ptq

Q1 2014

petroleum technology quarterly

refining
gas processing
petrochemicals

special features
Outlook 2014
Environmental
opportunity crude processing
Automation & Control

cover and spine copy 18.indd 1

10/12/2013 13:02

1211_e

Does your raw natural gas contain hydrogen sulde,

carbon dioxide, mercaptans or more?
Whatever the impurity, whatever the composition, Air Liquide Global E&C Solutions
has the right treatment.
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 ecient.

www.engineering-solutions.airliquide.com

air
liquide.indd 1
AIR_2362_008_HydrocarbonEngineering_0413_RZ.indd
1

11/12/2013
07.03.13 19:53
17:00

ptq
PETROLEUM TECHNOLOGY QUARTERLY

Q1 (Jan, Feb, Mar) 2014

www.eptq.com

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 refinery profitability

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 refinery equipment with multiphase CFD modelling

Grant Niccum and Steve White Process Consulting Services
139 High-fidelity 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 refinery in Spain enables the site to
process heavy and sour crudes.
Photo: Repsol
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.

ed com copy 3.indd 1

12/12/2013 14:45

KBC Adv - PTQ Q1 2014.pdf 1 12/10/2013 9:29:32 AM

CM

MY

CY

CMY

KBC.indd 1

10/12/2013 16:06

p
T tq

he European Union has arguably

been the global leader in biodiesel
production and use, with overall
biodiesel production increasing from 1.9
P
ETROLEtonnes
UM TECin
HN2004
OLOGto
Y Qnearly
UARTE10.3
RLY million
million
tonnes in 2007. Biodiesel production in the
US has also increased dramatically in the
past
fewNo
years
Vol 19
1 from 2 million gallons in
2000 to approximately 450 million gallons
Q1 (Jan, Feb, Mar) 2014
in 2007. According to the National Biodiesel
Board, 171 companies own biodiesel
manufacturing
plants and are actively
Editor
marketing
biodiesel.1. The global biodiesel
Chris Cunningham
editor@petroleumtechnology.com
market
is estimated to reach 37 billion
gallons by 2016, with an average annual
Production
growth
rateEditor
of 42%. Europe will continue to
Rachel Storry
be the major biodiesel market for the next
production@petroleumtechnology.com
decade, followed closely by the US market.

Although
Graphics
Editor high energy prices,
increasing
global demand, drought
Rob Fris
graphics@petroleumtechnology.com
and
other factors are the primary
drivers for higher food prices, food
Editorial
competitive
feedstocks have long
tel +44 844 5888 773
been and will continue to be a major
fax +44 844 5888 667
concern for the development of biofuels.
To Development
compete, Director
the industry has
Business
responded
Paul Mason by developing methods to
sales@petroleumtechnology.com
increase
process efficiency, utilise or
upgrade by-products and operate
Advertising
Sales quality
with
lower
lipids as
Bob Aldridge
feedstocks.
sales@petroleumtechnology.com

Feedstocks
Advertising Sales Office
Biodiesel
to a diesel-equivalent
tel +44 844refers
5888 771
fax +44
844 5888 662
fuel
consisting
of short-chain alkyl

(methyl or ethyl) esters, made by the

Publisher
transesterification
of triglycerides,
Nic Allen
commonly known as vegetable oils or
publisher@petroleumtechnology.com
animal fats. The most common form
uses
methanol, the cheapest alcohol
Circulation
available,
Jacki Watts to produce methyl esters.
circulation@petroleumtechnology.com
The
molecules in biodiesel are primarily fatty acid methyl esters
Crambeth Allen
Publishing
Ltd by trans(FAME),
usually
created
Hopesay, Craven Arms SY7 8HD, UK
esterification between fats and methatel +44 844 5888 776
nol.
Currently,
biodiesel is produced
fax +44
844 5888 667
from various vegetable and plant oils.
First-generation food-based feedstocks
are straight vegetable oils such as
ptq (Petroleum Technology Quarterly) (ISSN
soybean
oilUSPS
and
animal
fats such as
No: 1632-363X,
No: 014-781)
is published
quarterly plus
annual
Catalysisgrease,
edition by Crambeth
tallow,
lard,
yellow
chicken fat
Allen Publishing Ltd and is distributed in the US
and
the
by-products
of
the
production
by SP/Asendia, 17B South Middlesex Avenue,
Monroe
NJ 08831. Periodicals
postagefrom
paid at New
of
Omega-3
fatty acids
fish oil.
Brunswick, NJ. Postmaster: send address changes to
Soybean
oil
and rapeseeds
ptq (Petroleum
Technology
Quarterly), 17B oil
Southare the
Middlesex Avenue,
Monroefor
NJ 08831.
common
source
biodiesel producBack numbers available from the Publisher
tion
theincUS
and Europe in quantiat $30 in
per copy
postage.
ties that can produce enough biodiesel to be used in a commercial market
with
currently
applicable

A window
half-open

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
inflated 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 flagship 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 inflated 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 competitive 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
refineries 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 perhaps across the globe as local technical and environmental issues surrounding 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 synthesis 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 refinery, 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 refine 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 refining on its doorstep. If a window opens
for GTL, you need to be certain that it will stay open.

chris cunningham
PTQ Q1 2014

ed com copy 3.indd 2

12/12/2013 10:50

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.
As the R & D team lead in oil refining 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 profitability. These two seeds grown by Ferry and his team are now bearing fruit for our
refinery customers around the world.

Leading minds. Advanced technologies.

www.CRITERIONCatalysts.com

criterion.indd 1

09/12/2013 11:35

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 profitability
Maria van der Hoeven

over the medium term, the persistence of excess

refining capacity is expected to weigh on refining
margins, severely affecting the less efficient and older
refineries in the world. The overall upgrading ratio
he world is undergoing an
gradually increases from 44% in 2012 to 47% in 2018 as
era of economic rebalancing
new refineries focus mainly on heavy crude oil
toward emerging markets,
processing, mostly from the Middle East and Latin
and that is especially true within
America. The capacity expansions in Asia and the
the oil market, where shifting
Middle East, however, represent two different distinct
patterns in consumption and
investment strategies. Whereas Asian capacity
production
further
complicate
development is meant to service rapidly rising
the broader trends. The global
domestic demand, Middle East exporters (mostly in
energy map is not just shifting
joint venture with OECD refiners or Chinese
east. Traditional producers like Saudi Arabia are seeing companies, and mostly in Saudi Arabia) are climbing
booming demand, while heavily consuming markets the value chain and expanding into products and
such as North America are experiencing production petrochemicals.
revolutions thanks to new techniques and technologies.
This has put OECD refiners in particular
New trading patterns and technological advances European ones into a tight spot. Facing weak
also mean new scope for international market reform demand, tighter regulations, and an aging capital
creating a common incentive for market stability stock, these refiners have little choice but to increase
and investment promotion. The next five years will their competitiveness through restructuring and
consolidate the rise of the non-OECD in virtually consolidation. The refining sector in OECD Americas is
every aspect of the oil market, and most of the growth undergoing a complete restructuring, with each
(though by no means all of it) will come from East-of- regional district trying to optimise its crude slate
Suez economies. The idea that the emerging markets between light domestic crude oil and heavy imported
and developing economies would eventually overtake oil. High sustained margins and the outlook for
the OECD in oil demand is nothing new, but it is continued growth pushes US refiners to invest in deep
happening faster than expected.
conversion and light oil processing. Recent
One of the consequences of those changes is a developments in Spain and Portugal show that
transformed global product supply chain. New, non- investment into deep conversion units can indeed
OECD mega-refineries are challenging OECD refining present a profitable alternative to closures. Yet despite
economics, at least beyond the US. Their expanding these efforts, the overcapacity which exists today is
reach is accelerating the globalisation of the product bound to force a continuation of the capacity shut-ins
market, particularly in the case of refineries geared we have seen since 2008. Over that period, 15
toward export. With it come the benefit of greater European refineries have closed with a total capacity of
market flexibility in the dispatch of product supply, 1.7 mb/d, and more are expected in the coming years.
but also longer supply chains, higher
Japanese refiners will close a total of
reliance on stocks to meet demand,
around 800 kb/d by next year, in
Other Asia
diminished visibility in inventory
line with government regulations
14%
levels, increased disruption risks,
aimed at increasing conversion
reduced market transparency and, China
yields.
Latin
45%
America
possibly, greater price variation
Looking out to 2035, the World
14%
between key markets and also
Energy Outlook 2013 shows the
between seasonal peak and troughs
global refining sector set to
in demand.
experience continued turbulence
Middle East
The IEAs Medium Term Oil Market
amidst declining OECD demand and
22%
Report 2013, released in May, offers
Asian growth. Strains on the refining
Other
5%
projections to 2018. As refinery
system are amplified by the
capacity additions are forecast to
changing composition of feedstocks.
increase faster than world demand Figure 1 Regional share of CDU expansions A growing share of oil supply

Executive Director
International Energy Agency

www.eptq.com

outlook copy 2.indd 1

PTQ Q1 2014 5

11/12/2013 15:57

bypasses the refining system altogether, including most

natural gas liquids as well as oil products produced
directly from gas or coal. As a result, global demand
for refined products grows by only 10 mb/d to 2035
less than the growth in both overall liquids demand
(16.8 mb/d) and net refinery capacity additions (13
mb/d). Over-capacity in the sector is therefore a long
term phenomenon as well, but prospects for OECD
refiners vary.
The outlook for the United States is helped by the
increasing availability of local crude, but the benefits to
refiners 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 difficulties 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 refining 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 refining capacity
are still not sufficient to meet regional demand, and
both China and India are net oil product importers in
2035.
The changes occurring in the refining balance are
reflective of both the traditional refining model in
which increasingly consumptive countries prefer to
import crude and refine it for domestic production
and also a new phenomenon in the rise of global
refining centres. The increase of product trade reflects a
desire of some exporters to capture the added value of
exporting refined product, and also an effort to
diversify into highly specialised industries such as
petrochemicals. While some export oriented refining
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 refining 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.

PTQ Q1 2014

outlook copy 2.indd 2

Daniel M McCarthy

President, CB&Is Technology

Operating Group
Member, Board of Directors, Chevron Lummus
Global LLC

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 olefin or liquefied 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 refining, 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 benefit 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, refiners, 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 significant reduction in
new drilling and curtailment of supply growth.
Since 2003, the impact on refining is equally dramatic

www.eptq.com

13/12/2013 11:44

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.

For more information about UOP adsorbents, visit www.uop.com/adsorbents

2013 Honeywell International, Inc. All rights reserved

uop too late.indd 1

09/12/2013 11:33

2.2

175

2.0

150

1.8

125

1.6

100
75

1.4

20
02
Se
p

20
ec
D

Ja
n

14

2
01
Ap
r2

20
ly
Ju

06
20
ct
O

20

1
Ja

00
Ap
r2

99
l1

09

04

1.0
8

25

20
04
M
ay
20
05
O
ct
20
06
Fe
b
20
08
Ju
l2
00
9
N
ov
20
10
Ap
r2
01
2
Au
g
20
13
D
ec
20
14

50

1.2

Ju
8

NY reg. gasoline spot

USGC mixed xylenes contract

200

$/bbl

$/octane bbl

2.4

Figure 1 US retail Octane Index Barrel (OIB) price

Figure 2 Mixed xylene pricing vs gasoline

as light paraffinic crude oils from shale have displaced

more naphthenic crude oils in the refiners 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 olefin demand is driving producers to secure
low cost supplies of ethane and LPG from shale gas.
For petrochemical producers securing feedstock from
refineries, 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 olefinic

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 efficient 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 finite; 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.

PTQ Q1 2014

outlook copy 2.indd 3

ISODEWAXING is a registered trademark of Chevron.

Rajeev Gautam

President & CEO

UOP LLC, A Honeywell Company

ooking back on 2013, it

was
an
impressive
and
eventful year for the refining,
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

www.eptq.com

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 specifically designed to help optimize
your operation. UOP invented the refining 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

uop
restless.indd
1
SPM-UOP-35
Adsorbents
Ad_v6.indd 1

09/12/2013
11:32
2/27/13 2:11
PM

where refiners 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
specifications 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, efficient and sustainable solutions to
directly meet these challenges.
Within refining, 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 Uniflex 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, refiners 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
Oleflex 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
olefin feedstocks, cheaper, more abundant feedstocks
such as coal will help producers address demand in
new ways. The UOP/Hydro Methanol-to-Olefins
(MTO) process converts methanol derived from noncrude oil sources such as coal or natural gas to
ethylene and propylene at very high efficiencies.
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. Efficient contaminant
removal, purification and liquids recovery will be
essential steps to efficiently 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

10 PTQ Q1 2014

outlook copy 2.indd 4

fuels. UOP is continuously improving techniques for

gas treating and contaminant removal to ensure that
the gas can be cleaned to liquefied natural gas
specifications.
In addition to natural gas, new, more efficient
alternative energy resources are on the horizon. These
renewable fuels and chemicals will support growing
global energy needs, while also addressing
environmental concerns.
UOPs first Ecofining unit for renewable diesel
production is on-line, with more units in design and
construction. This technology is preferred by refiners
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 efficient, use cost advantaged resources and
provide compelling environmental benefits.

Eric Benazzi

Vice President Marketing & External

Communication
Axens

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 inflows, emerging
economies are facing a difficult situation. This
monetary abundance led to the formation of credit
bubbles in Brazil, Russia, India, China and Turkey,
which increased inflationary pressures.
The necessary resorption of financial 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 mid1980s 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.

www.eptq.com

11/12/2013 15:58

cbi.indd 1

10/12/2013 09:54

limiting factors love

limitless possibilities
Overcome limiting factors affecting refinery capacity and
operating flexibility 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
Additives

Co-Catalysts
Services and Solutions

www.catalysts.basf.com/refining

basf.indd 1

6/12/12 13:23:11

Indeed, Americas debt ceiling has only been raised for

a few months. Lack of a compromise in February
would lead to a destabilisation of the US economy that
would compare to the 2008 subprime mortgage crisis.
Such a scenario was deemed unthinkable given its
devastating consequences.
So, which indicators can we reasonably count on?
The global fuel and petrochemical demand should
experience growth, sustained by a growing world
population and its aspiration to a greater mobility and
lifestyle. Growth in the demand for refinery products
could increase by 1.1 million b/d, to reach a 2014
global oil demand of about 92 million b/d. On-road
diesel will continue to drive the demand, calling for
additional hydroprocessing capacities.
Product specifications and environmental quality
standards will continue to tighten worldwide calling
for advanced technologies and innovative solutions.
Impulse,
our
latest
hydroprocessing
catalyst
technology and Symphony, our latest reforming
catalyst, are the result of a long term commitment to
clean fuels and our answer to market trends.
Feedstock and streams produced in refining,
petrochemical sites and natural gas fields are
continuously
requiring
solutions
to
remove
contaminants subject to stringent requirements and
emission limits. As a result, we deliver a complete
range of technologies and products and Claus catalysts
enabling the removal of all kind of contaminants such
as Cl, F, As, Si, Hg, H2S, RSH, and so on.
Demand in main petrochemical products should
keep growing at a rate of 4% to 5%. In regions where
demand is high or resources abundant, the
construction of mega scale complexes integrating
refining and petrochemical units will allow recovering
valuable products from low value heavy feeds and
reaching higher profitability. Regarding projects where
multiple technologies are involved, Axens Integrated
Process Solution is nowadays a must.
Those integrated mega scale complexes will continue
to require the latest technologies to convert the bottom
of the barrel and to maximise motor fuel yields,
notably diesel. VR ebullated bed hydrocracking plants
combined with solvent deasphalting or cokers units
and VGO hydrocrackers will play a key role as well as
scheme combining AR/VRDS and RFCC units when
naphtha/gasoline production is targeted.
In Europe, where refineries are facing strong
competitive pressure, energy costs represent more than
50% of operating expenses, so any improvement in this
field will improve competiveness. Many new refining
entrants on the market have invested in the assets and
developed their ability to process a higher number of
crudes. In order to help EU refiners on this path,
Axens revamping offer helps improve and upgrade
their existing assets and assists in improving process
energy efficiency in order to increase competitiveness
at a minor risk.
Innovation will continue to be the best tools to
respond to market changes and to improve
profitability.

www.eptq.com

outlook copy 2.indd 5

Roberto Penno

Chief Executive Officer

Global Engineering & Construction Group
Foster Wheeler

e are certainly seeing

robust activity in all of
the hydrocarbon related
business sectors in which Foster
Wheelers
Global
Engineering
and Construction Group operates:
onshore and offshore upstream oil
and gas, midstream/LNG, refining,
chemicals, and power, as well
as seeing continued investment in pharmaceuticals,
power, environmental and minerals and metals, where
we also operate.
This year there has been a very strong focus in the
industry on North America, with more projects
moving forward in both petrochemicals and chemicals
and in LNG liquefaction and more emerging too. Asia
and the Middle East also continue to see strong project
activity.
I would like to highlight a number of key themes we
are seeing which also underpin our companys E&C
strategy going forward.
Local service delivery has always been important to
us and is becoming increasingly important to our
clients. We have continued to develop operations in
key local markets, both organically, in joint venture,
and through M&A. Key markets in which we have
recently strengthened our ability to deliver the global
Foster Wheeler brand and quality locally include Saudi
Arabia, Mexico, and China, to name just three. I will
add to this the US, where we are already strong and

Product specifications and

environmental quality
standards will continue
to tighten worldwide
well known but have continued to enhance our
presence and capabilities. Our Houston operation has
had a great year with major wins, especially in
chemicals, and we acquired construction specialists
Yonkers Industries at the end of 2012.
Many clients with upstream and downstream
businesses are allocating more and more of their
planned capital spend to the upstream sector. This
sector, both onshore and offshore, remains a
strategically important market for us and one in which
we have continued to develop our skills, service
portfolio and geographic presence, with acquisitions in
the last 12-15 months in additional early consultancy
capabilities, in SAGD with the acquisition of Three
Streams Engineering in Canada, and in on-going asset
support.
We see more and more clients looking to develop

PTQ Q1 2014 13

11/12/2013 15:58

long term strategic relationships with a limited

number of E&C contractors to ensure availability of
high quality resources. For a number of years we have
had a strong focus on developing long term
relationships with clients, working with them
wherever in the world they invest. We aim to match
our skills and geographic footprint development to
align with our clients long term investment plans, so
that we are well placed to meet their needs right
through the project lifecycle, across their entire project
portfolio and wherever in the world they plan to
invest. We have seen a significant number of key wins
in 2013 coming from both long standing relationships
and also strategic relationships that are more recently
established.
Projects also keep on growing in size and complexity,
strengthening the requirement for contractors like
Foster Wheeler who have the global horsepower and
track record, technical expertise, and regional and local
knowledge to design and execute these projects safely
and successfully.
Competition remains strong. We focus on
opportunities
where
we
believe
we
have
differentiators, such as our technologies; our knowhow; our client relationships; our global presence and
our ability to work with clients from the very earliest
phases of projects to help them shape their
investments; and our proven track record of safely
delivering technically complex and very large projects,
often in challenging locations.

Charles T Drevna

President
American Fuel & Petrochemical Manufacturers

rom my vantage point in

Washington, DC, I cannot help
but see the year ahead as one
filled with many policy challenges
for the refining and petrochemical
industry.
A year ago, I was cautiously
optimistic that US President Barack
Obama would work to advance the
all of the above energy strategy that he so often and
adamantly talked about during his run for re-election.
But during his February State of the Union address,
and then with the release of his climate change plan
in late June 2013, it became clear that the anti-fossil
fuel sentiments of this Administration will continue
into 2014, and throughout the remainder of his
presidency.
Among the many expected battles ahead is finding a
palatable solution to the sour Renewable Fuel Standard
(RFS). Progress was made in late 2013 when EPA
recognised the potential adverse effects on consumers of
increasing the ethanol volume in the fuel supply and
reduced the 2014 mandatory volumes. Even with the
cut, greater reductions in the biofuel mandate are
necessary if consumers are to avoid all the detrimental

14 PTQ Q1 2014

outlook copy 2.indd 6

impacts of the statute. Opposing sides in this battle will

undoubtedly continue to draw lines in the sand until
Congress repeals this broken law or finds another
solution that considers the impact of this pernicious
policy on consumers, today and into the future.
But, our challenges do not end there. The industry
will continue to face a growing number of climate
regulations that conflict with one another. Research has
shown that the RFS increases ozone and greenhouse
gas (GHG) emissions and is in conflict with CAFE
Standards, as well as other Clean Air Act regulations;
while Tier III regulations result in greater greenhouse
gases, thus contradicting EPAs GHG regulations.
These conflicts could leave many facilities across the
country questioning how they will meet some
requirements without violating others.
Finally, one of the largest challenges may come with

We are currently witnessing

the beginnings of a far reaching
manufacturing renaissance
the Obama Administrations recent Executive Order
(EO) that adjusted the metrics by which all Federal
agencies assess the costs of carbon emissions, known
as the social cost of carbon (SCC) calculation.
This recent EO is comprised of subjective metrics
that were developed to try and quantify the alleged
societal benefits of reducing only greenhouse gas
emissions.
Unfortunately, the Social Cost of Carbon scheme will
only add additional burdens on the consumer because
it inflates the potential benefits of regulatory actions in
a manner that hides the true economic and societal
costs of new rules. Applying these calculations to
assess the costs and benefits of regulations could
erroneously be used to justify discontinuing the use of
traditional, affordable energy sources through overly
stringent regulations in future rulemakings.
Although the challenges ahead are great, the future
holds enormous opportunity for the refining and
petrochemical industry and the United States. We are
currently witnessing the beginnings of a far reaching
manufacturing renaissance brought on by the recent
increase in domestic oil and natural gas production.
Today more than $100 billion in new manufacturing
infrastructure has been announced and is expected to
be developed throughout the United States during the
next decade.
We are the most fortunate manufacturing region in
the world today with advantaged raw materials,
advanced infrastructure and expertise along the entire
supply chain. The petrochemical industry is on the
verge of an exciting expansion, but it can only
materialise if the federal government stays out of the
way and allows the market to work. My industry will
work tirelessly to advance this development
throughout 2014 and well into the future.

www.eptq.com

11/12/2013 15:58

ExxonMobil
Champion Technology
for Lube Basestocks
MWITM, MSDWTM, MAXSATTM Sweep the Competition

Providing the World with Lube Basestock Technologies and Catalysts

www.exxonmobil.com/tsl

exxon.indd 1

13/12/2013 10:56

pcs myth.indd 1

8/3/12 09:55:39

ptq&a
Q

We are finding a build-up of organic fouling material

in a fractionator after the coker. It does not appear to be
asphaltene based. Is this some form of polymer build-up and
how can we deal with it?

Celso Pajaro, Manager of Refinery Applications, Celso.

Pajaro@sulzer.com, Sulzer Chemtech USA

For high severity cokers, this is not an uncommon

problem; polymer type material is found below the
LCGO draw. The key to minimise its occurrence is to
reduce liquid residence time within the column.
If trays are used, they should be equipped with antifouling features such as fixed valves, sloped
downcomers and flow enhancers to reduce stagnation
and areas of recirculation on the tray deck.
Grid type packings can be used in high temperature
pumparound areas to minimise liquid hold-up and
residence time. Also, spray nozzles should be used for
the feed device. Strainers for spray nozzles should be
well oversided to capture any coke type material that
accumulates in the pipe.
Any pumparound piping operating at temperatures
higher than 500F (260C) should be as short as possible and should be designed for high velocities in order
to minimise liquid residence time.

5. What is the PSD of the slurry fines very fine

indicates attrition, coarse indicates hole/cyclone issues?
6. Do you have historical slurry PSD from a period
where losses were normal?
7. Are the losses only on the reactor side or have
regen losses increased too?
8. What does the whole catalyst mass balance look like:
losses on RX-side, RG-side, CAR, e-cat withdrawal,
e-cat flushing, additives?
9. Is the stripper behaving normally? What is the H-incoke? Has the regenerator T increased with the losses?
10. What are the steam rates going into the FCC any
valves vibrated open (for instance, emergency riser
steam)?
11. What are the bed levels and superficial velocities in
the riser, regen, and so on?
12. Are cyclone inlet/outlet velocities within design
limits?
13. What is the superficial velocity at the feed injector
tips? Air grid nozzles? Too high can cause attrition.
14. Fresh catalyst: is attrition index in the correct
range? Is particle density too low? Are there excessive
fines (0-20%)?
15. Check for substantial changes in slurry yield (feed/
conversion/distillation).

We have problems containing the catalyst in our FCC unit

and this is leading to contamination of our slurry oil. Is there
a simple way of narrowing down the probable cause(s) of our
catalyst losses?

Paul A Diddams, Senior VP FCC Additives, Paul.Diddams@

matthey.com; Bart de Graaf, FCC R&D Director, Bart.deGraaf@
matthey.com; and Charles Radcliffe, Technical Consultant,
FCC and Refining, Charles.Radcliffe@matthey.com, Johnson
Matthey

Troubleshooting FCC catalyst losses is fairly routine for

FCC catalyst suppliers involve your catalyst
supplier in resolving the issue.
Important considerations:
1. Are chemical analyses showing catalyst losses (not
coke or iron scale)?
2. How did the losses start? Following a TAR, unit
upset, feed change or other unusual operational
situation?
3. Was there a step change in losses or did the problem
gradually get worse?
4. What is the magnitude of losses? Consistent with
primary or secondary cyclone issues or attrition
source?

Stuart Foskett, Regional Technical Service Manager for

Asia, stuart.foskett@basf.com, BASF Catalysts

Catalyst loss problems can be the result of a variety of

different issues. In this example the losses are specifically noted as occurring from the reactor side. The first
troubleshooting step is to seek a particle size distribution (PSD) of the catalyst losses by collecting a
representative slurry oil sample, ashing the slurry and
analysing the collected solids. The PSD may have one
or more peaks. A peak at around 1-2 indicates fines
generated by attrition, a peak at about 20-30
represents typical cyclone losses and a peak at around
50-80 indicates loss of full range particles. It is good
practice to obtain a PSD for slurry oil ash and regenerator third stage separator fines at least once per
quarter so there is historical data to compare against in
a troubleshooting situation. BASF offers a slurry and
fines analysis service at no cost for FCC catalyst
customers.
If the majority of losses are attrition fines smaller
than 5, then all potential attrition sources in the riser
and reactor should be investigated. Common problem
areas include feed nozzles (plugged nozzles or erosion)
and stripping steam distributors (erosion, holes,
cracks), and hardware damage from wet steam. In

Additional Q&A can be found at www.eptq.com/QandA

www.eptq.com

Q&A.indd 1

PTQ Q1 2014 17

13/12/2013 11:47

some cases, it is possible to identify the problem area

by reducing gas flow and observing a change in losses,
for example by reducing stripping steam. Excessive
attrition occurring in the riser or stripper will often
result in higher losses from the regenerator as well.
A change in attrition properties of the catalyst itself
is sometimes a contributing factor towards attrition
losses. In general, poor attrition resistance of the catalyst can be ruled out as a primary cause if there is a
sudden step increase in attrition losses. If the increase
in losses is more gradual, check whether the timing
corresponds to a change in catalyst grade, and also
review the fresh catalyst quality analyses to confirm
the catalyst properties remain in the normal range.
A large peak around 50-80 can indicate loss of full
range particles through a holed or cracked plenum, or
by a cyclone failure. Potential cyclone problems can
include diplegs plugged by coke deposits, often associated with incomplete feed vaporisation, failure of a
flapper valve or trickle valve, a holed cyclone and
cyclone dipleg flooding. A change in the cyclone pres-

Excessive attrition occurring in

the riser or stripper will often
result in higher losses from the
regenerator as well
sure drop relative to the expected value is one possible
signal of cyclone issues. Occasionally it is possible to
identify a missing or stuck-open flapper valve by raising the reactor bed level to submerge and seal the
cyclone dipleg.
Other information that can be helpful in identifying
loss problems includes a regular catalyst mass balance
around the reactor and regenerator (both daily and
monthly), historical slurry alumina and BS&W analysis, calculation and monitoring of cyclone velocity,
cyclone pressure drop, steam and air distributor pressure drop, and feed nozzle pressure surveys. As with
all troubleshooting, simply looking for changes in
operating conditions that coincide with the timing of
increased losses may provide further clues of a potential cause.

George Yaluris, FCC Technical Service Advisor, george.

yaluris@albemarle.com, Albemarle

Troubleshooting catalyst losses either from the reactor

or the regenerator side of the unit is not a simple
undertaking. A systematic approach must be applied
to identify the symptoms, quantify the losses, conduct
investigative work to troubleshoot the causes, and
identify the best possible course of action. The good
news is that the source of catalyst losses can be
narrowed down to a short list of possible causes. A
solution without having to shut down the unit may be
possible depending on the nature of the problem.
On the reactor side, catalyst losses manifest themselves as increased solids in the slurry. They can lead

18 PTQ Q1 2014

Q&A.indd 2

to increased exchanger fouling, erosion, slurry pump

failure and the need to dispose of oily fines.
Troubleshooting these losses must start before the
problem begins by routinely collecting data to establish the units normal operation baseline. These data
should include accurate accounting of the amount of
catalyst lost under normal operation, catalyst data
(including fresh, equilibrium, and fines particle size
distribution [PSD], fresh catalyst attrition, ABD, and so
on), collection system and unit design details, and
operating details (for instance, temperatures, pressure
surveys, steam rates).
Once the problem is identified and quantified,
pinpointing the cause involves a detailed review of
possible causes to determine if they can be excluded.
There are three common types of causes: operation
changes, catalyst attrition and collection system hardware failure. The unit operations should be carefully
reviewed for any changes which may be causing or
contributing to the increase in slurry solids. Examples
are increased feed and catalyst addition rates, changes
in the reactor/stripper level, and excessive vapour/
solids flow rates leading to cyclone flooding. The fresh
catalyst properties should also be reviewed to ensure
the attrition resistance and other relevant physical
properties of the catalyst have not changed.
The slurry fines should be carefully analysed for
changes in the PSD and morphology (for instance, by
SEM imaging). This analysis can provide important
clues. An increase in the proportion and absolute
amount of fragmented particles (typically in the 20 to
40 range) may indicate the existence of an attrition
source. On the other hand, a flattening of the fines PSD
distribution and an increase in the amount of larger
particles being lost may indicate a decrease in cyclone
efficiency for example due to cracks and small holes.
The presence of equilibrium size catalyst can indicate
larger holes in the catalyst separation hardware, or
dipleg plugging, flooding or breaking, including
dipleg valves not opening. If the unit includes a device
for separation of solids from the slurry, the operation
of this hardware should also be reviewed.
The data should then be evaluated in an attempt to
locate the area of the problem. In addition to examining the catalyst and fines data as well as the operating
data (such as changes in the unit pressure and
temperature profile), step tests can be conducted to
further troubleshoot the problem, identify the location
and potentially provide relief from the losses. Gamma
ray scans may also be done to identify changes in the
catalyst flow and to help locate the cause.
Troubleshooting a catalyst loss problem requires
detailed familiarity with the unit design, current configuration, hardware conditions and operations, as well as
experience in conducting this kind of work. It also
requires working closely with the catalyst vendor who
will typically provide much of the analytical work
required. If the problem is related to operating or catalyst changes, a solution can often be devised to address
it. If the problem is related to mechanical failure, a
remedy may be possible while the unit is running.

www.eptq.com

11/12/2013 15:54

pcs 1.indd 1

8/3/12 09:50:20

in PNA content from 0.48 to 0.55. Notice also that the

conversion decreases despite the fact that the FCC feed
nitrogen content is at the lowest level suggesting the
PNA feed content has the larger impact on FCC
conversion.
Clearly, driving the hydrotreater to remove nitrogen
and PNAs improves FCC product value, but this needs
to be balanced against the increased costs of higher
hydrogen consumption and shorter cycle length that
result from this mode of operation.
The shorter run lengths from operating in PNA
Q Is it advisable to hydrotreat heavy vacuum gas oil before it mode can be extended with the use of a sulphur reducis fed to the FCC unit? If so, why?
tion catalyst or additive in the downstream FCC unit.
FCC gasoline sulphur reduction of 20% or higher from
A Charles Olsen, DHT R&D and Technical Services Director, a product such as D-PriSM or GSR-5 additives, or
Chuck.olsen@grace.com, Advanced Refining Technologies
SuRCA catalyst would enable the pretreater to run
FCC feed pretreating offers the potential to signifi- beyond the point when product sulphur levels exceed
cantly improve refinery economics. Advanced Refining target levels, while FCC gasoline product sulphur
Technologies has previously described the benefits of remains within the limits needed for gasoline pool
hydrotreating FCC feed on the FCC product yields and sulphur compliance. Extending operation in PNA
sulphur content in earlier publications (see for example mode closes the gap in run length versus a hydroNPRA papers AM 02-39, AM 04-29, AM 07-55, AM treater in the HDS mode of operation, while
11-21). That work has demonstrated that both the continuing to provide FCC feed of higher quality than
hydrotreating catalyst system and the operating strat- HDS mode. Use of OlefinsMax and OlefinsUltra addiegy for the pretreater are critical to providing the tives during maximum LCO operation can assist in
highest quality feed for the FCC as well as the ability maintaining liquid yields and gasoline octane.
to tailor the FCC product slate.
In addition to optimising the hydrotreater catalyst
In general, NiMo based catalysts produce higher system and the hydrotreater operation, the FCC cataquality FCC feed, with both lower nitrogen and PNA lyst formulation can also be included as one of the
content. The operating mode of the hydrotreater can variables that can be adjusted to maximise profitability.
also be used to further improve the FCC feed. It is Adjusting FCC catalyst formulation as the hydrotreater
generally accepted that removing PNAs and nitrogen moves from start of run to end of run can also signififrom the FCC feed will improve FCC performance and cantly offset a decline in product quality and preserve
increase conversion. As an example, Figure 1 product value. In this case, using FCC products such
summarises FCC pilot plant data for several hydro- as Midas for maximum LCO selectivity with improved
treated FCC feeds. The left y-axis represents the FCC bottoms cracking or materials from the Genesis and
conversion and the right y-axis represents the FCC feed Achieve Catalyst technology platforms provide flexibilquality, shown as a ratio of either product nitrogen ity in catalyst formulations for conventional as well as
content to feed nitrogen content or product PNA unconventional feeds.
content to feed PNA content. Clearly, the PNA and
All of the combinations for optimising the FCC
nitrogen content of the FCC feed have a strong effect pretreater/FCC complex demonstrate the need for
on FCC performance. At the highest pretreater severity refiners to follow an integrated approach to managing
the FCC conversion actually decreases by about 1 wt% the catalysts and operation of the FCC pretreater and
relative to the previous severity in the chart. The FCC unit. Grace Catalyst Technologies is well suited to
decrease in FCC conversion coincides with an increase assist with this endeavour through its Advanced
Refining Technologies (ART) and Refining
Technologies businesses. The complexity of
78
0.8
PNA
combinations of catalyst design and operat77
Nitrogen
0.7
ing conditions for both the FCC feed
76
FCC
hydrotreater and the FCC unit presents a
75
0.6
significant optimisation opportunity for
74
0.5
73
refiners to drive the combined operation to
72
0.4
maximum product value.
71
70
69
68
67
66

0.3
0.2
0.1

Nitrogen or PNA content,

product/feed

FCC conversion, wt%

Changes in the catalyst physical properties or the use of

a fluidisation aid such as Albemarles Smoothflow may
help the unit run until the next turnaround, provided
that the failure does not worsen. Of course more serious
hardware failures will eventually require a shutdown to
repair the damage. In this case, troubleshooting the
problem will help the operator better prepare for the
shutdown and locate the damage faster, thus quickening
repairs and decreasing downtime.

