeatable data
Install on-line analysers to
actively and continuously monitor
critical stream properties, thereby
allowing for better responses in
operation to property changes
Prepare reference and training
materials for operations and engi-
neering that summarise a given
streams property, acceptable levels,
monitoring methods, and corrective
actions, thereby improving how the
staff respond to quality changes
Essentially, treat stream quality
impacts as one would any other
SOL
Include stream quality changes as
part of operations refresher train-
ing and situational/what-if drills.
kbc.indd 4 11/12/2013 13:01
www.eptq.com PTQ Q1 2014 107
Use of multiple components and
narrow cuts allows the process simu-
lation to represent and track the
distribution of properties and
contaminants more rigorously.
Therefore as new feedstocks are
analysed or as fractionation cutpoints
are adjusted, the process simulator
will seamlessly account for the prop-
erty shifts and do so more
thoroughly than any of the other
packages discussed in this article.
These same tools can be used to
model an individual process unit as
well. Through these individual unit
models, more detailed impacts on
unit performance can be analysed,
as the user can include very
detailed models of individual
process equipment. Also, these
same individual unit models can be
used to enhance the LP representa-
tion. The reactor model for an
individual process unit can be used
to run a series of cases, where indi-
vidual feed property and unit
operating conditions are pertur-
bated and the resulting predictions
used to update LP model shift
vectors. KBCs Petro-SIM includes
the LPU (LP Utility) tool, which
automatically completes these cases
and creates an updated LP
sub-model, which allows for rapid
updating and upgrading of a
plants LP model.
The primary disadvantages of a
process simulator include: solution
time for a facility wide simulation
is longer than a LP model, but
computing speed and technology
improvements are narrowing this
gap; and the model is typically run
on a case study basis, so the user
may need to run multiple cases to
reach an optimum solution.
However, this case study
approach does provide additional
insight into how the refnery
responds to a given change, which
an LP cannot easily replicate. The
process simulation technology is
proprietary, which requires licens-
ing from a software vendor,
although many refners already
license unit simulators and fow-
sheets for traditional uses and
require only incremental licences to
develop a whole plant simulation.
Table 4 summarises the advantages
and disadvantages of each tool.
capability to model the entire facil-
ity from crude through to fnished
products. Each process unit is
represented by a linear sub-model,
which takes feed fows and key
properties and generates product
rates and key properties. Each
sub-model is a linear representa-
tion. Complex linear programs
will use a base+delta approach,
thereby allowing the model to
more accurately predict unit
performance.
One key advantage of an LP is
that it solves quickly and allows the
user to evaluate tens and hundreds
of crudes, crude slates, operating
strategies and cases in a matter of
minutes. In addition, the LP is an
optimiser, such that the resulting
solution is the projected optimum
for a given set of constraints.
The challenge with most LPs is
that they are linear representations
of a non-linear process. Use of the
base+delta methodology helps miti-
gate this limitation to some degree,
but the LP is dependent on good
delta shift vectors which are gener-
ated by other tools and technology
licensors.
LPs can be useful tools for evalu-
ating different crudes and the
impact on refnery performance and
economics, but certain limitations
will impact the number of proper-
ties a refner can predictively track.
Process simulations
Over the last 10 years, process
simulation packages, such as KBCs
Petro-SIM, have evolved and
improved substantially in model-
ling refneries. Process simulation
packages not only include detailed
thermodynamic packages and the
ability to simulate pumps, compres-
sors, exchangers, separators, and
fractionators, but also include the
capability to model kinetic reactors,
track and predict stream properties
from crude through fnished prod-
uct, and simulate an entire facility
on a rigorous, non-linear basis.
Hence these models can help a
refner understand how the entire
facility will respond to given feed-
stock quality change and determine
how the distribution of contami-
nants will change throughout the
facility.
crude and vacuum fractionation
sectional effciencies are assumed
based on typical unit performance
(or anticipated generic perfor-
mance) and are essentially fxed.
Though reasonably accurate for
most users and tasks, this simpli-
fed distillation method may not
fully capture how a given crude
mixture and rate will respond in
the crude or vacuum unit. In addi-
tion, the CAMS tools cannot
capture limitations on heat recov-
ery, furnace performance, and
pumps limitations. Finally, the
CAMS tools do not simulate any
process units or product blending
downstream of the crude unit;
these tasks are left to other
methods.
Spreadsheet correlations
Some refners rely on Excel spread-
sheets to estimate stream properties
and the impact on unit perfor-
mance and yields. Often, these
property calculations are based on
open literature equations or simpli-
fed blending approaches.
Although these spreadsheets are
easy to use and can provide quick
answers, the tools are often
dependent on the user to maintain
integrity and accuracy. If the
spreadsheet is complex and lacks
documentation, other users may
improperly utilise the tool and
information, thereby resulting in
faulty recommendations and opera-
tional changes. In addition, as with
the CAMS tool, these spreadsheet
tools typically only cover a few
properties or potentially one units
performance, and do not have the
effective ability to track properties
throughout a facility.
Therefore, if a spreadsheet for a
downstream unit exists, it will
require input on stream properties
for the feeds, which may depend
on not only the crude and vacuum
unit but also other process units.
Hence the user may have to esti-
mate critical input properties to
complete the calculations.
Linear programs
Most refners rely on a linear
program (LP) to make crude, feed-
stock, and high level operating
strategy decisions. The LP has the
kbc.indd 5 12/12/2013 12:17
108 PTQ Q1 2014 www.eptq.com
There are several considerations
and options to manage crude
sulphur content and distribution.
Accurate assay information is crit-
ical to understand the potential
impact of processing a new crude.
Regular monitoring of not only
crude and product sulphur content
but also intermediate process stream
sulphur levels is an integral part of
refnery performance monitoring.
For those facilities processing
purchased intermediates (gasoil,
diesel, naphtha), consider periodic
sulphur distribution testing to prop-
erly understand how sulphur is
dispersed along the boiling range.
The purchased stream may be a
blend of multiple and divergent
streams such that monitoring a
total sulphur value may not give a
complete picture of how the stream
will impact the plant.
Metals
Metals such as nickel (Ni) and
vanadium (V) will have a signif-
cant impact on catalytic units. In
particular, FCC and gasoil hydro-
processing (gasoil HDS) units are of
particular interest. Often, the
cutpoint of gasoil within a vacuum
unit will be dictated by the amount
of metals contained. Normally, the
limit of Ni+V will be about 2 ppm
total for a gasoil HDS unit, though
the range can be higher if more
demetallisation catalyst is loaded.
FCC units can manage higher levels
of metals, as long as the catalyst
addition rates and formulation are
adjusted accordingly. Given that
most US FCC units process hydro-
treated feeds, the gasoil HDS unit
typically limits gasoil cutpoint.
Figures 2 and 3 illustrate the
metals distribution of back-end
gasoil distillation range metals for
several conventional and uncon-
ventional North American crudes.
Typically, vanadium content is
higher than nickel content, and
vanadium content ascends more
rapidly than nickel content. An
unconventional crude does not
automatically equate to a high
metals crude. Relative trends
between Canadian bitumen and
Maya crude illustrate this point.
Use of diluents with bitumen can
impact the amount and distribution
conventional and unconventional
North American crudes.
As expected, heavier crudes, both
conventional and unconventional,
have higher sulphur content
compared to lighter crudes.
Upgrading of bitumen to synthetic
crude can signifcantly improve the
sulphur content of the gasoil and
resid range material. In particular,
Sweet Syncrude can approach
sulphur levels at or lower than
lighter crudes such as WTI, Brent
and Bakken across the entire boil-
ing range.
Tight oil crudes can have a
sulphur advantage over sister
conventional crudes, thereby
making these feeds attractive to
sweet crude confgured refneries.
Property distributions
To help understand how property
and contaminant distributions can
be different for various crudes and
the resulting impact on refning
operation, KBCs Petro-SIM was
utilised to generate property curves
for some of the key properties
discussed above.
Sulphur
One of the primary crude properties
refners focus on is sulphur content.
Higher sulphur crudes require
additional assets to remove and
process the sulphur, such as hydro-
treaters, hydrogen plants, amine
systems, sour water systems and
sulphur plants. Figure 1 illustrates
the sulphur distribution for several
Figure 1 Sulphur for conventional and unconventional crudes
Albian heavy synthetic (unconventional)
Athabasca dillbit (unconventional)
Bakken (unconventional)
Maya (conventional)
Brent (conventional)
WTI (conventional)
Syncrude sweet (unconventional)
Athabasca bitumen (unconventional)
S
u
l
p
h
u
r

c
o
n
t
e
n
t
,

w
t
%
Cutpoint temperature
Tool Advantages Disadvantages
Crude assay Create standard crude cut estimates Estimates fractionation efciency
management system Ease of use No downstream impacts

Spreadsheet Ease of use Need estimate of input properties
correlations Open literature based for changing conditions
Linear programs Commonly used Linearisation of non-linear processes
Solution speed Limited number of cuts tracked
Facility wide prediction capability
Process simulations Facility wide, non-linear Solution time
representation Proprietary methods
Predict changes in properties from
separations/reactions
Detailed component property curves
Tool advantages and disadvantages
Table 4
kbc.indd 6 11/12/2013 13:01
www.eptq.com PTQ Q1 2014 109
at much lower distillation tempera-
tures than most conventional and
unconventional crudes
Brent and Syncrude Sweet have
similar concarbon distributions,
though they are produced by
different methods. Syncrude, by
defnition, is a synthetic crude
that is created by coking and
severe hydrotreating of Canadian
bitumen. This process signifcantly
reduces the amount of resid in the
crude but also impacts the amount
and distribution of concarbon.
Based on the assay data utilised,
Bakken crude actually has a
units, feedstock concarbon content
determines the coke yield, thereby
infuencing the capacity of the
unit as well as the liquid product
yields.
Figure 4 illustrates the concarbon
distribution of back-end gasoil
distillation range for several
conventional and unconventional
North American crudes.
As with metals, an unconven-
tional crude does not necessarily
mean a crude will have high
concarbon levels However, several
of the example conventional crudes
have concarbon levels that appear
of metals, even in the gasoil range.
The type of diluent will have a
signifcant impact on this effect.
Tight oil crude (Bakken) has
substantially lower vanadium
content than even a sister conven-
tional crude such as WTI, though
the nickel distributions are similar.
There are several considerations
and options to manage crude
metals content and distribution. For
instance, design of the vacuum
tower HVGO wash zone (distribu-
tors, wash rates, grid/packing),
fash zone (tower entry, vapour
disengaging), and transfer line
(velocity and fow regime) will
infuence entrainment, which can
impact gasoil metals as much as the
natural distribution of metals.
Regular metals balances among
HVGO, overfash (slop wax), and
vacuum resid can help validate
vacuum tower entrainment
performance.
Use of a process simulator can
help segregate the impact of metals
from cutpoint and metals from
entrainment, thereby providing
supporting evidence for improving
a wash zones design.
Active tracking of catalyst
activity and metals loading in the
gasoil HDS is necessary to ensure
the catalyst will meet the desired
cycle length.
As part of vacuum unit monitor-
ing and monthly production
planning backcasting, utilise a
process simulator and actual plant
data to help identify inaccuracies in
crude assay metals content and
distribution profle.

Concarbon
As with metals, concarbon is a key
contaminant that impacts several
units. Concarbon has an impact on
hydrotreater catalyst deactivation,
such that the combination of
metals, concarbon and coke forma-
tion from hydroprocessing
reactions dictate overall cycle
length. In the FCC, concarbon has a
signifcant effect on the reactor heat
balance, thereby infuencing cata-
lyst circulation rate (cat/oil ratio).
For FCCs processing resid range
material, the concarbon level
typically requires a catalyst cooler
to maintain heat balance. For coker
V
a
n
a
d
i
u
m

c
o
n
t
e
n
t
,

p
p
m

w
t
Cutpoint temperature
Arabian heavy (conventional)
Athabasca dillbit (unconventional)
Bakken (unconventional)
Maya (conventional)
WTI (conventional)
Athabasca bitumen (unconventional)
Zuata medium (unconventional)
Figure 3 Vanadium for conventional and unconventional crudes
Albian heavy synthetic (unconventional)
Athabasca dillbit (unconventional)
Bakken (unconventional)
Brent (conventional)
WTI (conventional)
Zuata medium (unconventional)
N
i
c
k
e
l

c
o
n
t
e
n
t
,

p
p
m

w
t
Cutpoint temperature
Figure 2 Nickel for conventional and unconventional crudes
kbc.indd 7 11/12/2013 13:02
110 PTQ Q1 2014 www.eptq.com
facility impact for these crude
modications, as long as the tool
has representative assay informa-
tion and can properly account for
downstream unit impacts on vary-
ing concarbon levels.
Carbon to hydrogen ratio
Carbon to hydrogen ratio (C/H) is
a property that is often not at the
forefront of properties to measure
and track, but this ratio essentially
dictates a renerys conguration
and capabilities. In simple terms, a
given crude will have a certain
C/H ratio and this property has a
distribution, just as other crude
properties do. On the product side,
gasoline, jet, and diesel have certain
ranges of C/H that allow the prod-
uct to meet quality specications,
such as specic gravity, cetane,
aromatics, and so on. These quality
specications can vary substan-
tially, depending on local, regional,
national, or global standards.
Therefore, not only is the renery
congured to separate and upgrade
streams and remove contaminants
such as sulphur and nitrogen, but it
is also arranged to change the C/H
ratio of streams. To impact this
ratio, one must reject carbon and/
or add hydrogen. Units such as the
FCC and coker reject carbon,
whereas hydroprocessing units
primarily add hydrogen. Selling
heavy products, such as fuel oil or
marine fuel, is a round-about way
to reject carbon and avoid having
to add hydrogen, though the prices
and demands for these products
are typically poor. Table 5 summa-
rises some representative ranges for
C/H for various crude grades and
typical rening products.
Although C/H ratio can be meas-
ured, most reners do not regularly
test for this property, and most
crude assays do not carry it,
although they do carry other prox-
ies for C/H such as Watson K and
PONA. Therefore correlations like
those typically found in a process
simulator are used to calculate C/H
ratio from bulk stream properties.
This property is critical for an accu-
rate process simulation, as the
kinetic reactors need to predict and
utilise this information to deter-
mine the amount of hydrogen that
lift limited. However, the unit may
not be able to increase gasoil
cutpoint if the tight oil concarbon
level is relatively high and the cycle
length of the gasoil HDS unit is
overly impacted. If the new crude
slate results in insufcient gasoil
material to keep the FCC at maxi-
mum capacity utilisation, the
rener may have to modify the
crude slate, evaluate externally
purchased feedstock, or understand
the overall impact on economics of
a slack FCC.
Given that most tight oils not
only have low resid yields but also
relatively low concarbon levels, the
coker unit may have insufcient
feedstock to maximise capacity
utilisation. Therefore, the crude
slate may require modication,
external feeds may be needed, or
the economic debit for a lower
coker rate should be included in
the overall protability analysis.
LPs and detailed process simula-
tors can help analyse the overall
relatively high level and rapid
ascent as regards concarbon.
As with metals, the same vacuum
tower conguration and operating
conditions will impact the amount
of concarbon in gasoil due to
entrainment. Therefore concarbon
should be part of the unit capability
review, tower design and opera-
tions monitoring information.
Also as with metals, specialised
concarbon destruction catalyst can
be installed in the gasoil hydro-
treater to help manage concarbon
and ensure the unit meets cycle
length targets. In addition, monitor-
ing of feedstock concarbon levels
should be part of the catalyst activ-
ity monitoring program.
For those reneries considering
processing of tight oils, several
points should be evaluated. These
tight oils typically have less gasoil
range material than most conven-
tional crudes. Therefore, the unit
may be able to increase cutpoint if
the vacuum tower is hydraulically
Arabian heavy (conventional)
Athabasca dillbit (unconventional)
Bakken (unconventional)
Maya (conventional)
Zuata medium (unconventional)
Brent (conventional)
Syncrude sweet (unconventional)
C
o
n
c
a
r
b
o
n

c
o
n
t
e
n
t
,

w
t
%
Cutpoint temperature
Figure 4 Concarbon for conventional and unconventional crudes
Crudes C/H range Products C/H range
Dilbit and bitumen 7.3 -8.1 LPG 4.5 -4.9
Synthetic crude 7.0 -7.2 Gasoline 5.7 -6.2
Conventional heavy 6.9 -7.3 Jet 6.2 -6.7
Conventional medium 6.3 -6.9 Diesel 6.4 -6.9
Tight oils 6.3 -6.5 Fuel oil >7.0
Crude and renery product C/H ranges
Table 5
kbc.indd 8 11/12/2013 13:02
www.digitalrening.com is an
easy-to-use website for the rening, gas
and petrochemical processing industries.
A simple search facility enables you to
nd and display relevant information
across the whole spectrum of articles,
literature, news, events and companies.
Readers can download technical articles
and other information from over 40
categories.


Find relevant industry events worldwide.


