Escolar Documentos
Profissional Documentos
Cultura Documentos
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
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
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
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!
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
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 designs that allow the ball to
rotate on all planes, and others that
restrict the ball, so it can rotate on
only one plane. The latter is
preferred. Each time a ball seals
against a seat, a line seal is created
where the seat meets the ball.
If the ball is free to locate on all
planes, it will develop multiple,
criss-crossing line seals, which will
create leak points. If the ball can
only rotate on one plane, it will
develop only one line seal, with
which it can seal repeatedly with-
out leakage.
In addition, a differential hard-
ness between the stem tip and the
seat is desirable in a needle valve
because it aids in the process of
forming a seal. The differential
hardness allows the hardened stem
tip to coin out imperfections in the
seat due to light contamination or
the build-up of a flm on the seat. A
low differential hardness between
the stem tip and the seat can cause
premature failure due to the inabil-
ity of the stem tip to coin out the
seat, or from excess torque when
closing the valve.
2
0
1
3
S
T
R
E
A
M
L
IG
H
T
,
IN
C
.
CONNECT WITH US WWW.STREAMLIGHT.COM
When you have a job to do,
youre not thinking about
your flashlights engineer or
the design, functionality or
keen ingenuity that offers
a premium product for the
GREATEST VALUE.
In fact, you dont have to think
about your flashlight at all.
Because its a STREAMLIGHT.
130 LUMENS
15,000 CANDELA
POLYSTINGER
LED HAZ-LO
200 LUMENS
1,150 CANDELA
KNUCKLEHEAD
80 LUMENS
4,000 CANDELA
TRIDENT LED
HEADLAMP
615 LUMENS
4,000 CANDELA
E-FLOOD
LITEBOX
swagelok.indd 4 11/12/2013 14:52
ABB Consulting 113
AFPM Annual Meeting 2014 146
Air Liquide Global E&C Solutions IFC
Albemarle IBC
AMACS Process Tower Internals 38 & 41
Axens OBC
BASF Corporation, Catalysts Division 12
Bilnger Water Technologies 65
Bryan Research & Engineering 103
Burckhardt Compression 35
CB&I 11
Criterion Catalyst & Technologies 4
CS Combustion Solutions 45
DigitalRening.com 111
DuPont Belco Clean AirTechnologies 79
DuPont Sustainable Solutions 37
Elliott Group 83
Enersul 122
European Fuels Conference 62
Everlasting Valve Company 89
ExxonMobil Research and Engineering Company 15
European Fuels Conference 62
Flexim 121
Foster Wheeler 92
Four Quest Energy 74
GE Oil & Gas, Surface Pumping Systems 73
Grabner Instruments 91
Grace Catalyst Technologies 21
Heurty Petrochem 32
Hoerbiger Compression Technology 132
IDTC & BBTC 86
ITW 80 & 149
Johnson Matthey Process Technologies 23
John Zink Hamworthy 129
KBC Advanced Technologies 2
Kurita Europe 67
Linde 25
Merichem Company 54
Metso Automation 119
OHL Gutermuth Industrial Valves 131
Onis International 68
Optimised Gas Treating 85
Process Consulting Services 16 & 19
Prosim 26
Rening China 2014 100
Sabin Metal Corporation 114
Samson 137
Sandvik Materials Technology 46
SOGAT 2014 138
Spectro Analytical Instruments 61
Streamlight 151
UOP 7, 9 & 31
Wood Group Mustang 28
World Petroleum Congress 150
Yokogawa Europe 53
Zeeco 105
Zwick Armaturen 145
Alphabetical list of advertisers
152 PTQ Q1 2014 www.eptq.com
For more information on these advertisers, go to www.ptqenquiry.com
ad index copy 2.indd 1 13/12/2013 15:52
want results? demand ACTION
.
If raising the alky feed and distillate from your FCC unit is high on the agenda, you demand more than a
conventional catalyst you demand ACTION