Increasing pretreater severity

Figure 1 Pretreater PNA and nitrogen conversion effect on FCC conversion

20 PTQ Q1 2014

Q&A.indd 3

George Anderson, Global Hydroprocessing

Specialist,
george.anderson@albemarle.com,
Albemarle

In general, it is advisable to hydrotreat

heavy vacuum gas oil (HVGO) before it is
fed to the FCC unit. HVGO typically
contains metals (Ni, V), contaminants (As,

www.eptq.com

11/12/2013 15:55

Grace FCC Catalysts

and Additives
Innovative Catalytic Solutions

grace.indd 1

Industrys Broadest Catalyst Portfolio

Flexible Functionality for Processing

Unconventional Feeds

Global Manufacturing Footprint

World-class R&D

Industry-leading Technical Service

10/12/2013 09:55

Typical properties of HVGO Feeds to FCC-PT units

Feed property
Boiling range, F
Sulphur, wt%
Nitrogen, wppm
Aromatics, wt%
CCR, wt%
Nickel + vanadium, wppm
Silicon, wppm

Typical values
650-1100
1.0-3.0
1000-3500
35-55
0-4
1-10
0-4

Table 1

Si), relatively high sulphur and nitrogen, and heavy

aromatics with 3 rings. The range of typical properties for HVGOs is shown in Table 1.
FCC units generally operate better on cleaner feeds.
Removal of Ni and V in an FCC pre-treat (FCC-PT)
unit helps the FCC catalyst (and e-cat) retain better
activity and reduces H2 generation in the FCC unit.
Basic nitrogen titrates the FCC catalyst and inhibits
catalytic cracking activity. Thus, HDN in an FCC-PT
unit has significant FCC unit performance benefits.
Heavy aromatics do not crack in the FCC unit unless
they are at least partially hydrogenated. Many 2-, 3and 4-ring aromatics will pass on directly to the LCO,
HCO and slurry products. The remaining aromatics
will form coke on the FCC catalyst and deactivate it.
For those refineries that do not hydrotreat HVGO feed,
use of FCC catalyst with high accessibility and high
metals tolerance (such as Albemarles Upgrader and
Amber catalyst families) is highly recommended.
Reducing HVGO sulphur content in FCC-PT does
not have a major impact on FCC unit operating efficiency, but it does have a major impact on FCC unit
product quality. FCC naphtha typically contains
sulphur levels that are ~5-7% of the sulphur levels for
hydrotreated FCC unit feed. On the other hand, FCC
naphtha contains sulphur levels that are ~10% of the
sulphur levels for unhydrotreated FCC unit feed. Thus,
the benefit of HDS in the FCC-PT unit is two-fold: 1)
the FCC unit feed has a lower sulphur level and 2) a
lower proportion of the FCC unit feed sulphur will
end up in the FCC naphtha product. As gasoline
sulphur limits are reduced (such as with implementation of Tier 3 gasoline regulations in the US beginning
in 2017), hydrotreating of HVGO feeds in FCC-PT
units becomes even more critical for many refineries.
FCC unit LCO product quality is also impacted by
FCC-PT operations. The benefit of FCC-PT is that the
LCO product from hydrotreated feed contains lower
sulphur than LCO generated from unhydrotreated feed
and generally has higher yield from hydrotreated feed.
This benefit may be partially offset by the fact that the
LCO from hydrotreated feed may be more aromatic.
This may result in higher H2 consumption during
diesel hydrotreating and lower ULSD product cetane
number.
A final benefit from hydrotreating the FCC unit feed
is the beneficial effect on environmental emissions and
related clean-up costs. Reducing FCC unit feed sulphur

22 PTQ Q1 2014

Q&A.indd 4

and nitrogen contents leads to lower SOx and NOx

emissions from the FCC unit regenerator. This translates into lower NOx stack emissions for almost all
refineries, and either lower SOx stack emissions, or
lower treating costs to meet SOx emission regulations.

Paul A Diddams, Senior VP FCC Additives, Paul.Diddams@

matthey.com; Bart de Graaf, FCC R&D Director, Bart.deGraaf@
matthey.com; and Tom Ventham, Senior Technical Service
Engineer, Tom.Ventham@matthey.com, Johnson Matthey

Resid FCC units are designed to process HVGO without necessarily hydrotreating it first.
Advantages of pre-hydrotreating HVGO:
1. Contaminant levels are decreased: V, Ni, N, S
lower delta coke, lower CAR, lower stack emissions
(SOx, NOx)
2. Coke precursor levels are decreased: CCR lower
delta coke, relieves regen T and air constraints
3. Partial aromatics saturation increases feed crackability lower delta coke, improves conversion, increases
gasoline selectivity.
Disadvantages of pre-hydrotreating HVGO:
1. Capex required to build a HVGO hydrotreater
2. Opex required to run a HVGO hydrotreated
3. Increased hydrogen consumption on the refinery
4. More H2S/NH3 to the Claus plant.
Outside the US, the only justification is the increased
conversion (volume expansion if this word would have
been used abroad); inside the US, environmental considerations will have a substantial input to the decision.
We have seen some refineries in Europe, particularly
Eastern Europe, switch to HT VGO FCC feed (or,
perhaps more precisely in many cases, MHC feed) as a
catch-all solution originally forced by a move to Euro
IV/V and the need to reduce gasoline sulphur. In
many of these cases, they then struggle to close the
heat balance. This could be seen as a significant disadvantage in addition to those given above.
So, it is great to have, but a full economic assessment
is required to see if the advantages outweigh the
disadvantages.

We want to send coker naphtha for hydrogen production.

Can you recommend a hydrotreating scheme to facilitate this?

Charles Olsen, DHT R&D and Technical Services Director,

Chuck.olsen@grace.com; Brian Watkins, Manager, Plant &
Technical Service Engineer, Brian.watkins@grace.com; and
Brian Slemp, Technical Service Manager, Brian.slemp@grace.
com, Advanced Refining Technologies

A properly designed coker naphtha treating system for

the preparation of hydrogen plant feed requires special
consideration to mitigate the impacts of contaminants
and olefins. A traditional moderate pressure hydrotreater with significant recycle gas flow and multiple
catalyst beds with sufficient quench to moderate the
exotherm from the large olefin content is preferred to
prepare coker naphtha for steam reformer feed.
Processing coker naphtha can have several undesirable effects on the performance of the hydrotreater if

www.eptq.com

11/12/2013 15:55

service,
support,
solutions.
At Johnson Matthey we dont just provide world
class hydrogen catalysts, we provide knowledge
and understanding of the hydrogen process
and catalytic interactions through a unique
range of services to enhance plant operation
and improve efficiency and reliability.

UK

Tel +44 (0)1642 553601

Fax +44 (0)1642 522542

www.jmprotech.com

j mathey.indd 1

11/12/2013 09:33

the catalyst system is not properly designed to deal

with the unique challenges of the feed. In general,
coker stocks have a higher level of olefins present from
the coking process. Once in the hydrotreater these
olefins will quickly get saturated (see Figure 1) resulting in high hydrogen consumption and generation of a
lot of heat. As a general rule of thumb, one mole of
hydrogen is required per mole of carbon-carbon double
bonds, or between 5-10 times the bromine number
reduction in standard cubic feet of hydrogen per barrel
(scf/b). This additional heat (130-160 Btu/scf of hydrogen consumed), if not managed properly, will initiate
additional reactions, quickly creating a very high
temperature rise. The high temperatures can accelerate
coking and can lead to olefin polymerisation, resulting
in a dramatic increase in pressure drop. This can set an
upper limit on how much coker naphtha can be
processed based on the need to limit the heat rise or
hydrogen consumption.
+H2
(C)n-C=C
(C)n-C-C + Heat
Easy

Figure 1 Olefin saturation

A system that is properly sized and activity graded

is extremely important when processing coker feeds.
ART recommends utilising a grading system to help
mitigate pressure drop build-up. A large inert hold
down ring (GSK-19) with a very high void fraction
used for trapping large particulates is placed at the top
of the catalyst bed. A smaller diameter macroporous
ring (GSK-9) that traps iron as well as other fine particulates is typically used as the next layer. After that,
two types of smaller rings are used as active grading.
These materials have a low level of active metals
which help to begin olefin saturation reactions as well
as provide additional external void space at the top of
the catalyst bed. Avoiding the use of any highly active
catalyst at this point is also recommended. Below the
grading system, it is recommended to use a layer of
larger size (1/10in or 1/12in) catalyst which provides
activity for olefin saturation and additional void space
for pressure drop mitigation. This layer is often a catalyst which is suitable for trapping silicon (and arsenic)
which is another concern when processing coker
naphtha.
Coker naphtha contains a significant amount of
poisons that need to be removed prior to being used as
feed to a hydrogen plant. The reactor must be large
enough to trap the silicon and arsenic normally found
in coker products. ART has extensive experience in
coker naphtha hydrotreating and has several products
designed to combat the contaminants found in coker
products. ART also has a portfolio of high activity
naphtha catalysts designed to not only remove high
levels of sulphur and nitrogen but also remain tolerant
to silicon contamination. ART utilises several grades of
silicon and arsenic trapping materials depending on
the level of contaminants present in the feed and the
need to balance HDS/HDN activity. These grades

24 PTQ Q1 2014

Q&A.indd 5

include: AT724G, which is used in applications requiring moderate activity and high silicon capacity;
AT734G, which has high silicon capacity and high
arsenic capacity; and AT714G which has low activity
and ultra high silicon pick-up capacity.
The potential polymerisation of the cracked naphtha
can be addressed with a diolefin reactor and/or appropriate activity and size grading to ensure a sufficient
run life. A diolefin reactor is operated at a much lower
temperature in order to selectively catalyse the diolefin
saturation reaction and avoid any excess heat generation from sulphur and nitrogen removal. Liquid
product recycle may also be considered as an additional method to provide a heat sink to control reactor
temperature rise.
Coker naphtha is known to vary dramatically
during the various portions of the coking cycle.
Special consideration of the feedstock design basis
needs to reflect the various portions of the coking
cycle. This will ensure the new hydrotreater design
can accommodate not only the average feed proper-

A system that is properly sized

and activity graded is extremely
important when processing
coker feeds
ties, but both the maximum and minimum naphtha
properties experienced during the coking cycle. Steam
reforming hydrogen production is an extremely high
temperature operation sensitive to carbon lay down
and coking. Naphtha end point control and carbon
speciation are important considerations to determine
if the hydrogen unit reformer is capable of handling
variable higher carbon number feeds. The use of the
pre-reformer may be required to facilitate the feed
change.

Q We are finding particulates and water in our product streams.

Is it best to trace back for the problem or is there some means of
separating the contaminants without further processing?

Paul A Diddams, Senior VP FCC Additives, Paul.Diddams@

matthey.com and Charles Radcliffe, Technical Consultant,
FCC and Refining, Charles.Radcliffe@matthey.com, Johnson
Matthey

Difficult to generalise without knowing which

product(s) but definitively identify and fix. The following are good general rules for troubleshooting:
a) Check levels in water boots of upstream vessels
b) Any aqueous treatment upstream check levels and
condition of demisters; check for foaming
c) Check upstream water coolers for leaks, with water
side the higher pressure
d) Analyse water for clues as to source, for instance
cooling/BFW/condensate water treatment chemicals
e) Analyse particulates for clues to source.

www.eptq.com

11/12/2013 12:04

Were local. Were global.

And were proud to be both.
For over 60 years, weve been there for our refining
and petrochemical customers. Linde Engineering
North America Inc. offers single source responsibility
for technology, engineering, procurement and
construction - T-EPC.

Selas Fluid refinery and petrochemical fired heaters

Oxidation and incineration technologies
Engineered revamps and rebuilds
Hydrogen and synthesis gas plants
Air separation plants

Headquarters: Five Sentry Parkway East, Blue Bell, PA 19422 USA Phone +610.834-0300
Texas Office: 3700 West Sam Houston Pkwy. South, Suite 425, Houston, TX 77042 USA Phone +281.717-9090
Al-Khobar: Linde Arabian Contracting Phone +966.3.887-0133
www.lindeus-engineering.com

linde NA.indd 1

12/09/2013 11:06

PTQ&A

copy 1

8/3/06

21:12

Page 3

QUESTIONS
A Celso Pajaro,
A Luigi
Manager
of Refinery
Applications,
Celso. from
3.5Bressan,
wt% toDirector
0.5 wt%
byTechnologies,
2020. The Marine
ofglobally
Process
and
Foster
challenges
remain
in process
configuration,
catalyst
content
of 5-8 wt%
largely
present
in the unstripped
Typical
nitrogen
balance:
wt%
feed
nitrogen
to
FCC
products
Pajaro@sulzer.com,
Sulzer
Chemtech
USA
Environment
Protection
Committee
(MEPC)
alsoto
Wheeler
Italy
development
and
materials
development
when
before
etherification.
A
TAME
unit
is
an
opportunity
corrosion occurs. Once sensors are corrosion also declined
products. substan- India Nagi-Hanspal is Lead Refinery Engineer:
want
tomajority
keep
the of
FCC
at pool.
fullthe
feed
rate.
An option
would
be
The
presence
of
water
is
typically
due
to
the
use
of
agreed
to
use
equivalence
technology
such
asto
In
the
cases,
feedstock
dictates
which
membranes
are
employed
for
optimisation
of
the
flow
Technical
Services
with
Dorf
Ketal
Chemicals,
into
reach
a
higher
RON
Oligomerisation
of
C
installed, the system collects data tially, illustrating the
importance
of
The
levels conversion
of sulphur
and nitrogen before
in cokethe
areFCC.
much
5
add
partial
hydrocracking
The
stripping
steam.
One
way
is
to
change
strippers
from
on-board
scrubbing
for
post-treatment
instead
of
using
Mumbai,
India,
and
is
primarily
responsible
process
is
more
attractive.
Generally,
where
it
is
possidiagram.
Oxygen
supply
(by
definition)
is
also
a
distillate
(PolyFuel
process)
is
an
opportunity
to
switch
and
reports trends over time.
corrosion
Product
Nitrogen, wt%FFmetallurgy selection
higher
than
in the
feed. In (HC)
the regenerator,
feedfor
rate
to the
hydrocracker
can be fixed carbonaso that the
steam
option
install
either
low
sulphur
bunker
fuels.
for
providing
refinerythis
improvement
ble
to
use catalytic
steam
reforming,
technology
is
requirement
for a another
partial oxidation
plant
(both
CPOa ceous
Fuel gasto
NHreboiler,
/HCN
5-15is to
gasoline
diesel,
particularly
inprocess
Europe,
control.
3
unstripped
hydrocarbons
are
FCC
cancoke
beinto
keptand
full,
but
there
is substantial
additionalwhere
diesel
strategies
through
unit tomonitoring.
She
Gasoline
1-5
salt
dryer
or
high
efficiency
coalescer
(like
Sulzer
In
reality,
it
will
likely
be
difficult
implement
preferred.
Feedstocks
ranging
from
methane
down
to
and
POX).
refiners
face
surplus
gasoline
production.
HTCI
case study
produced. Of
some
VGO isisrequired,
andadditional
H2O. Sulphur
oxidisedeither
to
combusted
to course,
CO2, CO
LCO/diesel
provides
chemical
solutions
to2015.
Dorf Ketal
site
Dusec).
These
devices
are refineries
able10-20
to
remove
submicron
post-treatment
scrubbing
in
the
ECA
Thus,
naphthas
can
be
readily
processed
inruns.
a by
catalytic
steam
In catalytic
steam
reforming,
technology
licensors SO
purchased
or through
additional
crude
One
of the
worlds
largest
Conclusion
,
SO
,
COS
and
H
S.
Nitrogen
behaves
differently;
HCO/bottoms
25-35
3
2
2
teams
globally
and
holds
a
MEng
degree
in
water
droplets.
Bothwith
needanthe
stream
to
be atproper
storage
clean
maritime
fuel
is feedstocks,
the
more realistic
solution.
This
reformer;
for heavier
partial
oxidation
isbe
offer
commercially
proven
solutions
that
are planning,
The
pressure
of the
partial-conversion
HC
needs
to
With
high-acid
processed
blends
average
Coke different
35-60
CO
as
oxygen
reacts
with
the
coke
matrix
to
form
CO2,College
chemical
engineering
from
Imperial
Q
temperature
to minimise
water
solubility
isoptimised.
expected
tohigh-pressure
increase
demand
for
marine
distillate
oil
A
unit
be
able to
make ultrathe
only
alternative.
based
on
furnace
technology,
andtoitmaximise
remains
What
arrangement
trayswill
andmay
packing
designs
will
can be processed
successTAN primarily
of 1.31
mgKOH/g
and
a crudes
O,viable
much
of London.
theofnitrogen
initially
form
and
H
Email: indianagi@dorfketal.com
2
low-sulphur
diesel
(ULSD)
directly.
However,
capital
costand
is
water
removal.
(MDO)
and
marine
gas
oil
(MGO)
by
approximately
the
most
widely
applied
technology
for
producing
give
me
the
best
efficiency
of
fractionation
in
my
crude
strategies are
sulphur
content of 1.28 wt%. The fully if appropriate
Table
1
HCN.
high.
A less
capital-intensive
option
is to use
a lower-pressure
A
Particles
are
typical
corrosion
products
that
can
be
500
000
b/d.
The
increased
demand
will
likely
be
Lorena
Mosca,
Technology
Project
Coordinator,
KThydrogen.
Steam
reforming
is
particularly
effective
vacuum
distillation
units?
adding
high- toMahesh
heavy vacuum gas oil loop was implemented before
Subramaniyam
is Director
of
This
is similar
the post-treat
chemistry
inResearch
coal
HC,
with
some
of observed
the
diesel integrated
minimised
by
metallurgy
in the
points
supplied
from additional
distillate
production
existing
Kinetics
Technology,
l.mosca@kt-met.it
when
methane
isupgrading
available
as
an economical
source
acid
crudes
toofthe combustion.
feed
Success
monitored
closely
& Development
withdiesel
Dorfusing
Ketal
Chemicals
and
compounds
present
in for
FCCcorrosion.
feedstocks;
these
are
measHCN
aother
thermodynamically
unstable
withdiet.
the HC.
If there is
are
difficult
stocks
that need
Aunderstanding
where
corrosion
is higher.
Another
option
is installing
hydrotreating
capacity
asthesuch
well
as
additional
capacity
If
consider
catalytic
oxidation
ofcycle
natural
gas
hydrocarbon
example,
from
shale
Pajaro,
of
Refinery
Applications,
Sulzer
leadspartial
companys
developments.
requires
a gas).
thorough
TANand
in reported
the feed
stream
was
approxitowe
be Celso
treated
inFCC
the Manager
refinery,
asconditions
FCCchemical
light
oilgiven
(LCO),
ured
as(for
total
nitrogen
and
basic
nitrogen.
species
under
regenerator
and,
particle
filters,
or to
just1.19
allow of
a product
storcoming
online
inCelso.Pajaro@sulzer.com
the near
against
steam
itfuture.
compete
economically
However,
in cases
where
heavier
areprocesses,
holds
awill
number
of patents
for
oil
treatment
USA,
itChemtech
may identificaalso
be reforming,
possible
to
economically
utilise
the
corrosion
mately removal
1.79
mgKOH/g
Typically,
about
30-50%
ofwith
the
feed hydrocarbons
nitrogen
is basic.
sufficient
time
and He
temperature,
HCN
will
bepost-treat
fully
age
tank
to
act
like
a
big
settler,
allowing
the
In
order
to
meet
the
2020
mandates
(S
<0.5
wt),
the
chemicals
and
engineers
products
to
address
removal
from
syngas
is
mandatory
when
CO
used
as
feed,
a
pre-reformer
is
usually
applied.
This
unit
for
them.
There
is
no
specific
arrangement
of
packing
and
trays
tionacid
ofsites
susceptible
systems
wt% nitrogen
sulphur. species
The keystrongly
performance
2andand
These
adsorb on
on converted
no trace of HCN will be found in(zero
the
accumulation
of
the
particles
which
will
be
removed
industry
will
require
approximately
2-3
million
b/d
of
emerging
issues
in
the
hydrocarbon
industry.
With
regard
to
FCC
LCO,
this
stream
is
an
inferior
emission
hydrogen
production).
In
all
other
cases,
CO
not
only
increases
costs,
but
also
the
susceptibility
to
for
getting
the
best
efficiency
of
fractionation.
The
choice
monitoring
to adapt
indicator
theare
trial
was corrosion
2
the
catalystfor
and
thereby
transportedconsistent
with the cataregenerator.
Under FCC regenerator conditions, N2 is
potential
diesel
component
with
its in
low
cetane
value,
even
Hecover
holds
a PhD
chemistry
from
Indian
during
tank
maintenance.
clean
bunker
fuel
to
increased
global
consumpit
depends
on
the
cost
of
the
oxygen
available
at
the
upsets.
In
these
cases,
partial
oxidation
becomes
a
in
column
internals
is
a
function
of
the
type
of
crude
as system
dynamics
and nitrogen species, followed by NO,
control
maximum of
8 mpy.
the
most
stable
lyst
into at
thea regenerator,
where
they treatment
are combusted
after
hydrotreating.
Partial
hydrocracking
of also
LCO
(at
Institute
of Technology,
tion.
This
may
athe
significant
shift
inused
quality
from
isMumbai.
industrial
site. cause
Refineries
where
O
competitive
being
processed
and
column
diameter
available.
crudeabout
diets half
change.
Dorf Ketal
HTCI
treatment
results
2
which
at
thermodynamic
equilibrium
should
reachfor
a
together
withoption.
the coke.
As a rule
of thumb,
moderate pressure) Email:
can drmaheshs@dorfketal.co.in
be
considered
to upgrade
the
heavy,
high
sulphurunits,
fuels
tobeen
lighter,
lower-sulphur
fuels
sulphur
recovery
wastewater
treatment
orof
Although
CPO
of methane
still
verycombusted
much in the
Structured
hasfor
used
in
every
are
in
Figure
6. endsisup
concentration
about
10
ppm.
Evidently,
ofQ
theshown
feed total
nitrogen
being
in stable
material.
Whilepacking
someofdiesel
molecules
will
be lostsection
tonitrolighter
A
big
drop
in
the
allowable
sulphur
content
of
marine
(MDO,
MGO).
For
this
to
occur,
there
will
need
to
be
a
gasification
are
candidates
for
low-cost
oxygen
development
phase,
it
remains
theoretically
favourable
crude
columns,
often
to
substantially
increase
efficiency.
is a mark
Cormon
and
by an independent
material,
the diesel
retained
is upgraded to aequilibrium,
higher-quality
gen
chemistry
is not
at thermodynamic
the Testing
FCC regenerator.
Table 1 thirdshows CEION
a typical
FCC of Teledyne
Parag
Shah
workscapacity
in Global Refinery
Technical
gas
oil
is
looming
(from
1%
down
to
0.1%
from
2015).
Is
it
considerable
increase
in
global
for
bottoms
balance
if
one
can
manage
production.
in
terms
of
energy
and
CO
Hydrosteel
is
a
mark
of
Ion
Science.
In
many
cases,
the
efficiency
within
a
given
section
can
diesel
blending
stock.control.
party laboratory
confirmed
the
2
but
is under
kinetic
nitrogen
balance.
Services withhydrotreating
Dorf Ketal Chemicals,
he
possible
that wecarry
meet
thisreaction
usingAs
existing
hydrotreating
andby secondary
by where
2020.
to
selectively
the
of on
methane
with upgrading
be increased
two to three
theoretical stages
each.
effectiveness
ofcould
the out
formulation.
Coke
composition
varies,
depending
feed propis
instrumental
in
software
development
for
capacity
in
the
refinery
orinteresting
will we
to invest inhave
new plant? Consequently,
fuels
will
needcan
to be
be
at a
Whatits
arebunker
some
of
the
most
attractive
isomerisation
oxygen
toseen
CO.
Some
concepts
However,
indesalter
the
top
section
asold
challenge
can be
in
Figure
7, need
HTCI
ofuse
coke
combustion
erties
and
stripper
efficiency.
CokeReferences
consistsbeen
of Investigation
adequacy available
testing and
monitoring
Qwith
configurations
and
catalysts
to
meet
the
premium
to
justify
the
substantial
refinery
investment.
We
are
maximising
propylene
in
the
FCC
and
would
also
presented
for
the
debottlenecking
of
existing
refinery
particularly
corrosive
crudes.
Monel
and
AL6XN
reduced corrosion
by more
Temperature
programmed
(TPO) may
beHis
carbon-rich
polycyclicrates
aromatic
structures
containing
fouling
inoxidation
preheatisomerisation?
exchanger
trains.
1 and
Houlton
G, Crude demand
toan
increase,
feedgrowing
demand
forshipping
light
paraffin
What
A
Hiroshi
Toshima,
Global
Application
Marketing
As
alternative,
industries
have
been
serilike
to
maximise
xylenes
production
from
our
FCC
gasoline.
steam
reformers
by
adding
a
CPO
unit,
which
is
materials
have
been
used
successfully
in
these
applicathan 80% inand
a laboratory
experience
includes
than
eight years
in
used
to
study
coke
combustion.
In themore
TPO
experiment,
heteroatoms
contaminantblend
metalsquality
as changes
well as
in store,
Oil and
Gas
Journal,
can
be
done
to lower
the equipment
cost,
such
as
the
Specialist,
HPC,
hiroshi.toshima@albemarle.com,
ously
examining
implementation
ofproperly
scrubbing
units
additional
process
step(s)
do
wecatalyst)
need?
always
investment
in terms of(for
sizeAlbemarle
and
capi- aWhat
tions,
but
care
needs
to
be
taken
to
match
the
with aa smaller
TAN
of
5.0 mgKOH/g.
the
worlds
largest
grassroots
refinery
and
he
sample
of
coked
catalyst
(spent
is
heated
in
unstripped
hydrocarbon
products
example,
recycle
hydrogen
compressor?
2010.
The
International
Marine inOrganisation
(IMO) has
using
currently
available
high
sulphur
fuel
oil (HSFO).
tal
when
reforming
or
correct
materialto
with
and
type
ofengineering
corrosivity.
holds
in chemical
from
Although
the reduction
corroinert
upthe
toalevel
aBEng
chosen
initial
temperature
10-30%
of compared
coke
may tobesteam
gasoline,
diesel
andadding
oila of an
2 Zhang
L,fuel
Shortage
crudes,
not atmosphere
products,
A
amended
MARPOL
Annex
VI
to
decrease
the
sulphur
The
technology
has
been
commercially
proven
to
Stuart
Foskett,
BASF
Regional
Technical
Manager,
Stuart.
POX
unit.
Mumbai
University.
In
such
corrosive
sections,
trays
have
anto
innate
Bruno
product
line
Axens,
bruno.
sion on
carbonthat
steelcould
is clearly
thestripped
drive Asian
refining
Oil 150C).
andDomergue,
Gas Journal,
(say,
The gas
supply
is manager,
then switched
the
range
products
not be
from
themarket,
content
of
bunker
fuel
from
1.0
wt%
to
0.1
wt%
in
the
reduce
SOx
emission
by
97%
or
higher.
It
also
prevents
Email:
paragshah@dorfketal.com
Foskett@BASF.com
domergue@axens.net:
Gasoline
specifications
are
getting
advantage
due
to
their
higher
thickness.
Some
heavy,
2011.a hydrogen combustion gas-containing oxygen (and optionally other
most ofdramatic,
410 Typically,
stainless steel
pores
the catalyst).
coke has
designated Emission Control Areas (ECA) by 2015, and There
increased
demand
for MDO
and MGO
supply.
tighter when
it
comes
to octane,
sulphur,
olefins the
and
only
limited
flexibility
when
adjusting
selectedis gases).
The temperature
is ramped
(at, say,
aromatics
yield
in
FCC
gasoline.
Increasing
the
catalyst
15C/min)
to 730C, then held isothermally.
A Charles
Radcliffe,results
Technical aConsultant,
FCC and
Refining,
rare
earth content
small increase
in aromatCombustion
flue gases areincontinuously
monitored
by
Charles.Radcliffe@matthey.com,
JohnsonUnfortunately,
Matthey
ics
by
hydrogen
transfer
reactions.
this
infrared and mass spectroscopy.
Difficult
to answer
without more
but there
is
opposite
the requirements
for specifics,
maximum
The
first to
nitrogen
species evolved
(at the propyllowest
are two
options
for possibly
meeting
this:
ene,
where
hydrogen
transfer
should
be
minimised
to
temperature) is HCN (see Figure 1). HCN generation
a) More olefins
active from
catalyst,
possibly
densetoload
if not
preserve
cracking
by
ZSM-5
propylene.
starts at very low temperatures, already around 450C
already done
so; more
frequent
change to
out;maximise
if using
Otherwise,
simply
operating
the
(780F).
At this
low temperature,
the unit
rate of conversion
mixed Co/Mo
&
Ni/Moboth
thenthe
perhaps
shifting
more
conversion
will
enhance
propylene
yield
is
of HCN to thermodynamically more stable NO or Nand
2
towards Co/Mo
will help.
aromatics
in
the
light
FCC
gasoline
fraction.
relatively slow. NO is not observed until the flue gas
b)An
Remove/reduce
the highest
sulphur occurs
feed compoincrease
in the
benzene
temperature
reaches
about
600Cyield
(1100F). Aswhen
the
nent (LCO the
andpropylene
coker gasoil).
increasing
yield,
as
shown
in
1. A
temperature is increased further, the HCN Figure
concentrapossible
explanation
for this are
cyclisation
reaction
declines
as NO increases.
Thethe
HCN
is no longer
tions
involving
light
olefins,
although
less
clear,
stable
to survive so it and subsequent N in
Q Weenough
suffer
occasional
corrosion
problems
with the
commercial
data
also
a .possible
increase
in
Despite being
thercoke
are converted
intosuggest
NO or N
2
exchangers
in ourassociated
crude unit with
overhead.
We assume
these
total
aromatics
a
higher
propylene
modynamically unstable under FCC regenerator
eventsand/or
depend ZSM-5.
on variations in crude quality. Is there some
yield
conditions,
N2O is also observed at similar temperaway
we aromatics
can monitor
and
treat flows
on-line? aromatics
Or is new
Since
not crack,
FCC
O may also
be a
tures to NO. Thisdoindicates
that
N2gasoline
metallurgy
aincrease
more economic
option?
will
also
with
a
higher
mono-aromatics
reactive intermediate in the formation of NO.
content
in the
feed.content
For units
with
a feed hydrotheFCC
oxygen
of the
combustion
gas
AIncreasing
Berthold
Otzisk,the
Consulting
Engineer,
Otzisk@kurita.de,
treater,
increasing
hydrotreating
severity
from 3 vol% to 4 vol% increases the rateorofseeking
HCN
Kurita Europe
GmbH
catalysts
with
greater activity
for traces
aromatics
saturation
conversion
substantially.
The first
of NO
in flue
The main
question
is whether there
is a permanent
will
increase
mono-aromatics
by
saturating
more
gas form at a lower temperature of about 550C
corrosion attack,
resulting
in occasional
heat
exchanger
multi-ring
aromatics,
although
hydrogen
consumption
(1000F).
leakages.
will
alsosimple
increase.
These
experiments show that HCN is readily
Crude refiners
oil contains
various
chloride
salts. The
initial
Some
believe
that higher
aromatics
producformed during coke
combustion
at low
temperatures,
corrosion
control
system
is
the electrical
desalter,
KIDExtractor
Ltd. gasoline
tion
can
be
achieved
by
processing
FCC
After
you
have
used
especially if oxygen concentrations
are
high at these
P.O. Box
11, Zebbug
where
corrosive
salts are
removed
from
crude
oil. The
it the
first
time,
the
through
the
reformer,
typically
atemperature
heart-cut
(170-300F,
MALTA
low
temperatures.
Once
the
is high
Kid
will
crude
oilExtractor
quality
can
be a prime
reason
for operational
Tel.
00356-21-462891
76-149C).
FCC
gasoline
is
a
poor
reformer
feed,
enough, the HCN becomes Fax.
unstable
and is readily
00356-21-462755
become
your
trusty
problems.
Some
crude
oils
tend
to create
very stable
with
higher
nitrogen
and
N+2A
than
straight-run
converted to NOx or N2.
MOBILE: 00356-94-20596
hydraulic-tubeemulsions,
in unwanted
desalter
upsets
feed,
and resulting
olefinsof that
must
be
saturated
in and
the
E-mail:
The
implication
this
is
that
theidrojet@videobank.it
points of greatest
bundle-extractor
Website:potential.
www.kidextractor.com
increased
overhead corrosion
Sodium
hydrotreater.
HCN formation in the FCC regenerator are likely to be