Read news items focusing on contracts,
company news and new products.
Download articles from our new White
Papers section
dig ref copy 2.indd 1 16/09/2013 10:02
112 PTQ Q1 2014 www.eptq.com
Case studies
To illustrate the observations, prin-
ciples, and recommendations
discussed within this article, two
case studies are presented.
Renery reconguration
A South American rener was in
the front end loading (FEL) 0 stage
of a browneld reconguration
project. The strategic goal was not
only to process internally produced
crudes but also allow exibility to
operate as a merchant rener,
thereby buying and processing a
wide range of opportunity crudes
and exporting high quality prod-
ucts to multiple regions. Therefore,
to select the correct renery cong-
uration and test multiple scenarios,
KBCs Petro-SIM and a linear
program model were selected as
the tools of choice.
Crude assays from open literature
sources, the reners own database,
and other licensed crude assay
databases were used to characterise
the potential crude slates. A full
renery owsheet was constructed
in Petro-SIM to examine potential
renery congurations and track
key stream properties from crude
through nal product blending.
The LP and Petro-SIM models were
run in tandem to evaluate various
crude baskets, with the LP selecting
the handful of economic crude
slates and Petro-SIM completing
the technical validation against unit
design capability and checking
product quality assurance. Once a
conguration was selected, the
various stream properties (feed and
product quality targets) were
provided to technology licensors
for development of design
packages.
This particular facility is now
under construction and should be
operational in the next several
years. During the other FEL stages,
the use of Petro-SIM helped stream-
line the design process, thereby
reducing overall project timetable.
In addition, the selected congura-
tion has been stress tested against
alternate crude slates and
operational scenarios, thereby
demonstrating the robust congu-
rations ability to meet future
processing objectives.
specications. Process simulators are
very effective in analysing these
impacts as they can calculate
and track C/H and have the
kinetic models to predict carbon
rejection and hydrogen addition.
LP models can be utilised but the
unit sub-models must have the
necessary shift vectors to account
for hydrogen consumption
requirements.
For heavy unconventional crudes
that are processed in coking rener-
ies, the resulting liquid products
have poor C/H ratios. Therefore,
additional hydroprocessing hydrau-
lic and thermal capacity is required
to add sufcient hydrogen to
produce on-specication products.
Since many tight oils are rich in
distillate material, these crudes t
nicely with diesel-centric markets.
Many of these crudes are highly
parafnic, so even though the distil-
late in crude may already meet
cetane requirements and require
minimal sulphur removal (see
above), the cold property (freeze,
cloud, and pour point) impact
should be reviewed. As an example,
if the jet product is drawn to a
freeze point specication, processing
tight oil may require dropping kero-
sene range material to the diesel
pool, which impacts cetane index
and hydroprocessing unit severity
and hydraulic requirements.
must be added or carbon rejected
to model crude through nished
products.
Figure 5 illustrates the C/H
distribution of kerosene and diesel
distillation range for several
conventional and unconventional
North American crudes. Kerosene
and diesel range material is a good
example of C/H ratio impacts as
the key product property for diesel
is cetane which is driven by API or
specic gravity. Diesel gravity can
really only be altered by addition
of hydrogen. Therefore, higher
C/H ratio diesels have lower
cetane indices and require hydro-
gen addition to meet specication.
Most Canadian unconventional
crudes are hydrogen-decient
such that signicant hydrogen
addition is required to meet cetane
specication. Though several of the
conventional crudes have better
C/H ratios compared to their
unconventional counterparts in the
kerosene range, the C/H ratios
begin converging towards the back
end of the diesel range. Some of the
example crudes, such as Bakken
and Syncrude, have relatively at
C/H proles.
Careful review of the hydropro-
cessing assets, as well as hydrogen
purchase and production options,
are critical to ensure a new crude
will meet the necessary product
Albian heavy synthetic (unconventional)
Arabian heavy (conventional)
Athabasca dillbit (unconventional)
Bakken (unconventional)
Maya (conventional)
Zuata medium (unconventional)
Brent (conventional)
Syncrude sweet (unconventional)
C
a
r
b
o
n
/
h
y
d
r
o
g
e
n

r
a
t
i
o
Cutpoint temperature
Figure 5 C/H ratio for conventional and unconventional crudes
kbc.indd 9 11/12/2013 13:02
www.eptq.com PTQ Q1 2014 113
Further reading
1 Aldescu M, Heavy Oil Upgrading, AFPM
Annual Meeting, March 2012.
2 Kuhl M A, Hoyle A, Ohmes R, Capitalizing on
Shale Gas in the Downstream Energy Sector,
AFPM Annual Meeting, March 2013.
3 Ecopetrol Website, Cao Limon Crude Assay,
August 2003.
4 http://www.ecopetrol.com.co/english/
documentos/40546_Assay_Cano_Limon.xls
5 Sayles S, Routt M, Unconventional Crude
Oil Selection and Compatibility, NPRA Annual
Meeting, March 2011.
6 Tucker M A, LP Modeling Past, Present, and
Future, NPRA 2001 Computer Conference, CC-
01-153.
7 Ohmes R, Sayles S, Analyzing and Addressing
the Clean Fuels and Expansion Challenge, NPRA
Annual Meeting, March 2007.
8 Sayles S, Unconventional Crude Processing
Part 2: Heteroatoms, Crude Oil Quality
Association (COQA), October 2010.
Petro-SIM is a trademark of KBC Advanced
Technologies.
Robert Ohmes is a Principal Consultant with
KBC Advanced Technologies and a Professional
Engineer in Louisiana.
understand the second and third
order effects of a crude slate change
on the refnery operation. Finally,
several capital investment options
were screened for viability more
rapidly than other methods would
typically allow, thereby permitting
the organisation to focus on those
investments that showed economic
and technical promise.
Summary
In conclusion, characterising and
tracking of specifc properties and
contaminants is critical in modern
refneries. These items impact not
only selection of crude slates for
existing facilities but also the
impact on revamp and new unit
designs. Several viable sources and
methodologies exist for measuring
and generating data, but the chal-
lenge is converting this raw
data into useful information. The
key consideration is applying the
tools and methodologies that are
ft for purpose and allow the
refner to meet economic, reliabil-
ity, environmental, and safe
operating targets.
Tight oil evaluation
A North American refner was in
the process of examining the
impact of processing tight oil in
their facility. The refner wanted to
understand not only the amount of
tight oil the existing facility and
confguration could process, but
also how strategic investment
options would impact the amount
processed.
A Petro-SIM simulation of the
existing confguration was created
and a series of case studies
completed. Some of the scenarios
included substitution of existing
light conventional crudes with tight
oil, addition of new towers and
processing units, and determination
of hydrogen addition requirements.
The resulting case analysis
demonstrated to the refner that a
certain amount of tight oil could be
processed in the existing facility
with minimal change in operating
targets. In addition, the case studies
highlighted constraints and oppor-
tunities that were not immediately
apparent, due to Petro-SIMs ability
to track key stream qualities and
ABB Consulting.
Safe, reliable, efficient
operations.
Operators aiming to achieve operational excellence in the face of increasing
competitive, legislative and economic pressures, often require expert
assistance. ABB Consulting assists customers in the global process
industries. We offer expertise in; inspection, integrity management,
operations improvement, process safety, project services, site and asset
closure, technical engineering, technical software, technical training and
competency. We benefit customers by reducing risk, optimising cost and
improving manufacturing efficiency. www.abb.com/consulting
ABB Consulting
Tel: +44 (0)1925 741111
E-mail: contact@gb.abb.com
PTQ Q1 2014 (125x178) Dec 2013.indd 1 09/12/2013 14:49:02
kbc.indd 10 11/12/2013 13:02
sabin.indd 1 27/02/2013 15:01
New process arrangements for upgrading
heavy oils and residua
I
n this article, we consider
the results of integrating residue
fuid catalytic cracking (RFCC),
hydrocracking (HCR) and hydro-
conversion (HRH). A target refnery
was simulated and calibrated, then
the HRH unit patented by the
Research Institute of Petroleum
Industry (RIPI) was added to these
units. After validating Hysys-
Refnery with actual and design
data, four integration strategies
were examined to increase the
yields of gasoline and diesel. These
cases, named simple series, series,
parallel and residue upgrading,
were compared to the base
(designed) one. The results showed
that by implementing the
mentioned cases, the production
yields of gasoline and diesel would
increase considerably. For the best
case residue upgrading strategy
the yields of gasoline and diesel
would increase to 6.98% and
53.96%, respectively, in comparison
to the base case. Moreover, with
this integration strategy, and no
change in operating conditions, the
bottom of barrel could lead to zero,
while fresh feed fow rates and
their impurities would remain
constant.
The markets demand for heavy
petroleum products such as heavy
oil is set to decline, while the
demand for lighter, more valuable
products such as gasoline and
diesel is expected to increase.
Projections of demand for refned
products indicate that middle distil-
lates (diesel fuel and jet fuel) will
grow at the highest rate. Demand
for diesel fuel is projected to grow
by 2% per year worldwide.
1
In
many countries, the need for
Simulation studies reveal increased gasoline and diesel production from extra-heavy
oil by means of RFCC and hydrocracking units with a mild hydrogenation process
SEPEHR SADIGHI, REZA SEIF MOHADDECY and KAMAL MASOUDIAN
Research Institute of Petroleum Industry
gasoline is a crucial subject, so any
improvement in the refnery
process to increase the yield of this
product is welcome.
HCR and RFCC are the major
processes to produce diesel and
gasoline, respectively. Moreover,
recently the HRH unit has been
developed by RIPI
2
to convert resi-
due and heavy cuts to lighter
products.
Hysys-Refnery is a simulation
tool commercialised by KBC
Advanced Technologies and AEA
Technology-Hyprotech. This simu-
lator has made signifcant advances
in detailed representation of reactor
sections. Mohaddecy and Sadighi
3
have demonstrated the ability of
this software to simulate the HCR
and catalytic reforming units. Lee
4

used Hysys-Refnery software to
study the integration of FCC and
hydrotreating units. The research
was conducted in two steps. First,
the simulation and calibration of
the two units was carried out, and
then the integration was performed.
The results revealed that increasing
the hydrotreating severity
decreased the production of SOx
and NOx in the FCC unit.
4
Dean, et
al, integrated FCC and HCR units
to improve the conversion of resi-
due to more valuable products in
the gasoline range.
5
A FCC unit
was assumed as the upstream unit,
and products above the gasoline
boiling range were sent to the HCR
unit. The HCR off-test stream was
fed to the FCC unit, and the FCC
cycle oil was desulphurised and
cracked. The simulated integration
scheme that was considered
resulted in a 60% reduction in
hydrogen consumption. Tallman, et
al, integrated the FCC unit with
a thermal cracking unit to maxim-
ise ethylene and propylene
production.
6
Initially, the target refnery with
HCR and RFCC units was simu-
lated using the HCR-Sim and
FCC-Sim reactor modules available
in the Hysys-Refnery simulator.
For validation purposes, actual data
for the HCR unit and design data
for the RFCC unit were used. This
validated case is termed the base
case hereafter. The HRH unit was
simulated in the Hysys-Refnery as
a yield reactor, to satisfy the yields
reported by the licensor. Finally,
while keeping constant the feed
quality and fow rate to the HCR
and RFCC units, these units were
integrated by various strategies to
increase the yields of diesel and
gasoline. The integrated cases that
were considered are termed simple
series, series, parallel and residue
upgrading strategies.
Hydroconversion unit
The HRH unit is a liquid phase
mild hydrogenation (at 60-100 atm
and 400-500C) process for upgrad-
ing extra heavy oil. The HRH
process recently developed by RIPI
produces higher valued products,
which alternatively could be used
as feedstock for refneries.
7
The
main objective of this process is the
break-up of high molecular weight
hydrocarbons to light and
medium molecular weight prod-
ucts. This process is a novel method
for upgrading heavy residue to
lighter products. In the process,
two types of reactions namely,
cracking and mild hydrogenation
occur simultaneously. The
www.eptq.com PTQ Q1 2014 115
ripi.indd 1 11/12/2013 13:08
116 PTQ Q1 2014 www.eptq.com
hydrogenation and operating
conditions allow higher conversion
without coking and excess polym-
erisation. The HRH process has
high fexibility with regards to the
type of feed and the amounts of
sulphur and heavy metal contents
in the feed. The process can elimi-
nate all the heavy metals and
almost 50% of the sulphur compo-
nents in the feed. The main
products are gasoline, diesel and
FCC feedstock, and the designed
yields for these cuts are 20, 39.6
and 30 wt%, respectively, on the
basis of the fresh feed.
Base case
The base case considered in this
study takes vacuum gas oil (VGO)
from the vacuum distillation tower
as the HCRs fresh feed, combined
with the recycle stream from the
bottom of the HCR fractionation
tower to make the combined HCR
feed. The RFCC feed is composed of
light vacuum gas oil (LVGO), heavy
vacuum slops (HVS), heavy gas oil
(HGO) and treated residue (TR).
Unstabilised LPG
Raw FCC feed
Recycle
feed
Recycle
input
HC off test
HC
residue
HCR
combined feed
LVGO
HVS
HGO
Treated
residue
OCT
LPG
CDU
LPG
A
HC yield
(mass)
HC yield
(vol)
FCC yield
(mass)
FCC yield
(vol)
Hydrocracker
Splitter
Residue fluid
catalytic cracker
Recycle
Recycle
adjust
HFeedM
FCC
Mixer 1
FCC
Mixer 2
FCC
offgas
RFCC
coke
FRCG
LCO
CSO
HC offgas
HC LPG
HC light naphtha
HC heavy naphtha
HC kerosene
HC diesel
R
HC
fresh
feed
Figure1 Arrangement of units and streams in the base case
Unstabilised LPG
Raw FCC feed
Recycle
feed
Recycle
input
HC off test
HRH feed
CSO
HC
residue
HCR
combined feed
LVGO
HVS
HGO
Treated
residue
OCT
LPG
CDU
LPG
A
HC yield
(mass)
HC yield
(vol)
FCC yield
(mass)
FCC yield
(vol)
HRH yield
(mass)
HRH yield
(vol)
Hydroconverter
Hydrocracker
Residue fluid
catalytic cracker
Recycle
Recycle
adjust
HRH feed
mixer
HFeedM
FCC
Mixer 1
FCC
Mixer 2
FCC
offgas
RFCC
coke
FRCG
LCO
HC offgas
HC LPG
HC light naphtha
HC heavy naphtha
HC kerosene
HC diesel
HRH offgas
HRH gasoline
HRH diesel
HRH FCC feed
R
HC
fresh
feed
Splitter
Figure 2 Arrangement of units and streams in the simple series strategy
ripi.indd 2 11/12/2013 13:08
The TR stream is the result of
hydrotreatment of vacuum residue,
which removes it from the sulphur
and metallic impurities, and thus
makes it suitable as the RFCC feed.
The RFCC units product quality is
directly affected by its feedstock
quality. In particular, unlike in
hydrotreating, RFCC redistributes
sulphur into its products.
Consequently, in all integration
strategies, the quality of RFCC feed-
stock has been kept the same as in
the base case. The block diagram
and feed fow rates of the base case
www.eptq.com PTQ Q1 2014 117
are shown in Figure 1 and Table 1,
respectively. Both units have been
simulated, calibrated and validated
using the actual data gathered from
the target refnery.
8,9