The Premium Alternative

in Process Simulation

Kidextractor

Quarter page

26
PTQQ4
Q1 2013
2014
6 PTQ
82 PTQ Q2 2013
www.eptq.com

ENQUIRY NO

214

www.ptqenquiry.com
8 PTQ Q3 2012

www.eptq.com
www.eptq.com
www.eptq.com
PTQ Q3 2013

& ANSWERS

high-sulphur c
deposition, w
Of the va
specially des
chlorinate
Sulzers
VG A
length
and
becausema
it
If corrosion
being by f
trays, such as
valve trays w
aromatics,
wh
widest
operat
accordingly. W
from these va
than the C5/C6
level
and mini
and olefins,
iso
est
possible
ef
a processing
For theare
kero
options
ava
most valve
simple)
UFM
tra
favourite), pa
ran
structured
separation
and
AGO
fractiona
geographic are
The wash o
chlorinated alu
rates
coupledp
the highest
structured
catalyst. It pac
als
designed
struc
costs, owing
to
(for recycle
high
yields un
of
of conditions.
Margaret Stin
structured
pac
equipment re
ciency and cap
uop.com: Th
The common
strippin
most
with
a variabl
for light
paraf
Structured
pac
chlorided cata
process
all f
some
success;
hydrocarbons
stripping.
Ho
coupled
with
somewith
diffi
reach high iso
For vacuum
efficient and d
choice
due toha
Par-Isom
mum
pressure
converting
an
vapourisation.
unit, to light p
is required,
be
section,
trays
major
equipm
tured packing
additionaldrop
fee
pressure
typical hydrotr
several vacuu
the highest-ac
structured
pac
market today.
case
fromwith
water cru
or
subject to foul

Stephan
Zu
isomerisation
zuijdendorp@
Q
Is there a
chlorinated alu
gasification ove
platinum inve
high conversio
A
Girish
K In
Ch
ATIS
range.
and
Licensing
design, and
improvements
Company,
giris
reduction. Dep
Hydrogen
dem
and economic
increase,
with
hydrogen recy
cracking units
interest. The lo
clean
fuels m
compensated
b
heavier
crudes
by cracking.

Historically
With
th
hydrogen
from
octane-e
gasoline
produ
feedstock
for
gen
demand.
what opport
has advanced
capacity and
hydrogen
similar to Figy
pressures,1,2 it

18
121

P T Q Q2 2006
Q&A.indd
6
Q&A
copy 15.indd
2
j matthey.indd 2

13/12/2013 13:55
11:48
16/09/2013
11/03/2013 13:02

chloride is thermodynamically stable and will leave

the fractionator with the bottom product. Magnesium
and calcium chlorides thermally decompose at relatively low temperatures. About 90% of the MgCl2 and
about 15% of the CaCl2 are converted to metal oxides
and hydrochloric acid (HCl) in the presence of water
vapour (hydrolysis). The HCl boils overhead in the
distillation column and dissolves in the overhead accumulator drum. Severe corrosion of HCl occurs at the
dew point and the rst drops of condensate have a
very low pH of 12. Corrosion protection can be
achieved by neutralising the acidic species with suitable neutralising amines to a pH of 5.5-6.5.
Ammonia is not a suitable amine because of its high
volatility. The disadvantage of many neutralising
amines is the formation of corrosive amine salts. They
have a low pH and require a good water washing
system for salt removal. Kuritas patented ACF technology is a perfect corrosion inhibitor alternative,
based on a very strong organic amine. The formed
ACF salts are liquid salts with a neutral pH and very
high water solubility. Additionally, a lming amine can
be injected to provide a good corrosion protection of
the metal surface.
A perfect corrosion control programme combines
corrosion inhibitor injection with the execution of
analysis methods. The installation of electrical resistance probes at several locations in the overhead
system may help to monitor corrosion rates continuously. Operations, which mature daily, are the
analysis of the overhead receiver water draw. A
continuous online pH monitoring should be done
there, because dew point pH measurement is not
feasible. The iron concentration in the accumulator
drum should not exceed 1 mg/l in the dened pH
range of 5.5-6.5. When the salt content of the desalted
crude oil is 3.5 mg/l, the chloride content of the overhead receiver water can be expected to be 15-20 mg/l.
Corrosion and fouling of the overhead system may
occur when the chloride content exceeds 20 mg/l. A
50 ppm concentration of hydrochloric acid is quite
corrosive to carbon steel.
The use of corrosion inhibitor programmes is a
cost-effective, attractive option. The
installation of dosing equipment can be
NOX
realised easily. In case of very severe
corrosion the selection of corrosion
resistant alloys (CRA) should be the
rst choice. They can resist corrosion
more effectively than carbon steel materials. CRAs are far more expensive but
longer lasting materials. These alloys
are mixtures of various metals such as
iron, molybdenum, nickel, chromium,
copper,
cobalt,
tungsten
and/or
titanium.

Charles Radcliffe, Technical Consultant,

FCC and Refining, Charles.Radcliffe@
matthey.com, Johnson Matthey

N2

temperature and pH, and tends to be localised in the

dew point zone. High TAN crudes and high sulphur
crudes are generally worst. Talk to chemical treatment
companies about corrosion monitoring and institute a
control programme, typically a combination of lming
amine to protect steel and neutralising amine to raise
pH.

What is the best approach to cut NOx emissions from an

ethylene cracking furnace?

Paul A Diddams, Senior VP FCC Additives, Paul.Diddams@

matthey.com; Charles Radcliffe, Technical Consultant, FCC and
Refining, Charles.Radcliffe@matthey.com; Wayne Armstrong,
Business Manager Europe, FSU & Africa, Wayne.Armstrong@
matthey.com; and Joseph McCarney, Business Development
Manager, Joseph.McCarney@matthey.com, Johnson Matthey

Growing concern over air quality and local pollution

has led to an increased interest in abatement of NOx
emissions. NOx may be a problem in a number of renery
processes including steam methane reforming, uid
catalytic cracking as well as renery heaters and boilers.
NOx emissions may be substantially reduced (over 90%)
via selective catalytic reduction (SCR). Ammonia (NH3)
is introduced into the exhaust gas where it mixes with
the NOx and as the homogeneous NOx/NH3 mix ows
through the channels of the honeycomb SCR catalyst the
NOx is neutralised at the surface producing nitrogen
and water (see Figure 1). SCR requires an exhaust gas
temperature in the range 200-480C. The process can be
optimised for minimum pressure drop.
In general, burner NOx depends on ame temperature. LoNOx burners are designed for staged
combustion to lower peak temperature. If not already
tted then this is the rst step; if tted, check operation for even burn and try adjusting primary/
secondary air ratios.
a) Eliminate nitrogen sources in the feed
b) Use low temp processes such as low NOx burners
c) Selective catalytic reduction needs an operating
temperature in the range 200-480C and needs reducing agent ammonia.
NH3

Basic reactions
4 NO + 4 NH3 + O2 4 N2 + 6 H2O
6 NO2 + 8 NH3 7 N2 + 12 H2O

Undesired parallel reactions

SO2 + O2 SO3
NH3 + SO3 + H2O NH4 HSO4

H2O

Crude overhead corrosion is related to Figure 1 Reactions in selective catalytic reduction

www.eptq.com

Q&A.indd 7

PTQ Q1 2014

27

11/12/2013 15:55

Were Big
without the burdens

As a Top 10-ranked firm in refineries 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 flexibility and responsiveness that belies our size
thats why were big without the burdens.

People Oriented...Project Driven

For more information, contact us at processplants@mustangeng.com,

call 713-215-8000 or visit www.mustangeng.com/process.
mustang.indd 1

09/12/2013 11:29

Processing Trends
Whose responsibility is reliability?
Robin Peatey Foster Wheeler

n todays world of increasing costs and challenged

profit 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 refinery facilities) the meaning of
reliability is frequently misunderstood. Using examples
from everyday life, the article will start with a definition
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 refineries, with
ever increasing pressure on margins, ensuring assets
remain operational has more significance now than
ever before. It is frequently the case that those facilities
that are reliable are also successful financially, environmentally and in terms of safety.
Reliability is often seen as a function of the maintenance 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 definition
from Lees Loss Prevention in the Process Industries: reliability 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 definition 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 specified time

www.eptq.com

proc trends copy 11.indd 1

period. Examples are tubes in fired 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 definitions, consider what
reliability can mean for an item with which we are
probably familiar, the family car. A possible definition
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 NCAP1
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 specified 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 considered 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/maintenance or that can suffer a random failure in service is
very common in process industries.

PTQ Q1 2014 29

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
Reliability theory
at end of their life.
Now that we have established that reliability is more
On the surface, this seems to be a sensible theory.
about maintaining a required function than on whether However a 1978 report for the aviation industry5
an individual component fails, let us consider what we showed that failures in real life seldom follow this
can do to protect the reliability function when an indi- pattern. They identied that failures could be categorised by six failure curves, with 11% of failures
vidual component within a system fails.
Classical reliability thinking believed that most aligning with a time based failure mechanism (see
components followed the bath tub failure pattern Figure 2).
This is important to understand as there are still
described in Figure 1.3
many facilities and companies that
base their maintenance strategies
on the belief that equipment mainly
fails based on time or time in
Useful working life
Infant
Falling
service.
(random failures)
mortality
apart
So if 89% of components do not
Optimum point
fail
based on time in service, how
to replace
do they fail? Figure 3 shows the
percentage of component failures
that align with a number of variaTime
tions of possible random failure
patterns.
Figure 1 Bath tub curve
It may be thought surprising that
components that exhibit a high failure rate during early service life
and then settle to a low random
7%
failure rate for the rest of their
service life account for almost 70%
of components.
4%
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 possible. However, strategies also need
14%
to be in place to manage both the
2%
random failures and the relatively
small number of failures that can
be predicted based on service life.

5%

Failure rate

Failure rate

Failure rate

Failure rate

Failure rate

Failure rate

Failure rate

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.

68%

Time
Time

Figure 2 Time-based failure curves

30 PTQ Q1 2014

proc trends copy 11.indd 2

Figure 3 Random failure curves

Reducing infant mortality

If a high percentage of equipment
failures are a result of infant mortalities, 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

www.eptq.com

11/12/2013 15:52

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 NOx 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.

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.

uop next gen.indd 1

09/12/2013 11:34

Creating engineering solutions since 1953

PARIS / BEIJING / BUCHAREST / BUZAU / DAMMAM / DUBAI / HOUSTON / JOHANNESBURG

LOS ANGELES / MOSCOW / MUMBAI / PUNE / RIO DE JANEIRO / SEOUL / ST. PETERSBURG / TULSA

heurtey.indd 1

10/12/2013 13:46

this article. However, in the authors experience, to

successfully reduce infant mortality failures an organisation 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 profile 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 infamously as the elk test2
Explosion at the Nypro (UK) Caprolactam Plant at
Flixborough due to failure following a modification.3
Let us consider in more detail an example of the
poor reliability of a pair of refinery 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 identified 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 significant 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 analysis (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, commissioning 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 corrosion and/or many other failure mechanisms or
non-compliance can occur. Such failures are often
unexpected, can present significant safety risks and
often result in long system outages whilst the correct
equipment is obtained. Robust processes for the

www.eptq.com

proc trends copy 11.indd 3

specification, 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 catastrophic 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 corrosion and/or mechanical damage, it can either cause
delays due to the part needing to be replaced because its
poor condition has been identified 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

Figure 4 Oil drum store

PTQ Q1 2014 33

11/12/2013 15:52

The
safe
mu
pro
fun

Infant mortality
restarted

Initial period of
infant mortality
Failure rate

Rainwater

Unnecessary
maintenance

Time

Figure 5 Water ingress into oil drum

Figure 6 Infant mortality reintroduced by unnecessary maintenance

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 construction 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, reintroducing 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

proc trends copy 11.indd 4

John Moubray4 states that for the failure modes experienced 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 signicant 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 failure of an item and supports the identication of
suitable condition monitoring tasks and the most sensible 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 variety 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

www.eptq.com

11/12/2013 15:52

YouR BeneFit:
lowest liFe cYcle costs

Full Range:
Rod load up to 1'500 kN/335'000 Ibs
Power up to 31'000 kW/42'100 HP

API 618
Reliable SwiSS Quality
You get moRe than Just a
PRocess gas comPRessoR
Lubricated up to 1'000 bar, nonlubricated up to 300 bar

burckhardt.indd 1

For highest availability: We recommend 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

7/3/12 21:39:05

Failure starts
here (P)

Vibration analysis
detects problem

100%

part of achieving high reliability and avoiding expensive equipment repairs.

Note that for components other than bearings the P-F Curve and the types of
condition monitoring tasks are likely to be
different.

(Typically) problem
detected 6 weeks
before failure

Condition

Operator
touch detects
problem

(Typically) problem
detected 2-3 days
before failure

0%

Time

Figure 7 P-F Curve for a bearing failure

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 bearing 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 benefit.
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
significantly less sensitive than vibration analysis,
would be able to detect a bearing problem maybe two
to three days before an actual failure. Hence, the relatively low cost of an operator checking all pumps on
the plant whilst on a walkabout during a shift is a vital

Assuming brake light

fails after 2 months
12-month check / PM = 17% availability

6-month check / PM = 67% availability

3-month check / PM = 92% availability

Weekly check / PM = 98% availability

Daily check / PM = 99.7% availability

e
0

10

11

12

Time, months

Figure 8 Effect of detective maintenance frequency on availability

36 PTQ Q1 2014

proc trends copy 11.indd 5

Detective maintenance
Where failure of a component happens
quickly or its condition cannot be monitored, a different approach is needed.
Examples of equipment that fail in this way
Equipment
fails here (F)
are electrical components such as light bulbs
or circuit boards. One minute they are working, 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 flicker to it actually failing is so short (maybe only a few
seconds) that it is not helpful in preventing the failure.
To understand detective (or failure finding) maintenance, 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, finds it
failed and then replaces it, the availability in that year
would be only 17%.
If the failure finding task of checking the brake light is
reduced from 12 to six months, as in case B, the availability of the brake light will have increased to 67%.
Cases C, D and E show how by increasing the failure
finding 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 redundancy 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 finding inspection task is typically 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 reduction 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
finding) in order to either prevent loss of the system
function or to avoid an unnecessary repair cost.

www.eptq.com

11/12/2013 15:52

Options
mitigation
Whilst for
identifying
suitable maintenance tasks can be

Vapour pressure, psia

and RCM will identify the appropriate detective,

There
arelower
several
options
to reduce
and
temperature
would
ethylene
yield predictive
increases; and
the recovery
of the steam
basedthe
on
operating
experience,
processes
such as failure
preventivesection
maintenance
tasks.
the
impact
ofeffects
glycols
coming (FMEA)
into
not
allow
complete
vaporisation
of propylene
yield Predominantly,
in the steamthese
cracker
FCC byunit
modes
and
analysis
and reliability
tasks areand
performed
either are
the
the
AGRU
with
the(RCM)
feed
gas.
To
main
feedstock,
leading
to cracker
decreases
with increased
integrated,
of the dilution
centred
maintenance
can provide
a robust
meth- operators
of the facility
or bybecause
the maintenance
and
decide
option
is most
appliunnecessary
coke
formation
andmaintenance
severity, whereas
increases in
the effect that the C4s from the FCC
odologywhich
for the
selection
of suitable
tasks. itinspection
departments.
cable,
of aware
the
usethat
of any
yields remain about unit have on the combined C4s
deposition
in the
feed
zone.
FCC unit.
C4 of
It is knowledge
worth
being
sometimes
the cost
glycol
upstream
of
theuse
AGRU
is and,
the where
same, this
butistheReferences
naphtha yield stream.
Another
option
is to
abenefit,
sepamaintenance
can exceed
the
very
important:
type
of glycol
withtoincreased
The main
options
for the2013,
producrate
riser,it where
temperature
1 Europeanseverity,
New Car Assessment
Programme,
November
www.
the case,
can which
be the
acceptable
to allowdecreases
a component
Water
was
applied
at which
temperature,
henceit, the
once-through
conversion tion of on-purpose
propylene are:
and
catalyst-to-oil
ratio
can
be optieuroncap.com
run until
failure
and
then
repair
or replace
assuming
Activator
what
isand
the
resulting
vapour
Direct
recycle
of C4s to
a separate
of naphtha on the 2steam
mised
where
the glycol
processing
Kotabecracker
M, HelsenisK, Global
Marketing
Management,
2010,
Wiley, 5th
the other
failure
consequences
areof
acceptable.
MEG
pressure
thenot
feed
gas, and
so the
on. higher than on the FCC
FCC riser
unit.
the LCN in
does
interfere
with
edition.
DEG
Not
considering
glycolfeedstock.
in the feed
Via metathesis
From a yields perspective,
cracking
of the main
In
Conclusion
3 Mannan disposS, Lees Loss Prevention
in the MDEA
Process Industries, 2012,
TEG combined with
gas
to
malperformance
ing be
of seen
the LCN
steam Oligomerisation
this
case,
the
increase
propylene
Fromcan
thelead
discussion
ininthis
articleofit can
that, through
Elsevier, the
4rd edition.
the
AGRU.
this(about
point
recycle
ofMaintenance,
the oligomerate
to Inc.
be
cracker versus
yield
isreliability
still Even
relatively
small
in fact,
isthough
everyones
responsibility.
4 Moubray riser
J, 1997,in
Reliability
Centred
Industrial
Press
50 a separate
400
may
be
after
several the
processed
in a Maintenance,
separate riser
the
could 5 Nowlan
yield F S,
better
2 Robust
wt%only
on strategies
top reached
of theand
one
achieved
procedures
duringFCC
the lifeHeap H F, Reliability-Centered
1978,ofreport
Temperature,
F
years
of
mitigation
measFCC unit.
economic
returns,
this needs
to Airlines
with
main
feedstock),
low
cycle the
of operation,
the asset
need towith
be
in place
to reduce
or but
No A066-579
by United
under the sponsorship of the Office of
ures
willthe
usually
be
required
sooner
TheWashington,
direct recycle
of the C4s cut
be further
eachof Defense,
conversion
per
pass,
aof high
gas
prevent
occurrence
infant
mortality
failures.investigated
To Assistant for
Secretary
US.
Figure
16
Vapour pressure curves for thermal reclamation
or
following,
several
involves a secondary high-severity
c case.
yield
andIna the
great
portion
of the
be later.
effective,
these
will
involve
almostspeci
all disciplines,
options
riser parallel to the main riser
ole
ns are
produced
being areas:
further
including
inpresented:
the following
2 C4s
backwash
the
Processing
designed tosection.
upgradeOmitting
the C4s into
converted
due
to secondary
reac- Option
Design for
lowest
life-cycle cost
Further reading
of increasing
the
capex, Process
Option
1into
backwash
section
is and
therefore
a
There are
severalCentre
commercially
ethylene,
tions
Another Instead
totalaromatics.
cost of ownership
Buy for
for Chemical
Safety, propylene
2007,
Guidelines
for catalytic
Risk Based
Consider
the
glycol
inis
the
is also
an option
toSafety,
go ahead
tolerate
proven
ways
to Process
upgrade
C4s means
naphtha.
Some
fuelhigher
gas isglycol
also
option
dispose
of
the
LCN
to there
Store to
equipment
tocontent
retain
its
reliability
(this
includes
American
Institute oftoChemical
Engineers.
feed
gas
during
thethe
design
phase; with
a standard
of and
the content
in the
gas. However,
it
produced
from the design
FCC unit
produced.
Thefeed
products
from both
send
it along
with
straight-run
preservation)
6
additional
margins
in discipline
the
solvent
AGRU
by accepting
increased
oper- will
leadattothe
increased
Figure
1 illustrates
steam crackers.
risersalso
merge
reactor amine
outlet
light
naphtha
to a with
steam
cracker.
Install/start-up
and precision
circulation
rate,careheat
exchanger
expenditures
losses
(higher
is theresome
of the
options.Robin(opex).
and travel
as aopex)
common
stream
to
Table 2 compares
the within
yields
that
Peatey isNot
Chief Engineer
Maintenance
andand
Commissioning
Operate
and
processational
limits
with
duties
absorber
height
can with
installing
a water-backwash
section
fore
notHe
theover most
Although
the
the main
fractionator.
can
be and
obtained
when
cracking
with Fostercontent
Wheeler, Reading,
UK.
has
25 yearspreferred
experience
the
Maintain
plant/equipment
precision
for butadiene
compensate
glycol
ingress
into in
the
absorber
will
slightly
measure.
this
s from
the top
steam
cracker
is mitigation
of C
An refinery
advantage
of holding
thisAlso,
process
is
naphtha
cracker
versus
at various
petrochemical
and
facilities,
a variety
of
reliability.on aforsteam
4
the
a acertain
range. increase
the vapour
pressure
losses
glycol
mitigation
measure
limited
signi
cantly
higher
than
of
theEngineer,
that Area
selective
hydrotreatment
for
in In
a AGRU
separate
on
FCC
asthat
Reliability
Maintenance
Engineer is
and
Design
order within
toriser
manage
the unit.
unavoidable
incidents
of roles
As
a consequence,
mainly
glycol
exiting
the
absorber
top Heto
a certain
glycol
feed gas
concens
FCC
the potential
to recover
the
diene
of
the
C
The
trends
area similar
incapital
both of
Engineering
Manager.
graduated
fromconversion
Loughborough
University
and
random
failures
good
understanding
ofunit,
reliability
4
expenditures
(capex)
compared
a iswhen
plant
with
a tration
only.
stream
as well
as the removal of
butadiene
isto lost
the
cases.
theapplication
severitywill
increases,
thetools
a Member
of light
the IMechE.
Email:range
solutions@fwc.com
theoryAs
and
of increase.
analysis
such as FMEA

www.eptq.com
www.eptq.com
14 PTQ Q3 2013

PTQ
PTQ Q4
Q1 2013
2014 77
37
www.eptq.com

q4
basf.indd
8 11.indd 6
proc
trends copy

12/09/2013
11/12/2013 16:18
15:53

fw.indd 3

10/06/2013 12:33

 Mesh-numerous alloys-plastics
Vane (Chevron)
Multi Pocket Vane Insertion Mistx

MIST ELIMINATORS

Wide variety of random packing

types, sizes and materials in stock!
Ask us how our patented SuperBlend
2-Pac can increase your capacity
and efciency.

RANDOM PACKING

Catalyst bed supports

Basket strainers
Nozzles Outlet/Inlet baskets
Distributors
Hub and header laterals

WEDGE WIRE SCREEN

Your Single Source for New or Retrot Internals

Trays (numerous options)
Sieve or perforated
Bubble cap trays
Cartridge trays
Dual ow
Bafe
Valve

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.

Visit our new website at

www.amacs.com

www.amacs.com 24hr EMERGENCY SERVICE (281) 716-1179

DISTRIBUTORS
& SUPPORTS

COALESCERS

STRUCTURED PACKING

Manufactured to customer
specications or engineered to
meet performance requirements.

Oil water separations Haze removal

from fuels Removal of tower wet reux
Caustic treater applications

Woven, sheet metal, and knitted

structured packing. Built to spec
or performance requirement.

amacs.indd 1

09/09/2013 17:14

HS-FCC for propylene: concept to

commercial operation
A FCC process provides a high light olefin yield from a wide variety of feedstocks
utilising high severity reaction conditions and a novel down flow reaction system
Nicolas Lambert Axens IWAO Ogasawara JX Nippon Oil & Energy
Ibrahim Abba Saudi Aramco Halim Redhwi King Fahd University of Petroleum & Minerals
Chris Santner Technip Stone & Webster Process Technology

he fluid catalytic cracking

(FCC) process has undergone
a long evolution of hardware
and catalyst changes, from bed
cracking with amorphous catalyst
to short contact time riser cracking
with sophisticated zeolite catalyst
systems. Improvements to the
process have provided a wide
degree of flexibility to selectively
target the production of distillates
or gasoline, or propylene from
VGO and residue feeds, thereby
making FCC the most widely used
conversion process.
More generally, the objective of
the process is to produce high
valued products, and increasingly
this includes fuels and petrochemicals, such as light olefins and
aromatics. At present, over 30% of
the worldwide propylene supply
comes from FCC-related processes
(FCC, RFCC, DCC). Fluctuating
product demand and price have
caused most new project developers to demand product flexibility
for long-term profitability and
process integration with petrochemical facilities for added
synergy and cost savings.
In order to respond to these
market demands, a new high severity down flow FCC (HS-FCC)
process has been developed by an
alliance of Saudi Aramco, JX
Nippon Oil & Energy (JX) and King
Fahd University of Petroleum and
Minerals (KFUPM), culminating in
a 3000 b/d semi-commercial unit in
operation since 2011 in Japan (see
Figure 1). The process provides a
high light olefin yield from a wide
variety of feedstocks utilising high
severity reaction conditions, a novel
down flow reaction system and

www.eptq.com

axens.indd 1

Figure 1 HS-FCC semi-commercial unit

proprietary catalyst. HS-FCC is

now available for licence from a
Global Alliance by Axens and
Technip Stone & Webster Process
Technology.

Features of HS-FCC

FCC utilises acidic zeolite catalysts

to crack heavy hydrocarbons into
lighter fuels such as gasoline and
distillate and, under more severe
conditions, into lighter olefins such
as propylene and butylene (and, to
a lesser extent, ethylene). Complex
secondary reactions that can

degrade the primary products to

less valuable components should be
limited to retain product selectivity
and refinery profitability. For
HS-FCC, the objective is to not only
improve the selectivity for normal
fuels production, but also to
maximise the potential of light
olefin and petrochemical production at high severity. HS-FCC
provides a total system to maximise
product selectivity and, in particular, propylene yield. Three key
elements are required to attain this
objective:

PTQ Q1 2014 39

12/12/2013 10:58

Typical operating conditions for FCC

and HS-FCC

Gasoline yield, mass%

60

FCC
Reaction T,C
500-550
Contact time, s
25
Catalyst/oil, wt/wt
5-8
Reactor flow
Up flow

50
40
30

Table 2

20
10
75

80

85

90

Conversion, mass%
B

Propylene yield, mass%

25
20
15
10
5
0
75

80

85

90

Conversion, mass%
HS-FCC catalyst
Commercial FCC catalyst (octane catalyst)
Commercial FCC catalyst (activity catalyst)
+10% ZSM-5 additive
+10% ZSM-5 additive
+10% ZSM-5 additive

Figure 2 Proprietary catalyst boosts ZSM-5 effectiveness for more propylene

Reaction conditions and advantages of HS-FCC in petrochemicals production

Advantages
Challenges
High temperature High conversion and olefins selectivity Increased thermal cracking, product

degradation
Short contact time Reduced secondary reactions and

thermal cracking

Reduced conversion, rapid mixing

and separation required

High catalyst/oil
Increased catalytic cracking

Very high catalyst circulation,

uniform flow, mixing and separation

Table 1

Highly selective catalyst and
additive system
Optimised reaction conditions
Down flow, short contact time
reaction system with rapid catalyst
separation.
The balance of these elements

40 PTQ Q1 2014

axens.indd 2

HS-FCC
550650
0.51.0
2040
Down flow

and realisation at commercial scale

is the key to success.

Catalyst system

The catalytic cracking reaction

pathways are complex, with the
primary formation of olefinic

products and parallel bi-molecular

hydrogen transfer reactions leading
to paraffin formation and aromisation of naphthenes. Managing the
acid site density of the catalyst can
suppress hydrogen transfer and
isomerisation reactions to maximise
olefins production. When coupled
with ZSM-5 pentasil cracking catalyst additives, the increased olefins
in the gasoline cut can be selectively cracked to further increase
the propylene yield.
The HS-FCC catalyst uses a high
USY zeolite content system with
very low acid site density, formulated to minimise hydrogen transfer
reactions for high olefin selectivity,
and low coke and gas selectivity.
This catalyst has been shown to be
more
effective
for
propylene
production when coupled with
ZSM-5 additives (see Figure 2).
Commercial catalysts and HS-FCC
catalyst exhibited a similar trend in
gasoline and propylene yield as a
function of conversion (severity),
but the customised HS-FCC catalyst
was much more effective in feeding the ZSM-5 additive with more
olefins, and more accessible linear
olefins, to produce more propylene.1

Optimised reaction conditions

When targeting maximum petrochemicals production, HS-FCC

operates under more severe conditions than conventional FCC. The
main reaction conditions applied
and the advantages and challenges
presented are shown in Table 1.
High
reaction
temperature
coupled with short contact time
increases the primary reactions
towards olefins, while limiting the
unwanted secondary reactions of
hydrogen transfer and thermal
degradation. A consequence of the
increased severity and short time is
the need for higher catalyst circulation (catalyst-to-oil mass ratio, or

www.eptq.com

11/12/2013 12:17

Gasoline, mass%

Gasoline, mass%

scan points representing one projects; revamps take place on

as-built pipe.
Evidently, that is only plants that may
contain hazardous
FCC
Feed + catalyst
HS-FCC
up flow
one clash,
butriser
some systems
will
chemicals,
temperatures
or
down flow
register it as thousands of clashes, pressures. Evidently, these considmaking it difcult to identify the erations imply the need for
real clash and to resolve it well-specied,
specialist
plant
Back
Plug
effectively.
design
software.
mixing
flow
Ensuring that an
Over
Low accurately
cracking
conversion
process
designed pipe spool
actually ts Business
correctly on site requires accurate The introduction to this article
fabrication. Leading 3D design referred to the need for efcient
Feed such
+ catalyst
solutions
as our PDMS or business processes. That is hardly a
AVEVA Everything3D can not only great insight, but, in the engineerReactor residence time
generate fully detailed fabrication ing industry, business processes are
drawings automatically, they can inextricably linked with the engiFigure
flow vs down
flow residence time profiles
neering and design technologies
also 3 Up
perform
manufacturability
checks at the design stage to help that generate project information.
maximise
fabrication quality. It is Best practice is therefore to select
A
60
also now possibleRiser
to scan a solutions that can share informacompleted fabrication
andflow
compare tion efciently and reliably. An
Down
it against the design model, to efcient revamp project workow
quickly verify
its accuracy and can thus be achieved using 3D
50
design technology that integrates
resolve any errors early.
Owner-operators
considering both with laser scan data sources
placing revamp projects should and with engineering data sources,
review a contractors
capabilities in so that engineering and design
40
these areas. The most capable typi- information can be kept synchrocally achieve less than 1% nised as the project progresses.
design-related rework costs, even From this, new design can automat30 projects. Our vision of ically generate accurate materials
on complex
75
80
requirements85 that feed 90into an
plant design70for lean construction
Conversion, wt%
goes further than this; our goal is to enterprise resource management
use B laser 60
scanning, among other (ERM) system.
Riser eliminate
Such engineering, design and
tools, to completely
Down
flow vision information management technolorework in 50
construction.
This
is discussed further in a business gies now exist and are in use on a
40
wide variety of new-build projects.
paper.2
Their ability to support efcient
30
Plan the installation
business processes becomes even
Scheduling the revamp installation more important on revamp projects
20
involves similar
considerations to with their need for on-time, rightplanning the original survey. EPCs rst-time, low-risk installation.
10
and owner-operators
must collaborate
closely
to
achieve
a
0
References
well-executed0 installation.10Here, the 20
30
40
1 Lighting the Way,
www.aveva.com/
power of 3D design can
add
C2-C4 olefins, mass%
publications
considerable value.
2 Plant Design for Lean Construction:
Reverse
engineering
objects
that
Figure 4 Selectivity benefits of a down flow reaction system4
Innovation for a new era in plant design, www.
are to be removed enables the creaaveva.com/publications
tion to of
accurate
demolition
C/O)
provide
the required
heat up flow FCC riser reactor system,
andand
thesufficient
determination
todrawings
the reactor
catalystof where the catalyst required for the
weight
and
centre
of
gravity,
which reaction is lifted up the reactor pipe
activity to achieve high conversion
Gary Farrow is Vice President 3D Data Capture
informs
the
correct
use
of
handling
at short contact time. The range of or riser by the vaporised and
with AVEVA in Cambridge, UK. He works with
equipment.
Model
animation
operating conditions for a conven- cracked
hydrocarbon feed. In up
customers in the use of 3D data capture
enablesFCCplanners
to evaluate
tional
and HS-FCC
are flow
fluid-solid
systems,
the solids
technology to increase
productivity
and to
proposed task
sequences
and to or catalyst are conveyed upwards
summarised
in Table
2.
advance the performance of AVEVAs LFM
check that, for example, items can against
force of engineer,
gravityhebyhasdrag
software.the
A mechanical
been
be moved
safely (DFR)
in thesystem
available forces
Down
flow reaction
the rising
involvedfrom
in 3D laser
scanninggases
from its(hydroinception
spacespecific
constraints.
Designconditions
review is carbons).
in the late 1990s,
undertaking
The
reaction
As initially
a result,
all projects
riser
arguably
even
more
important
for reactor
delivering data
and 3D models,
a huge
with very high C/O result in
systems
haveincluding
varying
scanning
project
for
Fluor/TCO
in
Kazakhstan.
revamps
than
for
new-build
certain challenges in a conventional degrees of catalyst back-mixing and

www.eptq.com
www.eptq.com

rev aveva.indd 3
axens.indd 3

AMACS
Compressor suction drums:

I think Ive got liquid carryover.

what can I do about it?

It happens in petrochemical
plants, refineries, and anywhere
else that the gas approaching
a compressor is wet. Traces of
aqueous or organic liquid escape
the inlet knockout drum, often
intermittently, and silently damaging the compressor. Telltale signs
include pitting corrosion, salt
deposits, and diluted lubricants.
Typical 1980s mist eliminator technology
Cooler

Stage 1

Stage 2

Knockout
drums

Instead of trying to repair

symptoms, look for the root cause,
which usually involves the mist
eliminator in the knockout drum.
Problems may include improper
mist eliminator specifications,
overloading, uneven velocity, waxy
deposits, liquid slugs, foaming,
incorrect installation, and several
other possibilities. New, highcapacity, high-efficiency mist
eliminator technologies pay off the
first time you avoid shutdown.

Read more on this topic at

www.amacs.com

Phone:+1-713-434-0934 Fax: +1-713-433-6201

amacs@amacs.com

Revamps
2013 41
31
PTQ Q1 2014

10/09/2013 11:20
11/12/2013 20:06

100

Dry gas

90

Yield, mass%

80

LPG

Propylene
10.5%

70

Propylene
10.6%

60
50

Gasoline

40
20

LCO+

10
0

RON
99

RON
98

30

Coke

Bench scale (0.1 bpd)

Demo plant (30 bpd)

Figure 6 Bench scale vs demonstration scale results on low sulphur VGO at high severity
without ZSM-5

Figure 5 HS-FCC demonstration unit

reflux along the walls, particularly

in the feed injection or catalyst
pick-up zone at the bottom of the
riser reactor. At very high C/O,
significant back-mixing is unavoidable. This problem is overcome in a
down flow reactor (DFR), where
both the catalyst and feed flow
downwards together (see Figure 3).
Down flow fluid-solid reaction
systems have been of increasing
interest in recent years to achieve

plug flow reaction conditions, as

summarised by Cheng.2 When plug
flow conditions are achieved, more
selective primary cracking results
in greater selectivity. FCC pilot
work demonstrating the effects of
short contact time and down flow
have been reported by Del Poso3
and Abul-Hamayel4 (see Figures 4a
and 4b). The general trend is that of
greater gasoline selectivity at short
contact time down flow, with a
maximum yield achieved at a
higher conversion level. This effect
is seen in Figure 4a, where the

maximum gasoline yield is about 5

wt% higher in the down flow
system. When olefins are of interest, the more selective down flow
reaction environment can produce
substantially more light olefins
(C2-C4) at the same gasoline yield
compared to a conventional up
flow system (see Figure 4b).
Although the idea of a controlled
high severity, short contact time
down flow reaction has been
considered for some time, achieving
this successfully on a commercial
scale has been elusive. Extensive

Air outlet
Catalytic
circulation
hopper

Injector

Main air

Air outlet
Separator
Catalytic
circulation
hopper
Total height: 35m (115ft)

Catalyst circuit

Catalyst inventory: 20T

Max. catalyst circulation: 1.0T/min

Lift air

Figure 7 500 b/d equivalent cold flow testing to scale up and optimise the reaction system

42 PTQ Q1 2014

axens.indd 4

www.eptq.com

11/12/2013 12:17

pilot work at the 0.1 b/d scale

demonstrated the principle, catalyst
system and operating conditions,
but did not address how rapid
mixing, reaction and efcient catalyst/gas separation can be achieved
at a large scale with a target residence time on the order of 0.5
seconds. On a commercial scale,
equipment design for very short
contact time with the mechanical
integrity to withstand high-velocity
catalyst circulation in a coking environment requires extensive research,
development and demonstration.

R&D history

The challenges of developing

this new technology required a
systematic research program undertaken by JX, KFUPM & Saudi
Aramco with the support of Japan
Cooperation Center, Petroleum
(JCCP). Early pilot work by both JX
and KFUPM in 1996-2000 demonstrated the benets of high severity
operation at controlled short
contact time in down ow mode.
Aramco became an active participant in the scale-up effort to design
a 30 b/d demonstration unit. JX
conducted large-scale, 30 b/d
equivalent, cold ow testing of the

Semi-commercial unit performance

Feed SG
Reactor T, C
Conv, w%
Light olefins, wt%
C2=
C3=
C4=
C5-220 gasoline, wt%
RON

VGO +
HC Btm
0.845
575
93.2
39
4
19
16
35
98.5

HDT
VGO
0.879
595
83.7
34
4
17
13
34
98.1

Propylene estimation, mass%

VGO+
AR
0.915
600
82.4
31
3
15
12
34
98.4

Table 3

catalyst circulation loop and

reactor-separator equipment to
validate the design of the demonstration unit.
The demonstration unit (see
Figure 5) was operated from 20032004 at the Aramco Ras Tanura
renery. Results from the demonstration unit validated the HS-FCC
concept, with good agreement
between 0.1 b/d pilot results and 30
b/d demonstration (see Figure 6).5, 6
A low sulphur VGO was cracked
at high severity in both the pilot
and demonstration units using only
the new HS-FCC catalyst without
ZSM-5 additive. A very high
propylene yield, over 10%, was

obtained along with a very high

octane gasoline.
Work immediately began on
scale-up to a commercial unit.
Important lessons were learned
concerning equipment design, and
larger-scale cold ow work was
undertaken by JX in Japan at the
500 b/d equivalent scale to optimise the feed injection zone and
separator design (see Figure 7).
This work was coupled with CFD
simulations to assist in larger-scale
equipment design.6

Semi-commercial unit

With the successful demonstration

of HS-FCC technology at the 30
Combined kinetic and
hydrodynamic model

Catalyst

27 lump kinetic model

VGO +
DAO
0.891
580
83.0
31
3
15
13
34
98.1

Feed

Feed

CFD
simulation

Assembly of a large
number of small reactors
Feed

30

Catalyst

Injectors

25
20
15
10
5
0

DFR
(down flow
reactor)

Feed A
Feed B
Feed C
Feed D
0

10

15

20

25

Experimental data,
mass%

30

Separator

Product

Product

Quench
Catalyst

Figure 8 Combined kinetic and hydrodynamic modelling assists design and scale-up

www.eptq.com

axens.indd 5

PTQ Q1 2014

43

11/12/2013 12:17

Propylene yield, wt%

25
23

DCC

21

HS-FCC

19
17
15
13

HP-FCC

11
9

R2P
Heavy residuals

Light residuals

VGO

H1-H2 VGO

Figure 9 Family of high-propylene catalytic cracking processes

HS-FCC unit

Petrochemicals
Steam
cracker

C2

Mixed C3

VGO
resid.

HS-FCC

Mixed C3
LCN
Oligomers
recycle

Polymer grade
propylene

PRU
Oligomerisation
Polynaphtha

Paraffinic
raffinate

Aromatics
complex
Paramax

HCN

Polymer grade
ethylene

Bz + PX + OX

Fuels

Figure 10 Integrated refinery-petrochemical complex

b/d scale completed, it was time to

look forward to scaling up to a fullsized commercial unit and to plan
for future licensing of the technology. Several FCC licensors were
interviewed and evaluated before
Axens and Technip Stone &
Webster Process Technology were
selected to assist in the design of a
3000 b/d semi-commercial unit,
plan for a larger commercial unit,
and serve as exclusive licensor for
the HS-FCC technology, relying on
its extensive knowledge in FCC
and RFCC design.
A complete 3000 b/d HS-FCC
unit with main fractionator, gas
plant and ue gas treatment was
designed for the JX Mizushima
renery.
Chiyoda
Engineering
performed the detailed engineering
and construction of the plant (see
Figure 1), which was put on stream
in early 2011.
Performance trials are on-going to

44 PTQ Q1 2014

axens.indd 6

evaluate yields and product properties for widely different feeds and
to demonstrate equipment reliability. Preliminary results showing
yields for several blends of VGO,
hydrocracker (HC) bottoms, DAO
and atmospheric residue are shown
in Table 3. Combined light olen
(C2-C4) yields of 30-40 wt% have
been demonstrated with 15-19 wt%
propylene and 4 wt% ethylene. The
yield of butenes is similar to propylene and offers opportunities for
greater petrochemical integration,
including oligomerisation and the
FlexEne conguration for even
higher
propylene
production.7
These results are without the use of
post-separator quench injection,
which will improve olen selectivity further. The catalyst system
continues to be optimised for the
various feeds.
When viewed from a petrochemicals perspective, the ethylene

produced becomes a signicant

boost to the economics. The gasoline also has value beyond fuels,
with an octane of 98-99, olen
content of 25-40 wt% and 35-50
wt% aromatics.
The testing programme will
continue, with 100% residue cracking trials to begin soon. With a
controlled short contact time, high
C/O and p lug ow reaction
system, HS-FCC is well adapted to
be highly selective for both light and
residue
feed
conversion
to
petrochemicals.
Throughout
the
programme,
equipment evaluation, inspection
and reliability data continue to be
gathered to guide further development and scale-up to a fully
commercial scale of at least 30 000
b/d. In parallel to this work, CFD
simulation of the DFR and separator
hydrodynamics are being combined
with a kinetic model to analyse the
results, validate the kinetic models,
and enable accurate predictions at
commercial scale for future feeds
and reactor congurations.

HS-FCC in the family of catalytic

cracking processes

The HS-FCC process expands the

operating window of catalytic cracking to encompass heavier feeds and
greater
propylene
potential.
Commercial processes for high
propylene production from light
distillate feeds and residue feeds
include DCC,8 high-propylene FCC
(HP FCC) and resid to propylene
(R2P). More severe conditions for
residue feeds to attain a higher
propylene yield have proven challenging in the past due to undesired
secondary reactions. High severity,
combined with an optimised catalyst system and a controlled short
contact time DFR reaction system,
allows the new HS-FCC technology
to provide selective conversion with
lower fuel gas production and a
greater olen and petrochemicals
yield even with heavy residue feeds.
Indeed, the selectivity of the system
presents opportunities to crack a
wide range of conventional and
unconventional feedstocks.
The technology mapping by
severity and feedstock is shown in
Figure 9.

www.eptq.com

11/12/2013 12:18

Absolute deltas, wt% of feed

also
integrated
solutions
forfor
steam/
ofWith
most Crude
refinery
off-gases,
reduc10
Very
Heavy
Upgrading
Long
Term
R&D
Wadsworth
D, LC-Fining
options
heavy
Technip
Stone
& Webster
Process
the
option
to
operate
at 11
Opportunities,
1994.
oil
upgrading,
Proceedings
of
the
NPRA
Annual
energy
generation
and
CO2
tion
in
natural
gas
and
water
conventional severity
or high sever- Technology are now offering
8
2consumption,
Yui S, Chung K H,and
Syncrude
upgrader
revamp
Meeting,
San Diego,
CA,Liquide
9-11 Marchcan
2008. also
handling.
Air
the
recovery
of
ity, the refiner6will have the ability HS-FCC technology on behalf of
improves
quality,
Oil Gas J, 2007, Vol. 12
Ordorica-Garcia
G, Croiset E, Douglas
P,
forproduct
EOR
sequestration.
an (over-the-fence)
supply
CO
the HS-FCC
Global Alliance team.
to 2select
an or
operating
mode and provide
105,With
46, 52.the configuration
Elkamel
A,
Gupta
M,
Modeling
the
energy
shown
in of air gases, power and steam,
4suited to the
prevail3feedstock
Chrones J, best
Germain
R R, Bitumen and
heavy demands
greenhouse
gas emissions of the
FlexEnecompression/liquefaction
isand
a mark
of Axens.
and
Figure
7,
the
feedstock
to
the
MPG
CO
2
ing
economic
A 1989,
high Canadian
2 conditions.
oil
upgrading
in Canada,
Fuel Sci Tech Int,
oil sands industry, Energy Fuels, 2007,
transportation.
unit
is
normally
reduced
to
that
References
product
slate rich in olefins 21, 2098.
7,severity
783.
0
amount
needed
tomakes
satisfy
the 1 Maghrabi
process forJ, maximized
aromatics
also
integra4and
Rana
M S, Samano
V, Ancheyta
J, Diaz J 13
MorawskiA,I, HS-FCC
Mosio-Mosiewski
Effects of
propylene
production,
10th Annual
Saudi-2
Conclusion
demand
of
the
complex.
If
syngas
withof petrochemicals
plants parameters in Ni-Mo catalysed
Ation
I, A review
recent advances on process
hydrocracking
Japanese
Symposium
on
Catalysis
in
Petroleum
In
the
context
of
ever
more
stringent
production
exceeds
the
amount
technologies
for upgrading
of heavy
oils and of vacuum residue on composition and quality
4 so that
more attractive
the natural
Refining
and Petrochemicals,
2000.
regulations
forTechnol,
refinneeded
required ofenvironmental
residua,
Fuel,to
2007,
86,
1216. the
obtained
products,
Fuel Dhahran,
Process
synergy
of produce
shared
intermediate
6
2
Cheng
Y,
Downer
reactor:
from
fundamental
ers,
there
is
a
trend
to
increase
(depending
first
of
all
on
5hydrogen
Speight
J
G,
The
Chemistry
and
Technology
2006,
87,
659.
products and recovery
schemes
A1
A2
A3 can A4
B1
B3
B4
study
to B2
industrial
application,
Powder
th
of
2007,
Danial-Fortain
P,2.28
Gauthier
T, Merdrignac
ed, CRC4.01
Press/Taylor
& 4.4314
residue
with
the
crude
9% quality/origin),
Coke,
4.32
3.38 the
1.44 conversion
1.57
1.49hydrocrackAn4 example
of HS-FCC
bePetroleum,
realised.
Technology,
183,
2008.
Budzinski
H, Reactivity
Athabasca
Francis,
Boca
Raton,
FL. 8.83 be
C syngas
8.77 used
6.52 for 3.00I,ers.
0.44
0.67
0.43
0.30 of
+
liquid, %
Depending
on study
the
conversion
surplus
integration
with can
a petrochemical
3 Del0.99
Poso
M,
Development
of ultra
selective
4.76 differences
3.14
1.42vacuum
2.25in hydroconversion
1.85
1.79
gas, % S, 4.51
residue
conditions,
6power
Sayles generation
S,Fuel
Romero
Understand
rate
of
the
units,
the
heavy
bottom
with
gas
turbines
complex is shown in Figure 10.
cracking
technology,
2nd
IFP
and
S&W
FCC
Today,
2010,
150,from
255. 10-20%. Using
between
thermal
and of hydrocracking,
Feed andCatal
formulation
yield
willWoodlands,
range
or the production
chemicals/
Forum,
The
Texas,
1996.
Ding F, Ng S H, Xu C, Yui S, Reduction of
Hydrocarbon Process, 2011, Sept, 37.
Air
Liquide Global
E&C Solutions
fuels. Alliance
Air Liquide
Global E&C 15
4 Abul-Hamayel
M A, Comparison
of
downer
for commercialisation
light
cycleof
oilOptiFuel
in catalytic
cracking ofpilot
bitumen7Global
Martinez
J, Sanchez
J L, Ancheyta
J, Ruiz
R application
Figure
3
Changes
in
pilot
plant
yields
with
Technology:
plant
Lurgi
MPG
technology-based
Solutions
can
provide
the
applicaand
riser
based
fluid
catalytic
cracking
process
HS-FCC
technology
is the derived crude HGOs through catalyst
selection,
S,The
A
review
of
process
aspects
and
modeling
runs
at
Penn
State
University
hydrogen
production
allows a
ble
technologies
(see
Figure
7).
at high
severity
conditions:
a pilot
of reactors
systematic
process
Process
Technol,
2007, 88,
833. plant study,
ofproduct
ebullated bed
for hydrocracking
of Fuel
Petroleum
Science
Technology,
22,
refiner
to
transform
these
Besides
MPG
technology,
including
research,
development,
16 Botchwey C, Dalai A K, Adjaye2004.
J,residues
Kinetics
heavy
oils, Catalcatalyst
Rev Sci Eng, 2010,
52, 60.
vessels
needed
for
additive
mixing
Albemarle
has
designed
an
addi5
Redhwi
H,
Meeting
olefins
demand
in aa
into
an
amount
of
hydrogen
that
air
separation,
CO
shift,
syngas
8pilot
Yui S,work,
Producing
crude oil of bitumen-derived gas oil upgrading using
30quality
b/d synthetic
demonstration
novel
FCC
technology,
18th
World
Petroleum
and
storage.
tion
system
to
ensure
proper
mixing
balances
the
refinerys
needs.
cleaning,
Lurgi
Rectisol,
PSA
and
commercial
NiMo/Al
from
Canadian
oil
sands
bitumen,
J
Jpn
Petrol
O
catalyst,
Can
J
Chem
unit testing, and ongoing semi2 3
Congress,
South
Africa,
2005. andseen
The
with
of 2008,
the solid
and
portions
MPG
is 478.
aimprovements
proven
reliable
methanation,
theliquid
company
also and
has
Inst,
51, 1. operation
2004,yield
82,
commercial
and testing
at Eng,
6
Okazaki
H,
High
severity
Fluidized
Catalytic
this
technology
are
hypothesised
to
to
avoid
solids
settling.
The
supply
9technologies
S, Athabasca
oil sands
produce
quality 17
Yui S, Sanford
Kinetics of
technology
for E,securing
thearomatics
hydrofor
sulphur
recovery
theYui3000
b/d scale.
These
successful
Cracking
(HS-FCC)

go
for
propylene!,
20th
be
the
result
of
reactions
in
both
the
consists
of
a
liquid
carrier
and
the
diesel
and jet fuels, Oil Gas J, 2000,
98, 47, hydrogenation
bitumen-derived
Can
supplyofto
a refinery. gas
It oils,
further
(preferably
sinceVol.oxygen
results andOxyClaus,
the modelling
tools gen
World
Petroleum
Congress,
Doha, 2011.
58.
Chem Eng,
1991,
69,
1087.
liquid
and
vapour
phases,
which are
additive
supplied
to the refinery
in Javoids
the
production
of
petroleum
is
available),
technologies
for
chemdeveloped for further scale-up 7 Ross J, (R)FCC product flexibility with FlexEne,
10
Yui S,
H, Processing
sands
Yui S,and
Sanford
E, Mild
hydrocracking
of
directly
influenced
the additive.
bags
or Chung
bulk Kshipments
foroilmixing
coke,
helps
to by
consume
much
icals
as
methanol
and 18
WRA Downstream
Asia,
2011
and
make is(such
the
technology
ready
bitumen
syncrudes
R&D focus,
Oil Gasfor
J, bitumen-derived
coker
andSingapore,
hydrocracker
heavy
The
active
sites
of
the
additive
are
on-site.
The
additive
injection
natural gas and water. In
propylene)
and fuelsAxens
production
www.axens.net.
commercialisation.
and less
2001,
Vol. 99,
17, 46.minimises
oils: kinetics,
yields, andcatalyse
product
intended
to product
preferentially
system
design
the size of gas
3

(for instance, Fischer-Tropsch), but

combination with the Lurgi Rectisol

technology,
as light
the
worldwide
properties,
IndIncrease
Eng Chem
Res,
1989,
28,
1278.
Projected
yields
forolefins
commercial
8 Dharia
D,
production,
19
Yui
S,
Removing
diolefins
from
coker
industrial
standard
for
clean
syngas
application
of OptiFuel
Technology
Hydrocarbon
Processing,
April 2004.
naphtha
necessary
before
hydrotreating,
Oil
production,
it
provides
a
clean
9 Roux R, Upgrading of heavy cuts into max
Gas
J, 1999,
97, 36.
hydrogen
product
and Predicted
enables
olefins
through
HS-FCC,Base
JPI Petroleum
Refining
20
Chang A-F,
Liu Y2012
A, capture.
Predictive
modeling of
economic
carbon
Conference,
Tokyo,
and www.axens.net.
Coke
32.2%
28.5%
large-scale
integrated
refinery
reaction
Dry gas
5.6%
4.4% and
Lurgi Rectisol,systems
MPG, OxyClaus
and
Purisol
are
fractionation
from
plant
data.
1:
C
+
liquid
yields
62.2%
67.1%
Nicolas
Lambert is Technologist in Part
Axens
3
marks
of Air processes,
Liquide Global
E&C
Solutions
hydrocracking
Energy
Fuels,
2011,
Middle Distillates & Conversion Business Line,
(Lurgi
GmbH). Selexol is a mark of UOP, a
25,
5264.
focusing
FCC technology. He is a graduate
Table 4 oncompany.
Honeywell
of Arts & Mtiers ParisTech.
Iwao
Ogasawara
is Facility
Planner
of
Technical
Max-Michael
Weiss
is over
Director
Innovation,
cracking
reactions
the
strictly
&
Engineering
Service
Department,
Refining
Clean
Conversion,
with
Air
Liquidethat
Global
E&C
thermal
coking
reactions
occur
Technology
&
Engineering
Division,
JX
Nippon
Solutions/Lurgi
GmbH. He graduated
in the traditional
delayedas Diplom
coker
Oil & Energy
Corporation.
Heengineering)
holds BS andfrom
MS
Chemie
Ingenieur
(chemical
operation.
During
development,
it
degrees
in chemical
engineering.
the
Technical
University
of Karlsruhe, Germany.
Anton
Alvarez-Majmutov
is
an
NSERC
Visiting
has
been
observed
that
OptiFuel
Ibrahim AHeurich
Abba is isChief
Technologist
of the
Helmut
Director
for Refinery
Fellow
at CanmetENERGY
working
on bitumen
Technology
shifts
the
delayed
coker
Chemicals
Research
Division
of
Saudi
Aramco
Applications in the HyCO Product Line in
upgrading
process
modelling
and simulation.
yieldsEngineering
towards
more
valuable
Research
&
Development
Center.
He
holds a
Global
& Construction
Solutions
He
holds
a
PhD
from
Mexican
Institute
of
products,
with
reduced
amounts
of
PhD
the University
British Columbia.
of
Airfrom
Liquide.
He studiedofprocess
technology
Petroleum
(IMP).
Halim
Redhwi
is
the
CEO
(A)
of
Dhahran
dry
gas
and
coke.
at the Technical University of Braunschweig.
Techno-Valley
Company
a Professor
in
Delphine
Roma
is the and
Global
Marketing
Jinwen
Chen is aEngineering
Senior Research
Scientist and
the
Chemical
Department,
King
Pilot
plant
verification
Manager in charge of the refining industry
Group
Leader atofCanmetENERGY.
He holds
Fahd
University
Petroleum
& Minerals,
Saudi
within
Air Liquide and
Global E&C
Solutions.have
She
Albemarle
OFTG
aArabia.
PhD inHechemical
engineering
from
Tianjinin
holds
BS,
MS,
and
PhD
degrees
holds
an MSc and
engineering
degree
from
the
conducted
a
series
of
pilot
plant
University.
chemical
engineering.
cole
Chausses
in Paris. (PSU).
runsdes
at Ponts
PennetState
University
Chris Santner
is Senior
Director
of Catalytic
Stefan
Walter
is Head
of Department,
The scope
of these
has
been
Mugurel
is studies
a Lead
Process
Cracking Munteanu
Refining
Technology
Gasification
Technologies,
withwith
Air Technip
Liquide
Engineer
at
CoSyn
Technology,
a
division
of
to
quantify
the
benefits
of
the
techStone & Webster
Process Technology.
He holds
Global
E&C Solutions/Lurgi
GmbH.
He
WorleyParsons,
in
Edmonton,
Canada.
He
holds
nology
as
a
function
of
additive
BS and MS as
degrees
in chemical
engineering
graduated
Diplom
Verfahrensingenieur
afrom
PhDthe
inUniversity
chemicalfeed
engineering
from Laval
composition,
properties
and
of
Houston.
(process engineering) from the Technical
University,
Canada.
operating
conditions.
University of Aachen, Germany.

The team of CS Combustion

Solutions is experienced in
engineering, supply and
commissioning of vertical
and horizontal burners
and combustors for
more than 20 years.

your experts
A Member of Unitherm-Cemcon Group
n
n
n
n

ACID/TAIL GAS
LIQUID SULFUR
SPENT ACID
CHEMICAL/PETROCHEMICAL RESIDUES

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

Variante 4a
www.eptq.com
www.eptq.com
www.eptq.com
www.eptq.com

albemarle.indd
air
liquide.indd 64
canmet.indd
axens.indd 76

PTQQ4
Q32013
2013 97
99
PTQ
PTQ
PTQQ2
Q12013
201443 45

07/06/2013 16:47
20:04
12/09/2013
08/03/2013
12/12/2013 13:04
10:58

sandvik.indd 1

09/12/2013 11:31

Optimising distillation column

product quality
Process simulation and random sampling were used to optimise product targets
for a propylene splitter unit
JOSE BIRD and DARRYL SEILLIER
Valero Energy Corporation

n an attempt to ensure that product specifications are always met

in distillation unit operations,
refinery operations will often run a
process unit at higher reboiler
duties than required, resulting in
excessive energy use and suboptimal product yields. Process
units are run in this less than efficient 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 operating
targets
when
product
specification 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 illustrate the application of the
methodology used, the optimisation of the operations of a
propylene-propane splitter distillation unit was chosen.1 Figure 1
provides a schematic of a typical
propylene-propane splitter distillation
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

www.eptq.com

valero.indd 1

specifications 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 specification 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
refinery
grade
propylene.
Accurately predicting the performance of the distillation unit at
lower propylene purities needs to
consider changes in feed composition, feed temperature and feed
flow 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 presence of process variability.
The following steps summarise
the methodology used in this
study:
Process simulation model calibrated to current unit operations
Probability distributions generated 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 resulting joint probability distributions
are used to assess the ability of

Cooling
water
Splitter
vent

1
10

Feed

Reflux

Side draw:
propylene

81
155

Steam
Bottoms:
propane

Figure 1 Propylene splitter schematic

PTQ Q1 2014 47

11/12/2013 12:21

50%

Distribution of product propene

25%

= 96.92
= 0.38

40%

20%

30%

15%

20%

10%

10%

5%

0%

0%

95.4 95.7 96.0 96.3 96.6 96.9 97.2 97.5 97.8

Distribution of feed temperature

= 93.25
= 4.90

88.8

83.8

Propene product
25%

Distribution of feed rate

25%

= 7536.72
= 450.12

20%

20%

15%

15%

10%

10%

5%

5%

0%

93.8

99.8

103.8

Feed temperature

0%

6625 6875 7125 7375 7625 7875 8125 8375 8625

Distribution of feed MA
= 17.87
= 5.34

4.5

7.5 10.5 13.5 16.5 19.5 22.5 25.5 28.5

Feed rate

Feed MA
25%

Distribution of feed propene

= 73.07
= 2.59

20%
15%
10%
5%
0%

67.5 69.0 70.5 72.0 73.5 75.0 76.5 78.0 79.5

Feed propene
Figure 2 Distribution of input variables

the process to meet product

specifications at different operating
targets and to determine the associated optimum unit operating
targets.
A detailed description of the
analysis methodology is provided
below, followed by results of analysis and conclusions.

48 PTQ Q1 2014

valero.indd 2

Analysis methodology

To examine the effect of process

variability on the ability of
the process to meet product
specifications, Monte Carlo simulation was used to generate a set of
process simulation cases that
captured existing process variability. The following model inputs

were considered to be primary

sources of process variability: feed
temperature, feed rate, feed propylene
composition,
feed
MA
concentration and propylene product percent purity. These inputs
were assumed to follow independent normal distributions. Historical
data with the APC system turned

www.eptq.com

11/12/2013 20:22

PP splitter optimisation

Profit, $MM

6
3
0
3
6

0
97.
00
96.
Pro 3.25 0
2
9
5
5.5
pen 2.5
94.
0
e in
75
75
94.
%
pro 1.
0
ity,
00
pan 1.0
Pur
e, %
4.0

Feed temperature

Feed C3 =

Feed rate

Process
simulator
Feed MA

MA

Propylene purity
30
25
20
15
10
5
0

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

90

92

94
96
Purity, %

98

Figure 3 Analysis methodology schematic

on were used to build the probability distributions of these inputs.