Integration of RFCC, HCR and
HRH units
Four case studies for the integration
of the RFCC, HCR and HRH units
have been surveyed. The main
objective of integration is to
increase the production yield of
gasoline while maintaining the feed
specifcations of the RFCC and
HCR units the same as in the base
case. The HRH unit patented by
RIPI is fexible towards a variation
in feed quality, so the variation in
the feed specifcation is deemed not
important for this unit.
Simple series integration
The block diagram and feed fow
rates of a simple series integration
strategy are shown in Figure 2 and
Table 2, respectively. The feed and
HCR feed ow rate
Fresh feed rate, BPD 24 090
HCR recycle, BPD 13 150
RFCC feed ow rate
LVGO, BPD 10 430
HVS, BPD 9568
Treated residue, BPD 60 860
HGO, BPD 5344
Feed ow rates in the base case
Table 1
HRH feed ow rate
HCR off-test, BPD 674.3
CSO, BPD 8951
RFCC feed ow rate
LVGO, BPD 5344
HVS, BPD 10 430
Treated residue, BPD 60 860
HGO, BPD 9568
Feed ow rates of HRH and RFCC in
simple series strategy
Table 2
RFCC feed ow rate
LVGO, BPD 10 430
HVS, BPD 9568
TR, BPD 60 100
HGO, BPD 5344
HCR off-test, BPD 758.7
HRH feed ow rate
TR, BPD 754.3
CSO, BPD 8927
Feed ow rates of HRH and RFCC in
series strategy
Table 3
Unstabilised LPG
Raw FCC feed
F treated
residue
H treated residue
Recycle
feed
Recycle
input
HC off test
HRH feed
CSO
HC
residue
HCR
combined feed
LVGO
HVS
HGO
Treated
residue
OCT
LPG
CDU
LPG
A
HC yield
(mass)
HC yield
(vol)
FCC yield
(mass)
FCC yield
(vol)
HRH yield
(mass)
HRH yield
(vol)
Hydroconverter
Hydrocracker
Residue fluid
catalytic cracker
FCC-ADJ
Recycle
FCC split
Recycle
adjust
HRH feed
mixer
HFeedM
FCC
Mixer 1
FCC
Mixer 2
FCC
offgas
RFCC
coke
FRCG
LCO
HC offgas
HC LPG
HC light naphtha
HC heavy naphtha
HC kerosene
HC diesel
HRH offgas
HRH gasoline
HRH diesel
HRH FCC feed
A
R
HC
fresh
feed
Splitter
Figure 3 Arrangement of units and streams in the series strategy
ripi.indd 3 11/12/2013 13:08
118 PTQ Q1 2014 www.eptq.com
Parallel integration
The block diagram and feed fow
rates of a parallel integration strat-
egy are shown in Figure 4 and
Table 4, respectively. The HCR unit
takes some LVGO as feed, and the
RFCC, in turn, takes some HCR
off-test stream instead.
Residue upgrading integration
The block diagram and feed fow
rates of a residue upgrading strat-
egy are shown in Figure 5 and
Table 5, respectively. The feed for
HCR was the same as for the paral-
lel case.
Results and discussion
The HRH unit yields valuable
lighter products with lower sulphur
and metallic impurities. It has been
demonstrated in this study that
optimum integration of the HRH
unit with the HCR and RFCC units
can lead to the production of the
highest amount of gasoline and
diesel. Four integration schemes
consisting of series, simple series,
parallel and residue upgrading
strategies have been studied using
the Hysys-Refnery simulator.
the same as in the base case. The
HRH unit takes some TR as well as
the CSO, and also the RFCC unit
takes the HCR off-test. The advan-
tage of this strategy over the simple
series is the direct feeding of the
residue treatment unit into the
HRH unit. In cases of deep catalyst
deactivation or complete shutdown
occurring in the residue treatment
unit, the fow of untreated residue
can be redirected to the HRH unit,
and this unit is capable of remov-
ing heavy sulphur and metallic
contamination.
product properties for HCR were
the same as in the base case. The
HRH unit takes the HCR off-test
stream and the RFCCs clarifed
slurry oil (CSO) as feed. The RFCC
takes the LVGO, HVS, TR and
HGO as feed.
Series integration
The block diagram, feed and prod-
uct fow rates of a series integration
strategy are shown in Figure 3 and
Table 3, respectively. The feed and
product properties for HCR were
Unstabilised LPG
Raw FCC feed
LVGO-FCC
LVGO-HCR
REC-FCC
REC-HCR
Recycle
feed
Recycle
input
HRH feed
CSO
HC
residue
HCR
combined feed
LVGO
HVS
HGO
Treated
residue
OCT
LPG
CDU
LPG
HC
fresh
feed
A
HC yield
(mass)
HC yield
(vol)
FCC yield
(mass)
FCC yield
(vol)
HRH yield
(mass)
HRH yield
(vol)
Hydroconverter
Hydrocracker
Residue fluid
catalytic cracker
LVGO-HCR
LVGO-ADJ
REC-S
Recycle
Recycle
adjust
HRH feed
mixer
HFeedM
FCC
Mixer 1
FCC
Mixer 2
FCC
offgas
RFCC coke
FRCG
LCO
HC offgas
HC LPG
HC light naphtha
HC heavy naphtha
HC kerosene
HC diesel
HC off test
HRH offgas
HRH gasoline
HRH diesel
HRH FCC feed
A
R
Splitter
Figure 4 Arrangement of units and streams in the parallel strategy
HCR feed ow rate
HCR recycle, BPD 11 530
LVGO, BPD 1690
RFCC feed ow rate
LVGO, BPD 8741
HVS, BPD 9568
TR, BPD 60 860
HGO, BPD 5344
HCR off-test, BPD 1688
HRH feed ow rate
TR, BPD 0
CSO, BPD 9021
Feed ow rates of HCR, RFCC and HRH
in parallel strategy
Table 4
RFCC feed ow rate
LVGO, BPD 8741
HRH FCC feed, BPD 4660
HVS, BPD 9568
TR, BPD 56 170
HGO, BPD 5344
HCR off-test, BPD 1688
HRH feed ow rate
TR, BPD 4686
CSO, BPD 8983
Feed ow rates of RFCC and HRH in
residue upgrading strategy
Table 5
ripi.indd 4 11/12/2013 13:08
Discover more at www.metso.com/oilandgas/fowcontrol
Neles Jamesbury Mapag
Look what goes into a Metso valve. It starts with a long track record of delivering engineered
performance and legendary reliability with premier products such as Neles, Jamesbury
and Mapag. But the numbers really paint the picture. In almost 90 years, Metso has delivered
globally millions of valves, control valves and on-off valves. We have also become one
of the leading suppliers of smart positioners. All backed by feld service expertise from over
55 automation service hubs and over 30 valve service centers around the world. We see it
this way: keeping oil and gas producers working safely and reliably protects investments,
people and the planet.
There is more
to Metso than
meets the eye.
metso.indd 1 10/12/2013 12:43
120 PTQ Q1 2014 www.eptq.com
lead to higher diesel production.
Comparing the series, simple
series and parallel strategies reveals
that the parallel strategy produces
the highest amount of gasoline and
the least amount of diesel. This is
because the HCR unit is fed the
lighter LVGO stream in the parallel
strategy, and therefore the HCR
unit effectively acts as a gasoline
booster in this strategy and less
diesel is produced. It can be seen
from the data in Tables 6 and 7 that
the sulphur and metallic contents
of the feeds to the HCR and the
RFCC unit do not vary considera-
bly in different integration
strategies. As was mentioned
before, the product quality of RFCC
unit is closely tied to its feed qual-
ity, and therefore the product
quality from the RFCC unit remains
constant under different integration
strategies. Moreover, the impurities
in HRH products are negligible, so
they can be directly added to the
gasoline and diesel pools with no
further treatment.
highest amount of gasoline and
diesel in comparison to the base
case. In this strategy, all of the
HRH-RFCC feed is sent to the
RFCC, and also all of the CSO and
the HCR off-test are fed to the
HRH. Consequently, in this
scheme, all the residue streams are
utilised effectively. Due to the oper-
ating conditions prevalent in the
HRH, this unit is geared. Therefore,
after comparing the presented
scenarios, we can conclude that the
introduction of a HRH unit will
Figures 6 and 7 show the percent-
age increase in the production yield
of gasoline and diesel with respect
to the base case for the integration
strategies considered. The gasoline
production yield used in Figure 6 is
the sum of light and heavy naphtha
from the HCR, full ranged cracked
gasoline (FRCG) from the RFCC
unit, and gasoline from HRH, while
the diesel yield is the sum of the
HCR and HRH diesel streams. It
can be seen that the residue
upgrading strategy produces the
Unstabilised
LPG
Raw FCC feed
LVGO-FCC
LVGO-HCR
REC-FCC
REC-HCR
Recycle
feed
FCC feed recycled
Recycle
input
HRH feed
CSO
HC
residue
HCR
combined feed
LVGO
HVS
HC
fresh
feed
A
A
HC yield
(mass)
HC yield
(vol)
FCC yield
(mass)
FCC yield
(vol)
HRH yield
(mass)
HRH yield
(vol)
Hydroconverter
Hydrocracker
Residue fluid
catalytic cracker
LVGO-SP
LVGO-ADJ
REC-S
HRH-REC
Recycle
Recycle
adjust
HRH feed
mixer
HFeedM
FCC
Mixer 1
FCC
Mixer 2
FCC
offgas
RFCC
coke
FRCG
LCO
HC offgas
HC LPG
HC light naphtha
HC heavy naphtha
HC kerosene
HC diesel
HC off test
HRH offgas
HRH gasoline
HRH diesel
HRH FCC feed
A
R
R
Splitter
F treated
residue
H treated residue
Treated
residue
TR-SP
HGO
OCT
LPG
CDU
LPG
Figure 5 Arrangement of units and streams in the residue upgrading strategy
6
5
4
3
2
1
G
a
s
o
l
i
n
e

i
n
c
r
e
a
s
e
,

%
0
Simple series Series Parallel Residue
upgrading
Integration cases
4.88% 4.96%
5.51%
6.98%
Figure 6 Comparison to base case of growth in gasoline production
ripi.indd 5 11/12/2013 13:08
www.eptq.com PTQ Q1 2014 121
of Heavy Hydrocarbonaceous Feedstock,
European Patent Application, EP 1754770A1,
Research Institute of Petroleum Industry.
8 Bahmani M, Sadighi S, Mohaddecy S R,
Mashayekhi M, Hydrocracker parametric
sensitivity study, PTQ, Q2 2009.
9 Sadighi S, Mohaddecy S R, Ghabouli O,
Rashidzadeh M, Optimisation of product yield
and coke formation in a RFCC unit, PTQ, Q2
2010.
Sepehr Sadighi is Project Manager, Catalysis
and Nanotechnology Division, Catalytic
Reaction Engineering Department, Research
Institute of Petroleum Industry (RIPI), Tehran,
Iran. He holds a PhD in chemical engineering
from Universiti Teknologi Malaysia.
Email: Sadighis @ripi.ir
Reza Seif Mohaddecy is Project Manager,
Catalysis and Nanotechnology Division,
Catalytic Reaction Engineering Department,
RIPI. He holds a MS in chemical engineering
from Sharif University of Technology.
Email: Seifsr @ripi.ir
Kamal Masoudian is Project Manager, Catalysis
and Nanotechnology Division, Catalyst
Characterization and Evaluation Department,
RIPI. He holds a BSc in chemical engineering.
Email: masoudiansk @ripi.ir
References
1 Bhaskar, M, Valavarasu G, Balarman K S,
Advantages of mild hydrocracking FCC feed
- a pilot plant study, Petroleum Science and
Technology, 21, 9 & 10, 1439-1451, 2003.
2 European patent application, EP
1754770A1, Process for Hydroconverting of
a Hydrocarbonaceous Feedstock, Research
Institute of Petroleum and NTI company, 2007.
3 Mohaddecy S R, Zahedi S, Sadighi S, Bonyad
H, Reactor modeling and simulation of
catalytic reforming, Petroleum & Coal, 48 (3),
28-35, 2006.
4 Lee R, Leunenberger E, Powell R, Optimizing
the cat feed hydrotreater/FCCU complex with
detailed simulation tools, Process Technology
Update, Desulphurization Process, Word
Rening, Jul/Aug 2001.
5 Dean R R, Mauleon J L, Combined Fluid
Catalytic Cracking and Hydrocracking Process,
US patent No: 4426276, 17 Jan 1984.
6 Talman J, Jonsgma B, Thamprajmachit B,
Cackett S, Wijk R, Synergistic integration of FCC
and hydroprocessing facilities for upgrading
bottom of the barrel, Asia Pacic Rening
Technology Conference, 2001, Kuala Lumper.
7 Kadiev K M, Mezhidov V K, Zarkesh J,
Masoudian S K, Process for Hydroconverting
60
50
40
30
20
10
D
i
e
s
e
l

i
n
c
r
e
a
s
e
,

%
0
Simple series Series Parallel Residue
upgrading
Integration cases
39.89%
41.27%
37.53%
53.69%
Figure 7 Comparison to base case of growth in diesel production
HCR feed Ni+V, ppmwt Sulphur, wt% IBP, C FBP, C
Base 2.98 1.222 309 516.4
Simple series 2.98 1.222 309 516.4
Series 2.98 1.222 309 516.4
Parallel 3.18 1.323 285.5 516.3
Residue upgrading 3.18 1.323 285.5 516.3
Feed quality of HCR unit in base case and integration strategies
Table 6
RFCC feed Ni+V, ppmwt Sulphur, wt% IBP, C FBP, C
Base 14.723 1 264.6 592.6
Simple series 14.723 1 264.6 592.6
Series 14.615 1.006 264.6 599.5
Parallel 14.874 0.9667 267.3 592.6
Residue upgrading 13.734 1.088 267.3 592.1
Feed quality of RFCC unit in base case and integration strategies
Table 7
www.eptq.com Revamps 2013 63
processes might increase in the
future, and the fouling of units
might become a more serious
concern, potentially impacting a
large number of production days.
By regular application of online
cleaning, the unit can always be
operated under clean conditions. In
addition, the vast majority of
mechanical work, which takes up a
signifcant number of days during
unit shutdowns, can be replaced
with an online cleaning process that
requires no opening of equipment
or man entry, and can be carried out
in as little as 24 hours oil-to-oil.
Acknowledgement
Special thanks to all the Lotos team for the
support and valuable assistance provided
during the entire time that ITW Online
Cleaning has been applied on site.
Mariusz Hoowacz has been Crude Oil
Distillation Complex Manager with Grupa
Lotos in Gdansk, Poland, since 2008 and, since
2010, Manager of the new distillation unit 120.
Rafa Zaprawa was VDU and propane
deasphalting unit Shift Team Manager with
Grupa Lotos in Gdansk, Poland, from 2001-
2008, then Deputy Manager of the crude oil
distillation complex.
Marcello Ferrara is the Chairman of ITW. With
27 years experience in the petroleum business,
including oil exploration and production,
rening, petrochemicals, transportation,
and energy production, he holds a PhD in
industrial chemistry and international patents
for new processes and additive compositions
for environmental control and for improving
petroleum/petrochemical processes.
Email: mferrara@itwtechnologies.com
market considerations and by the
downtime required for a mechani-
cal cleaning turnaround. With a
downtime of 15-20 days, it is much
more economical to run the units
under non-optimised conditions
rather than lose production. This,
however, results in energy losses,
giveaway and capacity reduction,
which negatively impact unit
economics.
The introduction of online clean-
ing, by cleaning the unit in 24
hours on an oil-to-oil basis, allows
for the recovery of losses and the
operation of units under improved
and more reliable conditions.
Furthermore, turnarounds can be
avoided or rescheduled with
reduced downtime.
In the case of turnaround
improvement, an additional
reduction in downtime can be
achieved by applying ITWs
improved degassing/decontamina-
tion technology.
Conclusion
The results of ITW Online Cleaning
have opened up new possibilities
for Lotos, whereby online cleaning
can be applied during a plant run,
to recover a units performance and
improve the level of operational
excellence, and in preparation for a
turnaround, to reduce downtime
and the turnarounds scope of work.
These options are increasingly
important, as the amount of oppor-
tunity crudes the company
Figure 7 HGO crude exchanger
Heavy crude Oil
Atmospheric Distillation
Vacuum Distillation
Coker & Visbreaker Feed
Fluidized Catalytic Cracker
Bitumen

Non-intrusive ow
measurement
up to 400C
Field-Proven at Reneries
Trouble free operation at
extreme pipe temperatures
No clogging, no abrasion,
no pressure losses
Installation and maintenance
without process interruption
Independent of uid or pressure
Hazardous area approved

renery.exim.com
itw.indd 5 10/09/2013 17:06
ripi.indd 6 11/12/2013 20:18
enersul.indd 1 6/12/12 13:25:03
Increasing conversion and run length
in a visbreaker
V
isbreaker economics are
mainly based on achieve-
ment of maximum
conversion. The main barrier to this
goal is in the loss of stability
reserve of asphaltenes, causing
their precipitation to give fouling in
the form of coke. These foulants
can severely shorten unit run
lengths by deposition at the heater,
pre-heat exchangers and columns.
Maximum conversion can be
obtained by setting the proper
process severity for any processed
feed (typically by controlling heater
outlet operating temperature) while
the use of antifoulants/anticoke
chemicals mitigates the fouling
rate, particularly when the unit is
set at optimal severity as a result of
monitoring. This approach results
in the best trade off between foul-
ing control and conversion.
This article presents several
advances that were put in place for
the ISAB Priolo refnery visbreaker,
resulting in improved performance
after successful results were
obtained in the past
1
and met all of
the desired targets. This was possi-
ble due to Baker Hughes VisTec
technology and the capability of the
ISAB refnery to use this technol-
ogy in a very effective way.
Visbreaking (thermal cracking)
limitations due to fouling
The main impact as a result of ther-
mal cracking reactions is a
progressive destabilisation of
asphaltenes contained in the unit
feed residuum. Visbreaker bottom
resid (vistar) has a higher fouling
tendency than the unit feed due to
the stability reserve of the
asphaltenes and their potential to
Combining anticoke/antifoulant treatments with monitoring technology enabled
a renery to keep its visbreaker unit at the best process severity for any feed
MATTEO VIRZI ISAB
MARCO RESPINI Baker Hughes
precipitate and result in coke and
deposits.
This loss of stability is related to
the thermal cracking of asphaltenes
and their associated stabilising
resins.
Asphaltenes de-alkylate to give
lower molecular weight but less
soluble, aromatic carbon rich, free
radicals, sometimes called cores.
This is shown schematically in
Figure 1.
Apart from this key driver to
fouling, the generation of lighter
paraffns by thermal cracking of
heavier molecules contributes to
the decrease in stability. Thermal
cracking reactions that occur in
visbreaking can be schematically
simplifed (see Figure 2).
The asphaltene cores (low molec-
ular weight asphaltenes) tend to
oligomerise (terminate free radicals
by recombination). This generates
coke precursors, which are barely
soluble asphaltenes that are rich in
aromatic carbon.
As long as these are dispersed in
the oil, their further polymerisation,
resulting in coke particles, is
controlled as their contact is limited
and is due to some saturation of
free radicals by the naphthenoaro-
matics present in the residuum and
vistar (hydrogen donors).
Once their concentration exceeds
the solubility limit, they separate
into a second liquid phase, called
mesophase, and rapidly polymer-
ise, resulting in coke particles.
From a kinetic point of view, the
generation of coke precursors has
high activation energy and is
favoured at higher reaction temper-
atures. As a result of the process
temperatures used, visbreaking
tends to produce high levels of
coke particles.
A very interesting insight and
confrmation on the impact of ther-
mal cracking comes from the direct
measurement of changes in the
www.eptq.com PTQ Q1 2014 123
Figure 1 Thermal cracking of asphaltenes during visbreaking
Saturates
Resins
Asphaltenes Low MW
asphaltenes
Lighter
saturates
Carboids
coke
Figure 2 Thermal cracking reactions in
visbreaking
baker hughes.indd 1 12/12/2013 21:14
124 PTQ Q1 2014 www.eptq.com
solubility blending number (SBn)
and insolubility number (In) from
visbreaking units.
2
A typical trend
is presented in Figure 3.
Figure 3 shows that there is no
major change in the SBn from the
visbreaker feed resid to the
visbreaker bottom atmospheric
distillate (vistar). However, the In
of the asphaltenes increases due to
thermal cracking of asphaltenes
into more insoluble ones along with
the loss of stabilising resins.
Stability reserve of asphaltenes is
proportional to the ratio of SBn to
In. The wider the gap, the lower is
the tendency of asphaltenes to
phase separate and precipitate to
give either coke (above 400C) or
deposits.
Asphaltene classifcation captures
a broad range of molecular struc-
tures. The In is an average value of
the asphaltenes; when the stability
reserve is reduced by visbreaking
to low values (small difference
between SBn and In) a portion of
the less stable asphaltenes (the ones
with higher In) begin to phase
separate to give fouling.
Therefore, although the
visbreaker resid has a stability
reserve, asphaltenes are on average
still soluble/dispersible in the
matrix without precipitation, and a
relatively small portion of them are
at the solubility limit and give foul-
ing. The lower the stability reserve,
the higher is the fraction of
asphaltenes likely to result in
fouling.
It is also clear how feeds with the
highest stability reserve can be
processed at higher severities
(conversions).
Limitations in visbreaker conversion
For any blend of processed crudes
there is an optimal conversion for
the downstream visbreaker.
Fouling rates have an exponen-
tially increasing trend with severity
and conversion. At low severities,
the impact of the conversion
increase is very limited. As conver-
sion limits are reached, any further
minor increase in severity has dras-
tic consequences on fouling rates,
resulting in severe reduction in unit
run length. This is shown schemati-
cally in Figure 4.
Sbn
In
S
b
n