Figure 2 shows histograms and
the calculated normal density functions 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 concentration in the propylene product.
Figure 3 highlights the analysis
methodology. First, 200 independent 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

valero.indd 3

the mean of the propylene purity

distribution. To generate the different
propylene
purity
target
scenarios, the mean of the propylene 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 combination of the model inputs based on
the 200 independent random
samples generated. The SAS
Analytics
procedure
PROC
SIMNORMAL2 was used to generate the random samples. The SAS
Analytics scalable software environment was selected for this
project due to its extensive statistical and charting capabilities.
The 200 cases were then evaluated 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 PetroSIM 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 incremental 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 concentration in the propylene product.
The contour maps of the joint probability distributions were generated
using the SAS Analytics procedure
PROC KDE2, which uses kernel
density estimation.
The process simulator was congured
to
use
three
process
specications: condenser temperature, 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

PTQ Q1 2014 49

11/12/2013 20:22

[97.0, 2.5]
200 cases

Splitter
vent

[97.0, 3.6]
200 cases

[97.0, 1.4]
200 cases
[95.5, 1.0]
200 cases

QC

[95.5, 4.0]
200 cases

[95.5, 2.5]
200 cases

Splitter
feed

[94.0, 3.6]
200 cases

[94.0, 1.4]
200 cases

T-100
Refinery
plant to
Crude-100

[94.0, 2.5]
200 cases

Propene
product

QR

Figure 4 Central composite design scheme

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 profit
response surface, the experimental
central composite design configuration shown in Figure 4 was used.3,4
This configuration allows for
consideration of second order terms
in the construction of the profit
response surface. The different
scenarios were defined by the
combination of average propylene
purity and the propylene concentration in the propane.

Base model description

To build the propylene-propane

splitter model, data historian information including flow 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 flow rates and compositions were performed to test the
validity of the data. Propylene and
propane product compositions and
flow 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

Stream conditions
Conditions
Temperature, F
Pressure, psig
Flow rate, bbl/day

Splitter feed
76
272
7094

Propane bottoms
137
283
1432

Propylene product
118
271
5561

Table 1
Stream compositions
Composition
Propane, wt%
Propylene, wt%
M-acetylene, wt%
Ethane, wt%
i-Butane, wt%
1-Butene, wt%
i-Butene, wt%
n-Hexane, wt%
Cyclohexane, wt%

Splitter feed
21.27
77.86
0.00
0.47
0.31
0.02
0.03
0.02
0.02

Table 2

50 PTQ Q1 2014

valero.indd 4

Propane bottoms
95.48
2.50
0.01
0.00
1.56
0.10
0.15
0.10
0.10

Propylene product
2.92
97.01
0.00
0.07
0.00
0.00
0.00
0.00
0.00

Propane

Figure 5 PetroSIM simulation model of the

distillation unit

Excel spreadsheet interface and

capability to run multiple cases.
To model the propylene-propane
splitter, a preconfigured distillation
column unit consisting of a reboiler
and a condenser was used. Figure 5
is a simplified process flow diagram
of the distillation unit used.
The output streams consisted of
an overhead vent stream, a propylene side draw stream and a
bottoms propane product stream.
The feed stream flow rate, temperature and pressure were specified as
inputs. Compositions of each
component were specified according to the reconstituted feed stream
calculated from the output streams.
Table 1 provides feed flow rate,
pressure and temperature, and
Table 2 the calculated feed
compositions.
The distillation column was
configured 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 pressure ranged from 265-282.6 psia
from the top to the bottom of the
column. The variability of the overhead temperature was minor, so it
was kept constant at 100F (38C)
for all simulations. The process
simulator was configured to calculate reboiler duty as well as flow
rates and compositions for the three
output streams. Table 2 provides

www.eptq.com

11/12/2013 20:22

Figure 6 Distillation column configuration

Profit, $MM

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 represents
the
average
condenser
temperature based on actual operations. 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 corresponding to a given propylene
purity scenario. The propane weight
percent in the propylene product
was computed for the given propylene 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.

www.eptq.com

valero.indd 5

3
0
3
6

0
97.
00
96.
Pro 3.25
25
95.
pe
.50
2
5
94.
ne
0
75
75
in
%
94.
pro 1.
0
0
ity,
0
0
pa 1.
Pur
ne
,%
4.0

Results and analysis

The analysis assumed a hypothetical base case with an average 97%

propylene purity and 2.5% propylene 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, renerygrade propylene and propane.
Table 3 summarises the incremental
prot and the calculated percentage
of off-spec results for the nine
scenarios considered. The incremental
prot
gures
shown
represent the average of the 200
cases of the corresponding propylene purity scenario. Note that cases
with a combination of low propylene 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

PP splitter optimisation

Figure 7 Profit response surface

Incremental benefits and off-spec results

Purity, %
94.0
94.4
95.5
95.5
95.5
96.6
97.0

Propylene in propane, %
2.5
1.4
3.6
1.0
2.5
4.0
1.4
3.6
2.5

Incremental benefit , $MM/yr

4.6
(4.1)
4.4
(5.8)
2.9
2.2
1.1
0.3
0

Off-spec, %
4.0
37.5
0.0
37.0
0.0
0.0
0.5
0.0
0.0

Table 3

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

prot response surface constructed

using the results of the nine
scenarios. To generate the response
surface, the SAS Analytics procedure PROC G3GRID was rst used
to generate a ner grid covering the

PTQ Q1 2014 51

11/12/2013 20:22

15
10
5
92

94

95

96

97

10
0

98

93

94

95

96

97

20
15
10
5
0

98

93

94

95

96

97

95.5% target and

2.5% propene

95.5% target and

4.0% propene

25

15

15

10

0.24
0.22
0.20
0.18
0.16
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00

10
5

0.000

93

94

95

96

97

98

92

93

94

95

96

97

25

0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00

15
10
5
0

98

92

93

94

95

96

97

Purity, %

Purity, %

94.4% target and

1.4% propene

94.0% target and

2.5% propene

94.4% target and

3.6% propene

25

0.14

25

20

20

0.12

20

15

15

0.08

10
5
0.000

93

94

95

96

97

10

0.06

0.02

98

0.10

0.04
0.00

92

93

94

95

96

97

98

Purity, %

Purity, %

MA

MA

0.095

98

20

Purity, %

25

92

92

95.5% target and

1.0% propene

20

0.65
0.60
0.55
0.50
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00

Purity, %

0.095

92

25

Purity, %

20

92

96.6% target and

3.6% propene

Purity, %

MA

15
5

MA

MA

25

93

0.55
0.50
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00

20

MA

MA

20

25

0.20
0.18
0.16
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00

MA

25

97.0% target and

2.5% propene

MA

96.6% target and

1.4% propene

98

0.22
0.20
0.18
0.16
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00

15
10
5
0

92

93

94

95

96

97

98

Purity, %

Figure 9 Joint probability density graphs

prot response surface shows that

the optimum falls in the top left
corner of the operating range highlighted 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 G3D2 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

PP splitter optimisation

2.5
0
2.5

0
2.5

ne in p

ty
,

3.2

Prope

.75

ropan 1
e, %

Figure 8 Suggested process optimisation path

52 PTQ Q1 2014

valero.indd 6

97.00
96.25
95.50
94.75
94.00
%

94.4, 3.6

97.0, 2.5

1.0

u
ri

5.0
0
4.0

96.1, 2.9
95.3, 3.2

Profit, $MM

5.0

optimum operating targets. The

path was designed with three incremental
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 highlight the product specication

www.eptq.com

11/12/2013 20:22

efficiency
trays
must
beline.
usedThe
to conventional
reduce
refluxanalyratio
constraints.
contour
maps
generated
using
forming
inside
the transfer
who are
not The
desk-bound
and
canwere
benefit
from access
to
and
lower
the
overhead
cooling
load.
Flue
gas
2
, which
uses
SAS Analytics
procedure
Proc
KDE
sisthe
showed
plug trends
flow
and
flow
in horizontal
and
real-time
data,
and slug
alerts
anytime,
anywhere.
turbines,
energy
recovery
turbines,
energyKernel
density
estimation.
The
panel chart
vertical
pipes
respectively
for hydraulic
Case3 3.x 3engineers
Plant
managers
and production
can was
use
saving
motors,
frequency
conversion
motors
and for
air
constructed
using the
SAS
Analytics
SAS
procedures
Figure
shows
the
volume
mobile 10b
BI anytime,
anywhere
infraction
order
tocontours
first underflow
regulation
systems
of
compressors
should
be
Proc
GSlide
Proc
GReplay.
Note
thatwith
the
andthen
Case
4, an
which
to
thesolutions
start-up
stand
issuecorresponds
and
propose
tocase
problems
employed
to
recover
pressure
energy
and
reduce
elecjoint
densityflow.
distribution
is experienced
veryin compact
attradithe
higher
The
flow regime
Case 4with
is similar
in a vapour
time-frame
not
previously
tric
energy
consumption.
High-efficiency
intensified
purities
as
long
as
the
propylene
concentration
to higher
slug
flow.
Even
though
the
gas
volume
fraction
has
tional desktop solutions. The prevalence of mobile
burners
need
toproduct
be
used
furnaces
improve
effiin
the propane
is in
above
2.5%.
also
that
increased
to
0.933
from
0.8
in
Case
1,toNote
the
slug
flow
devices
is transforming
the
process
industries.
Mobile
ciency.
The
exhaust
temperature
of
furnaces
should
be
when
theempower
propylene
propane
product
was
regime
still
prevailed
in in
the the
transfer
line.
Conventional
solutions
decision-makers
to have
immediate
reduced
to
improve
thermal
efficiency
by
2-3%.
New
either
1%
1.4%,
variability
of thethem
MA and
increased
analysis
slugthe
flow
in both
horizontal
access showed
to orimportant
data,
enabling
to vertimake
insulation
material
needs by
to be
employed
to of
reduce
as
reflected
the
wide
span
the
calsignificantly,
pipes
for
Case
4.
informed and quick decisions to improve profitability.
the understand
heatcontour
loss of equipment
and
pipelines.
density
map.
This
effect
is
more
noticeable
To
the
effect
of
variation
of
vapour
Easy, digestible analysis of plant information, even in
Low-temperature
heat
should
be
utilised
byexerted
applywhen
propylene
purity
was
reduced,
as is
volume
fraction
in the
start-up
casealso
on
remote
locations,
helps
industry
to pressure
react
to adverse
ing
low-pressure
steam
generation
technology,
from
the
(95.5%,
1%)
and
(94.4%,
1.4%)
onevident
the
wall,
pressure
fluctuations
in
the
transfer
line
changes and keep the operation performing to targets.
low-temperature
Organic
Rankine
Cycle
(ORC)
scenarios.
over
time
were
studied
(see
Figure
11).
It
was
found
The ability to access and analyse real-time plant data
systems
andfluctuations
Kalina Cycle
systems.
that
pressure
forthe
Case
3 are
low
compared
has
enormous
benefits. In
past,
users
needed
to be
Energy
optimisation
of
the area should
be carried
to Conclusions
Case
4.
This
is
due
to
liquid-dominated
operation
in
in the control room or in front of a monitor to track
out.
Heatline.
integration
between
the refinery
and 4local
article
illustrates
the
application
of
a
methodoltheThis
transfer
The
pressure
fluctuations
for
Case
are
and manage manufacturing performance. Mobile BI
co-generation
power
plant
to be
realised.
A
ogy
developed
tobe
optimise
propylene
unit
very
Case
1.more
This the
isneeds
because
thesplitter
slug
hassimilar
provento to
effective
when
usersflow
are
large
quantity
of
low-temperature
heat
in
a
refinery
product
quality
in the
presence
of
regime
prevailed
the transfer
line
for Case
4. varying
provided
within targets
visualisation
tools
(charts,
graphs,
cannot be
recoveredThe
because
a heat sink
is not
availaprocess
conditions.
methodology
uses
a process
portals and so on).
ble,
and
this
portion
of
low-temperature
heat
may
be
simulator
along
with Monte
Carlo random
sampling
Conclusions
and
recommendations
For todays
engineers,
the message
is simple:
mobile
used
as
a
heat
source
for
demineralised
water
and
to
optimise
product
quality
outside to
theachieve
current
operCFD
analysis of
a two-phase
vapour-liquid
mixture
flow
intelligence
provides
the platform
greater
boiler
feed water
inproduct
a delayed
cogeneration
power
plant.
Also,
range.
Two
specification
constraints
in ating
the
transfer
line
of
a
coker
unit
was
carried
profitability. State-of-the-art mobile software, such as
integration
with
thethis
local
chemical
plant
ispropylene
important
considered
in
1) minimum
outwere
at different
modes
ofstudy:
operation.
The decision-makflow
during
Aspen
InfoPlus.21
Mobile,
enables
faster
to
achieve
material
exchange
and
optimisation,
as
well
purity
and
2)
maximum
MA
concentration.
The
methstart-up
operation,
wherein
vibrations
were
observed,
ing and troubleshooting and displays critical,
as
optimum
energy
use.
Integrated
energy
optimisaodology
caninformation.
also normal
be applied
to situations
with a
was
compared
with
operation
where
vibrations
up-to-date
The
software
functionality
tion
ofnumber
the
area
should be
carriedlarge
outThe
rather
than
higher
of
product
specifications.
example
were
absent.
While
simulation
showed
slugs
during
improves employee efficiencies by simplifying routine
energy
optimisation
of
a
single
refinery.
here
assumed
thatsuch
inputasprocess
variables
thepresented
start-up operation,
mist
flow
prevailed
during
normal
engineering
analysis
tasks,
examining
and
followed
independent
normal
distributions,
but
other
operation.
Large
slugs
of
liquid
lead
to
large
pressure
comparing process data, reducing root cause analysis
Conclusions
probability
distributions
awill
multivariate
fluctuations,
ultimately
resulting
vibrations
during
time and easily
finding
KPIsuch
datainas
that
enable
the
The priority
ofcorrelation
refining
technology
development
normal
with
structure
or
independent
start-up.
Further,
CFD
simulations
were
carried
outand
to
engineer to respond to changing process conditions.
the configuration
of can
refining
units avoid
in China
have
gamma
distributions
be would
considered
for
model
determine
the
conditions
which
vibrations
distinct
characteristics.
During
asteam
long
period
the
The
same
methodology
be flow
used
to inoptiin inputs.
the transfer
line. Lowering
the can
rate
was
Conclusion
future,
the
FCC
unit
will
still
be
the
main
secondary
mise
refinery
and
petrochemical
units,
where
found
be the
only
solution
available
for reducing
transOvertoother
the
past
30 years,
MES
technology
has
evolved
to
conversion
unit
for need
gasoline
and
diesel
production.
product
specs
to
be
met
in
the
presence
of
ferfinal
line
vibrations.
Even
at
higher
steam
flow
rates,
a
slug
help refiners survive in highly competitive markets.
Therefore,
improving
theThe
quality
ofofFCC
gasoline
varying
process
conditions.
results
the
flow
regime
prevailed.
Good
agreement
wasanalysis
found
Real-time
data
and decision
support
tools
provide
and
diesel
is
very
important
for
oil
product
quality
were
used
to
determine
optimum
operating
targets
between
flow
regimes
obtained
from
CFD
and
convenaccess to plant information to allow quick and timely
upgrading
technology
in the future.was to reduce the
system.
for
the
unitto
APC
tional
analysis.
The
recommendation
responses
production
issues that negatively influWith
the
trend
towards
poor-quality
crude
oils,
steam
to avoid quality
vibrations
during
start-upcompliance.
operation.
enceflow
efficiency,
and
regulatory
future
refineries
should
further
optimise
process
flows
PROC
and PROC
G3GRID are
of SAS Analytics
Inc.
MESSIMNORMAL
is essentially
the nucleus
of marks
the operation,
which
References
and change
unit
configuration
so as toExcel
improve
Petro-SIM
is a mark
of KBC
Advanced Technologies.
is a markthe
of
linksD W,
allPerry
capabilities
of
the business.
ItHandbook,
is an integrated
1 Green
H B, Perrys
Chemical
Engineering
8thand
Edition.
comprehensive
utilisation
rate
of resources,
to
Microsoft
Corporation.
setSchepper
of production
activity
and B,
support
software
2 De
C K, Heynderickx
G J, Marin
CFD modeling
of all
meet the Srequirements
of
energyG savings
and emisgasliquid
and vapourliquid
flowand
regimes
predicted
the Baker chart,
designed
to harmonise
optimise
thebyplant.
References
sions reduction.
Chemical
Journal,
349357,
2008. production
TheEngineering
bottom
lineG E,
is138,
that
effective
drives
1 Gary
J
H,
Handwerk
Kaiser
M
J,
Petroleum
Refining:
More
detailed classification
ofInc,
crude oil Technology
fractions
3 Fluent
6.3 Documentation,
User
Guide,
Ansys
operational
excellence,
enabling
betterUSA.
and faster deciand
Economics,
5th
ed,
CRC
Press,
2007.
and the consideration of various narrow-cut processsions.
Software
technology
helps and
refiners
achieve
2
SAS
Customer
Support,
Rahul
C technologies
Patil
is Senior
Manager
ofhttp://support.sas.com.
the modelling
simulation
group
ing
will2005,
actively
promote
the
low carbon
consistent
across
allJamnagar,
assets.
It also India.
defines
3
Castillo
E, performance
Process
Optimization:
A
Statistical
Approach,
Springer,
of refining
R&D
at
Reliance
Industries
Ltd,
Gujarat,
He
emission of refineries. Narrow-cut processing will
2007.
holds
in chemical
IIT Kanpur performance
and a degree in
thea Masters
importance
ofengineering
real-timefrombusiness
become
the development
trend in the
overall process
4
Montgomery
C, Myers
R H, Response
chemical
engineeringDfrom
ICTexecute,
Mumbai.
management:
plan,
monitorSurface
and Methodology:
respond to
flow
optimisation
of
refineries
in
the
21st
century.Wiley,
Product
Optimization
Experiments,
Dr Process
Randheer
Yadav
specialises
in the
application
of computational
changeandimmediately
on
allUsing
timeDesigned
horizons.
History fluid
has
1995.
dynamics
systems
involving multi-phase
flows.
He holdshave
a PhD
in
showntothat
manufacturing
execution
systems
laid
5 Montgomery
D from
C, Peck
E A, Vining G G, Introduction to Linear
chemical
engineering
ICT
Mumbai.
the foundations to help refiners across the globe
Analysis,
5th ed,Vice
Wiley,
2012. and leads the modelling and
Dr Regression
Ajay Gupta
is Assistant
President
strengthen
their
competitiveness
and build
upon the
Sun Liligroup
is Vice
of Sinopec
Engineering
in
simulation
of President
refining R&D
at Reliance
IndustriesIncorporation
Ltd, Jamnagar,
pillars
profitability.
Jose
Bird
isprocessing
Director Energy
System
and
Metrics with
Valero Energy
charge
ofof
solution
studies,
engineering,
construction
and

cause sediment
fouling tendenc
tially with conve
the value benefi
to fouling and s
version relations
advantage in dec
are acceptable o
result in lost con
improvement in
and run length
concept of optim
rate of fouling an

A New Standard
in Pressure
Measurement

Fouling problem
Problem areas

High Performance

U1>Vi`iV>`i>`
>VV>V
U1V`>}i>Li
/>
v"i/
"
Ui}vV>

Safety

Uiiv>>vii
`Vii`v>i`i
U
iwi`
xn>>`>`v}i
i->``>i-

Robustness

U,}}i`]iV>VVV
Ui>i>v>`>
ii
ULiiiiV>v
Vi>i`i>L

Gujarat, India. He has Bachelors, Masters and Doctoral degrees in chemical

Corporation.
He is responsible
for implementing statistical solutions in
startup of
grassroots
refineries
Engineering
from
IIT, Delhi,
India. and refinery modification projects. She
the
areas
of
energy
efficiency,
process
optimisation,
process
has
over
20
years
experience
design
and
of
Marty
is thePresident
MES Manager
atengineering
AspenTech.
During
30 monitoring
years
in the
Dr Asit
K Moran
Das is Vice
andinheads
the Refining
R&D
atstartup
Reliance
and
ethanol
manufacturing
operations.
hydroprocessing
units,
has
won
several
Prizes
of
National
Scientific
processLtd,
industries,
he He
hasholds
worked
as a consultant
in over
65 refineries,
Industries
Jamnagar.
a degree
in chemical
engineering
from
Darryl
Seillier
Technology
Advisor
with
Valero
Energy
Corporation.
and Technological
Progress,
and
a BS
in petroleum
refining
from
chemical,
gas isplants
and
other
manufacturing
environments.
He
Jadhavpur
University,
West
Bengal,
a holds
Masters
in chemical
engineering
from
He
is responsible
for
company-wide
process
improvement
in
the
areas
China
University
of
Petroleum.
holds
a
US
patent
for
multivariable
control
and
a
degree
in
chemical
IIT, Kanpur, and a doctoral degree in chemical engineering from Ghent
of
energy
efficiency,
systems
and process optimisation.
Email:
sunlili.sei@sinopec.com
engineering
from thehydrogen
University
of Illinois.
University,
Belgium.

www.dpharp.com

The sections mo
column, vacuum
very high conv
also suffer from
ing of the se
unplanned shutd
The same tren
sediment genera
of conversion t
certain limits th
time cannot be c
From the abov
the proper oper
enables the best
sation and prod
rates of fouling
continuous con
stability is relat
deposits and gen
Optimal sever
feed being proc
refinery feed q
plant feed rate
factors that may
stability reserve
(often reported
and the intrinsi
stability reserve
solubility will a
to generate coke
fouling deposits
The amount
sodium and iro
process perform
and, in some c
favouring dehy
precursors
asphaltenes i
To avoid dep
plays a major ro
process can hav
tors, the catalys
catalyst activity
The effect of the
fouling can be

www.eptq.com
www.eptq.com
www.eptq.com
www.eptq.com

EJA-E-270x86.indd 1
valero.indd
7 8 3
ril.indd
6
sinopec.indd
aspentech.indd

baker hughes.indd 2

07/09/12 13:50
11/12/2013
12:22
13/09/2013
11:42
10/12/12
13:02:15
07/06/2013
19:12

Merichem Company.pdf 1 5/22/2013 8:09:34 AM

Treating Solutions

Proven
Results
C

CM

MY

CY

CMY

Merichem Company provides a diversified portfolio of products and services to a wide range of
operations, including the refining 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.

merichem.indd 1

06/06/2013 10:23

Extending the treatment of highly sour

gases: cryogenic distillation
Cryogenic bulk removal of H2S or CO2 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

ompanies willing to produce

large gas fields with very
high amounts of CO2 have to
face a constraint related to the
essential need to reduce the atmospheric emissions of greenhouse
gases.
The economics may also be
improved by the growing acceptance of the re-injection of separated
H2S and CO2, 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 recovery to a diminishing sulphur
market, or to limit atmospheric
emissions of greenhouse gases,
increases the number of highly sour
gas fields that can be reconsidered
as exploitable to produce much
needed natural gas.1,2
These new constraints lead to the
development of more energy efficient 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 preceding 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
benefits of the Sprex and SprexCO2
processes.

Cryogenic distillation for the

sweetening of super sour gases

Some gas fields contain very high

amounts of H2S (more than 30
vol%) or CO2 (up to 70 vol%) in

www.eptq.com

prosernat.indd 1

natural or associated gases. Even

though amine processes can be
optimised to treat very sour gases,
the high cost associated with sweetening 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 CO2
from gases with a very high CO2
content down to levered CO2
content generally between 5% and
10%. With the current status of this
technology, permeation membranes

Separating CO2 or H2S

down to a commercial
specification
requires a very low
temperature in the
reflux drum
can only be used to treat gases with
very limited H2S 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 CO2
specifications.
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 (H2S and/
or CO2) are recovered in the liquid
state under pressure. Producing the
acid gases as a high pressure liquid
saves
expensive
and
energy
consuming compression requirements, 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 separation. Separating CO2 or H2S down
to a commercial specification
requires a very low temperature in
the reflux drum, corresponding to a
very
significant
refrigeration
requirement. Furthermore, CO2
separation down to commercial
levels suffers from limitation due to
CO2 freezing conditions in the top
section of the cryogenic distillation
column. Some processes address
this limitation by adding, for example, a suitable hydrocarbon solvent
to the top section of the demethaniser to stay outside the CO2
freezing conditions, or by using a
column with a specific frozen CO2
remelting zone.
The
Sprex
and
SprexCO2
processes, jointly developed by
Total and IFP Energies nouvelles/

PTQ Q1 2014 55

11/12/2013 20:20

Very sour
natural gas

Sales gas

Pre-ex

AGRU
Acid gas
removal unit

Acid gas liquid

high pressure

Remaining
acid gas

Figure 1 Staged acid gas separation with

the Sprex process

Prosernat, are bulk fractionation

processes. The Sprex process for
bulk H2S removal does not require
upstream dehydration of the gas.
The refrigeration requirements of
Sprex and SprexCO2 are limited
compared to those of other cryogenic processes and, as the
temperature in the SprexCO2 is not
as low, the operating conditions are
far away from the CO2 freezing
region, therefore avoiding the need
for an additive fractionation and
recycle or for a remelting zone.
When pipeline or LNG specification is required, the Sprex or
SprexCO2 unit is easily and
economically combined with a
solvent based acid gas removal
unit.

Bulk H2S removal with the

Sprex process
The Sprex
developed

process was jointly

to
improve
the

TEG

Fuel
gas
Sprex

prosernat.indd 2

30C

energised
MDEA

50-65 bar

Acid
gas

Refrigerant

Sour gas
30-35C

H2S soup
60-70C

Re-injection well

Figure 2 Sprex simplified process flow diagram

economics of the production of

ultra sour gas with high H2S
content when the separated H2S is
re-injected. The process was
improved and several other patents
were filed in the 2000s.3 Figures 1
and 2 show the principles and
process flow 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 H2S content in the

Figure 3 The Sprex demonstration plant at Lacq in France

56 PTQ Q1 2014

Pipeline
gas

gas by producing a partially sweetened gas, which can then be

processed by a smaller, conventional amine sweetening unit,
capable of meeting the most severe
H2S and CO2 gas specifications
(pipeline gas or feed for a LNG
plant)
It produces a high pressure liquid
H2S 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 demonstrated in an industrial context at
the Lacq plant,4 in the southwest of
France (see Figure 3), the H2S
content in the gas leaving the Sprex
unit is about 10-12%. The reflux,
consisting of cold, dry H2S, dehydrates to some extent the incoming
gas in the upper zone of the Sprex
column. The minimum temperature
reached in the units low temperature reflux 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 H2S separated out in the
process is produced in a high
pressure liquid phase, requiring
considerable less energy for reinjection into an underground
reservoir, as the large, multistage
acid gas compressors can be
replaced by injection pumps.

www.eptq.com

13/12/2013 12:02

100

Acid gas
re-injection

Base case: Sprex + amine

Sprex
(H2S cryogenic
separation)

Amine unit
(MDEA)

Acid gas
re-injection

100%

90

Compression power, %

Amine unit
(MDEA)

Relative steam consumption, %

Reference case: full amine

80
70
60

60%

50
40
30
20
10
0

Sprex +
amine

Amine
only

100

100%

90
80
70

69%

60
50
40
30
20
10
0

Sprex +
amine

Amine
only

Figure 5 Sprex economic evaluation: acid

gas compression power

Figure 6 Sprex economic evaluation: steam

consumption

Now that the amine sweetening

unit located downstream has less
H2S 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%
H2S and 10 mol% CO2 at 70 bar.
The sales gas specication has been
considered at 4 ppm mol of H2S
and less than 2 mol% CO2. The acid
gas containing H2S and CO2 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 overall capex is decreased using Sprex.
Depending on the overall ow rate

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 H2S
increase with time, due to H2S
re-injection in the producing reservoir, 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 H2S 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 simplies 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.

www.eptq.com

prosernat.indd 3

Pressure, bar

Figure 4 Sprex economic evaluation:

considered cases

120
110
100
90
80
70
60
50
40
30
20
10
0

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-injection, 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 compositions and operating conditions of
the process: the Gasweet model was
calibrated from literature and

Liquid-liquid
Liquid-liquid-vapour
bubble curve

Liquid-liquid-vapour

Liquid-vapour
Liquid-liquid-vapour
dew curve

Vapour
Liquid-vapour
dew curve

25

50

75

100

125

150

175

200

225

250

Temperature, C
Figure 7 Phase diagram of H2O/H2S/ CH4/CO2 system. Comparison between experimental
data and the Gasweet model

PTQ Q1 2014 57

11/12/2013 12:33

35
Hydrate prediction temperature
Lab data in-house model

In-house model, C

30
25
20
15

CH4/CO2/H2S system

10
5
0

10

15

20

25

30

35

40

Lab data, C

Figure 8 Hydrate prediction temperature with very low water content: in-house model

Cas 30% yH2S exp

Cas 30% yH2S cal

Tray number

4
6
8
10
12
14
16
18
20

10

20

30

40

50

60

70

80

H2S content, %
Figure 9 H2S concentration profile: pilot results vs Gasweet model

50
MeOH
injection

45

Temperature, C

40

Temperatures
inversion

35
30

TR9
TR11
TR13
TR15

Hydrate
melting

25
Start of hydrate

20 formation
15
10
5
0

250

500

750 1000 1250 1500 1750 2000 2250 2500

Time, seconds
Figure 10 Forming and melting hydrate in a Sprex pilot plant: temperatures on selected trays

specifically acquired isotherms, for

fluid compositions including light
and heavy hydrocarbons, H2S, CO2,
nitrogen, water, aromatics, mercap-

58 PTQ Q1 2014

prosernat.indd 4

tans, COS, CS2 and methanol, in a

wide range of temperature/pressure conditions, from -85C to
150C and up to 150 bar.

Figure 7 shows that the developed

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% H2S and with low water
contents. The model that was eventually 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 dehydrated with cold liquid H2S in the
upper zone, and light hydrocarbons
and CO2 are stripped from the H2S
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 operating conditions of the Sprex unit,
and acid gas injection pipelines and
wells, to allow the safe selection of
materials of construction
Develop adequate risk management studies and mitigation
measures to manage the risk
involved in handling dense phase
H2S rich mixtures in large processing units
Select and qualify specific pieces
of equipment, such as liquid H2S
pumps.
The decision was made to build
and operate an industrial demonstration 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,

www.eptq.com

11/12/2013 12:34

Temperature, C

such as variations in raw gas flow,

conditions and quality, production
upsets, and so on.3 The unit
processed sour gas of variable H2S
content (from 18% to 40%), operated
by
conventional
plant
personnel, from the beginning of
2005 to 2006.
Operation
has
demonstrated
excellent agreement between the
thermodynamic models (L/V equilibriums [see Figure 9] and
hydrates formation conditions);
measured the actual efficiency of
the trays used in the column; validated the process and safety
engineering studies; and validated
the adequacy of the choice of materials 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 adjustments in the operating conditions
and methanol injection (see Figure
10).
The process was eventually
formally qualified by Total for
implementation in the development
of large super sour gas fields, using
the companys formal internal qualification procedure for innovative
processes.

20
10
0
10
20
30
40
50
60
70
80
90
100

50 bara
40 bara
30 bara
CO2 crystallisation limit
Gas
Liquid

10

20

30

40

50

60

70

80

90

100

CO2, mol%
Figure 11 CO2/ C1 distillation chart

gases containing a very high CO2

content (40% or so), possibly
together with high concentrations
of H2S.
The SprexCO2 process is based
on an optimised CO2 distillation,
which is performed at conditions
outside the CO2 crystallisation
domain. The parameters and limits
to avoid crystallisation are illustrated in the distillation chart based
on the main binary mixture (CO2/
CH4). Figure 11 shows a typical gas

specification achievable in the

distillation outlet around 20%
CO2 mol at 30/40 bar.
The SprexCO2 process is competitive
in
comparison
with
separation
conventional
CO2
processes (amines) or other hybrid
solutions (membranes + amines)
when CO2 concentrations are
higher than 40%, and CO2 reinjection is required for enhanced
oil recovery and/or carbon sequestration. Carbon dioxide is recovered

To amine
unit

Refrigeration
loop

CO2 soup
Refrigerant
Precooling
Thermal
integration

FC
Cold box

Cryogenic bulk removal of CO2

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 efficiency of the H2S
separation process and allows the
separation of CO2 from the gas.
However, the limit then becomes
the temperature at which CO2 crystallises. The enhanced version of
the Sprex process, called SprexCO2,
allows bulk removal of CO2 from

www.eptq.com

prosernat.indd 5

Sprex
column

From
pre-treatment

Pre-treatment

To
re-injection

Figure 12 SprexCO2 simplified process scheme

PTQ Q1 2014 59

11/12/2013 12:34

Auto-consumption
HC losses
Valorised gas
100

Material balance, vol%

90
80
70
60
50
40
30
20
10
0

Sprex CO2 Membranes

Figure 13 Overall performances

comparison: SprexCO2 + amine vs
membranes + amine

60 PTQ Q1 2014

prosernat.indd 6

Fluid composition

Methane
CO2

Mol%
24-27%
68-71%

Table 1

amine combination has been evaluated as an alternative, using

commercial membrane vendors
design data.
Removing CO2 from 70% to 20%
(our reference case) represents
more than 90% removal of the CO2
from the feed gas. Subsequent
amine units, using the AdvAmine
technology, remove all the remaining CO2 down to the gas
specification required.

The acid gas loading

of amine solutions
needs to be
maximised to reduce
the required amine
solution flow rate
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
significantly
reducing
overall
hydrocarbon losses with the separated CO2.
As Figure 14 shows, there is
roughly an order of magnitude
difference in power required for
CO2 recompression. From an energetic point of view, one of the
interests of SprexCO2 is, indeed,
that the CO2 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 benefits in terms of
waste heat recovery optimisation.
One of the advantages of the
proposed CO2 separation scheme is
the possibility to increase process

flexibility to a certain extent by

means of bulk and finishing separation units in series, in order to cope
with fluctuations in feed gas
composition or temporary offset
conditions.
The level of complexity of the
proposed CO2 separation scheme
remains moderate: no prototype
equipment is required, minimum
acid gas compression system is
required (only for the acid gas separated 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 efficiently and economically treat
any type of highly sour gas.
To be truly competitive, the
removal of acid gas components
H2S and CO2 from highly or super
sour gases requires the optimum
choice of process. Within the
AdvAmine series of processes, the
energizedMDEA process can be

100

Power consumption, %

in a liquid soup phase, so injection

is realised by pumps instead of
multistage
compressors,
with
considerable energy savings.
In order to match final gas export
specifications (pipeline or LNG
plant feedstock), the SprexCO2 unit
can be associated with a finishing
AdvAmine unit.
A simplified SprexCO2 scheme,
including upstream pre-conditioning units and downstream CO2
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 final
specifications 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 combination. The range of methane and
CO2 in the raw gas composition is
shown in Table 1.
The gas contains a few per cent
of C2+ and inerts as well.
The base case for natural gas
sweetening includes the SprexCO2
technology for bulk removal of CO2
and a finishing gas treatment based
on AdvAmine technology (energizedMDEA). The membranes +

90
85%

80
70
60
50
40
30
20
10
0

10%

Sprex CO2 Membranes

Figure 14 Power consumption dedicated

to CO2 recompression: SprexCO2 + amine
vs membranes + amine

www.eptq.com

11/12/2013 12:34

optimised
to reduce
energycoverage
requireand,
of course,
bad media
ments,negative
taking publicity.
advantage Sudden,
of the
and
ability of MDEA
to are
release
unplanned
shutdowns
nevera
substantial
partreliability
of the absorbed
acid
good
for the
of a re
ngases operating
in a lowunits,
pressure
flash the
to
erys
especially
produceones.
the semi-lean
in a
larger
Motors stream
and rotary
split-flow process
configuration.
equipment
are damaged
when
The acid gas
loading
suddenly
taken
from of
fullamine
speedsoluto a
tions stop,
needs and
to be
maximised
to
dead
consume
a great
reduce
required
solution
deal
of the
energy
while amine
restarting.
But
flowsteps
rate. Operating
plants at
if
are takenamine
to minimise
high loadings
has been
by
unplanned
failures,
all proven
of these
Total throughcan
extensive
operating
consequences
be reduced.
experience, and a liquid side cooler
Finding the
reliable
scheme
on most
the acid
gaselectrical
absorber
equipment
was
developed to further increase
Reneries
need
work the
with
electhis
loading
andtoreduce
cost
of
trical technology
vendors to design
highly
sour gas sweetening.
and
construct
the most reliable
When
the mercaptans
contentgrid
of
for each
unit.
cooperation
highly
sour
gasThis
becomes
an issuecan
to
help re
improve
theirsulphur
energy
meet
a neries
stringent
total
security andgasefspecification,
ciency. Tablethe
1
commercial
shows a process
list of from
selected
HySweet
Totalmajor
may
companies
from those that were
be
considered.
surveyed
Even thoughby
amine Hydrocarbon
based sweetPublishing
in preparation
ening
has Company
proven to
be a very
of its report.
The survey
coversbulk
120
efficient
technology,
cryogenic
or CO2 the
offers
an
companiesof from
world
removal
H2S around
economic
advantage
when the
that provide
various technologies

separated acid gases from super sour gas processing: H2S removal with the Sprex
sour gas are re-injected. The Sprex process, SPE paper IPTC 10581-PP, International
Energy processes, jointly
Financial
Plant
safety held in
Petroleum Technology
Conference
and SprexCO2
efficiency
performance
and
liability
Doha,
Qatar,
21-23
Nov
2005.
developed by Total, IFPEn and
Prosernat, can be combined with an 4 Huygue R, Lemaire E, Lecomte F, Deixonne
M, Lajeunesse H, Operational validation of the
amine unit to reduce the cost of the
Sprex process for bulk H S removal, Sulphur
production of gases containing Conference, Vienna, Austria,2 Oct 2006.
more than 25-30% H2S or over
Negative
an failure
40-45%Reliability
CO2. They providePower
publicity
Francois
Lallemand
is
Senior
Advisor with
impacts
economic and easy to operate alterProsernat.
native to amines, to membranes
Gauthier Perdu is Deputy Process Director
and to other more complex cryowith Prosernat and Head of the Amine Process
genic fractionation processes for the Technology Group within Prosernats Process
treatment of such highly sour department. He holds a Dipl. Engineer and Post
gases. The Sprex
Asset series of processes Graduate certificateEnvironmental
in chemical engineering
concern
and
management
was demonstrated in a unit from ENSIGC Toulouse. liability
installed at the Lacq sour gas plant, Laurent Normand is Business Development
and the technology is now ready to Manager with Prosernat. He holds a PhD in
Figure
2 Impacts of at
refiindustrial
nery power scale.
failures material science and graduated as an engineer
be implemented
References
that
are further discussed in the
1 Lallemand F, Aimard N, Rocher A, Sour gas
report.
production: moving from conventional to
advanced environmentally friendly schemes,
Six Sigma
Sogat Conference, Abu Dhabi, 29 Apr3 May
Strategies for process improvement
2007.
such
as SixMinkkinen
Sigma rely
on identify2
Lallemand,F,
A, Processes
combine
ingassist
andacid
removing
defects
within
a
to
gas handling,
reinjection,
Oil and
process
out proper
Gas
Journal,and
21 Jancarrying
2002.
3
Lallemand F, Lecomte
F, Streicher
C, Highly
operational
practices
by quali
ed

Elemental Analysis
of Fuels and Oils

in physics from the cole Nationale Suprieure

des
Mines de Paris.
personnel.
This starts with the initial
Claire Weiss works for Total EP in the process
design of the power system by the
department within the Acid Gas group.
electrical engineer, where all factors
Julia Magne-Drisch is Gas Treatment Project
such as steady-state load requireManager at IFP Energies nouvelles. She holds a
ments,
ambient conditions, and
PhD
in hydrotreatment.
quantiedGonnard
fault conditions
thorSebastien
is ProcessareEngineer
oughly
considered.
at
IFP Energies
nouvelles, Management
in charge of
and engineers
must
work with
process
development
andalso
techno-economic
evaluations.
the vendors to be certain that all the

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.

Determination of Sulfur and other

elements at-line and in the laboratory
Discover more exciting details at
www.spectro.com/fuels

www.eptq.com

prosernat.indd
7
hpc.indd 5

Q1 2013
2014 109
61
PTQ Q4

13/12/2013
13/09/2013 12:03
11:51

00000 EFC A4 Advert_Layout 1 13/12/2013 11:55 Page 1

Book before the 20th of

December and save 400
Use VIP code DF10

18th March 2014 // Rome, Italy

Europes number 1 strategic

event for the fuels industry
An unparalleled line up of
reners and industry experts:
Giacomo Rispoli
Executive Vice President
RSP/R&M - Ricerca, Sviluppo
Tecnologico e Progetti Refining & Marketing Division
ENI
Philippe Marchand,
Directeur of Biotechnology New Energies Division,
Total
Hugh Tucker,
Technical Director,
UKPIA

Alessandro Bartelloni,
Policy Director,
Europia

Organised by:

Supporting Partner:

Media Partners:

www.energyexchange.co.uk/europeanfuels
euro fuels.indd 1

marketing@theenergyexchange.co.uk
+44 (0)207 384 8012
13/12/2013 12:41

Optimising hydrogen sulphide scavenging

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

cavenging hydrogen sulphide

(H2S) from petroleum products
may be easy, but doing it efficiently 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.

H2S risk

Risk is known to be the product of

the probability of an event and its
consequences, and H2S is a poisonous
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.
H2Ss specific features make it
worthy of special attention. Being a
colourless gas, its presence can be
identified only by smell which can
be deceptive. H2S can be detected by
a characteristic odour of rotten eggs
when present in very low concentrations (1.4 g/m3 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 concentration becomes more dangerous.
And exposure at a concentration
equal to a few hundred mg/kg may
already be lethal.
H2S is present, in extremely variable amounts, in crude oils.

www.eptq.com

chimec.indd 1

Advanced refining processes generally allow for high, if not complete,

removal of H2S from light and
middle distillates, with rare but
possible exceptions.
The situation is different when it
comes to heavy fuel oils. Generally
speaking, H2S content in residues is
the result of the sum of what
remains trapped during the distillation processes, what is formed as a
result of cracking phenomena, and
what is removed by desulphurisation and/or stripping processes.
The nature of the sulphur
compounds present in the crude,
and the severity of cracking applied,
influence the amount of H2S that
can be produced as a result of these
phenomena. Conditions sufficient
for the formation of H2S, for example, can be reached by the bottom of
the vacuum distillation columns.
The nature of residue refining
processes is more oriented towards
the formation of H2S (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 thresholds for human health but there is
an important difference: the phase
to which these concentrations refer.
When we speak about petroleum
products, H2S content is reported as
concentration in the liquid phase.
From this liquid, H2S is then
released into the vapours above the
oil, where it becomes dangerous to
humans. The amount of H2S developed in the vapour phase depends
on several factors, including the
characteristics of the oil, conditions

of temperature and pressure, and

obviously the H2S concentration in
the liquid.
For residues and heavy fuel oils,
the levels of H2S in the vapours are
normally 50-100 times those in the
liquid. It is therefore clear why even
a few parts per million of H2S in the
liquid phase can be so dangerous.

H2S specification

On 1 July 2012, a H2S specification

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 specification set some new
operating requirements and new
challenges for all companies
producing marine fuels.
Refineries
have
means
of
controlling H2S content in the residues 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 conditions can lead to a H2S content as
much as an order of magnitude
higher than the minimum required.
Achievement of the H2S specification in ISO 8217 therefore starts
with the proper management of
fuel oil production processes. When
this is not possible, or the H2S level
is not sufficiently low, the specification has to be met through the
implementation of other solutions.
The main and most common
among these is the use of specific
additives able to remove H2S by a
chemical reaction, so-called H2S
scavengers. The alternative is the
installation of stripping columns
for the residue, but this solution is
seldom applied due to high instal-

PTQ Q1 2014 63

11/12/2013 12:39

lation and operating costs. In some

cases, there may be a competitive
solution through the use of
additives.
Even though some refineries were
already dealing with H2S scavenging before 2012, the number of
applications sharply increased after
the implementation of the new
specification, 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 H2S
scavenger treatment and the impact
they can have on the overall results.
A deeper treatment of the issue
would require the use of confidential information that cannot be
provided in these circumstances,
but Chimec is available to provide
further clarification of our experiences and technologies.

Case study

Refinery scheme, operating conditions and H2S content are among the
fundamental variables to be taken
into account for the proper setup of
an H2S scavenger application.
Monitoring, on the other hand, is a
key aspect of treatment management and, to achieve real cost
optimisation, specific programmes
should be set up based on the characteristics of the site.
The best way to show this is
through real experience. It is not
the specific case that this article
wants to focus on, as it may differ
from the situation in other refineries, but the basic principles applied.
These principles have wider
applicability.
The refinery described in this case
study produces marine residual
fuels through the blending of a
vacuum residue and a mix of fluxants. The main source of H2S is the
vacuum residue while its content in
the lighter components is negligible.
When we approached the refinery
for the first time, an H2S 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.

64 PTQ Q1 2014

chimec.indd 2

Even though the refinery was

applying
a
good
monitoring
programme, the dosage rates were
unusually high. Two very important
aspects were that the starting H2S
content in the heavy fuel oil was
quite variable, ranging between 10
mg/kg and 40 mg/kg, and that the
refinery had very flexible production planning which made it
necessary to have the H2S scavenged
in a short time to make the fuel
always available when needed.
In particular the heavy fuel oil
needed to be on specification by the
time it reached its intermediate storage tanks, before final blending, and
the minimum storage time in these
tanks was only a few hours, the
time therefore available for the H2S
scavenger to complete its action.
The high variability of the H2S
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 refinery, aware of the importance of these factors, was
monitoring the blank H2S 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 fulfillment of the
H2S spec.

Monitoring scavenger treatment

There are two main targets in

monitoring H2S scavenger treatment: reliable measurement of the
H2S 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 necessary to identify the critical variables
and to define the best procedures.
Continuous assistance and collaboration during the process make
these procedures effective.
With regard to reliable measurements, 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 H2S scavenging
applications, this means taking due

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 H2S 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
field 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 injection point, but the time between
sampling and testing varied, differing by up to a few hours (see
Figure 1). That was causing sometimes misleading results because,
oddly enough, when the refinery
was more efficient and the fuel was
tested a very short time after
sampling, high H2S content could
be found. This was wrongly
ascribed to poor efficiency or insufficient dosage rates.
The consequences of this erroneous evaluation were magnified by
the corrective measures applied in
case of off-specs: a doubling of the
scavenger dosage rate until the H2S
content returned below the 2 mg/
kg level.
As a consequence of the particular operating conditions, the needs
of the refinery, the applied procedures and the absence of any
technical assistance from the H2S
scavenger supplier, the dosage
rates applied were high and outside
the standard.
When Chimec was asked to set
up an alternative treatment, we
performed, first of all, a laboratory
study to select the most suitable
product and the theoretical dosage
rate necessary to fulfill the specification. The tests were performed by
simulating plant operating conditions and carrying out a first
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 concentration in comparison to the previous

www.eptq.com

11/12/2013 12:39

JOHNSON SCREENS BRAND OF PRODUCTS

AND SOLUTIONS FOR THE HYDROCARBON
PROCESSING INDUSTRY

As the manufacturer of the Johnson Screens brand, Bilfinger Water Technologies provides:
A wide range of vessel internals in down flow processes to
retain the often costly media and to provide a collection area for
the process flow across the entire vessel diameter or length

Scallops to meet the various conditions of your process and to

ensure best flow distribution, optimum mechanical resistance
and long service life

All assemblies required in a radial flow reactor vessel, custom

built to satisfy every installations needs

Expertise and global field service from initial product design

through proprietary fabrication, inspection, testing, installation
and start-up

BILFINGER WATER TECHNOLOGIES

www.water.bilfinger.com
Australia - Asia Pacific
Phone +61 7 3867 5555
Fax +61 7 3265 2768
asiapacific.water@bilfinger.com

France
Phone +33 5 4902 1600
Fax +33 5 49021616
france.water@bilfinger.com

Japan
Phone +81.45.661.3575
Fax +81.45.661.1921
japan.water@bilfinger.com

North & South America

Phone +1 651 636 3900
Fax +1 651 638 3171
usa.water@bilfinger.com

PTQ_Issue Q1.indd
1
biflinger.indd
1

11/12/2013
15:22:43
11/12/2013
14:57

H2S
scavenger

VDU

= sampling point
Fluxant
30 minutes

St. mixer
HFO
4-6 Barg
130-140C

~4h

Figure 1 Heavy fuel oil rundown line

scavenger, but we could not verify

this. The following evidence, in any
case, showed us that most probably
this was not the case.
The next step was to verify the
laboratory results in the field
through
an
industrial
test.
Management of the scavenger
dosage rates appeared from the
beginning more difficult than
expected. The high variability of the
H2S content and the excessive
corrective measures the refinery
used to apply did not permit keeping the dosage rate at the intended
levels. In particular, dosage rate
management was unbalanced by a
fear
of
producing
off-specs,
regarded as a bigger threat by
comparison to the extra costs arising
from overdosage. Initially, it was
possible to apply an average dosage
ratio only 15% lower than the one
applied by the previous supplier.
Our approach was to put more
50

H2S content, mg/kg

45
40

focus and effort into standardisation

of the monitoring procedures and in
particular into the contact time
experienced by the fuel samples
before testing. More care was also
given to the time between a variation in H2S level and the follow-up
adjustment in dosage rate. Each
delay had a potential impact on
chemical consumption due to the
consequent
overdosages
and
off-specs that could result from it.
The better results obtained helped
also to increase the confidence of
refinery personnel in dealing with
H2S levels near the specification.
This allowed us to reduce the
corrective measures and therefore
provided further benefit through a
reduction in chemical consumption.
At the end of the first optimisation period, the average dosage rate
had been reduced by 35%
compared to the level applied by
the previous supplier. Still, we had

Measured (moving averages)

Predicted (moving averages)

35
30
25
20
15
10
5
0
10

11

12

01

02

Month
Figure 2 Prediction system results

66 PTQ Q1 2014

chimec.indd 3

03

04

05

not reached the target identified in

the laboratory study, and in our
opinion the response time to the
variations in H2S content, still too
high, was one of the main reasons.
In order to reduce this response
time, we needed to anticipate a
correction to the dosage rate before
the result of an H2S test was available. In other words, it was
necessary to develop a tool able to
predict with a certain accuracy the
vacuum residue H2S content from
the plant operating conditions.
First, we identified the process
variables that could have a major
impact on H2S content and, among
these variables, those that could be
more easily monitored. Then we
developed a model able to correlate
these variables with H2S concentration and optimised it using data
available from the plant.
It appeared from the beginning
that the prediction system was able
to give a good indication of the H2S
trend (see Figure 2). The model
enabled preliminary optimisation
of the dosage rate at the same time
as a variation in operating conditions (and hence H2S content)
occurred, so that only fine tuning
was necessary after identification of
the new H2S level through laboratory analysis.
The prediction system could not
be used for a full replacement of
laboratory analysis; this was not
the scope for which it was developed. An improvement in its
precision was later obtained
through the implementation of new
models and more variables, but still
its intrinsic error suggested that it
could only be used in combination
with laboratory analysis.
The prediction system was then
implemented in the form of a
simple mask, available in the
control room PCs, so that operators
and supervisors could be constantly
more involved. Implementation of
the prediction system enabled
further improvement in the average
dosage rate, reducing it to approximately 50% of the level applied
under the previous regime.
Figure 3 shows the trend of
dosage rate reported as parts per
million of additive for each part per
million of H2S. The explicit numbers

www.eptq.com

11/12/2013 12:39

and competitive configuration. potentially to the revamp of existby DN-3551. The points enclosed in customer conveyed the need to
Certainly, if theTreatment
objective is to Further
ing facilities.
optimisationThe zero gasoline
more oil sands-derived
the highlighted areasstart-up
labelled process
after software
maximise revenues with proven achieved
production
refinery is a practical
coker gas oil. Consequently, the
DN-3651 represent samples tested implementation
technology, in a reliable environ- and viable reality.
initiative
as part of scale-up and manufactur- rst dosage
rate: was to plan a transiment, the configuration that has Average
tionrelative
to toaour more metals-tolerant
ing optimisation. Use of high 37.5%
been developed by Foster Wheeler current start-up
system,
and
Ascent
throughput
experimentation catalyst
utilising its SDA + DCU combina- References
catalyst
was
selected
for
allowed rapid development of a 1DN-3551
Gillis D, Unique opportunities with proven
tion offers a competitive advantage.
Cycle 2. toThis
catalyst
has
been
reliable manufacturing process technologies
maximise residue conversion
tomargins,
have superior
feed poison
yielding the highest practical cata- &shown
refinery
Asia Technology
forum,
How to make the right decision
tolerance
and
was
considered
to be
lyst performance.
Bangkok, 10-11 Oct 2012.
The integration of refining and
accommodate
Applying technology improve- 2better
Edwardssuited
S M, et to
al, Relative
economics the
of
petrochemical
facilities
offers
C4s processing
routes,
PTQ, would
Q1 1998, 1.be
expected
poisons
that
ments such as DN-3651 requires a mixed
several opportunities for adding
holistic joint approach to ensure associated with processing increvalue to refinery streams by Time
that the full benets of the technol- mentally higher percentage of oil
exploiting synergies between differ- Blasis Stamateris is Downstream Business
ogy
can
be taken
advantage of by sands-derived coker gas oils.
Figure
3 Dosage
trend
ent
process
units
while rationalising
in the Business Solutions Group
The second
initiative included
the operator, taking into account Consultant
investments that can positively of Foster Wheeler, UK, participating in refinery
optimisation
of
the demet catalyst
the
site-speci
c
constraints
and
key to project
avoiding
unwanted off-specs configuration
With broad experience
in the studies
management
studies, feasibility
and
impact
economics.
package
to
enhance
catalyst
stabilobjectives.
and
unnecessary
extra
consumption
of
treatments
applied
to
bunker
fuel,
his
FEED
work
for
grassroots
refin
eries,
CTL
Different refinery configurations
ity.
Prior
toactivity
Cycle
1s Aflow
completion
In
this example, Criterion utilised facilities
of
chemicals.
main
areas
of
include
improvers
and
refinery
upgrades.
graduate
in
can be tailor-made to suit a range
and
the
DN-3551
loading,
we
a collaborative
approach
with
its chemical
for middle
distillates
andhas
residues.
holds
engineering,
he
over 25Heyears
achieve these
targets,
collaboof To
processing
objectives,
including
advised
the
customer
that
the
operacustomer
to
improve
catalyst
a
masters
degree
in
chemistry
from
Rome
ration between
and experience in oil refining and upgrading.
variations
to the customer
propylene-totion
was Email:
beingggrande@chimec.it
subjected to a higher
University.
performance
and facilitate
more Email:
blasis_stamateris@fwuk.fwc.com
supplier
is
fundamental
to
identify
ethylene ratio, and the production
Alessandra
Berra
is
Technical
Dan
Gillis
is
Director,
Refining
forcatalyst
Foster
feed
endpoint
and
theDevelopment
dif
cult
oil
sands-derived
feed
available
tools
ofcritical
middlevariables,
distillates and
aromatics.
ManagerUS.
with
Chimec
Fuel
Additives
Wheeler,
He
provides
technical
direction
system
was
at
risk
of
accelerated
processing
capabilities.
Three
initiaand options
areas forunits
improveCapacities
of and
the process
can and
Technological
Units.
She has overfor
15 years
applications
development
Fosterof
tives
were
implemented.
ment.
And
it
is
only
through
such aging due to feed resid and metals
be adjusted to accommodate differexperience
in
fuel
additives
development
and
Wheelers
proprietary
heavy
oil
technology.
He
Subsequent detailed
The originalthatcycle
utilised
collaboration
new 1ideas
and entrainment.
ent
crude oils.
in thegraduate
downstream
renewable
ismarketing
an engineering
of theand
University
component
analysis ofat
Criterions
Centinel
cata- feed
tailored
solutions
canDN-3100
come
fuel areas and
holds a masters degree in
Canada.
The concepts
outlined
are alive.
appli- Saskatchewan,
Research
facility
lyst as the primary conversion Criterions
chemistry
from Rome University.
dan_gillis@fwhou.fwc.com
Giacinto
is Product
Managerand
with Email:
cable
to Grande
grassroots
projects
con
rmed
elevated
feed
poisons
system.
During
the
cycle,
the
Chimec Fuel Additives Technological Units. Email: aberra@chimec.it
Ratio additive: H2S

confidentialtotal
information
the
ofaremagnitude
installedbutcost
tubular
ow
reactors
with
autotrend
may
be
easily
seen.
The
aver(TIC) estimates that have been
mated
process control
sampling.
age dosage
rate and
during
our
prepared,
based
on
Foster
The
use
of
this
multi-tube
reactor
management
was
reduced
by
37.5%
Wheelers in-house cost estimating
system
allows
signicant acceleraas a result
of implementation
of the
database.
tion
of
catalyst
development
relative
predictive
software.
As can be seen, the investments
to conventional
techniques.
result oftesting
the
good
for As
thea upgrading
section
ofresults
the
Leads
generated
with
theturned
high
achieved,
the
industrial
test
refinery configurations are of the
throughput
equipment
into a order
continuous
application With
towere
the
same
of magnitude.
con
rmed
by
conventional-scale
full
satisfaction
of
the
customer
and
similar product slates and investpilot
plant testing.
the supplier.
The model
algorithm
ments,
the differences
in economic
The
target performance
level and
for
and
parameters
are
controlled
indicators between the refinery
DN-3651
was
an
improvement
in
updated when will
necessary
by signifiChimec
configurations
not be
HDN
activity
of
at
least
10F
personnel,
maintain
effectivecant
enoughtoto
clearly the
favour
one
(5.5C)
relative
to DN-3551 (~20
ness
of
the
software.
configuration
over
another.
RVA) with equivalent or better
Therefore, when it comes to the
HDS
activity with a challenging
Conclusions
decision
of which refinery configudesign
feedstudy
containing
bitumenThe case
presented
here
ration
to choose,
other factors
derived
vacuum
gas
oil
and
heavy
S
scavenger
applishows
how
an
H
2
besides the product
slate and
coker
Figure 14 illustrates
cation gas
mayoil.
need
care to
investments,
suchparticular
as operating
the
highreal
throughput
experimentaprovide
cost
optimisation.
Not
complexity and associated reliabiltion
HDN and
HDS
activity simidata
all
refineries
have
conditions
ity of the units, should be taken
obtained
during
the prototype
lar consideration.
to those
discussed
and the
into
development
phase
as differ
well as
necessary
solutions
may
Ultimately, the selected option but
is
during
the
scale-up
and
commerit is the
that must
be the
based
onapproach
an individual
refiners
cialisation
phase.
The activity as
same in allThe
theSDA
cases.
objectives.
+ DCU options
presented
in
these
plotsis imporis the
Good additive
selection
represent
well-proven
technologies.
decrease
in
temperature
required
to
butis generally
sufficient.
Iftant,
there
a suitablenot
outlet
for theA
achieve
the
target
S
or
N
level
in
proper
treatment this
setupisand
coke
produced,
a tailored
viable
the
product
relative
to
that
required
monitoring procedures provide the

We keep a close eye on your plant efficiency.

The trouble-free operation of your plant is the precondition for maximum economic efficiency. Kuritas patented ACF technology inhibits the formation
of chloride and ammonium salts, thereby preventing fouling and corrosion. This is only one example of how Kurita protects your production unit against
unscheduled shutdowns, extends its runtime and lowers process costs while at the same time increasing plant and operational safety. Particularly
for refineries and petrochemical plants, Kurita offers innovative and sustainable process technologies as well as a broad portfolio of products for water
management. From consultation through to implementation and maintenance, we support you with a comprehensive range of services.
As you can see, we keep a close eye on the effcient operation of your plant.
Process technologies for refineries and petrochemical plants
Water treatment technologies for cooling water, boiler water, waste water and reverse osmosis

Kurita Europe GmbH . Industriering 43 . 41751 Viersen . Germany

Phone:+49 2162 9580110 . www.kurita.de

RZ_Anzeige_Quer_24052013_MQ.indd 1

www.eptq.com
www.eptq.com
www.eptq.com

chimec.indd 4
f wheeler.indd 8
cri.indd 8

24.05.13 15:24

PTQ Q1 2014 67
PTQ Q4 2013 27
PTQ Q3 2013 81

11/12/2013 12:39
12/09/2013 13:41
10/06/2013 16:38

Flare Netware

Onis lnc. | One Riverway | Suite 1700 | Houston | Texas 77056 | USA
1 (713) 840 6377 | sales@onislineblind.com | www.onislineblind.com

H2S Service

www.onislineblind.com

Representation in : Europe (Germany, Great Britain, Finland, Netherland, Spain)

Latin America (Argentina, Brazil, Mexico, Venezuela)
Asia and Middle East (China, India, Kuwait, Russia, Saudia Arabia,
Singapore, South Korea, Taiwan)
onis.indd 1

onis.indd 1

10/06/2013 13:10

11/12/2013 10:35

Overcoming high conductivity in process

condensate
Analysis of design and operations enabled a refiner to eliminate contaminants
from hydrogen plant process condensate by installing a high pressure stripper
TAPAS Kanti SAMADDAR and GREGORY M LILBURNE The Bahrain Petroleum Company BSC(c)
CALUM MCINTOSH and MILETA BABOVIC Johnson Matthey

n 2007 the Bahrain Petroleum

Companys (Bapco) Low Sulphur
Diesel
Production
(LSDP)
Complex came onstream. The
complex centres on a 60 000 b/d
heavy vacuum gas oil (HVGO)
hydrocracking unit, which is
supplied with hydrogen from the
No.2 Hydrogen Plant (2H2P). 2H2P
is a conventional unit, licensed
from Technip USA, consisting of
zinc oxide (ZnO) purification, a
top-fired Steam Methane Reformer
(SMR), High Temperature Shift
(HTS) conversion and a pressure
swing adsorption (PSA) unit. The
ZnO, SMR and HTS catalysts were
all supplied by Johnson Matthey.
The feed to 2H2P is sweet natural
gas and the plant design capacity is
100 million scfd of 99.9% purity
hydrogen. 2H2P is a large exporter
of 600 psig superheated steam to
the rest of the refinery.
After start-up, Bapco found that
the water in the refinery condensate
systems had high conductivity and
had to be dumped to sewer to
prevent it from entering the boilers
in the Bapco operated power plant.
The cause was ultimately attributed
to the presence of ammonia (NH3)
and methanol (MeOH) in the 2H2P
process condensate, which was
recycled within 2H2P for the generation of steam that was then
exported, condensed and recycled
to the power plant. The high
conductivity masked the presence
of any leaks from the seawater
cooling system into the refinery
condensate systems and was also a
threat to the integrity of these
systems due to corrosion. The 2H2P
process condensate, equivalent to
200 US gal/min, therefore had to

www.eptq.com

bapco.indd 1

be routed to the oily water sewer

(OWS). Consequently, the production of make-up water to the
refinery, the flash evaporated distillate (FED) from the desalination
units, had to be increased and a
significant financial penalty of the
order $100 000 per month was
incurred. Hence a permanent solution was required.
Bapco undertook a detailed
review and evaluation of different
options for reducing NH3 and
MeOH levels in the process
condensate. This also involved
discussions with several experts
including the catalyst vendor,

High conductivity
in the refinery
condensate return
systems was caused
by NH3 in the 2H2P
process condensate
process licensor, other hydrogen
plant operators and water treatment consultants. Ultimately the
company decided that the most
cost effective approach was to
install a high pressure (HP) Steam
Stripper.
The new stripper was designed
by Technip USA. Johnson Matthey
was again closely involved to
provide
catalyst
performance
predictions for the revamped
process flow scheme. The stripper
was successfully commissioned in
Q1 2011.
The process engineering and

catalytic issues associated with the

new HP Steam Stripper, as well as
the concerns, problems and limitations that arose during project
development and unit start-up, and
how they were overcome, are
presented in this article.

Process condensate flow scheme

(original design)

The original process flow scheme

before the installation of the new
HP Steam Stripper is shown in
Figure 1. About 200 US gal/min of
process condensate is produced in
2H2P during normal operation. The
condensate system design that
Bapco selected was the same as that
used in the older No.1 H2 Plant
(1H2P), which had never had any
problems.
In the 2H2P flow scheme, the
desulphurised natural gas is first
mixed with high pressure superheated steam and this combined
stream is further heated in the feedsteam superheat coil in the SMR
and flows into the top of the
reformer tubes. The reformer effluent is cooled to about 650F (343C)
in the process gas boiler and flows
into the HTS reactor. The HTS reactor effluent is then cooled to 110F
(43C), which also causes the steam
in the effluent to condense. The
process condensate is separated
from the hydrogen-rich process gas
in the PSA feed knock out (KO)
drum. The gas is routed to the PSA
unit for final purification.
In the original as-built unit
configuration, the process condensate from the KO drum was mixed
with FED make-up water and
routed to the deaerator for degassing, using low pressure (LP) 25

PTQ Q1 2014 69

11/12/2013 14:55

Vent

Deaerator
Fuel
gas

Steam
Methane
Reformer

Steam
Drum
LP
steam

Process Gas Boiler

Natural
gas

Zinc
Oxide

FI

H2 to
PSA

HTS

FV

PSA Feed
KO Drum

Export
Superheated
Steam

Oily water sewer

Flash
Evaporated
Distillate

Figure 1 Original process flow scheme

psig steam. The deaerator operating

pressure is 15 psig. The condensate/FED from the deaerator was
heated and recycled to the steam
drum for the generation of 600 psig
steam. The steam is superheated
and part of this stream is routed to
the SMR as feed process steam,
with the remainder being exported
to all parts of the refinery.
Condensate from the refinery
steam users is collected in the High
Grade (HG) and Low Grade (LG)
condensate systems and returned to
the power plant, where it is mixed
with FED and flows into the boilers
as boiler feed water (BFW).
Conductivity meters are installed
on the condensate return headers to
detect leaks from the seawater cooling systems into the steam/
condensate systems. When the
conductivity goes high, above
25 S, all the condensate is automatically dumped to the OWS to
ensure that any hardness salts do
not enter the boilers. They would
cause scale formation, corrosion
and hot spots. Dumping of conden-

70 PTQ Q1 2014

bapco.indd 2

sate is a very costly emergency

procedure because the FED production rate from refinery desalination
units
must
be
substantially
increased to meet the refinerys
FED and steam demand.