a
n
d

I
n
Feed residuum Vistar
Process severity/conversion
Sbn >> In
High stability reserve
Figure 3 Solubility blending number (SBn) and insolubility number (In) changes from
visbreaker feed resid to vistar
F
o
u
l
i
n
g

i
n
c
r
e
a
s
e
s
Conversion increases
Optimal operational
window
Fouling
Increased maintenance cost
Shorter run lengths Conversion loss
Figure 4 Rate of fouling vs conversion
C
o
k
e
Furnace outlet temperature
Figure 5 Coke generation vs furnace outlet temperature (from experimental data on
industrial visbreaker)
baker hughes.indd 2 12/12/2013 21:15
Fouling can severely impact the
heater run length due to coke depo-
sition. Visbreakers typically show a
clear exponential increase of coke
generation with heater temperature
(see Figure 5).
Heat exchanger fouling is another
area of severe problems for most
visbreaker units due to the high
fouling tendency of vistar. This
problem is related to the low stabil-
ity reserve of asphaltenes in vistar,
which can be measured using the
VisTec Stability Index (VSI).
Normally, fouling tendency
increases exponentially with a
decrease of VSI and this trend
limits, together with coke particle
generation, maximum achievable
conversion and severity (see
Figure 6).
Anticoke and antifoulants are
able to extend unit run lengths in
our experience by 50% to 75% with-
out a reduction in severity or
conversion; however, they give the
best results in high fouling situa-
tions below the exponential fouling
range. The extent of fouling in the
exponential range is clearly too
high to be reasonably controlled,
even by anticoke and antifoulants
at reasonable dose rates and treat-
ment cost.
As the conversion increase
requires running the unit at high
severities, close to the onset of
exponential fouling, a detailed
knowledge of the minimum accept-
able vistar stability reserve limit is
necessary to run the unit at maxi-
mum conversion, together with an
effective method to measure the
stability.
This was achieved at ISAB by
following Baker Hughes proprie-
tary guidelines and by continuous
and prolonged experience and opti-
misation on this unit.
Another very important aspect is
the refners technology and the abil-
ity of the operations staff to adjust
process parameters related to sever-
ity in order to keep the unit at the
optimal severity stability reserve.
This requires a continuous and
careful daily adjustment and
considers frequent changes in
visbreaker feed composition and
the frequent processing of new
opportunity crudes.
www.eptq.com PTQ Q1 2014 125
ISAB refnery staff developed
deep knowledge and best practices
on how to use the Baker Hughes
monitoring data to adjust the
process. This was a key aspect to
the successful results achieved at
this visbreaking unit.
Monitoring visbreaking
Baker Hughes has developed an
advanced approach to visbreaker
optimisation by developing the
patented VisTec Coke Index (VCI)
and VisTec Stability Index (VSI).
3
The VSI is used to determine the
stability of vistar/heavy fuel as
well as the SBn and In parameters.
It uses an automated instrument
that provides very high levels of
repeatability and removes any
subjectivity with an error of 2% or
less. This test error is 10 times less
than the traditional optical micros-
copy based p-value. The method
uses light absorbance in the
near-infrared range to detect
asphaltene destabilisation.
The VSI has four main advan-
tages over other methods:
It is extremely accurate and
repeatable
It is designed specifcally for the
determination of visbreaker feed
and tar stability
It gives rapid results, allowing
refners to react quickly and main-
tain optimum operating conditions
It has been developed and
proven in use on many visbreaker
units throughout the world.
F
o
u
l
i
n
g

f
a
c
t
o
r
s

i
n
c
r
e
a
s
e

r
a
t
e
,

h
.
m
2
.

C
/
K
c
a
l

d
a
y
0.0e
+0
5.0e
7
1.0e
6
1.5e
6
2.0e
6
2.5e
6
3.0e
6
3.5e
6
4.0e
6
Vistar stability reserve, VSI
Figure 6 Heat exchanger fouling rates vs stability reserve of vistar
500
800
700
600
400
300
200
100
0 p
r
o
b
e

s
i
g
n
a
l

d
e
r
i
v
a
t
i
v
e
+100
0 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9
VisTec stability index
Destabilisation point
Figure 7 Asphaltene destabilisation point
For any blend of
processed crudes
there is an optimal
conversion for
the downstream
visbreaker
baker hughes.indd 3 12/12/2013 21:15
126 PTQ Q1 2014 www.eptq.com
Two or more year run length for
the heater
One year minimum cycle for the
visbreaker vacuum tower
No limitations in maximum
throughput
Higher conversion than in the
previous run (conversion as the
sum of atmospheric and vacuum
distillates).
The achievement of targets was
based on more severe control limits
compared to fouling indicators.
This was possible as a result of
previous experience during the rst
run and by the improved capability
to rapidly set this visbreaker unit at
the optimal severity conditions
controlled by VisTec fouling
monitoring.
All of the above targets were
achieved. The vacuum tower run
length was doubled compared to
the one year target, allowing the
ISAB renery to avoid a long turn-
around of this section of the
visbreaking unit.
Figure 9 illustrates visbreaker
conversion vs target, showing how,
on average, conversion was higher
than the target and was kept close
to the target even during the last
portion of the run. Figure 10
shows atmospheric and vacuum
conversions.
While atmospheric conversion
was rather constant, vacuum
conversion exhibited some decrease
over time due to a progressive
pressure drop increase across the
vacuum wash grid, resulting in a
lower extraction of vacuum gasoil.
This run showed very good
results in terms of conversion and
run length. An increase in conver-
sion of more than 2% was obtained
while run length was further
increased to more than 24 months
compared to the previous run (the
rst with application of VisTec
technology). To date, this is the
longest run length this visbreaker
has accomplished.
ISAB visbreaking unit constraints
The main constraints to unit run
length and conversion for the rst
run were the increase in heater
pressure drop due to foulant
deposition and the progressive
coking of the vacuum wash grid.
Visbreaker tar and heavy fuel oil
stability.
ISAB visbreaking unit targets
The rst run at the ISAB visbreaker
with the VisTec programme
had very positive results.
1
This
article refers to the second run
at the ISAB visbreaker, managed
with the VisTec programme, which
had more difcult targets:
Figure 7 shows a typical output
with the asphaltene destabilisation
point highlighted.
The VisTec Coke Index (VCI) uses
a patented technique to measure the
concentrations and size of coke
particles in the vistar. This measure-
ment gives useful indication of the
tendency to produce fouling inside
the furnace tubes at different
conversions. An example of the VCI
particle size population analysis is
shown in Figure 8.
The main applications for VisTec
analytical tools are in predicting
furnace coking tendency and the
control of:
Visbreaker preheat exchanger
fouling
Visbreaker bottom column fouling
Atmospheric and vacuum column
fouling between ash and gas-oil
extractions
6
10
9
8
7
5
4
3
2
1
N
u
m
b
e
r

o
f

p
a
r
t
i
c
l
e
s
,

t
h
o
u
s
a
n
d
s
0
1 2 3 4 5 6 7 8 9 10
Particle size, m
Figure 8 Example of the VCI particle size population analysis
C
o
n
v
e
r
s
i
o
n
,
%
2
6
/
2
/
2
0
1
1
6
/
6
/
2
0
1
1
1
4
/
9
/
2
0
1
1
2
3
/
1
2
/
2
0
1
1
1
/
4
/
2
0
1
2
1
0
/
7
/
2
0
1
2
1
8
/
1
0
/
2
0
1
2
2
6
/
1
/
2
0
1
3
6
/
5
/
2
0
1
3
Target
Figure 9 Conversion vs target
The achievement of
targets was based on
more severe control
limits compared to
fouling indicators
baker hughes.indd 4 12/12/2013 21:15
www.eptq.com PTQ Q1 2014 127
This resulted in an increased delta
P across the tower and reduced
ash and extraction of vacuum
gasoil, which in turn reduced the
contribution of vacuum gasoil to
the total conversion.
This run indicated that extending
the cycle and increasing severity
resulted in some differences
compared to the previous run. The
main limitations for this run were:
The unit was close to skin
temperature limits at the end of the
run, even if skin temperature rates
were increased reasonably little due
to the extremely long run. In the
previous run, the main limitation to
heater run length was pressure
drop
There was an increase in fouling
rates for the preheat exchangers
Vacuum column wash grid foul-
ing was, as observed in the previous
run, a concern, resulting in limita-
tions in conversion due to lower
yields of heavy vacuum gasoil.
Clearly these aspects are related
to the fact that running the unit at
very high conversion results in
very high fouling rates. However,
this is controllable through the use
of VisTec antifoulants.
There were also times, due in
part to the wide range of crude
blends processed and the frequent
switch between tanks of very
different composition, when the
unit ran for very short periods at
excessive severity.
These situations, if not properly
managed, can produce drastic and
irreversible fouling in a short time
(one or two days), compromising
the unit run length.
Renery operations developed
in-depth know-how on ways to
adjust the unit severity rapidly and
keep the unit close to optimal
conditions.
This involves optimal manage-
ment of the heater burners, allowing
the avoidance of excessive uctua-
tions in heater outlet temperatures;
rapidly managing changes in feed
rates or feed composition; and a
rapid setting of the unit heater
outlet temperature in order to meet
optimal VisTec vistar stability and
particle count parameters.
Figure 11 shows the stability
reserve measured as VSI and Figure
2
6
/
2
/
2
0
1
1
6
/
6
/
2
0
1
1
1
4
/
9
/
2
0
1
1
2
3
/
1
2
/
2
0
1
1
1
/
4
/
2
0
1
2
1
0
/
7
/
2
0
1
2
1
8
/
1
0
/
2
0
1
2
2
6
/
1
/
2
0
1
3
6
/
5
/
2
0
1
3
C
o
n
v
e
r
s
i
o
n
,
%
Atmospheric distillates
Vacuum distillates
Figure 10 Atmospheric and vacuum conversions
V
S
I

s
t
a
b
i
l
i
t
y

r
e
s
e
r
v
e
2
9
/
9
/
2
0
1
0
1
7
/
4
/
2
0
1
1
3
/
1
1
/
2
0
1
1
2
1
/
5
/
2
0
1
2
7
/
1
2
/
2
0
1
2
2
5
/
6
/
2
0
1
3
Lost conversion
Excessive fouling
Figure 11 VSI stability of visbreaker residue
V
C
I
,

w
B
P
I

u
n
i
t

v
o
l
u
m
e

c
o
n
c
e
n
t
r
a
t
i
o
n

c
o
k
e


1
8
/
1
1
/
2
0
1
0
1
7
/
4
/
2
0
1
1
1
4
/
9
/
2
0
1
1
1
1
/
2
/
2
0
1
2
1
0
/
7
/
2
0
1
2
7
/
1
2
/
2
0
1
2
6
/
5
/
2
0
1
3
Figure 12 VCI coke particles
12 shows coke particle generation
by VCI.
Running at severe conditions in
terms of heater outlet temperatures
at the limits indicated by the stabil-
ity reserve (VSI) and coke particle
generation (VCI), the throughput
was on average rather high and
baker hughes.indd 5 12/12/2013 21:15
128 PTQ Q1 2014 www.eptq.com
resulted in high required duties for
the heater and impacted skin
temperatures.
Full throughput on the unit could
however be maintained at the end
of the run (see Figure 13).
Another factor that potentially
impacts heater coking concerns the
periods with low feed rates. These
result in less than optimal velocities
and more resistance to heat transfer
with hotter oil lm temperatures
and higher rates of coke generation
on heater internal coils. (This typi-
cally happens when feed rates are
below 60-65% of maximum design
rates.) The ISAB visbreaker unit has
limitations in design that do not
allow the use of velocity steam
to compensate for reduced
throughputs.
Despite these short but critical
periods at low feed rates, the skin
temperatures were on average kept
under control.
The rate of increase in skin
temperature was modelled by the
Baker Hughes VisTec proprietary
multiple regression analysis
method that allows the effects of
changes in process condition to be
separated (feed rates and heater
outlet temperatures) from the real
rate of skin temperatures due to
coke deposition.
Using this approach, it is possible
to normalise the skin temperatures
to standard conditions of feed rate
and heater outlet temperature.
Using this model showed trends
indicated in Table 1 for this run for
the most critical skin temperatures.
The highest average rate of
increase was 0.216C/day, indicat-
ing good results. Our research
indicates that typical rates for units
without anticoke treatment and
poor management range between
0.5-1.5C/day, while results below
0.3C/day are considered to have
an optimal control over heater
coking.
When compared with the results
of the previous run (see Table 2), it
can be seen that there is a clear
improvement over skin tempera-
ture control in this run.
The good results obtained in the
latest run are also due to excellent
optimisation of burner operation
throughout the run.
300
350
250
200
150
100
50
U
n
i
t

f
e
e
d

r
a
t
e
0
1
/
4
/
2
0
1
1
1
/
6
/
2
0
1
1
1
/
8
/
2
0
1
1
1
/
1
0
/
2
0
1
1
1
/
1
2
/
2
0
1
1
1
/
2
/
2
0
1
2
1
/
4
/
2
0
1
2
1
/
6
/
2
0
1
2
1
/
8
/
2
0
1
2
1
/
1
0
/
2
0
1
2
1
/
1
2
/
2
0
1
2
1
/
2
/
2
0
1
3
Figure 13 Unit feed rate
VKTR1_9M.PV VKTIM5.PV VKTIM1.PV VKTIM3.PV
Skin modelling result soaking heating heating heating
Average coking rates, C/day 0.179 0.135 0.169 0.216
Impact of feed rate, C/(ton/h) 0.134 0.094 0.103 0.191
Heater outlet temperature impact, C/C 2.214 2.157 2.008 2.089
Skin temperature model results
Table 1
VKTIM6.PV VKTIM5.PV
Skin modelling result soaking heating
Average coking rates, C/day 0.467 0.237
Impact of feed rate, C/(ton/h) 0.206 0.111
Heater outlet temperature impact, C/C 3.324 3.791
Skin temperature model from previous run
Table 2
550
650
700
600
500
450
T
e
m
p
e
r
a
t
u
r
e
,

C
400
2
6
/
2
/
2
0
1
1
6
/
6
/
2
0
1
1
1
4
/
9
/
2
0
1
1
2
3
/
1
2
/
2
0
1
1
1
/
4
/
2
0
1
2
1
0
/
7
/
2
0
1
2
1
8
/
1
0
/
2
0
1
2
2
6
/
1
/
2
0
1
3
6
/
5
/
2
0
1
3
VKTIMS.PV heating
VKTIMS.PV heating normalised
y = 0.132x 4878.4
R
2
= 0.3851
Figure 14 Optimisation of burners and impact on skin temperature
baker hughes.indd 6 12/12/2013 21:16
KALDAIR
j zink.indd 1 12/12/2013 09:39
130 PTQ Q1 2014 www.eptq.com
Some critical skins showed an
increased trend in the rst period
and the renery was able to prop-
erly adjust burner ring and ame
D
e
l
t
a

P

t
r
e
n
d
,

m
m
H
g
1
3
/
1
1
/
2
0
1
1
1
2
/
1
/
2
0
1
2
1
2
/
3
/
2
0
1
2
1
1
/
5
/
2
0
1
2
1
0
/
7
/
2
0
1
2
8
/
9
/
2
0
1
2
7
/
1
1
/
2
0
1
2
6
/
1
/
2
0
1
3
7
/
3
/
2
0
1
3
Figure 15 Wash grid delta P trend
60
70
50
40
30
20
10
H
V
G
O
,

t
o
n
/
h
0
1
/
4
/
2
0
1
1
1
/
6
/
2
0
1
1
1
/
8
/
2
0
1
1
1
/
1
0
/
2
0
1
1
1
/
1
2
/
2
0
1
1
1
/
2
/
2
0
1
2
1
/
4
/
2
0
1
2
1
/
6
/
2
0
1
2
1
/
8
/
2
0
1
2
1
/
1
0
/
2
0
1
2
1
/
1
2
/
2
0
1
2
1
/
2
/
2
0
1
3
Figure 16 HVGO yield
1
3
/
1
1
/
2
0
1
1
1
2
/
1
/
2
0
1
2
1
2
/
3
/
2
0
1
2
1
1
/
5
/
2
0
1
2
1
0
/
7
/
2
0
1
2
8
/
9
/
2
0
1
2
7
/
1
1
/
2
0
1
2
6
/
1
/
2
0
1
3
7
/
3
/
2
0
1
3
35
40
30
25
20
15
10
W
a
s
h