The problem: impurities in process

condensate

During commissioning of 2H2P, the

process condensate was initially
routed to sewer. However when
2H2P was brought fully onstream,
the condensate was routed to the
deaerator and then to the steam
drum, with some of the 2H2P superheated steam being exported into the
refinery steam system. After about 30
minutes, the refinery condensate
conductivity shot up from a typical
value of < 10 S to 140 S. As a
result, the condensate recovery
systems high conductivity alarms
were activated and all the condensate
was automatically dumped. The situation returned to normal when the
process condensate was again routed
to OWS in 2H2P.
Through a series of plant trials,

Bapco established that the high

conductivity in the refinery condensate return systems was caused by
the presence of NH3 in the 2H2P
process condensate, at typical levels
of 500 wppm, which had ended up
in the superheated steam system,
via the steam drum. In addition to
NH3, the condensate also contained
other dissolved gases like carbon
dioxide (CO2), carbon monoxide
(CO) and other by-products including MeOH and formic acid.
Bapco had previously experienced elevated concentrations of
NH3 in the process condensate in
the No.1 Hydrogen Plant (1H2P)
but this was a short term phenomenon attributed to high activity fresh
catalysts and it disappeared within
four to six weeks after start-up.
However on 2H2P, the problem
persisted and even after six months
of operation the NH3 had only
decreased to 300 wppm. The MeOH
concentration was 200 wppm.
During
this
period,
Bapco
worked very closely with Johnson
Matthey, the 2H2P catalyst vendor,
to identify and understand the
reaction mechanisms that were
occurring.
Johnson Matthey confirmed that
at the 2H2P operating conditions,
the concentrations of NH3 and
MeOH would persist throughout
the four-year operating cycle.

Ammonia and methanol formation

The NH3 and MeOH (and other

by-products) that ended up in the
process condensate are formed,
respectively, in the SMR and HTS
reactor.

Ammonia formation

The key reactions related to NH3

formation in the SMR are:
N2 + 3H2 2NH3 H -92 kJ/kg-mol N2

(1)

CH4 + H2O CO + 3H2 H +206 kJ/kg-mol CH4

(2)

The overall NH3 production is a

trade-off between several competing factors, as follows:
Higher feed N2 content increases
the NH3 (reaction #1). Bapcos
natural gas contains 11 mol% N2,

www.eptq.com

11/12/2013 14:55

which is unusually high and so

drives up the NH3 make compared
to most other natural gas fed H2
plants
Increasing the steam to carbon
(S/C) ratio reduces the N2 partial
pressure, which decreases NH3
levels
Higher operating pressure drives
reaction #1 equilibrium to the right,
hence more NH3 is produced
Operating temperature also has
an impact, but the situation is
complex, with competition between
equilibrium and kinetics. Higher
temperature and pressure increase
any reaction rate, but whether a
higher reformer outlet temperature
gives more or less ammonia is also
affected by a balance between reaction #1 equilibrium and the steam
reforming equilibrium (#2). A
higher temperature produces more
H2 from the steam reforming reaction, which pushes the NH3
equilibrium to the right, but it also
pushes reaction #1 equilibrium to
the left because it is exothermic.
Each case needs to be assessed
individually.
The difference in the NH3 levels
in the process condensate from
Bapcos 1H2P and 2H2P is due to
the relative operating conditions
prevailing at the outlet of the
reformer in each case. The impact
of N2 content and the SMR operating conditions is summarised in
Table 1.
The initial high activity of fresh
SMR catalyst means that NH3
production activity is also high.
Experience shows that, in practice,
the actual start-of-run (SOR) NH3
levels in the condensate are
around 30% of predicted equilibrium values, while at end-of-run
(EOR) they are about 10% of the
equilibrium predictions due to
lower catalyst activity. For 2H2P,
the SOR and EOR estimated NH3
levels were, respectively, 420
wppm and 140 wppm. The predictions are consistent with Bapcos
measured SOR concentrations of
about 500 wppm. For 1H2P, the
predicted NH3 levels at SOR and
EOR were, respectively, 140 and 47
wppm. The corresponding field
measurements were 45 and 8
wppm.

www.eptq.com

bapco.indd 3

Effect of N2 content and SMR operating conditions on NH3 formation

1H2P
Inlet S/C ratio
6
Reformer inlet N2, mol% wet
1.9
1480 (804)
Tube outlet temperature, F (C)
Tube outlet pressure, psig
384
Predicted equilibrium NH3 levels in process condensate, wppm
470

2H2P
3
3.2
1580 (860)
436
1400

Table 1

Since equilibrium levels of

ammonia are not attained in either
1H2P or 2H2P, the limiting factor
on NH3 formation is kinetics.
2H2P also operates with half the
S/C ratio of 1H2P, leading to
higher inlet N2 concentration, and
at a higher operating pressure and
tube exit temperature. All three
factors
drive
the
rate
of
NH3 formation, resulting in higher
NH3 levels in 2H2P compared to
1H2P.
The short term phenomenon of
elevated concentrations of NH3 in
the process condensate at SOR in
1H2P, and the subsequent drop
within four to six weeks after SOR,
is likely a result of the high S/C
and steam promoted passivation
of the catalyst towards the NH3
formation reaction over that
period. The high S/C ratio used in
1H2P reduces the partial pressures
of nitrogen and hydrogen, and
therefore reduces the potential for
ammonia formation. It should also
be noted that the two reformer
units are of different design, as
well as operating at different
conditions, which results in different
temperature
and
gas
composition profiles.
The NH3 dissociation reaction,
which is the reverse of the NH3
formation reaction, also influences
overall NH3 production:
2NH3 N2 + 3H2 H +46 kJ/kg-mol NH3

(3)

This reaction has been studied at

various temperatures, pressures
and with various catalyst types.1
As well as being the catalyst of
choice for reforming, nickel is also
effective in the cracking of NH3,
which is favoured by high temperatures and low pressures. In
this respect the temperature is
favourable in 2H2P, but the pressure is not. The reverse is the case
for 1H2P.

Formation of methanol and other

by-products2

Methanol is produced in the HTS

converter, and especially in the Low
Temperature Shift (LTS), which is
installed in 1H2P. It is formed
through copper promoted side reactions of H2 with carbon monoxide
(CO) and carbon dioxide (CO2). (In
addition to iron oxide and chromium oxide, the HTS catalyst
contains a small amount of copper
oxide.2 The LTS catalyst is copper
based.) The relevant reactions are:
CO + 2H2 CH3OH H -90 kJ/mol

(4)

CO2 + 3H2 CH3OH + H2O H -50 kJ/mol (5)

The high operating temperature

in the HTS converter kinetically
favours
MeOH
formation.
However, the high temperature
also drives the equilibrium for reactions #4 and #5 to the left i.e. lower
MeOH formation. The NH3 from

HTS: effect of gas composition and operating conditions on MeOH formation

S/dry gas ratio
Inlet concentrations CO/CO2/H2, mol% wet
Temperature, F (C)
Pressure, psig
Predicted equilibrium MeOH levels in process condensate, wppm

1H2P
1.062
3.9/5.9/35.3
630 (330)
380
66

2H2P
0.513
8.9/5.7/45.3
630 (330)
430
447

Table 2

PTQ Q1 2014

71

11/12/2013 14:55

Process condensate
Flow:
255 Std US gpm
Temperature: 445F
Estimated composition
H2O:
99.04 wt%
CO2:
0.80 wt%
NH3:
1260 ppmw
MeOH:
360 ppmw

Neutralising amine

Stripper overhead
Flow:
40.4 Mlb/hr
Temperature: 475F
Pressure:
525 psig
Estimated composition
H2O:
96.97 wt%
CO2:
2.53 wt%
NH3:
0.39 wt%
MeOH:
0.11 wt%

HP steam saturated
Flow:
45 Mlb/hr
Temperature: 490F
Pressure:
605 psig
Stripped condensate
Flow:
263 Std USgpm
Temperature: 475F
Pressure:
525 psig
Estimated composition
H2O:
100.0 wt%
CO2:
0.0 wt%
NH3:
<0.50 ppmw
MeOH:
<10.0 ppmw

Figure 2 HP Steam Stripper details

the SMR passes through the HTS

catalyst bed effectively unchanged.
There is also potential for MeOH
to react with NH3 to form amines
exhibiting the well known sh-like
odours. However, due to the
prevailing conditions in the bed,
monomethylamine formation does
not occur to a signicant extent.
Ethanol can form at about 10% of
the methanol concentration from
condensation-type
reactions
of
methanol. Formic acid and acetic
acid can also be formed over the
HTS bed, with typical levels of tens
of
wppm
in
the
process
condensate.3
The effect of feed composition
and operating conditions on MeOH
formation in the HTS converter is
shown in Table 2.
MeOH formation across the HTS
catalyst bed does not decrease
signicantly over the life of the catalyst. The dominant factors are the
operating conditions: temperature,
pressure and reactant concentrations. Although the model assumes
that an equilibrium level of MeOH
is generated across the converter,
some of the available evidence
shows MeOH levels in the condensate are less than equilibrium levels.

72 PTQ Q1 2014

bapco.indd 4

The actual MeOH concentrations

in the condensate that were measured in Bapcos 2H2P were about
200 wppm, as compared to the
predicted equilibrium value of 447
wppm.

Technology selection for ammonia

and methanol removal

The main objective was to select a

process for the removal of the
ammonia, methanol and other
impurities
from
the
process
condensate, to ensure that they did
not end up in the 2H2P export
steam to the rest of the renery.
Only proven technologies, which
could be practically and costeffectively integrated into the existing plant, with minimum downtime
and
minimum
environmental
impact, were considered.
Several technologies for removing
NH3 and MeOH from process
condensate are available, ranging
from steam and air stripping to
biological processes.4 Each has its
own advantages and limitations,
and their efcacy depends on the
system being treated and the
process requirements. There is no
common design solution. Bapco
also
reviewed
low
pressure

stripping and absorption processes,

as well as HP steam stripping,
although the experience of at least
one other rener showed that LP
stripping alone was not a viable
option.
Through discussions with other
hydrogen plant operators, technology licensors and water treatment
consultants, Bapco established that,
on large scale hydrogen plants, HP
steam stripping was typically used
for NH3 and MeOH removal. This
was a proven approach, which
would ensure that only clean
condensate would be recycled to
the steam drum for steam generation, and also allow the stripping
medium to be returned to the
process and minimise the loss of
water.
The process condensate is a
complex equilibrium system involving the ionic species such as NH4+,
CO32+ and OH- ions, as well as the
actual gas molecules. Fairly severe
operating conditions are needed to
maximise the concentrations of the
molecular form of the impurities,
for instance, NH3 and CO2, which
can then be stripped out of the
condensate.4 High temperatures
also reduce the solubility of these
gases and facilitate stripping.
The use of high pressure saturated stripping steam ensures high
enough operating temperatures to
achieve the required stripping
performance. In Bapcos case, the
operating pressure was selected so
that the overhead steam containing
the impurities could be routed
back to the process. The MeOH
would be destroyed in the
reformer, and the NH3 would be
recycled and reach an equilibrium
concentration.
A sketch of the new HP Steam
Stripper with the key operating
conditions is shown in Figure 2.
The column has two beds of
random packing and was installed
during the shutdown for turnaround and inspection in Q1 2011.
The details are as follows:
Diameter: 4ft 6in (1.4m)
Height: 72ft (22m) (tan-tan)
Materials of construction: carbon
steel with 3mm type 304L SS
cladding
Two packed beds: IMTP #40 type

www.eptq.com

11/12/2013 14:55

GE Works to redefine pump efficiencies.

In processing industries, GEs SPS pumping

systems work to increase efficiency, reduce
environmental impact and provide versatile
pumping solutions. For example, our multi-stage
centrifugal SPS pumps are a widely-used, efficient
solution for circulation, transfer boosting and
disposal applications. Our SPS systems provide
enhanced pumping efficiencies by lowering noise
and vibration levels, decreasing construction
lead times and delivering improved reliability and
extended runtimes. With GE, you have a worldwide
support system ready to provide engineering,
field service or sales support.
Take a look at our results, and see how
GE can work for you.
Call +1 713 260 6731 or e-mail sps.inquiry@ge.com

www.geoilandgas.com/sps

ge.indd
GE SPS 1Refining Ad_PTQ_1-18-13.indd 1

07/03/2013
13:37
1/22/13 9:50
AM

quest .indd 1

11/12/2013 09:36

304 SS random packing (0.25mm

thickness)
Bed height: 24ft (7.3m) each
Distributors: Inlet Model 116
Intalox Deck Gravity Distributor
and Redistributor Model 117.
The revamped process flow
scheme is shown in Figure 3. The
process condensate from the PSA
feed KO drum is now pumped
through the stripper feed-effluent
plate heat exchanger (PHE) into the
top of the column. As it flows
down through the packed beds, it
contacts saturated 600 psig steam,
which strips out the impurities. The
clean stripped condensate heats the
stripper feed in the feed-effluent
PHE, is combined with make-up
FED, and flows through the deaerator and then into the steam drum
for steam generation. The column
overhead steam, containing NH3,
MeOH, CO2, CO and other impurities, is routed to the reformer
process steam feed system, and the
combined stream ties into the sweet
natural gas feed line. The gas-steam
feed flows through the superheat
coil and into the reformer tubes.

Design issues

The key design issues were related

to the feed impurities levels, the
stripping performance efficiency,
the specification of the suction
conditions for the new process
condensate pump, and the new
stripper feed-effluent PHE.

Impurities and performance

specifications
Process condensate feed

It was vital to specify the correct

ammonia concentration in the
process condensate feed to the
stripper. Measurement of NH3
levels in the condensate being
routed to sewer was easy.
However, the stripper overhead
steam would be routed to the SMR
as feed and, although MeOH is
decomposed in the reformer, the
NH3 is essentially recycled, eventually reaching an equilibrium
concentration in the process
condensate. Modelling of this
system was therefore done by the
process licensor using the model
that had been set up originally for
the design of 2H2P and their

www.eptq.com

bapco.indd 5

Stripper design and actual impurity concentrations

Component1
Process condensate feed
Stripped condensate product

Design basis
Typical
Design basis
Actual
Ammonia, wppm
1260
1000 to 1400
0.5
<1.0

<3.0 guarantee
[Methanol + ethanol], wppm
360
100 to 300
<10
Not detectable
Note 1: Bapcos water treatment consultant advised that the CO2 and O2 should be, respectively, less than
0.5 wppm and 0.007 wppm at the deaerator outlet. Bapco adopted a conservative approach and used
these specifications for the stripped process condensate ex the HP stripper; that is, the inlet to the deaerator.
In normal operation, these parameters are not measured.

Table 3

proprietary software. The predicted

equilibrium concentration of NH3
in the condensate was 1260 wppm,
which was significantly higher
than the measured concentration in
the original flow scheme.
The methanol concentration of
360 wppm was also estimated by
the licensor using their thermodynamic equilibrium model. The
actual maximum concentration

The concentrations
of ammonia and
methanol in the
stripped condensate
were equally
important
measured during normal operation
was 335 wppm.
Stripped condensate product

The concentrations of ammonia and

methanol in the stripped condensate were equally important.
However, there is no generally
accepted specification for either
component in FED (BFW) or steam.
Bapco canvassed several water
treatment specialists and other
refiners, and decided on a specification of <1 wppm NH3 for the 2H2P
export steam. This limit was set to
avoid yellow-metal corrosion in the
steam and steam condensate
circuits.
The licensor would not accept
this specification of <1.0 wppm
NH3 for either the export steam or
the process condensate. However, it
was eventually agreed that the

stripper would be designed to

achieve <0.5 wppm NH3 in the
stripped condensate, although the
process guarantee would be for
<3 wppm. This was acceptable to
Bapco, since it meant that after
dilution with the FED make-up
in the deaerator, the specification
of <1.0 wppm for the export
steam would
be
met.
The
actual performance in 2011/2012
was consistently less than 1.0
wppm NH3 in the stripped
condensate.
The major concern with the methanol was that it could form organic
acids, which would require higher
dosage rates of sodium phosphate
for neutralisation. As a result, the
foaming tendency in the steam
drum would increase, causing
carryover of sodium salts to
the steam superheat coils and
steam turbines, which was not
acceptable.
Based on feedback from water
treatment
specialists,
Bapco
concluded that MeOH levels of <10
wppm in the stripped condensate
would not cause significant problems. The licensor advised that this
limit was achievable in practice and
that it was generally used in their
designs. However, they would not
provide a guarantee of performance. Bapcos water treatment
consultant also confirmed that <10
wppm was typical of actual industry
performance,
and
Bapco
specified this as the basis for
design.
Lower MeOH levels could also
have been achieved, but only if a
low pressure stripper were to be
installed downstream of the HP
Steam Stripper. However, in this
case, there was no operational

PTQ Q1 2014

75

11/12/2013 14:55

justification for the additional

investment.
Since start-up, little or no methanol has been detected in the
stripped condensate. The stripper
design impurities levels and actual
performance are summarised in
Table 3.

Condensate pump net positive

suction head available (NPSHA)

Specification of the NPSHA for the

new process condensate pump was
a serious challenge. The condensate
in the PSA feed KO drum contains
0.5 mol% dissolved gases. A decision had to be made on how much
free vapour, if any, would be
allowed at the inlet to the centrifugal pump. Too much vapour would
lead to cavitation and mechanical
damage or loss of pumping capability. Too little or no vapour could
lead to a very low NPSHA specification and a costly pump, with
special design features to obtain the
extremely low NPSH required like
lower speed, double suction inlet,
larger impeller inlet (eye) area and
a flow inducer.5,6
For example, zero vapour would
require the specification of an
NPSHA of just 3.7ft (1.13m). In
the unrealistic case of pure water
with no dissolved gas, the liquid
vapour pressure would be the
KO drum operating pressure and
the NPSHA would be 903ft (275m).
After considerable review and
evaluation, the effective vapour
pressure was selected to allow
2 vol% vapour at the pump suction,
which would have no adverse
effects on the pump i.e. no vapour
lock or cavitation.5,6 The corresponding
NPSHA
that
was
specified was 402ft (122.5m), and
the actual NPSHR of the new pump
was 7ft (2.13m). The pump has now
been in continuous trouble-free
service for 18 months.

Stripper feed-effluent plate heat

exchanger (PHE)

To improve efficiency and minimise

cost and energy consumption, a
stripper feed-effluent heat exchanger
was used in the overall system
design. However, this presented
another unique challenge because
the approach temperatures between

76 PTQ Q1 2014

bapco.indd 6

the process condensate feed and the

stripped condensate effluent were
very low, of the order of 30F (17C).
Rigorous in-house simulations
showed that if conventional shell
and tube heat exchangers (SHE)
were to be used, 10 single pass
shells or four F-type shells would be
required. This approach would have
been very costly, resulted in high
pressure drops, and required a more
complex piping arrangement and a
significantly larger equipment footprint area. This conclusion was also
confirmed by the licensor. The solution was to use a PHE.
Although SHEs are widely used
in petroleum refineries, PHEs are
less popular. Bapco is no exception,
although a few PHEs are used in
low pressure, non-hydrocarbon
applications. The refiners main
concerns were reliability, maintenance and plate sealing issues at

Accurate and reliable

measurement of the
total steam flow and
the feed gas flow to
the reformer is vital
for control of the
S/C ratio
high pressure. However, the licensor recommended the use of a
single PHE with an all-welded
plate construction, which would
address these concerns. Such equipment had been successfully used in
similar services on other hydrogen
plants, was considered to be reliable, and would have an operating
cycle of several years. This
approach would minimise investment and maintenance costs, and
simplify the piping configuration.
Moreover, the footprint area would
be 85 to 95% less than the SHE
arrangements. The use of a PHE
would therefore facilitate installation and integration into the
existing flow scheme.
Bapco accepted the licensors
recommendation and installed a
single PHE. Given the advice on

reliability and experience of others,

there was no justification for a spare
exchanger. The equipment operated
without any problems for almost 18
months, but then leaks developed.
Unfortunately, the PHE could not
be inspected or repaired locally and
had to be returned to the overseas
manufacturer. Consequently, this
forced a long term shutdown of the
stripper and the costly routing of
the process condensate to OWS
again. The reason for the leaks is
not yet known.

Reformer steam flow measurement

Accurate and reliable measurement

of the total steam flow and the feed
gas flow to the reformer is vital for
control of the S/C ratio and to
ensure that it does not fall below
3.0, which is the minimum safe
operating limit for 2H2P.
As Figure 1 shows, in the original
configuration the flow meter (FI) on
the reformer process steam feed
line was upstream of the flow
control valve (FV). Hence, the pressure at the flow orifice was constant
and did not need pressure or
temperature compensation.
However, as Figure 3 shows, in
the new flow scheme there would
be no flow or pressure control on
the overhead steam from the stripper, and the column pressure
would float on the reformer system
back pressure. Since the stripper
overhead steam line would tie into
the fresh process steam feed
system, the total combined steam
flow had to be measured and
controlled. Therefore, the locations
of the existing steam FV and FI
were swapped. The stripper overhead product steam now enters the
system between the FV and the FI,
and the total steam flow is essentially controlled by the FV, which
regulates the flow of the fresh
steam feed.
As a result of this control scheme,
the operating conditions at the
steam FI varied, and pressure and
temperature compensation of the
flow measurement was required to
obtain an accurate total steam flow
rate and steam to carbon ratio.

SMR furnace duty and temperatures

As

was

described

above,

the

www.eptq.com

11/12/2013 14:55

Vent

Deaerator
Fuel
gas

Steam
Methane
Reformer

Steam
drum
LP
steam

High Pressure
Stripper

Process Gas Boiler

Natural
gas

Zinc
Oxide

Stripping steam

FI
FV

H2 to
PSA

HTS

Plate Heat
Exchanger
Export
Superheated
Steam

PSA Feed
KO Drum
New equipment

Flash
Evaporated
Distillate

Figure 3 Revamped process flow scheme

stripper overhead steam ties into

the fresh process feed steam
system. As a result, the combined
process
steam
temperature
decreased after start-up, which
reduced the temperature at the
inlet to the gas feed-steam superheat coil and therefore increased
the firing duty. However, the steam
superheat coil load also decreased
because a portion of the HP saturated steam leaving the steam
drum is diverted to the stripper as
stripping steam, thereby reducing
the total steam for superheating.
Overall, there was a marginal
increase in reformer absorbed
heat duty and reformer firing,
the main effect being an increase in
the cross-over temperature, which
is a critical process variable trip
setting.

Start-up operating issues

Ammonium bicarbonate
salt formation

The formation and deposition of

ammonium bicarbonate salts in the

www.eptq.com

bapco.indd 7

stripper overhead system was a

serious and unexpected development, which caused operational
and control problems.
The ammonium bicarbonate salts
formed from the reaction of the
NH3 and CO2 in the overhead
steam at temperatures below 170F
(77C). This meant that they
precipitated out in the instrument
impulse lines, causing blockages
and erroneous readings. The salts
also blocked impulse lines on the
emergency shutdown (ESD) flow
transmitters on the reformer inlet
process steam feed flow line,
resulting in unreliable readings
and a high risk of a plant
shutdown.
There were several unsuccessful
attempts to prevent salt formation
by insulating the instrument
impulse lines and the associated
instrument condensate pots. The
problem was eventually solved by
removing the condensate pots and
installing electrical heat tracing on
the impulse lines from the take-off

point on the process line to each

instrument. After these modifications, the problems of ammonium
bicarbonate
salt
deposition
disappeared.

Use of neutralising amine

In the original plant configuration,

neutralising amine was used in the
deaerator. Amine injection was also
required for the new HP Steam
Stripper to achieve a pH of 8.5 to 9
in the stripped process condensate
and prevent corrosion in the downstream carbon steel piping.
Initially, it was intended to use
the same amine for both services.
However, the amine used in
the deaerator needs to be more
volatile than that required for
the stripper, to enable it to partition
more towards the vapour phase to
maintain the pH of the condensed
steam. The stripper amine must be
less volatile to stay in the liquid
phase to maintain the pH of the
stripper process condensate.
Therefore, a new amine storage

PTQ Q1 2014

77

11/12/2013 14:56

tank and two high pressure

injection pumps were installed. The
amine is injected into the stripper
bottom head for better mixing
rather than into the bottom outlet
line. Directionally, this may also
elevate the pH of the steam-water
system inside the column, which
would shift the equilibrium in
favour of NH3, which can then be
stripped out.
The amine injection system works
very well and there have been
no pH control or corrosion
problems.

Lessons learned and concluding

remarks

Bapcos HP Steam Stripper has

been very successful in operation.
The problems of high conductivity
condensate have been eradicated,
ensuring that clean HP steam flows
to the superheat coils and steam
turbines, and saving 200 USgal/
min of condensate from going to
drain. The financial benefit is about
$100 000 per month. Several process
design and operating challenges
had to be overcome to achieve this
success, out of which came the
following key learning points
related to the original hydrogen
plant design and the stripper
project itself:
In the design of a new H2 plant
or any revamp facilities, do not
forget the utilities. Always verify
the quality of the process condensate and evaluate the potential
impact on the utilities systems. The
onus is on the client to decide on
the need or otherwise for process
condensate purification, but the
question must also be asked of the
licensor
Work closely with the catalyst
vendor to identify and determine
the nature of the side reactions and
the concentrations of impurities,
which may form and accumulate in
the process condensate and other
systems
Having a technically strong and
experienced
in-house
process
engineering, mechanical and metallurgical engineering capability was
extremely beneficial, especially
given the uncertainties that arose.
These subject matter experts had an
intimate knowledge of 2H2P,

78 PTQ Q1 2014

bapco.indd 8

including the original design basis

and the catalytic processes, as well
as the operation, control and
constraints, and ensured that
Bapcos requirements were properly met
It was vital to involve the right
people in the evaluation and
decision making process. Diagnosis
of the problem and establishing the
product
purity
specifications
required input from Johnson
Matthey, the catalyst vendor,
water treatment specialists, process
licensors and hydrogen plant
operators
Throughout the project, it was
important to question and analyse
everything, even though this sometimes frustrated the licensor/
contractor. Bapco had to be
convinced, and able to minimise
the uncertainty and risk. Good
licensors/contractors understand,
respect and even welcome this
approach

In the design of a
new H2 plant or any
revamp facilities,
do not forget the
utilities
Bapco made time for in-depth
PFD and P&ID reviews with the
licensor and the EPC contractor,
which paid dividends in highlighting problems such as increasing the
stripper T/T length to ensure sufficient boot space for the packing
hold-up liquid, and installing a trip
on the stripping steam inlet to
prevent liquid carryover in the
event of a reformer trip
Despite Bapcos attention to

detail, there were some surprises
that needed to be addressed
quickly and efficiently, including
the need for flow compensation on
the reformer steam flow and the
ammonium
bicarbonate
salt
formation.
As a result of these actions, the
HP Steam Stripper was successfully
commissioned and all specifications
are being met.

References
1 Hacker V, Kordesch K, Ammonia Crackers,
Handbook of Fuel Cells Fundamentals,
Technology and Applications, 2003, ISBN
0-417-49926-9, Vol 3, Pt 2, 121-127.
2 Pach J (Johnson Matthey Catalysts), Causes
and consequences of by-product formation in
ammonia plants, 54th Symposium on Safety in
Ammonia Plants and Related Facilities, AIChE,
2009, Vol 50, 263-274.
3 Madsen J (Haldor Topse), Process
condensate purification in ammonia plants,
Symposium on Safety in Ammonia Plants and
Related Facilities, AIChE, 1991, Vol 31, 227-240.
4 Odenbrand I (Lund University) and Anderson
K H (Haldor Topse), Investigation of process
condensate treatment of a methane steam
reforming plant, Department of Chemical
Engineering, Lund University, 29 May 2009,
www.chemeng.lth.se/ket050/Finalreport2009/
Topso.pdf
5 Tsai M J, Accounting for dissolved gases in
pump design, Chemical Engineering, Vol 89, No
7, 26 Jul 1982, 65-69.
6 Chen C C, Optimal system design requires
the right vapour pressure heres how to
calculate it, Chemical Engineering, Vol 100, No.
10, 1 Oct 1993, 106-112.

Tapas Kanti Samaddar is the Process Specialist

Lubes in Technical Services with Bapco,
Bahrain. He has 20 years of experience, mainly
in technical support for petroleum refining,
specialising in hydrocracking, lube base oil and
fluid catalytic cracking, and holds a Bachelor
of Technology degree in electrochemical
engineering from Madurai Kamaraj University,
India.
Gregory M Lilburne is the Manager, Technical
Services with Bapco, Bahrain. He has 33 years
of experience, mainly in technical support for
petroleum refining, holds BE and PhD degrees
in chemical engineering from the University
of Melbourne, Australia, and is a Fellow of the
IChemE and a Chartered Engineer (UK).
Calum McIntosh is Technical Sales Manager,
Middle East Region with Johnson Matthey. He
joined Johnson Matthey in 2002 on acquisition
of the ICI Catalyst Business, holds BSc and
MPhil degrees in chemistry from the University
of Heriot Watt in Edinburgh, Scotland, and is
co-author of several patents and academic
papers in the field of catalysis.
Mileta Babovic is a Senior Process Engineer
with Johnson Matthey. He joined Johnson
Matthey in 2007 after several years in
academia and related R&D industry, focusing
on heterogeneous catalysis and process
intensification. He holds a MSc degree in
chemical engineering from the University of
Belgrade and a PhD from Newcastle University,
UK, and is co-author of several patents that
have seen successful industrial application.

www.eptq.com

11/12/2013 14:56

TM Belco Technologies Corporation

Reducing refinery emissions has never been so easy.

EDV Technology provides refiners with an economic and reliable means for controlling
emissions of pollutants (PM, PM10, SOx, HAPs, NOx ) from FCCUs and other stationary sources.
Write, call or fax Belco Technologies Corporation.

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.indd 1
Belco_Ad_7_2010_210x297mm.indd
1

06/06/2013
9/29/10
4:58:2710:16
PM

ITW

Innovative
Technologies
Worldwide

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

Now hiring
Professionals
Worldwide

Join ITW Team worldwide and send your Curriculum Vitae to : jobs@itwtechnologies.com

itw.indd 1

11/9/12 11:45:38

Mercury treatment options for natural

gas plants
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

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

www.eptq.com

uop.indd 1

60
50

Log K equilibrium

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 throughout gas plant assets, making it
difficult 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 practicable is the best choice.

40
2CuO + 2H2S + Hg = HgS + Cu2S + 2H2O
Cu2(OH)2CO3 + 2H2S + Hg = HgS + Cu2S + 2H2O + CO2
2CuS + Hg = HgS + Cu2S
Cu2S + Hg = HgS + 2Cu

30
20
10
0
10

10

20

30

40

50

60

70

80

90

100

Temperature, C
Figure 1 Thermodynamic driving force favours cupric sulphide over cuprous sulphide for
reaction with mercury

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

The common and

traditional approach
to mercury removal
has been the use of
sulphur impregnated
carbon beds
capillary condensation of water and
heavier hydrocarbons in the micropores of the carbon sub-structure.
Sulphur impregnated carbon products 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 fixed bed, metallic based
MRU products find greatest success
in their ability to treat wet gas
streams in up-front positions.
UOPs non-regenerative metal
sulphides (UOP GB series of products) are a set of versatile
non-regenerative mercury removal
adsorbents. They are operationally
flexible 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 containing thousands of micrograms or
parts per billion levels of mercury
can be treated to extremely low
effluent levels using GB adsorbents.
These high capacity mercury adsorbents are engineered using a
copper based active component
finely dispersed across an alumina
substrate. This high capacity leads

PTQ Q1 2014 81

11/12/2013 12:41

GB adsorbent

Mesopores enable rapid

mercury scavenging

Macropores are preferable for

handling liquid condensation and
minimising mass transfer limitations

10

100

1000

10000

100000

Pore width,
Figure 2 Pore features of UOP GB products

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 flexibility, 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 maximising mercury adsorption as these
beds are designed to operate effectively at short contact times and are
exposed to hydrocarbons close to
their dew point. Therefore it is critical to specifically engineer the pore
distribution of the adsorbent. As
Figure 2 shows, an engineered mix
of
microand
mesoporous

structures ensures rapid mercury

adsorption with the ability to withstand liquid condensation and
minimise mass transfer limitations.