o
i
l

r
a
t
e
,

m
3
/
h
5
Figure 17 Wash oil rates
distribution in order to minimise
the coking at those locations. An
example of starting from the
middle of the run is shown in
Figure 14. Heater optimisation
resulted in a reduced increase trend
for one of most critical skins.
Some critical skins showed an
increased trend in the rst period
and the renery was able to prop-
erly adjust burner ring and ame
distribution in order to minimise
the impact of bad heat ux distri-
bution on skin temperature increase
rates.
Another important aspect of the
unit is the vacuum wash grid. This
is a critical section in terms of foul-
ing. Wash grid fouling limits the
vacuum at the ash zone due to
increased pressure drop, resulting
in less extraction of heavy vacuum
gasoil (reduced conversion), and
nally it requires the grid to be
replaced with a clean one. This is a
long operation, typically requiring
a one month shutdown of the
vacuum section.
The ISAB vacuum section is
managed with a lower than typical
level of wash oil that is sent back to
the wash grid in order to maximise
conversion.
In order to mitigate fouling, an
asphaltene dispersant is dosed with
the wash oil, resulting in partial
dispersion of vapour entrained
asphaltenes deposited on the grid
before they convert to coke.
As the dispersant also has some
cleaning effect, it also improves the
removal of asphaltenes/coke by the
wash oil in addition to stabilising/
dispersing asphaltenes.
The delta P trend, normalised to
constant heavy vacuum gas oil
(HVGO) yield, is shown in Figure
15. It is interesting to note how
onset of fouling leads to rather
rapid and irreversible fouling. This
happened around the beginning of
May 2012.
Despite the increase in delta P,
the renery was able to maintain a
good yield of heavy vacuum gasoil
(see Figure 16). This result was
achieved due to a progressive
reduction of wash oil. When the
HVGO grid is partly fouled, wash
oil can produce ooding and may
not be effective in removing depos-
its, increasing delta P and possibly
allowing deposits to degrade on the
grid. In these situations, a
controlled decrease in wash oil can
baker hughes.indd 7 12/12/2013 21:16
www.eptq.com PTQ Q1 2014 131
3 Respini, Jones, Spanu, Sesselego, Avoiding foul
play, Hydrocarbon Engineering, Nov 2006.
Matteo Virzi is Senior Technology Manager
with ISAB Priolo renery in Italy. He is an
expert in distillation and thermal conversion
processes. With more than 20 years of
experience in technology, operations and
automation, he holds a degree in chemical
engineering from the University of Palermo.
Email: mvirzi@isab.com
Marco Respini is a Senior Technology Expert
with Baker Hughes Downstream Chemicals,
specialising in renery and petrochemical
process improvements in fouling control.
He has 15 years of rening experience and
is currently involved in developing new
technologies for improving renery conversion
processes. With extensive experience in
asphaltene related problems in oil production
and rening, he is an inventor of ve US patents
and has published 10 technical papers and
seven conference papers on visbreakers and
heavy fuel oil stability problems. A graduate
of Milan University with a degree in industrial
chemistry, he has been a Research Fellow in
the eld of organometallic catalysts and is a
registered professional chemist in Italy. He is
also a member of ACS and NACE.
Email: marco.respini@bakerhughes.com
which provided refnery personnel
with the right information to be
able to constantly keep the unit at
the best process severity for any
processed feed.
VisTec is a mark of Baker Hughes Incorporated.
References
1 Petralito G, Respini M, Achieving optimal
visbreaking severity, PTQ, Q1, 2010.
2 A Phase Separation Kinetic Model for Coke
Formation, Preprints ACS, Div. Pet Chem, 38,
428-433, 1993.
result in better handling of fouling,
and at the same time increase
HVGO directly (less gasoil recycled
as wash oil). Figure 17 shows the
trend of wash oil rate.
The decrease in wash oil rate
was continuously optimised. The
HVGO was analysed on a daily
basis with respect to the level of
contaminants, using the VCI tech-
nique that measures the coke
particles entrained within the
HVGO (thus not removed by wash
oil).
Also, the rate of antifoulant on
the wash oil was optimised and
increased when needed.
Conclusions
This article describes an example of
high conversion visbreaking with
increased run length when
compared to typical visbreaker run
lengths.
This successful result was made
possible by coupling Baker Hughes
VisTec anticoke/antifoulant treat-
ments with monitoring technology,
A controlled decrease
in wash oil can result
in better handling of
fouling, and at the
same time increase
HVGO directly
B
e
s
t
V
a
lv
e
s
s
in
c
e
1
8
6
7
baker hughes.indd 8 13/12/2013 11:38
Long-time customer and partner JX Holdings Inc.
from Japan orders anniversary HydroCOM
www.hoerbiger.com
For more information:
compressor-mechatronics@hoerbiger.com
Experience real capacity control and energy savings
HydroCOM is an efcient, stepless, dynamic and fully-automated control system. Several different methods
are available for capacity control for reciprocating compressors.
Only HydroCOM, however, achieves best results in terms of energy savings, speed and accuracy. It improves
process control, optimizes performance and pays for itself within a short period of time.
HydroCOM
the 1000th!
HydroCOM_CE_Buyers_guide.indd 1 11.12.2013 14:45:40 hoerbiger.indd 1 11/12/2013 14:59
Troubleshooting renery equipment with
multiphase CFD modelling
M
uch of the common process
equipment in refneries
today was designed and
built according to traditional empiri-
cal design methodologies that were
developed decades ago. Without an
intimate understanding of the
complex fow patterns present
within a given system, designers
had to rely on conservative assump-
tions and trial and error to ensure
that equipment met design require-
ments. Modern computational fuid
dynamics (CFD) tools allow design-
ers to pull back the veil on complex
internal fows, but their use has
been limited by available computing
power. As computing power contin-
ues to increase, CFD is becoming a
practical tool for industrial scale
problems. Through a deeper under-
standing of standard process
equipment, it is now possible to
identify opportunities to improve
both the function and the capacity
of installed systems.
In many cases, small modifca-
tions can eliminate the need to
design and fabricate new equip-
ment, resulting in signifcant cost
savings without compromising
performance. This article will
discuss several cases demonstrating
the application of CFD to tradi-
tional process equipment. Each
case presented will discuss the
motivation for the use of CFD, the
assumptions required to yield a
practical and robust CFD simula-
tion, some details pertaining to the
CFD modelling itself, and most
importantly the practical
outcome of the simulation exercise.
Central to the growing popularity
of CFD for industrial scale prob-
lems is an ability to simplify a
Computational uid dynamics is a useful and increasingly practical tool for improving
the design of and increasing the understanding of common process equipment
GRANT NICCUM and STEVE WHITE
Process Consulting Services
simulation. An extremely detailed
simulation incorporating all of the
relevant physical minutiae is of
little value if the results cannot be
interpreted and applied to solve a
real world problem. Furthermore,
the additional complexity and
computational expense required to
perform an extremely high fdelity
simulation is often unjustifable or
unattainable for many industrial
problems. In most cases, a simpli-
fed modelling approach
specifcally developed to examine
the variable(s) of interest is the
most effcient practice. By carefully
considering all of the independent
and dependent variables relevant to
the design question at hand, the
pain and expense of a CFD project
can be greatly reduced. Just because
an engineer can solve for every
possible variable throughout an
entire domain doesnt mean that
he or she should. Unnecessary phys-
ics complicates the setup
of a simulation, signifcantly
increases the computational time
required, and may decrease the
stability of the simulation to the
point where a converged solution is
impossible.
Case 1: liquid knockout drum
After performing a detailed
dynamic process simulation study
of a particular unit, it was discov-
ered that, given the right
circumstances, vapour/liquid rates
could be far above the design
capacity of an existing liquid
knockout drum. In addition to
incurring signifcant expense,
replacement of the new drum
would have been diffcult due to
space restrictions. It was hypothe-
sised that internals could be added
to the drum to adequately increase
the vapour-liquid separation. For
verifcation, CFD could be used to
confrm the effectiveness of any
design changes that would see the
drum operate while signifcantly
under-sized according to traditional
sizing methods. Particle sizes of a
certain critical diameter were
considered the break point for
effective operation of the separator.
Therefore, the CFD analysis was
used to develop and test modifca-
tions to allow the drum to
effectively trap particles with
larger than acceptable diameters
within the drum.
A brute force CFD approach to
this design problem would have
been to model all of the relevant
physical phenomena at the same
time: multiphase vapour/liquid
fow at the inlet, breakup/coales-
cence of the liquid droplets,
formation of a liquid flm on the
walls of the drum, collection/
movement of free liquid in the
bottom of the drum, and so on. The
modelling task was greatly simpli-
fed, however, by carefully
considering the variables of inter-
est. The functions of the liquid
www.eptq.com PTQ Q1 2014 133
A simplied
modelling approach
developed to examine
the variable(s) of
interest is the most
efcient practice
pcs.indd 1 11/12/2013 14:15
134 PTQ Q1 2014 www.eptq.com
fow rates proved more diffcult
than trapping the initially entrained
particles. The high velocity vapour
fow tended to re-entrain signifcant
quantities of liquid. However, a
design was developed that could
satisfy both requirements.
The use of these two complemen-
tary CFD simulations led to a robust
design and confdence that the
knockout drum can perform
adequately under the given set of
operating conditions. The use of
CFD in this case allowed many
design iterations to be evaluated
within a matter of days to arrive at
an optimal solution. For compari-
son, a sizing exercise using the
traditional sizing methodologies
was performed for a new drum to
handle the same loads. According to
traditional sizing methods, a new
drum for the same service would be
approximately three times the diam-
eter of the existing drum and would
have a tangent-tangent length
roughly equal to the current drum.
The CFD optimised internals
support the continued use of the
existing drum.
Case 2: fractionator overhead
receiver
In the case of an overhead receiver
for an FCC fractionator, CFD was
used to predict the consequences of
an existing design that appeared
intuitively fawed. The overhead
receiver was designed with the
inlet at the bottom of the drum,
presumably to reduce the complex-
ity of the large diameter piping to
the inlet. Although there was a
short inlet riser inside the drum,
the normal liquid level could easily
exceed the height of this riser.
Intuitively, it was hypothesised that
an inlet jet agitating the fuid in the
drum would interrupt the separa-
tion process and lead to high levels
of water in the hydrocarbon outlet.
CFD was used to examine the
effects of this inlet design in detail.
Horizontal gravity separators are
designed based on cross-sectional
fow velocities and fuid residence
time. Traditional methods assume
plug fow through the vessel, mean-
ing that each phase has a uniform
cross-sectional velocity equal to the
volumetric fow rate divided by the
knockout drum are to separate and
collect liquid particles, and to
prevent re-entrainment of the free
liquid phase that has collected in
the bottom of the drum. The varia-
bles that needed to be evaluated to
verify that the drum would
perform as required are the fates of
liquid particles entrained with the
gas at the inlet to the drum and the
shape/size of the stable liquid area
in the bottom of the drum.
Variables such as liquid wall flm
thickness are not signifcant to the
overall function of the drum and
were not modelled, as their omis-
sion did not signifcantly affect the
variables of interest.
Modelling was further simplifed
by segregating the variables of
interest, as they are independent of
one another. Each design option
was evaluated using one model for
particle tracking and a second for
monitoring the free liquid phase.
Although it may seem counterintui-
tive that two models would be
more effcient than one, this
arrangement allowed the designer
to perform several design iterations
using the less computationally
intense particle tracking model
before running the more complex
gas/liquid interface tracking
model. Furthermore, the separation
allowed the two models to be set
up quite differently to give the best
answers for the variables that each
was tasked with solving.
The frst of the two simulations
was used to track particles
entrained with the vapour at the
drum inlet. The discrete phase
model (DPM) was chosen in this
situation for its ability to track
particles through the domain and
because the volume fraction of
liquid entrained within the vapour
fow was low. Small, light particles
follow vapour fow streamlines
more closely than larger particles
because they have less momentum
relative to the drag caused by the
bulk vapour fow (lower Stokes
number). With this principle in
mind, a conservative particle size
should be smaller than the maxi-
mum allowable droplet size. These
conservatively sized particles were
injected with the vapour at the
separator inlet and tracked
throughout the domain as the
vapour travelled from the separator
inlet to the outlet. A simplifying
assumption was that the particles
underwent partially elastic colli-
sions (some energy lost) when they
encountered walls within the
vessel. Thus, wall collisions tended
to slow the liquid droplets down
until they separated from the
vapour fow and settled in the
bottom of the drum. In reality,
some of these collisions would have
splashed to create multiple smaller
droplets, but this phenomenon was
ignored because escape of smaller
droplets was acceptable and there-
fore not consequential to the
design. The design was modifed
and simulated iteratively until no
particles escaped through the drum
outlet.
The second simulation was
designed to model the stability of
the free liquid phase in the bottom
of the drum. This simulation
employed the volume of fuid
(VOF) model to track a well defned
vapour-liquid interface. Liquid
volumetric fow rates were signif-
cantly lower than gas volumetric
fow rates, and liquid entering and
exiting the drum was not signifcant
to the problem of maintaining a
stable liquid layer. The model was
therefore built with no liquid fow
in/out, and a mass of liquid was
manually placed within the drum at
the start of the simulation and
allowed to slosh around due to
interaction with the vapour fow. If
the fow agitated the liquid layer to
the point where liquid mass
escaped through the outlet, the
design was modifed. Stabilising the
free liquid layer with high vapour
Use of computational
uid dynamics
allowed many design
iterations to be
evaluated within a
few days to arrive at
an optimal solution
pcs.indd 2 11/12/2013 14:15
cross-sectional area through which
that uid ows. In a given uid, a
droplet of a certain size has a xed
terminal velocity. The separator
must be sized with sufcient resi-
dence time for that droplet to
traverse into the proper phase at the
outlet from any starting position at
the inlet. If the inlet design leads to
heavy agitation within the drum,
the most basic design assumption
plug ow through the cross-sec-
tion of the drum fails to hold. If
the plug ow assumption is not
met, some fraction of the liquid in
the separator may experience resi-
dence times signicantly below the
design value, and good separation
will not be achieved.
A CFD simulation was designed
to test the validity of the plug ow
assumption with the bottom inlet
design. The volume of uid (VOF)
model was chosen for this
multiphase problem. The CFD
model contained a set liquid
volume inside the drum, and the
vapour load was applied at the
inlet of the drum, escaping
through the gas outlet. Since the
design velocities within the liquid
phase of the drum are extremely
low, inclusion of liquid mass
entering and exiting the drum
would not have affected the varia-
bles of interest (anomalous
cross-sectional velocities), so this
was omitted from the simulation.
Any regions with signicant move-
ment can be assumed to be
operating away from the plug ow
design assumption. Signicant
recirculation or shortcutting within
the liquid region of the drum
should be viewed as an indicator
of poor separation that will result
in signicant amounts of water in
the hydrocarbon outlet.
The results of the CFD simulation
explained the problems that this
FCC unit main fractionator had
been experiencing. The CFD
showed that the vapour jet leaving
the riser entrained a signicant
amount of liquid. Figure 1 shows
the intense agitation of the liquid
within the drum. This created a
large recirculation current within
the liquid portion of the drum,
completely violating the plug ow
assumption used in the separator
www.eptq.com PTQ Q1 2014 135
sizing process. Based on this result,
it was condently predicted that
the liquid-liquid separation was
poor, and there was likely a large
amount of water in the hydrocar-
bon outlet, which feeds the
fractionator reux pumps. This
water contains ionic compounds
that will deposit in the tower as
salts as the water evaporates (see
Figure 2). The tower was, indeed,
experiencing issues related to salt
in the overhead system, and the
circulation of large quantities of
water in the reux is a contributing
factor. Lab samples have conrmed
that the weight percentage of water
in the reux stream is very high.
The next application of CFD to
this drum will be to correct the
deciencies in the inlet design. By
incorporating CFD into the revamp
process, modications can be tested
and an iterative design developed
to ensure that the recirculation
zones are eliminated and that the
ow in the drum more closely
resembles the plug ow assumed in
the initial vessel sizing calculations.
Correction of this problem will
eliminate the salting problems
occurring in the tower.
Case 3: FCC fractionator feed
inlet device
A common feed inlet device for
FCC fractionators is a series of
pipes intended to break up and
distribute the ow within the
column (see Figure 3). Experience
has shown that, when installed
properly, this design improves
tower performance and prevents
coking on the tower wall opposite
the inlet. A CFD simulation was
performed to understand the mech-
anisms by which this device works
and to look for simple improve-
ments to the design that could
increase performance or reduce
complexity and cost. The intuitive
theory going into the study was
that the ow was redirected
upward by the pipes, thus creating
a more uniform distribution across
the tower according to the place-
ment of the pipes. By fully
understanding the mechanisms that
allow this inlet device to achieve
good results, better design deci-
sions can be made in future
installations.
Bottom inlet
agitates liquid
in drum
Water in reflux
carries salts
Salt deposition
plugs trays
Figure 1 Contours of the volume fraction of liquid within the drum. The red region is
100% liquid, and the blue region is 100% vapour. With the liquid level above the short
inlet riser, liquid is entrained into the jet of vapour entering the drum
1.00e
+000
0.00e
+000
9.47e
001
8.95e
001
8.42e
001
7.89e
001
7.37e
001
6.64e
001
6.32e
001
5.79e
001
5.26e
001
4.74e
001
4.21e
001
3.68e
001
3.16e
001
2.63e
001
2.11e
001
1.58e
001
1.05e
001
5.26e
002
Raw gasoline
volume fraction
Figure 2 Poor separation in the overhead
receiver leaves water in the reux, which
carries ionic compounds that will deposit
in the tower as salts when the water
evaporates. These salts can cause corrosion
and plugging of tower internals
pcs.indd 3 11/12/2013 14:16
136 PTQ Q1 2014 www.eptq.com
deected near a pipe was easily
corrected back into the mean jet
ow by the momentum of the
surrounding jet that had not
directly impacted one of the distrib-
utor pipes.
If the pipes did not effectively
redistribute the feed within the
tower, how did they decrease coke
formation on the opposite wall of
the tower (the stagnation region)
as seen in the eld? The CFD
shows that the pipes acted to scat-
ter the jet on the wall of the tower
rather than to distribute the feed.
As the jet owed around the pipes,
it became less dened, meaning
that the stagnation region was not
as apparent as it was without the
pipes. The stagnation region was
broken up over the back wall of
the tower rather than remaining
stably focused on a single point.
The disruption of the stagnation
region greatly reduced coke forma-
tion for two reasons: 1) the liquid
impacting the back of the tower
did so over a larger area, concen-
trating less material on one spot;
and 2) the area immediately
surrounding the impingement
point had the highest wall shear
stresses of anywhere on the tower
wall. As the impingement region
was less dened, the shear stresses
were more evenly distributed,
contributing to the shearing of any
coke.
Droplet breakup is another factor
that may contribute to a reduction
in coke formation with the inlet
device installed. As the vapour
owed around the pipes, larger
droplets with high momentum
could not change direction fast
enough to avoid a collision with
the pipes. High impact velocities
and high wall shear forces around
the pipes caused the droplets to
shatter and break up into many
smaller droplets. These smaller
droplets had lower Stokes numbers,
meaning that their momentum was
less signicant in relation to the
drag caused by the vapour, and the
droplets were more likely to follow
the path of the bulk vapour ow
rather than impacting the rear wall
of the drum. By impacting and
shattering the droplets on the
pipes, where wall shear is at its
high liquid impact rate combined
with near zero vapour velocities,
and low wall shear stress created
favourable conditions for coke
formation in the stagnation region.
Contrary to expectations, the
addition of the inlet distributor to
the model did not greatly improve
tower cross-sectional ow distribu-
tion (see Figure 4). The simulation
nonetheless provided evidence to
explain the decreased coke forma-
tion on the wall opposite the inlet.
The inlet device caused the inlet jet
to widen more in the horizontal
direction, but had little effect on the
vertical size of the jet signicant
vapour quantities were not redi-
rected upwards into the tower with
each impact. This was due to each
pipe only blocking a small part of
the ow. Any vapour that was
The vapour was modelled as
incompressible due to the relatively
low pressure drop in the domain.
Detailed simulations were run with
and without the inlet distributor,
and the results were compared in
an effort to nd meaningful differ-
ences that would explain why the
distributor had such a noticeable
effect when installed. The discrete
phase model (DPM) was used to
simulate the fate of liquid droplets
entrained with the feed. With no
inlet distributor, the CFD model
predicted that the inlet vapour
would form a well dened jet
before impinging on the opposite
wall of the tower. Above the jet, the
ow upward through the tower
was far from uniformly distributed.
Furthermore, the stability of the jet
suggested that the impingement
point on the wall opposite the inlet
was unlikely to move over time.
This created a stagnation region at
the impingement point with essen-
tially zero ow velocity and no
wall shear stress.
While the impingement point
saw little vapour ow velocity, the
rate at which it was impacted by
any liquid particles in the feed was
high. If these particles were suf-
ciently large (high Stokes number),
the particles tended to detach from
the vapour ow and follow a trajec-
tory dictated by their momentum.
Thus, as the vapour streamlines
turned abruptly at the stagnation
point, the liquid continued along
the path of the jet and impacted the
wall in the stagnation region. The
Reactor
effluent
Quench
Inlet device
Figure 3 Field experience has shown that
installing a feed inlet device similar to the
one shown improves performance and
reduces coking on the rear wall of the
tower
0.00e
+000
1.07e
+002
1.00e
+002
9.33e
+001
8.67e
+001
8.00e
+001
7.33e
+001
6.67e
+001
6.00e
+001
5.33e
+001
4.67e
+001
4.00e
+001
3.33e
+001
2.67e
+001
2.00e
+001
1.33e
+001
6.67e
+000
Velocity, ft/s
Figure 4 Velocity prole on a vertical plane within the tower. The case with the inlet
device (left) is similar to the base case (right). Signicant quantities of vapour are not
directed upward by the inlet device
pcs.indd 4 11/12/2013 14:16
www.eptq.com
maximum, rather than in the stagnation region at the
point of jet impingement, coke formation is mini-
mised. Armed with an understanding of the
fundamental mechanisms behind this style of inlet
device, future designs can be adjusted to maximise
the positive functions while minimising negative
consequences.
Conclusion
CFD has been applied to these case studies to gain a
better understanding of the phenomena that govern
the performance of examples of common process
equipment. The designer can take advantage of this
better understanding of the underlying physical
phenomena to improve performance. For example, in
the case of the FCC unit fractionator inlet distributor,
understanding that droplet breakup and jet scatter-
ing are the two key principles that allow this device
to prevent excessive coke formation will lead to
future designs that specifcally maximise these
phenomena rather than designs that attempt to
improve the device through misguided attempts to
affect the distribution of the fow within the tower.
CFD is an extremely useful and increasingly practi-
cal tool for improving the design of and increasing
the understanding of common process equipment. By
carefully defning a problem and the variables of
interest, simulations can be built that give good infor-
mation where necessary without incorporating
unnecessary complexity. Finally, it is worth mention-
ing here that CFD inputs, assumptions and results
should be carefully scrutinised. Commercial CFD soft-
ware will work dutifully on any problem that it is
given, and will often reach a converged answer, but
the software has no notion of correctness. Incorrect
inputs or the failure to include important physical
aspects of a simulation will lead to incorrect answers.
Part of the engineers responsibility in running a CFD
simulation is to consider all of the trade-offs between
complexity and accuracy so that an adequate solution
can be developed with the minimum level of compu-
tational expense. The ultimate usefulness of a
simulation lies in an engineers ability to use the
information to draw concrete conclusions and develop
solutions that improve performance.
Grant Niccum is a Process Engineer with Process Consulting Services,
Inc. in Houston, Texas. Process Consulting Services provides grassroots
and revamp front-end process engineering to the renery industry
worldwide.
Steve White is a Chemical Engineer with Process Consulting
Services. He has more than 37 years of process design experience in
renery revamps and grassroots units including crude/vacuum, FCC,
hydrotreater, alky, butamer, reformers and others. He previously worked
for Jacobs Engineering, UOP and ARCO.
SAMSON AG MESS- UND REGELTECHNIK
Weismllerstrae 3
60314 Frankfurt am Main Germany
Phone: +49 69 4009-0 Fax: +49 69 4009-1507
E-mail: samson@samson.de
Internet: www.samson.de
SAMSON GROUP www.samsongroup.net
A
0
1
1
2
0
E
N
Partner with
the Best
With over 50 independent subsidiar-
ies and more than 220 engineering
and sales ofces spread across the
world, SAMSON ensures the safety
and environmental compatibility of
your plants on any continent.
To offer the full range of high-quality
control equipment used in industrial
processes, SAMSON has brought
together highly specialized compa-
nies to form the SAMSON GROUP.
pcs.indd 5 11/12/2013 14:16
C
M
Y
CM
MY
CY
CMY
K
SOGAT2014 Ad1 Full Page Bleed 303mmX216mm for PTQ.pdf 1 12/3/2013 12:06:10 PM
sogat.indd 1 09/12/2013 11:32
High-delity operator training simulators
T
he current trend in operator
training simulators (OTS) is to
connect a high-fdelity
dynamic simulation model to a
copy of the actual plant distributed
control system (DCS) that is
running the same software as the
plant and uses the same
human-machine interface (HMI).
This approach is attractive and
usually offers the best opportunity
to produce a high-fdelity OTS that
will maximise training value.
However, success is not guaranteed
by the project model, and there are
often some complex and diffcult
issues to solve as part of the project
delivery, particularly for existing
plants where the process and
control dynamics are well known.
Project phases
Project award only arrives after a
lot of preliminary work by the sales
team and the buyer. Issues are
discussed and decisions taken.
Project award typically marks the
end of this phase and is usually a
happy occasion. Those involved
(vendor and buyer) will congratu-
late themselves on their decisions
and look forward to an OTS that
meets all their expectations.
Thereafter, the project implementa-
tion team will be put to work
following a fairly standard project
model (see Figure 1).
The frst major implementation
task is model development. A new
dynamic process model will require
an enormous amount of data to be
processed and confgured within a
model framework. This needs the
attention of simulation experts in
order to ensure robustness, accu-
racy and adequate simulation
Integrating a high-delity simulation model with a real plant control system
requires some nessing to produce a true high-delity operator training simulator
MARTIN SNEESBY
APESS
speed. Several months (or more)
are likely to be needed. Short-cuts
are rarely available or effective
because they almost inevitably
reduce robustness and/or accuracy
and, therefore, compromise the
projects objectives.
A pre-existing simulation model
can be used, but it is still likely to
need signifcant work to adapt it to
the specifc OTS requirements,
current plant confguration and
operating conditions, and/or to
make it consistent with the current
simulation technology. Reuse of an
existing model also risks depriving
the project of the deep process
understanding that comes from
building a high-fdelity model.
Process understanding is almost
always crucial to any problem solv-
ing on an OTS, including issues
with the DCS and ESD functional-
ity, and it is critical this is not
ignored. Common mistakes
include:
Separating the model builders
from those with real plant
experience
Trying to apply a one size fts all
execution model
Ring-fencing parts of an existing
process model against any change.
These approaches should be
avoided, even if they appear super-
fcially cost effective.
Pre-existing models often come
with baggage, too. Rose-tinted
www.eptq.com PTQ Q1 2014 139
Delivery
(FAT and SAT)
Build high-fidelity
dynamic
simulation model
Configure plant
DCS for OTS
Validate
model
OR
Integrate
components
Test
integrated
system
Configure or
emulate plant
ESD
Modify existing
high-fidelity dynamic
simulation model
Figure 1 Standard project model for an operator training simulator
140 PTQ Q1 2014 www.eptq.com
will conclude that the tuning is
right so that the model is wrong.
The simulation engineer will then
be asked to fx the high-fdelity
dynamic simulation model to better
match actual plant performance.
And this is where the problems
really start.
Process control engineers tend to
have the best understanding of the
offset between a model and reality.
They are often reluctant to use the
results of a dynamic simulation
model to build or tune their
controllers. Instead, they prefer to
use the plant as a test bed and tune
their controllers to the actual plant
responses. In most cases, this
approach is pragmatic, sensible and
effcient (although it is often possi-
ble to get good, noise-free data
from a high-fdelity simulation,
too). What should be clear is that
combining plant controller tuning
constants with a simulation model
built from theory is not guaranteed
to work perfectly. However, despite
this, it often comes as a surprise to
the project team and is unlikely to
have ever been discussed back at
the post-award dinner.
Operators (and engineers) tend to
look at the overall system response
when comparing a simulator to
their reality. Both the model and
the control system contribute
substantially to the overall
response. They are complementary
rather than supplementary. When
tuning a controller, the process
control engineer will look for a set
of tuning constants that work best
with the plant responses, with a
goal of achieving a quarter-decay
response or some other (arbitrary)
target that makes sense at the time
and for that particular system.
There is not a right answer to
controller tuning. It is a problem to
be optimised against a wide range
of constraints. If the simulation
model differs from plant behaviour,
a different set of tuning constants
will be optimal.
At this stage of the project, it
might be time for some introspec-
tion. The use of the actual plant
controllers (or simulated version
thereof) and actual plant tuning
constants is a good thing, but the
real goal should be the overall
glasses can make old models unre-
alistically attractive, even though
the project team accepts that there
have been changes to the plant,
changes to simulation software and
the need for a new project at all.
Too many projects have foundered
on the unrealistic assumption that a
pre-existing model will be good
enough to meet elevated standards.
The dynamic process model must
then be validated. Tests are usually
set up on the standalone model to
prove its accuracy to one or more
steady state conditions before test-
ing the models dynamic behaviour.
The model responses should be
reasonable and realistic for a wide
range of transients. Judging reason-
ableness and realism is diffcult.
Typically, experienced operating
staff are asked to make this assess-
ment and, while this is usually the
best option, it is not quantifable
and is subject to preconceptions
and other human factors. For exam-
ple, it is very common to have
experienced operators reporting
that a model reacts more slowly
than the real plant. This misconcep-
tion is well known by the idiom
that a watched pot never boils. Real
plant data that are free of excessive
noise and unmeasured disturbances
would be ideal, but is diffcult to
collect and is not usually included
in the project planning.
Model validation is, occasionally
(and wrongly), considered an
unnecessary luxury. Maybe those
involved are suffciently confdent
of the models fdelity, or they feel
the schedule is more pressing than
quality control. However, skipping
model validation (or failing to take
suffcient notes during the process)
is usually the prelude to a projects
disaster.
In parallel with model develop-
ment (or starting slightly later) is
the preparation of an OTS version
of the real plant DCS. This step is
necessary to incorporate basic OTS
functionality such as start, stop,
save, load and possibly faster than
real-time operation. It may also be
a necessary (or cost effective) mech-
anism to ensure the project
hardware costs are minimised so
that redundancy (unnecessary in an
OTS) is not built in accidentally.
Not all vendors have the same
approach to preparing the OTS
version of the real plant DCS, and
it is not guaranteed that all aspects
of the DCS functionality will be
fully supported. In particular, logic
sequences and alarm monitoring
systems should be scrutinised to
ensure that they function correctly
under all circumstances, including
save and load. This phase is not
trivial, and the effort involved
should be measured against the
likely statement of requirements
that it should be possible to quickly
and easily import that current plant
DCS database into the OTS.
So far, so good. The project has
successfully developed a high-
fdelity dynamic simulation model
of the process and an OTS version
of the real plant DCS. The integra-
tion of these two systems will
typically take a few weeks to a few
months. This task is predominantly
mechanistic, although there will
always be a set of more taxing
problems to solve, and the diff-
culty should not be dismissed
lightly. Care needs to be taken to
ensure that signals have the right
sense (on/off or healthy/
unhealthy), control actions and
valve actions match, and signals
arrive at the correct landing site
with the DCS. The complexity of
this task is often multiplied by the
need to include an ESD system that
is tightly coupled to the DCS, or to
manipulate process signals to meet
plant formats. Aspects of this
project phase can be automated,
but only after establishing the rules
that apply for this particular
project.
Surprises
The fnish line would seem to be in
sight. However, it is at this stage of
the project that someone will
suddenly discover that a particular
control loop on the integrated
system (model plus simulated DCS)
is unstable or slow to converge.
They will be disappointed and an
investigation will begin. The frst
port of call is likely to be checking
the controller tuning constants,
perhaps by ringing the control
room to make sure the project team
has the latest data! Next, someone
system response. Exactly how good
is the high-fdelity dynamic simula-
tion model?
Sources of model error
High-fdelity dynamic simulation
models are now widely accessible.
Computing power has increased
and the tools have become easier to
use. The unit operation models
typically contain the best available
solutions. The old adage of garbage
in equals garbage out still applies,
but simulation software, when used
by the right people and in the right
way, is more powerful than ever.
Engineers should be able to trust
the results from well built models.
Despite this, there remains a (small)
gap between a model and reality.
The size of the gap (the model
error) might be decreasing, but it is
foolish to ignore it.
The process modelling engineers
that build and confgure these
models are required to accept and
make many assumptions and esti-
mates. The better modelling
engineers will be inherently aware
of the assumptions they are making
(either explicitly or implicitly) and
will be well equipped to make
appropriate engineering estimates
where necessary. However, this is
not always suffcient to match plant
behaviour for many reasons (see
Figure 2). These include:
Imperfect unit operation models
that do not capture all of the real
process behaviour and/or the real
process dynamics
Invalid modelling assumptions,
either explicitly made by the
modelling engineer, or implicit to
the modelling approach
Inadequate physical properties
and/or thermodynamic methods
Inaccurate design performance
data (ie, the unit does not operate
exactly according to design), or
degraded system operation (fouling
or other limitations)
Inaccurate data estimates
Unmodelled malfunctions and
phenomena (fooding, side reac-
tions, and so on)
Stability and robustness
constraints imposed by the simula-
tion mathematics
Misleading plant data that
suggest correlations between
www.eptq.com PTQ Q1 2014 141
actions and effects where no direct
connection exists or responses
infuenced by unmeasured (or
unobserved) disturbances.
The basic unit operation models
have changed very little over the
last 50 years. This is not to say that
they are already perfect, but it
suggests that we have found ways
to live with their weaknesses,
particularly for equipment design,
which remains the dominant appli-
cation of process simulation
technology.
The equilibrium stage model is a
good example of both the strengths
and weaknesses of the established
approaches to unit operation. It
includes assumptions about perfect
mixing and instantaneous
vapour-liquid equilibrium, which
we know to be dubious, and can
use effciency factors that are hard
to justify or specify a priori. Despite
the limitations, the equilibrium
stage model has a solid track record
in process design and it remains a
cornerstone of process simulation.
However, there is less evidence that
it meets all the demands of a high-
fdelity dynamic model. For exam-
ple, variation in throughput with
stage effciency (or weir height) is
often seen on process plants, but is
not predicted. Similarly, fooding
(like surge on a compressor)
involves pseudo-random phenomena
that are beyond most models. Even
the hold-up on a tray can be diff-
cult to predict accurately because of
the need to calculate frothing
factors and so on. Nevertheless,
distillation models based on the
equilibrium stage model usually
predict time constants that are good
enough for most purposes, except
possibly advanced controller
design, where step testing is still
the de facto approach.
It is easy to overlook the thermo-
dynamic models used in a
simulation as a potential source of
uncertainty. Hopefully, the model-
ling engineering will be suffciently
skilled to choose an appropriate
method, but the results still contain
a margin of error. Thermo will
generally affect process gains rather
than time constants, but may affect
both in distillation and some other
systems.
Many models are based on
assumptions of lumped (homoge-
neous) volumes even though we
know that perfect mixing is
unlikely to be realised. Pure dead-
time is often poorly captured
because of this assumption.
However, lumped volumes are
necessary to meet speed and
robustness demands.
Equipment is often designed and
delivered with a margin of error.
We can never be absolutely sure
how an item of equipment will
perform. Even a simple valve is less
Unskilled
simulation
engineer
Misleading
plant data
Hidden
assumptions
Inadequate unit
operation models
Equipment
design margins
Poor engineering
assumptions
Estimated
data
Inaccurate
VLE
Model
error
Figure 2 Sources of plant model mismatch with high-delity OTS models
142 PTQ Q1 2014 www.eptq.com
gain and low integral time (high
reset rate). Other loops are more
complex and require the right
combination of gain and integral
time to match the process (see
Figure 3).
Figure 4 shows the effect of the
process gain and time constant on
the overall loop performance for a
specifc set of controller tuning
parameters (responses are indica-
tive only). The situation is the
inverse of the controller tuning. If
the model time constant is much
less than the actual process time
constant, the loop may be unstable,
slow to settle or very sluggish.
Mismatches in gain also affect the
response, but less critically if the
process gain is predicted
accurately.
It should also be noted that loops
with little apparent dead time are
Henry Kister are especially note-
worthy. The prevalent message in
these books is that things can go
wrong on an operating plant.
Sometimes, process equipment
does not operate the way it was
intended or designed. So why
would we expect all models to be
perfect over all operating condi-
tions? Of course, we do not. We
just occasionally get seduced by the
accuracy of high-fdelity models
and forget that all model predic-
tions are subject to uncertainty.
Controller tuning parameters
Process control engineers under-
stand that all control loops have to
be tuned for their specifc charac-
teristics. Some loops are simple and
need little attention. For example, a
simple fow controller is almost
guaranteed to work well with a low
than defnitive different valve
manufacturers will use different
equations to calculate fow, and the
results can vary by more than 10%.
Heat exchangers and distillation
columns have similar degrees of
uncertainty associated with the
design and subsequent
performance.
Plant data are usually held up to
be fact. However, transmitters and
measurements are not infallible.
Uncertainty is always present.
Sometimes plant measurements can
mislead and suggest phenomena
that are not actually there. Data
reconciliation is one technique for
guarding against these errors, but it
is not always viable.
Many excellent chemical engi-
neering books have been written on
troubleshooting plant problems.
Books by Norman Lieberman and
0.5
0
R
e
s
p
o
n
s
e
0.5
0 1 2 3 4 5 6
Time
1
Kc = 5
Kc = 3
Ti = 4 Tp = 20
td = 4
Kp= 1
Ti = 11 Ti = 20
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
3
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
2
0.3
0.4
0.2
0.1
R
e
s
p
o
n
s
e
0.1
0.2
0
0 50 100 150
Time
4
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
6
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
5
0.3
0.4
0.2
0.1
R
e
s
p
o
n
s
e
0.1
0.2
0
0 50 100 150
Time
7
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
8
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
9
Kc = 1
Figure 3 Predicted control responses with various controller parameters for a FOPDT model with process gain (K) = 1, process time
constant (Tp = 20) and process dead time (td = 4)
www.eptq.com PTQ Q1 2014 143
In short, tuning the model to
match a desired performance is
diffcult to impossible.
The pragmatic engineering solu-
tion is to use a modifed set of
controller tuning constants so that
the overall process response
(process dynamics plus control
action) is closer to reality. This adds
a small amount of complexity to
any future upgrades to the simu-
lated DCS, but that process is
already suffciently complicated
that it should only be realistically
done no more often than once per
year.
For the examples in Figure 3 and
4, the central trend (#5) represents
a controller that is well tuned to the
process. The overall response is
good. However, if there is a
mismatch between the process time
constant predicted by the model
and that seen on the actual plant
tempted to modify the tested model
with a gain or bias on a particular
instrument, or to add a lag (or lead)
to the system to meet the immedi-
ate need. These actions, even if
superfcially effective, are unlikely
to address the real problem and
often create further issues in them-
selves because a high-fdelity OTS
needs to work over a very wide
range of operating conditions, and
this type of tuning at one operating
point could be detrimental at
another.
If quick fxes are excluded, the
modelling engineer must review
the process model with a fne-
toothed comb. This involves
checking and rechecking all the
input data (thousands and thou-
sands of parameters) and assessing
all the implicit assumptions. A
solution may or may not be availa-
ble via this route.
more robust against model control-
ler mismatch than those with more
dead time. Unfortunately, model
predictions of apparent dead time
tend to be much less accurate than
those for gain and the process time
constants.
Solutions
A high-fdelity process simulation
is built from unit operation models
that are based on accepted theory
and confgured with design perfor-
mance parameters. These models
are said to be predictive because
they do not rely on operating data
for their development. Results are
not regressed into the model. There
are no tuning parameters to quickly
change the gain or time constant of
a particular loop. This, in itself,
might come as a surprise to some
members of the project team.
The simulation engineer may be
Kc = 3
Ti = 1
0.5
0
R
e
s
p
o
n
s
e
0.5
0 1 2 3 4 5 6
Time
1
Kp = 1.5
Kp = 1.0
Tp = 10 Tp = 20 Tp = 30
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
3
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
2
0.3
0.4
0.2
0.1
R
e
s
p
o
n
s
e
0.1
0.2
0
0 50 100 150
Time
4
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
6
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
5
0.3
0.4
0.2
0.1
R
e
s
p
o
n
s
e
0.1
0.2
0
0 50 100 150
Time
7
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
8
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
9
Kp = 0.5
Figure 4 Predicted control responses with xed controller parameters and variable process responses
144 PTQ Q1 2014 www.eptq.com
might make the model unstable
under some conditions!
Transmitter time constants are
typically estimated. In fast loops,
these time constants may be signif-
cant in the overall loop response.
The approach to equilibrium (for
example, effciency factors) can also
affect responses. Again, these
values are typically estimated and
some tuning can be justifed.
Metal masses are important for
start-up and cool-down, but do not
generally have a substantial impact
on process dynamics around the
normal operating point. Heat
losses to ambient should be treated
similarly there is uncertainty in
these calculations, but excessive
tuning is not warranted because
they only affect long term dynam-
ics and has negligible impact on
the model mismatch at normal
operating conditions.
Tuning should always be sensi-
ble. Changing a value by 10% is
almost always okay. Changing a
good estimate by 50% might be
acceptable sometimes. Changing
estimates by orders of magnitudes
suggests either very poor engineer-
ing in the model construction phase
or that the engineer is tuning the
wrong input altogether.
Conclusions
Model building is a skilled exercise
and should be treated as such. A
high-fdelity model requires deep
process understanding and experi-
ence. It is not an off the shelf item.
Modelling engineers must recog-
nise the sources of modelling error
and act to limit them. There can be
a reluctance to tune predictive
models (those built only from data
inputs), but some selective tuning,
if done with sound engineering
judgement (that is, choosing the
right things to tune), can be benef-
cial. Engineering is reliant on
theory, but the best engineers are
also pragmatic and results oriented,
and should recognise occasions
where empirical evidence (data)
trumps the accepted theory.
The overall system behaviour
should be considered as the highest
priority, and is certainly more
important than any specifc data
set. Plant tuning constants will
process dynamics is likely to be
small.
If the problem is an unstable
controller, it is necessary to look at
the model parameters that directly
affect the process dynamics. First
and foremost are the system
volumes. Volumes are often esti-
mated inaccurately because they do
not affect the steady state solution
and larger volumes help model
stability. It is tempting to over-
estimate volumes in order to make
the model robust. A corollary of
this is that reducing model volumes
and we insist on using the actual
plant tuning constants, the
responses will certainly differ from
the plant and will possibly show
signs of instability if the plant
controller is tuned aggressively.
However, if the simulated control-
ler can be recalibrated to match the
model dynamics, the overall
response can be signifcantly
improved.
It is important to realise the limi-
tations of retuning controllers. If
the model is signifcantly slower to
respond than the real plant, no
amount of controller tuning is
going to result in a response as fast
as the plant. However, if a loop is
stable on the real plant and unsta-
ble in the OTS (with plant tuning
constants, see Figure 5) then retun-
ing the OTS controllers can
certainly help. Of course, this
should always be considered in
conjunction with the possibilities of
retuning the model, as discussed
below.
Model tuning opportunities
Despite all of the above, which
hopefully comes across as logical
and reasonable, some project teams
still insist that the plant tuning
constants should be treated as
sacrosanct and that it is up to the
modelling engineer to fx the
model. What, realistically, can be
done?
Some data are more solid than
other data. Recognising the more
questionable data is a good starting
point for model tuning. Avoid
adjusting anything that is held to
be true, including the laws of phys-
ics and the frst law of chemical
engineering (the mass balance).
The operating plant often
provides clues: an exchanger seem-
ingly operating at better than
design performance, a compressor
not delivering the expected head,
possible fooding in a column, and
so on. These issues can often be
addressed by tuning the equipment
parameters, for instance, (increas-
ing the specifed UA, adjusting
compressor factors) so that a
closer match to the steady state
condition is achieved. The process
gains can be manipulated via
this route, but the impact on the
0.5
0
0.5
0 1 2 3 4 5 6
Time
Plant behaviour
Response of high-fidelity model
with plant tuning constants
Response of high-fidelity model
with calibrated tuning constants
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
0.3
0.4
0.2
0.1
0.1
0.2
0
0 50 100 150
Time
K
Figure 5 Potential benets of retuning
controllers to work with a high-delity
simulation instead of the actual plant
dynamics
ww.eptq.com PTQ Q1 2014 145
responses that a high-fdelity model
with a non-complementary control
layer imported from the operating
plant. Use the high-fdelity control
emulation, but be tolerant of the
process model and accept that
some controller tuning is likely to
be benefcial.
OTSs will continue to become
more accurate, but this will not be
achieved through software devel-
opments alone. The more detailed
the simulation, the more data and
understanding that it holds, and,
consequently, the more engineering
effort required. As with any engi-
neering activity, the time and effort
should be assessed against the
value. The best OTS will not be the
best choice for everyone.
Martin Sneesby is an independent dynamic
simulation consultant with more than 20
years of experience in process simulation and
modelling, including many operator training
simulators and detailed engineering studies.
Email: martin.sneesby@apess.co.uk
system is complementary to a plant
or simulation. A low-fdelity
process model with a complemen-
tary (well tuned) control layer
might provide better overall
often need adjustment to work well
with even the most accurate simu-
lation models. Failure to allow for
controller tuning on the simulator
system may be detrimental to the
overall projects success.
Prognostications
High-fdelity process simulation
models are already very accurate, if
built well with the right data and
good engineering assumptions, and
offer a lot of value, particularly for
OTSs. However, not everything can
be modelled perfectly. Modelling
tools and principles can still be
improved. There is still room for
further development to reduce the
granularity of models (such as
smaller volumes, smaller time-
steps, more detailed unit operation
models), although there are dimin-
ishing returns.
The increase in fdelity of DCS
emulation that is achieved by using
real hardware and control confg-
urations is laudable, but it should
be remembered that a control
Green-eld projects share many of the
same characteristics as OTS projects
built for existing plants, but they tend
to be less technically challenging
and more schedule driven. This is
because there is not the same wealth
of detailed knowledge about the
operating plant and assumptions and
estimates become more acceptable.
Data are usually more readily available
and reasonableness will be judged by
less stringent criteria. The project team
generally recognises that value is more
closely linked to the delivery date
(to maximise training time) than it is
to accuracy. Indeed, the incremental
benet of a high-delity OTS for
training is small much can be
achieved with relatively simple, stable
and robust models.
WWW.ZWICK-ARMATUREN.DE
HIGH STANDARD VALVES
FOR NON-STANDARD CONDITIONS.
Dont miss out
on the industrys
most important
event of the year!
Hyatt Regency
Orlando (formerly
The Peabody Hotel)
Orlando, Florida
March 23-25, 2014
Be part of the action at the 112th
AFPM Annual Meeting where the
rening industrys most
important players gather to
network for three days of insight,
innovation and solutions. Explore
the challenges and opportunities
that are on the rening industrys
horizon at AFPMs Annual
Meeting in Orlando!
Registration is now open.
afpm.org
EXTEND
YOUR REACH
ATTEND THE 2014
AFPM ANNUAL MEETING
AFPM_2014_AM_PTQ_PRINTER.indd 1 11/18/13 2:47 PM 000_afpm.indd 1 09/12/2013 11:25
What your transmitter may not be
telling you
I
n process instrumentation appli-
cations measuring pressure, ow
or level, the engineer or techni-
cian is focused on accuracy as a top
priority. And a critical piece of
equipment is the transmitter.
Engineers will usually buy the most
accurate one they can and dedicate
a great deal of attention to it.
However, the transmitter is only
as accurate as the inputs provided
to it. The process instrumentation
loop the set of tubing and
components that connect the
process to the transmitter is just
as important. The role of this loop
is to present a set of process condi-
tions to the transmitter. These
conditions must be precisely the
same as those in the process. If they
are not, the transmitter will not
provide useful measurements.
Further, it is often hard to know
when the process instrumentation
loop is not performing well. There
is no alarm that sounds. So while
the engineer or technician may be
focused on the transmitter, the
instrument loop may be the cause,
undermining any possibility of
success.
Therefore, there is good reason to
be educated about the possible
issues in a process instrumentation
line, including those related to
overall design and layout, as well
as individual component quality
and installation.
Close coupling
Before we review the process instru-
mentation loop in detail, let us look
rst at a recently developed alterna-
tive. It is an elegantly simple
solution, and if your application
allows you to employ it you should.
If you value accurate process measurements and specify a premium transmitter,
your instrument loop requires the same level of attention
ERIC MOORE Swagelok Capital Projects Company
SAM JOHNSON Swagelok
The usual process instrumentation
setup entails at a minimum a
process interface valve, impulse
lines, a manifold and a transmitter
(see Figure 1). Impulse lines can be
costly to install and maintain, with
challenges like clogs, leak points,
temperature control and corrosion.
Close coupling eliminates the
impulse lines. The process interface
valve and manifold become one unit
(see Figure 2), and the transmitter
mounts directly onto it. Then, the
entire assembly attaches to the
process line. Almost everyone who
learns about this alternative likes it.
The challenge is nding the right
places to use it.
One limitation is temperature.
One reason for the traditional setup
with impulse lines is to protect the
transmitter from the high tempera-
ture of the process line. If the
process is too hot, the transmitter
may not be able to operate only a
few inches away in a close coupled
installation.
A second limitation is access. If
you need to get to the transmitter
for calibration, it needs to be acces-
sible, so mounting a close couple
on a process location 50ft in the air
does not make a lot of sense.
The only other obstacle is cost.
Close coupling requires an initial
investment, but in the long run it
may be less costly, especially if you
gure in the low cost of maintaining
a close coupled system, as compared
to the traditional alternative. If you
have an opportunity to employ this
solution, you should do it.
Standardisation: a step towards
simplied maintenance
If your goal is an optimal design,
there are a limited number of ways
to set up the process instrumenta-
tion loop. And yet, in most plants,
there is a multitude of variations,
and many of them are not optimal.
These variations have been devel-
oped over time, by different
engineers and/or contractors for
www.eptq.com PTQ Q1 2014 147
Impulse
line
Process line
PIV x 3
Manifold
Transmitter
Figure 1 The process instrumentation loop
consists mainly of a process interface valve
(PIV), impulse line and ttings, a manifold
and the transmitter
Figure 2 A close couple eliminates the
impulse lines. The process interface valve
and manifold become one unit, and the
transmitter mounts directly to it
swagelok.indd 1 11/12/2013 14:52
148 PTQ Q1 2014 www.eptq.com
a single, self-contained unit, which
has these advantages:
Fewer leak points
Flexibility to congure the prod-
uct using different sized orices
and anged connections
Reduced size and weight, reduc-
ing the need for and expense of
structural support of the instru-
mentation system (see Figure 3)
Reduced cost
Faster and easier installation
Availability in a wide range of
pressure classes and materials for
compatibility with specic piping
system designs
Fire safety ratings.
An alternative to the DBB shown
in Figure 3 is a monoange, which
is smaller and more compact,
consisting of three needle valves
two block valves and a bleed valve.
Needle valves have a smaller orice
than ball valves, so the monoange
is appropriate for lighter, less
viscous uids.
The classic DBB design is appro-
priate for all uids, but especially
for those with higher viscosity
when using ball valves. In fact, a
good DBB design should be rodda-
ble when system uids are dirty or
contain wax. Roddable means that
you can run a rod through the
middle of the DBB to clear a clog.
Impulse lines
Impulse lines connect the process
interface valve to the manifold and
transmitter. Their purpose the
same with all process instrumenta-
tion components is to convey the
precise process conditions to the
transmitter. When laying out
impulse lines, three main objectives
come into play:
Prevent corrosion or scaling
Reduce leak points
Maintain temperature within a
certain range.
The rst two are best achieved by
using tubing and tube ttings made
from an appropriate alloy (stainless
steel as a minimum), as opposed to
carbon steel pipe and threaded
connections. Stainless steel tubing
can be bent and shaped, which
reduces the number of mechanical
connections. When mechanical
connections are necessary, high
quality, two ferrule, mechanical
different projects and reasons. Such
a situation can be a drain on your
time and energy. Each system has
different needs in terms of mainte-
nance, and when things go wrong
there are a multitude of possibilities.
Ideally, all systems should be
designed using a consistent set of
criteria, including budgets and
allowances for downtime, mainte-
nance and accuracy. The result is a
high degree of standardisation. For
example, before standardisation, a
rening plant may have 30 differ-
ent congurations for process
instrumentation loops. After stand-
ardisation, the same plant may
have only ve, with each contain-
ing the same basic components: a
transmitter mount, manifold system
and redundant pressure measure-
ment. The only variations might be
the tubing runs and the type of
process interface valve (based on
temperature, pressure and location
of the valve).
With standardisation, many
things become simpler, including
maintenance, installation, training
and diagnostics. Error is also
reduced. In addition, the facility
can stock fewer replacement parts,
reducing overheads.
Basic building blocks
For each of the basic building
blocks in a process instrumentation
loop the process interface valve,
the impulse lines and the manifold
there are critical choices in terms
of materials and design that can
affect accuracy.
Regarding materials, stainless steel
or another corrosion resistant alloy
is strongly preferred in most appli-
cations. Still, many industrial plants
employ carbon steel for process
interface valves, for some piping
and even for some manifolds (or
parts of manifolds). In certain low
moisture applications, like oil,
carbon steel is acceptable, but for
most other applications it can be a
risk. The scaling that commonly
builds up on carbon steel can break
away, ow downstream, lodge in a
valve seat and prevent a positive
shut-off. The result is an inaccurate
transmitter calibration and/or inac-
curate transmitter readings (as we
discuss in more detail below). If you
employ carbon steel components in
the instrument loop, they will
require very close monitoring to
ensure that scaling is not affecting
the operation of the valves in the
system.
Process interface valve
The process interface valve (PIV) is
the rst valve off the process line.
Historically, the PIV of choice has
been a single gate valve or ball
valve. Both continue to be used
today, especially in the US, but the
best practice is a double block and
bleed (DBB) valve, which consists
of two isolation valves and one
bleed valve in between. The main
reason for employing a DBB is
safety. If you need to shut down
the process instrumentation line for
maintenance, you would close both
block valves and open the bleed
valve. If for any reason the rst
block valve were to leak, the second
block valve would prevent pressure
or uid build-up in the process
instrumentation loop.
Although the function of a DBB
can be constructed by using three
separate valves, the better option is
26.5 in
(673mm)
9.5 in
(241mm)
Figure 3 A double block and bleed valve can
be constructed from three separate valves
or can be purchased as a single, self-
contained unit, reducing size and weight
swagelok.indd 2 11/12/2013 14:51
grip-type tube fttings will not back
off with thermal cycling or vibra-
tion, unlike the threaded fttings
used with carbon steel.
The third objective maintaining
temperature within a certain range
is achieved by heating the
impulse lines. You can insulate your
impulse lines manually feld
tracing or purchase tubing that
has been insulated in the factory
and encased in a polymeric jacket.
Pre-insulated bundled tubing comes
ready to install in coiled lengths. It
can be heated with steam or left
unheated, as the application
requires. It is important with pre-in-
sulated tubing bundles to follow the
manufacturers instructions for seal-
ing the insulation when splicing into
or cutting into the pre-insulated
tubing bundle.
Manifold
The manifold consists of a set of
valves whose bodies are machined
from a single block of metal,
usually stainless steel. The mani-
fold mounts to the side of the
transmitter and serves a critical
function, enabling calibration or
service of the transmitter.
Quality is especially important in
a manifold. During calibration or
normal operation, at least one of
the valves in the manifold is in the
off position. If the shut-off is less
than complete, the result is an inac-
curate reading from the transmitter.
For example, when calibrating a
www.eptq.com PTQ Q1 2014 149
transmitter in a standard differen-
tial pressure setup, the technician
closes the two isolation valves and
opens the equalisation valve (see
Figure 4). This procedure results in
the transmitter being zeroed out
for calibration. If one of the two
isolation valves is leaking, however,
the calibration will be unsuccessful
and all subsequent readings will be
inaccurate.
Further, after calibration, the
operator will close the equalisation
valve and open the two isolation
valves, placing the transmitter back
in service. If the equalisation valve
does not achieve a positive shut-off,
high pressure will leak across the
seat to the low pressure side,
rendering the differential reading
inaccurate.
Why would these valves leak?
One cause is debris, either scaling
HP LP
To
transmitter
To
transmitter
Isolation
valves
Equalisation
valve
Figure 4 A manifold, consisting of three
needle valves, enables the technician to
isolate and calibrate the transmitter in
the eld
yield selectivity, in particular the
propylene yield. The higher the
hydrogen content, the higher the
propylene yield.
For the types of crude oil that
have been considered here, the
typical CCR content of atmospheric
residues is about 12-14 wt%, and
metals are 80-90 wtppm. Thus,
pretreatment of the feed is required.
The ARDS unit is a specially
designed hydrotreater that pretreats
the feed to the RFCC unit to reduce
contaminants including sulphur,
nitrogen, organometallic metals and
CCR. Additionally, it increases the
hydrogen content of the atmos-
pheric residue, thus improving the
crackability and selectivity of the
residue. For a typical ARDS HDS
rate of 90%, the CCR reduction is
about 65-70% and the metals reduc-
tion (HDM rate) is about 85-90
wt%. The HDT RFCC feed will
then have a CCR content of less
than 6 wt% and metals content of
less than 10 wtppm.
Due to the feed rate, nature of the
feed, and processing objectives, a
relatively high pressure and low
space velocity design is required
compared to other hydroprocessing
technologies. Even so, the cycle
length of the ARDS unit is typically
only one year. For the case evalu-
ated and ARDS operating
objectives, at least two reactor
trains, each with three to fve very
large and thick-walled reactors, will
be required. Consequently, both the
20 PTQ Q4 2013 www.eptq.com
initial investment and on-going
catalyst replacement cost will be
much higher than for most hydro-
processing units.
The other key features of this
confguration to maximise the
production of petrochemicals are:
Process straight-run light naph-
tha along with light catalytic
naphtha (LCN) in the steam cracker
Process heavy straight-run naph-
tha in a catalytic reforming unit
(CRU) to produce reformate
Process heavy catalytic naphtha
(HCN), pygas and reformate
through the aromatics complex to
produce paraxylene and benzene
The C
4
s with relatively high
olefns content are either sold or
recycled to the second riser of the
FCC (directly or through oligomeri-
sation) to maximise propylene
production
Products from the RFCC unit and
steam cracker are combined in a
single light ends recovery section to
produce ethylene and propylene,
minimising the capital investment
Hydrogen from the CRU and
steam cracker is supplemented by
further hydrogen production from
steam reforming to be used in the
hydrotreaters.
Complex conguration with
HCU + VRDS + RFCC
The intention of this refnery confg-
uration is to increase the production
of middle distillate at the expense of
petrochemicals. Relative to the frst
confguration of ARDS + RFCC, the
LVGO fraction of the crude is
routed to a VGO hydrocracker
(HCU) to shift the yields in the
desired direction. Figure 2 shows a
schematic of the refnery.
The key features of this refnery
confguration are:
A vacuum tower is installed to
produce a diesel-type cut to be
processed in the distillate hydro-
treater, a LVGO cut to be processed
in a two-stage, almost full-conver-
sion hydrocracker unit oriented
towards the production of middle
distillates
The HVGO is combined with
vacuum residue to be pretreated in
the vacuum residue desulphurisa-
tion (VRDS) unit before processing
it in the RFCC unit
Depending on the quality of the
combined vacuum residue and
HVGO feed to the VRDS unit, a
slip stream of vacuum residue
could be taken and blended with
diluents such as RFCC decanted oil
(DCO) and light cycle oil (LCO) to
produce some fuel oil. The VRDS
unit would then be able to reduce
the CCR to the levels required by
the RFCC feed, especially when the
CCR content of the feed to the
VRDS unit exceeds 24 wt%.
Complex conguration with
SDA + DCU + HCU + RFCC
The refnery confgurations illus-
trated above, where residue
hydrotreaters reduce the feed
Max Ovchinnikov is a Senior Research Chemist
with Criterion Catalysts and Technologies
based in Houston, Texas. He is primarily
engaged in the research and development of
catalysts for hydroprocessing applications and
has 12 years of experience in heterogeneous
catalysis and rening technologies. He has co-
authored over 20 technical publications and
holds a PhD degree in organic chemistry from
Iowa State University.
www.eptq.com PTQ Q3 2013 83
2 Dekkers C, Daane R, Oil & Gas J., 1999, 97,
145.
3 Puri B K, Irgolic K J, Environ. Geochem. Health,
1989, 11, 95.
4 Nielse B, Villadsen, Appl. Cat., 1984, 11, 123.
5 (a) Internal communication, Criterion
Catalysts & Technologies; (b) Merryeld R N,
Gardner L E, Parks G D, Catalyst Characterization
Science, ACS Symposium Series 1985, 1.
6 Bhan O K, Arsenic removal catalyst and
method for making same, US Patent 6759364.
requires increased process unit
capabilities in a number of refning
process units, particularity hydro-
processing assets. The magnitude
of impact, of course, depends on
the particular feed processed, the
inherent fexibility of the facilitys
operations and the ability to
offset this increased processing
requirement.
In making this move, proper
selection requires a good under-
standing and modelling of the
facility as well as a strong knowl-
edge of the sources and processing
requirements of the available
crudes. With the continued devel-
opment and increased availability
of such feeds, this knowledge needs
to be continuously updated to
ensure minimal operational
surprises. Reviewing TBP curves
and bulk properties of whole crude
or even the individual cut-point
ranges does not
necessarily characterise the feed
diffculty or indicate the full impact
on the individual processing units.
References
1 Henke K R, Arsenic, Environmental Chemistry,
Health Threats and Waste Treatment, Wiley,
2009, 186.
30
50
60
40
20
10
N
o
r
m
a
l
i
s
e
d

H
D
S

t
e
m
p
e
r
a
t
u
r
e
,

F
Base
Alternative feed Current feed
Temp. 1 Temp. 2 Temp. 3 Temp. 3 Temp. 2 Temp. 1
check-
back
Temp. 1
Catalyst age
DN-3651 alternative feed DN-3551 alternative feed
DN-3651 current GO DN-3551 current GO
Figure 15 DN-3651 and DN-3551 normalised HDS temperature
increased water in the sulphur
plant feed and thus more condensa-
tion in the sulphur plant feed NH
3

gas knock-out drum (and thus
water that has to be recycled back
to the sour water stripper).
The design shown in Figure 2
was common in the 1960s.
However, it also suffers from the
same heat balance drawbacks and
needless complications as seen in
www.eptq.com PTQ Q2 2013 101
Figure 1. In other words, a lot of
equipment is added to generate
reux when no fractionation is
required between the feed and
overhead product. Again, the only
purpose of the tower is to strip out
the NH
3
and the H
2
S.

Correct stripper design
In 1969, while working for the now
vanished Amoco International Oil
(1)
(8)
(15)
Sour
water
feed
90F
Low NH
3
water
to desalters
Steam
250F
E-1
P-1
13 PSIG
Ammonia gas
to sulphur plant
190F
Medium content
NH
3
water to
hydrotreaters
Figure 4 Two-stage sour water stripper design without feed preheat
Company in the UK, I designed a
sour water stripper that eliminated
the unnecessary features of the unit
shown in Figure 2.
Figure 3 shows the essentials of a
correct sour water stripper design.
Feed is brought in at ambient
conditions (70-100F, 21-38C) from
the sour water feed tank. To heat
the feed from 90F (32C) to 250F
(120C) requires about 16 wt%
steam ow, or about 1.3-1.4lb of
steam per gallon of stripper
bottoms, which is close to a typical
design stripping steam ratio for
sour water strippers. The E-1 feed
preheater, reux pump (P-2) and
the reux cooler (E-2) shown in
Figure 2 are all eliminated. How,
then, does one know that the
design shown in Figure 3 will
work? Because it was built this way
(at the Amoco renery in Milford
Haven, Wales, UK) in 1970, where
it worked just as well as the
conventional design shown in
Figure 2.

Two-stage sour water stripper
Figure 4 shows a sour water strip-
per with a side draw-off. The partly
stripped sour water is extracted
from tray 8 and directed to the
hydrotreaters for use as make-up
water in the salt (NH
4
HS) removal
step of the reactor efuent.
Completely stripped water from
the sour water stripper bottoms is
sent to the crude desalter. While
102 PTQ Q1 2013 www.eptq.com
communicate where the module
will be installed on the plot plan.
Connections between the modules
are designed to be similar in confg-
uration so that construction is
relatively straightforward. Ventech
estimates that, with modularisation,
approximately 70% of a project is
already complete even before the
modules are shipped from their
facility. This greatly decreases feld
construction time to deliver an
operational facility (see Figure 1).
These methods also facilitate easy
disassembly and relocation, if
necessary, at some point in the
future. For example, a remotely
located gas processing facility could
be easily taken apart and moved to
a new natural gas source if an exist-
ing supply was depleted in its
current location.
Applying modularisation to refn-
ery construction has advantages
with regard to productivity, prod-
uct quality and ensuring the safety
of construction personnel. Since the
modules are built in a well-lit,
climate-controlled environment,
work can continue around the clock
regardless of weather conditions,
for greater productivity and easier
quality control. Since module height
is restricted, safety is enhanced, as
workers build at limited heights
within the fabrication facility.
Modularising GTL
The same advantages of modular
construction of refneries are being
applied to the construction of
distributed GTL plants. The GTL
process involves two operations: the
conversion of natural gas to a
mixture of carbon monoxide (CO)
and hydrogen (H
2
), known as
syngas, followed by a Fischer-
Tropsch (FT) process to convert the
syngas into paraffnic hydrocarbons
that can be further refned to
produce a wide range of
hydrocarbon-based products, includ-
ing clean-burning, sulphur-free
diesel and jet fuel. Speciality prod-
ucts including food-grade waxes,
solvents and lubricants can also be
produced from the paraffnic
hydrocarbons.
Large, commercial-scale GTL
plants, including the Sasol Oryx and
the Shell Pearl plants (both located
in Qatar), have been built at enor-
mous capital cost. The Oryx plant,
designed for production levels of
34 000 b/d, cost around $1.5 billion
to build. The Shell Pearl plant, with
an ultimate design capacity of 140
000 b/d of GTL products and 120
b/d of natural gas liquids, cost
around $18-19 billion. Conventional
GTL plant designs rely on econo-
mies of scale to drive positive
fnancial returns and are viable only
where there are large supplies of
low-priced natural gas.
However, another option being
developed smaller-sized and
distributed GTL plants shows
promise for deriving value from
smaller accumulations of unconven-
tional gas that would otherwise be
left underground, such as shale gas,
tight gas, coal bed methane and
stranded gas (gas felds located too
far from existing pipeline infra-
structure). A small, modularised
GTL plant has the fexibility to be
installed close to the trapped
resource and then used to process
that resource locally. Associated gas
(gas produced along with oil) is
another area of opportunity for
modularised GTL plants. This gas
is typically disposed of either by re-
injection, at considerable expense,
back into the reservoir or by the
wasteful and environmentally
damaging practice of faring, which
is subject to increasing regulation.
Modularised GTL plants enable this
otherwise wasted gas to be
converted into additional revenue.
In the larger economic picture, a
modular GTL capability can be the
key factor that enables the construc-
tion of upstream projects that
would otherwise be cancelled
because of poor results derived
from economic models. For exam-
ple, some shale gas discoveries are
being hampered by high develop-
ment costs, which result in marginal
economics due to gas prices that
are often low. These projects can be
enhanced by converting the gas to
higher-value clean fuels produced
in the GTL process.
ITW
Innovative
Technologies
Worldwide
ITW is a fast growing Company, marketing and implementing unique and patented Production Units Online
Cleaning, Tank Cleaning, Decontamination and Reclamation technologies, along with Specialty Chemicals.
Given the considerable success, ITW is expanding its markets and activities and is looking for experienced
professionals worldwide. The candidates should have minimum a 5 years experience in at least one of these
fields:
refining/petrochemicals process specialty chemicals sales
refining/petrochemicals process technology, operations, maintenance, turnaround
and should be willing to travel nationwide and worldwide, along with being prone to hard working and sales.
Having a technical degree and good market knowledge is also required.
The available positions will cover: technical sales, implementation of ITW technologies on the field,
sales and operations management.
Interesting compensation plans will be given along with serious career possibilities. Please contact:
ITW S.r.l. S.Cusumano 96011 Augusta - Italy jobs@itwtechnologies.com www.itwtechnologies.com
HIRING EXPERIENCED PROFESSIONALS WORLDWIDE
velocys.indd 2 10/12/12 14:27:56
lieberman.indd 3 08/03/2013 16:02
cri.indd 10 10/06/2013 16:38
f wheeler.indd 3 12/09/2013 13:40
from corrosion (as mentioned
earlier) or other debris, such as
burrs. Burrs are a byproduct of the
machining process during manufac-
turing. A critical step during
manufacturing and installation
should be the removal of all burrs
from all wetted surfaces. Otherwise,
they can dislodge when the mani-
fold is in service and catch in or
scratch the valve stem or seat,
preventing a positive shut-off. Burrs
may also be an issue in monofanges
and other globe pattern needle
valves.
Seat seals in needle valves
Another reason why valves leak is
poor design. Most needle valves
employ a metal to metal seal. The
metal tip of the stem moves down-
ward to seal against a metal seat of
matching shape. The tip of the stem
may be shaped like a ball or a vee.
In either case, it is important that
the tip does not rotate with the
stem. If it does, it may grind into
the seat, scoring it and creating a
pathway for leakage.
There are two main design
approaches to achieving a non-
rotating stem tip. In the case of a
vee tip, there may be a knuckle
joint that enables the upper stem to
turn while the lower stem remains
stationary, except for axial (up and
down) movement.
In the case of a ball tip, a ball
foats in the stem tip, like in a ball-
point pen. Among ball tips, there
swagelok.indd 3 11/12/2013 14:52
moscow.indd 1 09/12/2013 11:28
www.eptq.com PTQ Q1 2014 151
the same level of attention.
Measurement accuracy is as
dependent on the transmitter accu-
racy as it is on the quality of the
instrument loop components, as
well as how they are installed and
maintained. Standardising your
facility on a core set of instrument
loop details, quality components
and quality transmitters will raise
the accuracy of your measurement.
More accurate measurements will
yield welcome dividends in terms
of time, effciency and plant
proftability.
Eric Moore is a Swagelok Capital Projects
Company Technical Manager, based at Swagelok
in the Isle of Man, and a Swagelok subject
matter expert for process instrumentation
products and applications, providing guidance
on technical matters relating to construction
projects. He holds a diploma in engineering
from Isle of Man College.
Sam Johnson is a Product Manager at Swagelok
Company. He is responsible for application
and product knowledge, and developing sales
strategies for process instrumentation, piping,
and medium and high pressure products. He
holds a bachelors degree in history from
University of North Carolina at Charlotte.
Seat seals in ball valves
The ball in a ball valve usually seals
against a fuoropolymer seat. High
pressure in the fuid stream pushes
the ball against the seat, resulting in
a seal. However, in a low pressure
situation, additional force may be
required for an effective seal. This
force sometimes referred to as
live loading comes from a
spring or O-ring inserted between
the end screw and the seat. Not all
ball valves contain these additional
components, but they are valuable
in achieving a positive shut-off in
less than ideal conditions, when
temperature and pressure are fuctu-
ating. In valves that do not contain
these components, an interference ft
is usually relied upon to provide a
low pressure seal between the seat
and the ball. Interference seal-type
seats typically do not offer low
pressure gas sealing abilities for
more than a few actuations.
Conclusion
If you value accurate measurements
and specify a premium transmitter,
your instrument loop requires
are desig