Regenerative adsorbent for mercury

removal
UOP HgSIV adsorbents are regenerative molecular sieve products
that contain silver on the outside
surface of the molecular sieve pellet
or bead. Mercury from the process
fluid (either gas or liquid) amalgamates with the silver and a
mercury-free dry process fluid is
obtained at the bed outlet. An additional layer of HgSIV adsorbent, to
an existing dryer configuration,
results in removal of design water
load and the mercury from feed
gas. Mercury and water are both
regenerated from the HgSIV adsorbents using conventional gas dryer
techniques where the mercurysilver amalgam on the molecular
sieve is regenerated thermally. The

Raw
natural gas

Feed gas
separator
UOP MOLSIV
adsorbents

UOP GB MRU

CO2 removal

Dryers

Figure 3 MRU upstream of the acid gas removal section and dehydrators

82 PTQ Q1 2014

uop.indd 2

mercury desorption profile from

the HgSIV adsorbent is similar to a
typical water regeneration profile,
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 adsorbents 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 pressure 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-regenerative 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 significantly reduces

www.eptq.com

11/12/2013 12:41

Customer:

Challenge:

Result:

Ethylene plant, Saudi Arabia.

Overhaul and repair four critical steam turbines in 12 days.

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?

C O M P R E S S O R S

T U R B I N E S

G L O B A L

S E R V I C E

The world turns to Elliott.

www.elliott-turbo.com

elliott.indd 1

10/12/2013 09:53

Raw
natural gas

Feed gas
separator
UOP MOLSIV
adsorbents

CO2 removal

Dryers

UOP GB MRU

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

Figure 4 MRU downstream of the acid gas removal section and dehydrators

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 condensate 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 downstream of the dryers (see Figure 4).
Placement of the MRU downstream of the dehydrator might not
appear to be a logical choice for a

user. However, when plant operations decides to use sulphur

impregnated carbon as the adsorbent for mercury removal from gas,
the MRU location becomes critical.
Sulphur impregnated carbon based

For a regenerative
mercury removal
option, HgSIV
adsorbent can
be loaded with
the dehydration
molecular sieve
mercury adsorbents are highly
sensitive to moisture and thus need
dry gas to effectively remove

Raw
natural gas

Feed gas
separator

UOP MOLSIV
adsorbent

UOP HgSIV
adsorbent
CO2 removal

Dryers

Figure 5 Regenerative mercury removal along with dehydration

84 PTQ Q1 2014

uop.indd 3

For a regenerative mercury removal

option, HgSIV adsorbent can be
loaded with the dehydration molecular sieve. This helps protect the
brazed aluminum cold box in the
cryogenic section. HgSIV adsorbents can be used in natural gas
plants to meet the processing
requirements of a mercury content
of less than 10ng/Nm3 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

www.eptq.com

11/12/2013 12:41

DHP units.
holds a degree
in chemical
The new
technology
is pro
table
It is in Europe where the and
rener
and He
particularly
the
ability
being
cooled
and example
passed
through
shows
a simplified
of how
engineering
from
Middle
East inTechnical
Maximum cut S for piping condition-based approach
taking
into
account
todays
US
difference
between
re
nery
yield
to
anticipate
market
needs
differH
from
2
a separator,
is sent
through a
the
equipment
condition-based
reduction
Dedicated
University,
Turkey,
and is a certified
Energy
market
prices
even
if
the
middle
structure
and
market
demand
is
ent
regions
as
constraints
evolve.
small GB might
adsorbent
bed. to
Since
approach
be applied
set
chlorine
trap
Supervisor
for
industrial
plants.
especially since convendistillate pricegasis flow
lower
than the critical,
regeneration
is relatively

Max allowable predicted CR
Maximum cut S
UOPosman.karan@tupras.com.tr
GB
limits
based on piping condition.
In
PolyFuel
andMRU
PolyNaphtha
are trademarks of
Email:
tional
re
ning tools
do not
have the
gasoline
price.
Indeed,
as
a
result
of
Line
Material
Op temp,
To next T/A, To 2nd T/A, To next T/A, To 2nd T/A,
small
compared
to
the
feed
gas,
Axens. Asim Ay is CCR/NHT/ISOM Units
this example, the last
inspection
Mehmet
Raw
Reduction

F
mpy
mpy
wt%
wt%
exibility
to reduce excess
gasoline
shaleGB
gas adsorbent
production,requirement
LPG prices are
To booster
the
isRecycle
gas gas
Process Superintendent
with
Kirikkale
natural
was
done
March 2011;
the new
chamber
6
Heavy dieseland to
5Cr increase
625
29
11
>5 Tpras Manager
2.1
H2 toin LPG
compressor
and/or
C
)
cut
Marielle
Gagnire
is
Technology
production
the
low.
Adding
(C
relatively
small
when
compared
to
3
4
refinery.
He
holds
a
degree
in
chemical
reduction
12
Atmospheric
resid
9Cr
690
33
12
>8
4.4
crude slate is expected to start in
for hydroprocessing and6.1
olefins technologies
in a PolyFuel
unit lowersfeed
feedstock
of middle
distillates.
the
requirement
gas amount
8 Vacuum resid
9Cr
675
12
engineering 4from Middle
East Technical
Separator
2013;
and there are for
planned turnadownstream
FCC,
especially
oligomerisation
Reaction
costs
and
contributes
to
increased
Moreover,
with
European
re
neries
drum gas
Feed
University,
Turkey.
adsorbent.
rounds in 2016 and 2021.
and etherification technologies,
in Axens
UOP MOLSIV
separator
protability,
while
maximisingsection
facing increasing
difculty in nd- Email:
MehmetAsim.Ay@tupras.com.tr
The first step using this approach Table 5
adsorbents
Marketing,
Technology
and
Technical
middle distillates
production in the ing export markets for their excess Koray Kahraman is CCR/NHT/ISOM
Units
Natural
gas liquids
is
to estimate
the treatment
condition
of the
Assistance Department. SheUOP
is an HgSIV
engineering
Catalyst
stream
Process
Chief
Engineer
with
Tpra
s
Kirikkale
re
nery.
gasoline
and
given
the
tensions
in
C4, C4at
+ and
some
instances
Process
streameven
graduate
from the Ecole
Nationale
Suprieure
predicted
corrosion
rates,
the
planned maintenance shutdown. In than
pipe
the in
time
when
processing
adsorbent
experience
Tostreams
reach 15%
IRRthefor cryogenic
PolyFuel middle distillate supply, PolyFuel refinery. His six years of refinery
from
C
de Chimie de Paris,
and holds a post-graduate
allowable
predicted
the
new crude is expected to this example, the next two turna- maximum
3
includes
the
process
side
of
hydrocracker
with
prices
based
in 2012gas
in the
US should full a primordial role in engineering degree from the IFP School. and
section
of after
a adsorber
natural
plant
Figure
3 PFD
drum departrates plants,
are then
rounds are being considered. Using corrosion
commence.
The
inspection
production
sulphurreduced
recovery,
Gulf
Coast,
the middle distillate adjusting the gasoline-distillateCO2hydrogen
removal
Annick
Pucci is Dryers
Deputy
Product the
Line Manager
exhibitshould
mercury
based
ment
be able contamination.
to provide the this approach, the maximum appropriately
NHT, ISOM, CCR and
DHP on
units. He uncerholds a
price
can be $96/t
lower thancan
gasoproduction to better t market in the field of light ends hydrotreatment
and
This mercury
be acceptable
measurements
are
takenfrom
at the
year without
a major rate
interruption,
inchemical
the corrosion
model.
corrosion
is that tainty
measured
pipecontamination
thickness
degree in
engineering
from Middle
line.
If
the
middlewhere
distillate
price Figure
demand.
a mercury
specialist removal
in refining olefins processing,
6
Configuration
showing
regeneration
gas
high
in
situations
upstream
outlet
stream
of
this
adsorber
and
indicating
that
the
vibration
probEast
Technical
University,
Turkey.
The
maximum
allowable
most recent inspection as well as which would put the pipe at its
particularly for FCC effluent upgrading.
were
equal
to theis gasoline
price
In other regions, new tendencies Email:
mercury
treatment
either
ineffeckoray.kahraman@tupras.com.tr
the
results
are 0 ppm
HCl,Based
whereas
lem was correctly
identified.
corrosion
rates
from
thickness
just as This
the predicted
historical
corrosion
rates.
on retirement
She holds
a bachelors
degree
in Marketing
chemical
($1129/t)
and
LPG priceMercury
kept at mercury
such as shale
oil and options
shale gas and
are Satyam
Mishra
is Axens
GlobalTechnology
Strategic
treatment
tive last
or concentration
not thepresent.
Arnaud Selmen
is
Manager
inlet
case shows
that a goodis
licensor
5
can
then
be
used
to
estiturnaround
starting. If Table
the
known
conditionaverages
of the planned
engineering
Ecole
Nationaleproduct
Suprieure
Manager
forfrom
UOPs
Adsorbents
line
$636/t,
thefrom
IRR these
would reachstreams
28%.
revolutionising The
the decision
US market,
for Naphtha
Hydrotreatment
and Reforming
arrangements.
treatment
30
ppmthe
HCl.
refinery
relationship
essential
mate
a
maximum
percentage
of
piping
only has isto
make it for
to in
pipe,
time andliquid
the historic the
des
Industries
Chimiques
de
Nancy,
France.
Honeywell He
UOPs
Dubai
office.
In with
this
providing
additional
light
products
Technologies.
has
worked
mainly
using
one
orturnaround,
morethat
of these
is done principally
thetosame
way the
solving
problems
require
both role,
sulphur
for
each
stream
similar
to
next
a schemes
higher
corrosion
rate, it is in
easy
estimate
Emilie
Rousseau
is
a
Strategic
Marketing
Satyam
is responsible
for coordinating
Conclusion
and
consequently
inuencing
bottom-of
the-barrel
technologies,
specialising
would
generally
depend
the what
as for
feed gas
treatment.
Options
Conclusion
technological
operational
done
in the first
example.
rate and
would
beonmore
the
thickness
of the
pipe at the
time corrosion
Engineerwas
in Axens
Marketing
Department.
She
technology
development
and
promotion
With
the world removal
market forare
middle
market balance
prices.requireToday in heavy crude oil upgrading. He has also been
operational
andand
for mercury
the acceptable
Currently,
the H2-rich
compresexperience. than
holds
a chemical
engineering
degree
from
the
In
this
case,
the
9Cr
vacuum
resid
ifcapital
it cannot
be and
when
processing
the gas
new
crude
implementation
of
the
UOP
involved
in
the process
design
ofAdsorbents
aromatics
distillates
growing
and
a
reduced
in
the
US,
as
a
result
of
the
impact
ments of until
the user.
However,
for portfolio
non-regenerative
or regenerative
Ecole Nationale
Suprieure
des petrochemical
Ingnieurs
en
sors
are running
smoothly
without
piping
would
not as
make
it through
some
turnaround
commences.
The inspection
depart- replaced
to
the natural
gas
and
Osman
Kubilay
Karan
is
the
Hydroprocessing
complexes
and
NHT,
well
as
reforming
demand
for
gasoline
infor certain
of shale gas
on the
costand
of enviLPG, Arts Chimiques et Technologiques de Toulouse,
addressing
health,
safety
adsorbent
beds
designed
liquid
any
problem.
The
compressors
Satyam
holds
a bachelors
in
a second
turnaround
cycledegree
downSuperintendent
the road. Inwith
this Tpra
case, industry.
ment can also supply the retirement further
Units
Process
startup
and
troubleshooting.
He even
holds
regions,
the new process for olenic ronmental
PolyFuel
is already
pro
table
for as units
a masters
in chemical
engineering
from
concerns,
the
treatment.
from
Indias
Maharaja
were
recently
opened
by the it
with
theengineering
current
feedstock.
By
planhas been
assumed
predicted
thickness
for the pipe,
allowing
Kirikkale
refinery.
His 25 that
years
ofscheme
refinery chemical
an engineering
degree
from
the
ENSGTI
cut.
gasoline oligomerisation allows the involving
mixed feed of upstream
LPG and C5/C
Imperial College
in London
a masters
in
mercury
6
of
been
mechanical
maintenance
group and
experience includes
the operational
and Sayajirao
ning
toUniversity
replace
ita and
in and
the
next
rates using
the modified
remaining corrosion
allowance
to corrosion
engineering
school
andBaroda
DEA
inhas
refinery
The

exibility
of
the
new
process
re
nery
scheme
to
be
adapted
to
a
energy economics
andand
corporate
management
involved
with
the
oil
gas
industry
for
the
Summary
removal
offers
maximum
process
sides
of
crude,
vacuum
units,
no
green
oil
formation
was
found,
process
modelling
from
IFP
School.
McConomy curves are +/- 50%. To turnaround, and the heavy diesel
be calculated.
from IFPyears.
School.
many
advantages
to CCR
the Email:
maximum
distillate
mode. reliable offers
UOP
offers
various
insurance.
hydrocracker,
hydrogen
production plants,
although
theypiping
ran
forreplacements
almost one decrease
and15Arnaud.SELMEN@axens.net
atmospheric resid piping at
the likelihood
of an last
Normally,
would be done during a regularly unplanned failure due to higher the following turnaround, the

Get Real

SyngaS, Lng and Carbon Capture

Dont Guess!

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.

ProTreat Simulation Software

Taking the Guesswork Out of CO 2 Removal

Learn how state-of-the-art simulation science can benefit you.

www.ogtrt.com

+1 281 970 2700

OGT-0117 PTQ Half Horizontal Ad.indd 1

www.eptq.com
www.eptq.com
www.eptq.com
www.eptq.com

axens.indd 7
uop.indd
tupras.indd4 4 5
corrmat.indd

info@ogtrt.com

11/13/12 10:40 PM

PTQ Q1
Q2 2013
21
PTQ
2014
85
PTQ Q4
Q3 2013
2013 65
33
PTQ

08/03/2013 11:49
11/12/2013
07/06/2013 20:13
17:58
12/09/2013
14:33

Co-hosted by:

IDTC & BBTC 2014

LISBON, 6-9 MAY

Sponsored by:

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

Speakers include:
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

epc.indd 1

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
Register now at www.europetro.com
E: conferences@europetro.com

T: +44 20 7357 8394

12/12/2013 16:29

Role of FCC catalyst in refinery

profitability
By switching to a rare earth-free FCC catalyst, a refiner 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

he FCC unit is the main

contributor (51-52%) to the
refinery gasoline pool of the
Lukoil Neftohim Burgas refinery in
Bulgaria, with the next largest
contributor being the reformer
(about 27%). The gasoline sensitivity (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
refinery gasoline pool. Considering
MTBE has a MON of 94-97, alkylate
a MON of 93 (both of which are
produced from C4 olefins) and FCC
gasoline a MON of about 82,
increasing the production of
C4 olefins 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 subsequently led to an increase in
gasoline MON by 0.5 numbers.
This allowed the refinery to
increase the production of automotive gasoline by 1.3%, and to
increase the share of premium
automotive gasoline from 68% to
73%. This resulted in an annual
six-figure improvement in refinery
economics.
The Reid vapour pressure (RVP)
of gasoline from the FCC unit
correlates to the content of C4
hydrocarbons in the gasoline, with
lower RVP values obtained by
producing more C4 at the expense
of gasoline. Optimising the FCC
gasoline RVP during the winter

www.eptq.com

grace.indd 1

Physical and chemical properties of

hydrotreated vacuum gas oil feedstock
for Lukoil Neftohim Burgas FCC unit
Density at 20C, g/cm3
0.895
0.3
Sulphur content, wt%
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
509
90 vol%
95 vol%
532
12.13
K-Factor
Molecular weight (Goosens)
369
Hydrogen content (ConocoPhillips)
12.5
Aromatic carbon content
17.02
(ConocoPhillips)
79.5
Gasoline precursors (LNB)

Table 1

season (RVP was reduced from 60

to 50 kPa) and increasing the C4
olefins yield (leading to a higher
MON in alkylate production)
resulted in an additional improvement in refinery economics by a
five-figure 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 regenerator, 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 properties are shown in Table 1.

Optimisation of FCC catalyst

technology

In response to rapidly inflating

rare earth metal prices in 2011,
Grace developed the REpLaCeR
series of rare earth-free FCC catalysts. 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 independently adjust the activity and
selectivities of zeolite and matrix,
as well as the ratio of zeolite/
matrix activity, enables maximum
formulation flexibility. For low
metal applications, REsolution
catalysts are proven to match, and
even improve on, the performance
of traditional rare earth based catalysts. To date, there have been
more than 15 successful applications 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 significantly from the previous 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 catalyst, the e-cat still contains a small
amount of rare earth, as the inventory was not 100% changed out.
Typical FCC unit operating conditions from periods using Catalyst
X and the Grace catalyst are shown
in Table 3.
Table 4 shows the yield structure

PTQ Q1 2014 87

12/12/2013 12:15

Physical and chemical properties of e-cats used in the study

E-cat properties
Supplier 1

Catalyst X
Apparent bulk density (ABD), g/cm3
159
Total surface area, m2/g
55
Matrix surface area, m2/g
Zeolite surface area, m2/g
104
Particle size distribution, %
0-20 m
0.1
0-40 m
1.8
46
0-80 m
0-149 m
Average particle size, mm
83
Unit cell size,
24.29
Al2O3, wt%
40.5
1.35
RE2O3, wt%
Na2O, wt%
1.15
Fe, wt%
0.61
V, wt ppm
206
Ni, wt ppm
25
25
Cu, wt ppm

Grace
REsolution
0.85
147
51
96
2
5
45
94
84
24.27
44.2
0.38
0.28
0.42
208
63
8

Table 2
Operating conditions in Lukoil Neftohim Burgass FCC unit
Operating conditions
Supplier 1

Catalyst X
Flow rates
Hydrotreated feed, t/h
220
11
Unhydrotreated feed, t/h
Recycle, t/h
2
Dispersion steam, kg/h
3000
2000
Riser steam, kg/h
Stripping steam, kg/h
5881
134
Air flow rate, kNm3/h
Fresh catalyst flow, t/d
2
Catalyst to oil ratio
8.1
Temperatures, C
Combined feed temperature
317
535
Riser temperature
Regenerator dense phase temperature
668
Regenerator dilute phase temperature
678
Air temperature
179
Steam temperature
262

Grace
REsolution
229
7
5
3435
2000
6200
118
2
7.9
331
533
668
677
184
247

Table 3

for FCC products obtained using

Catalyst X and the REsolution catalyst. 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 feedstock. 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

88 PTQ Q1 2014

grace.indd 2

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 olefins and
naphthenes to paraffins and
aromatics. Conventional understanding 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

Yield structure in Lukoil Neftohim

Burgass FCC unit
Product, wt%
Supplier 1 Grace

Catalyst X REsolution
3.8
3.5
Dry gas (C2-)
Total C3s
8.0
8.2
Total C4s
14.0
13.8
Gasoline (C5-195C)
51.1
51.2
LCO (195-296C)
9.8
10.3
8.9
HCO (296-360C) & slurry 9
.0
Coke
4.3
4.1
Conversion
81.2
80.8

Table 4
Product properties in Lukoil Neftohim
Burgass FCC unit
Product, wt%
Supplier 1

Catalyst X
Gasoline
MON
81.7
RON
94.0
RVP, kPa
55.1
FIA hydrocarbon
Composition, vol%
Saturates
37.8
Olefins
39.2
23.0
Aromatics
C3s, wt%:
Propylene
80.8
Propane
17.9
C4s, wt%:
i-butane
33.5
i-butylene
17.3
6.7
n-butane
Total olefins
56.3

Grace
REsolution
82.2
94.0
57.9

41.7
32.5
25.8
78.8
20.9
39.5
16.0
7.6
53.3

Table 5

higher hydrogen transfer activity,

resulting in a lower olefin content
of FCC gasoline, as well as the C4
and C3 fractions. This is due to the
Z-21 technology utilised in this
catalyst.
Figure 1 shows how the proportion 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

www.eptq.com

11/12/2013 12:48

Theres more
than one
way to skin the

CAT!

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 specifications 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.

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 refi ning processes now
specifies 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 refi ner
for continuous catalyst reforming. A preeminent
catalyst additive supplier has standardized on this
valve concept for all their chemical injection units.

Call us today at 1-908-769-0700 or visit our website at (www.everlastingvalveusa.com)

Valve Manufacturers Since 1906

Boiler Blowoff Valves

Bulk Material Valves

Process Valves

American Boiler Manufacturers Association

everlasting.indd 1

Everlasting

VALVE COMPANY, INC.

09/12/2013 11:27

Regular A-92
Gasoline A-93

FCC yields at different gasoline

RVP values

Premium A-95
Super A-98

Product, wt%

Catalyst X
2%

Dry gas
C3s
C4s
Gasoline
LCO
HCO
Slurry
Coke
Conversion

29%

1%

68%

2%
25%

73%

Figure 1 Refinery gasoline grades produced

by Lukoil Neftohim Burgas using Catalyst
X and the REsolution catalyst

requirements
of
standard
EN228:2012 for RON but not MON.
Switching to the REsolution catalyst

increased the MON of FCC gasoline from 81.7 to 82.2, resulting in a

nal MON of 84.5 for the gasoline
pool with the RON unchanged.
Since FCC gasoline accounts for the
majority of the gasoline pool (over
50 vol%), the FCC catalyst clearly
has a signicant impact on the
octane number of the total gasoline
pool.
Based on operational data from
the FCC unit, the economic efciency of using both catalysts from
different suppliers was determined
by applying a renery model,
which uses software for linear

Catalyst X

REsolution

1.9 vol%; RON 94.0; MON 81.7

Prime-G
unit

50 kPa
5.30
7.15
13.25
50.01
11.04
4.92
4.10
4.23
79.94

Table 6

REsolution

51.7 vol%
RON 93.1
MON 81.7

RVP
60 kPa
5.32
7.10
12.65
50.61
11.01
5.10
4.00
4.21
79.89

53.6 vol%
RON 94.0
MON 81.7

26.7 vol%; RON 99.9; MON 89.0

7.0 vol%; RON 94.0; MON 82.2

FCC gasoline

Reformate

51.7 vol%
RON 94.0
MON 82.2

Prime-G
unit

44.7 vol%
RON 93.1
MON 82.2

27.7 vol%; RON 99.9; MON 89.0

programming
by
Honeywell
(RPMS). The results from RPMS
demonstrate that replacing Catalyst
X with the Grace catalyst results in
an increased prot of 4% for the
overall oil rening operations of
Lukoil Neftohim Burgas. This
demonstrates the important role
that the FCC catalyst plays in rening operations protability.

Optimisation of FCC gasoline Reid

vapour pressure

A signicant parameter of FCC

gasoline is Reid Vapor Pressure
(RVP),1 which increases linearly
with an increasing amount of C4
components.
Lukoil
Neftohim
Burgas performed a commercial
test operating with two different
RVP values of FCC gasoline. Table
6 shows how the yield of FCC
products changed when switching
from an RVP of 60 kPa to 50 kPa,
with the yield of C4s increasing by
0.6 wt% at the equal expense of
gasoline.
The effect of operating with two
different RVP values on the renery gasoline grades produced is
shown in Figure 3. The total
amount of automotive gasoline
produced by the renery decreased
by 0.18 wt%, while overall LPG
production increased by 3.7 wt%.
This resulted in an increase of
Premium A-95 production from
71% to 73% at the expense of the
Regular A-92 grade. Reducing the
RVP from 60 kPa to 50 kPa
improved renery protability by
about 2.5%.

Summary
11.1 vol%; RON 97.2; MON 93.2

4.2 vol%; RON 63.2; MON 60.2

4.4 vol%; RON 112.0; MON 93.0

0.01 vol%; RON 108.0; MON 93.0

Gasoline pool
RON 95.0
MON 84.2

Alkylate

Straight-run
naphtha

10.9 vol%; RON 96.8; MON 93.1

5.1 vol%; RON 63.2; MON 60.2

Methyl tertiary 4.2 vol%; RON 118.0; MON 94.0

butyl ether

Bioethanol

0.1 vol%; RON 108.0; MON 93.0

Gasoline pool
RON 95.0
MON 84.5

Figure 2 Structure of gasoline pool at Lukoil Neftohim Burgas using Catalyst X and the
REsolution catalyst

90 PTQ Q1 2014

grace.indd 3

It has been demonstrated that the

FCC catalyst is an excellent tool for
improving overall renery protability. Optimising FCC technology
by switching to the Grace
REsolution catalyst led to an
increase in the FCC gasoline MON
by 0.5 points, which allowed the
renery to produce 5% more
premium grade gasoline.
The overall renery margin was
improved further by optimisation
of the FCC gasoline RVP. A reduction in the RVP from 60 to 50 kPa
resulted in an increase in LPG
production by 3.7% at the expense
of gasoline pool reduction. The

www.eptq.com

11/12/2013 12:48

higher FCC C4 yield led to higher

production of alkylate, which
resulted in the production of 2%
more premium grade gasoline.
References
1 Watanbe K, Nagai K, Aratani N, Saka Y,
Chiyoda N, Mizutani H, Techniques for octane
enhancement in FCC gasoline, 20th Annual
Saudi-Japan Symposium, Dhahran, December
2010. 17. Montgomery J A, Guide to Fluid
Catalytic Cracking, Part 1, 1993.

Ivan Chavdarov is a Chemical Engineer in

the Process Engineering department of Lukoil
Neftohim Burgas, Bulgaria. His activities are
focused on guiding the operation of the units
of the FCC complex, troubleshooting support
and optimisation of the performance of the
FCC complex.
Email: Chavdarov.Ivan.S@neftochim.bg
Dicho Stratiev is Chief Process Engineer with
Lukoil Neftohim Burgas. He holds a MS in
organic chemistry engineering, and a PhD and
a DSc in petroleum refining from the Burgas
University Assen Zlatarov. He has authored
more than 130 papers.
Email: Stratiev.Dicho@neftochim.bg

Ivelina Shishkova is R&D Department

Manager with Lukoil Neftohim Burgas. She
holds a MS in organic chemistry engineering
and a PhD in petroleum refining from Sofia
Chemical and Technological and Metallurgical
University, and has authored more than 20
technical papers.
Email: Shishkova.Ivelina.K@neftochim.bg
Rosen Dinkov is the Quality Manager in the
Process Engineering department of Lukoil
Neftohim Burgas. His research interests include
crude oil characterisation, bio/conventional
fuels blends characterisation and modelling of
refinery distillation processes. He holds a MS
in organic chemistry engineering from Burgas
University and a PhD in the technology of
fossil and synthetic fuels from the University
of Chemical Technology and Metallurgy, Sofia.
Email: Dinkov.Rosen.K@neftochim.bg
Vladimir Jegorov is the Sales Development
Manager for Grace in the CIS region. Prior to
joining Grace, he was an FCC process engineer
at the Mazheikiai refinery in Lithuania.
Petko Petkov is a full professor and rector
of the Burgas University Assen Zlatarov. He
teaches in the social science department in
the field of oil refining and lubricants, and has
authored more than 180 scientific papers and
five books.
Email: PST_Petkov@abv.bg

Regular A-92
Premium A-95
Super A-98

RVP = 60kPa
1%
28%

71%

RVP = 50kPa
1%
26%

73%

Figure 3 Effect of changing the RVP on

refinery gasoline grades produced during
the Resolution catalyst period

QUICK, ACCURATE, PORTABLE

DISTILLATION
MINIDIS ADXpert

Portable, True Atmospheric D7344 Distillation

Excellent D86, ISO 3405, IP 123 Correlation
Fully Automated Measuring and Cleaning
6 mL of Sample
No Glassware, no Open Flame
Distillation of Gasoline, Jet Fuel, Diesel, Solvents,
Unknown Samples, Raffinates and Biodiesel
Specified for Use by US Pipelines

Grabner Instruments Messtechnik GmbH | A-1220 Vienna, Austria | Dr. Otto-Neurath-Gasse 1

Phone +43-1-282 16 27-0 | Fax +43-1-280 73 34 | info.grabner-instruments@ametek.at
ADXpert half page 1.indd 1

www.eptq.com

grace.indd 4

06.12.2013 11:46:37

PTQ Q1 2014 91

11/12/2013 12:49

blow...
Our

heater doesnt

...the flames downward, against the natural buoyancy of the

flue gases. Our steam methane reformer design fires upward,
reducing fan power consumption by 40 - 50%, to save
operating costs.
Choose the Foster Wheeler Terrace Wall Steam
Reformer for your hydrogen plant.

Learn more at www.fwc.com/terracewall

foster wheeler.indd 1

08/03/2013 10:49

Field experience with a Claus furnace

checker wall
A new design of Claus furnace checker wall delivers improved mechanical
performance and process flow characteristics
JEFFREY BOLEBRUCH Blasch Precision Ceramics
MOSSAED Y AL-AWWAD Saudi Aramco
MENG-HUNG CHEN CPC Corporation

epending on the licensor,

there are a number of different architectures that may be
placed in the Claus furnace to influence performance. Each has its pros
and cons. Regardless of design,
mechanical stability has traditionally been problematic.
Blasch was tasked to develop a
checker wall that would be easy to
install, mechanically stable under
all conditions, allow for design
with varying degrees of open area,
and could incorporate the use of an
integral manway, if desired.
Concerns at this point were strictly
mechanical; there were no requests
for mixing or flow management.
The goal was simply to put in a
checker wall that could survive a
campaign intact.
The result was the Blasch
HexWall checker wall. The blocks
are designed to be stacked dry and
are mechanically engaged through a
series of tabs and slots, so the wall
is quite stable, even when several
metres in diameter. No mortar is
used between the blocks, allowing
the assembly to accommodate the
thermally driven expansion and
contraction that comes with reaction
chamber operation. The blocks may
also be reused.
The initial checker wall blocks
were designed to be 9in wide, the
same as a row of bricks in the
hotface. As larger furnaces were
encountered, additional designs
utilising longer blocks were developed so as to retain a reasonable
aspect ratio of wall width to height.
The largest walls (>5m) utilise 18in
wide blocks.
The walls may either be erected
into a slot in the existing lining, or

www.eptq.com

blasch.indd 1

installed against the hot face brick

with a course of brick on either side
to lock it in. The interlocking nature
of the blocks makes provisions for
expansion management straightforward and effective. The openings in
the blocks are round, so that any
portion of the block not directly
supported by other blocks is in the
form of an arch, and, as such, has
greater mechanical stability than a
flat span.
Over the next several years, as

Time, temperature
and turbulence were
the variables that
played the largest
roles among the
process constituents
the stability and durability of the
checker walls was proven out,
some operators, who previously
spent their time rebuilding walls,
began to question what they were
actually designed to do other
than remain standing for an entire
campaign.
After considerable consultation
with process licensors, engineering
companies, end users, and miscellaneous stakeholders, it became
apparent that almost everyone
agreed that some combination of
time, temperature and turbulence
were the variables that played the
largest roles on the process constituents. That makes sense, as
the three Ts of reaction kinetics,
and their effects on process

efficiency, are well documented.

Blasch was determined to find a
way to use the HexWall checker
wall to influence these parameters.
We selected turbulence, thinking
that if we could improve mixing in
the furnace, we could improve efficiency, but soon learned that they
are all inter related, and we ended
up learning a lot more about residence
time,
residence
time
distribution, and plug flow vs
stirred tank design along the way.
A vectoring hood was developed
that would fit into the outlet ends
of the existing blocks, and could be
oriented and secured in the field in
such a way as to redirect, or
vector, the downstream flow
allowing each individual block in
the assembly to contribute to the
creation of the overall desired flow.
The vortex was initially selected
because it was felt that this configuration made the best use of the
furnace volume and created a flow
pattern that yielded a long path
length with a very tight residence
time distribution. This variation of
the HexWall checker wall was
christened the VectorWall.
There are now nearly two dozen
VectorWalls installed worldwide,
as well as approximately 100 of the
earlier
HexWalls.
Highlighted
below is a pair of VectorWall installations where it was possible to
contrast stability related details
from
before
and
after
the
installation

Field experience with theVectorWall

checker wall: mechanical

One of the larger diameter

VectorWalls done to date is
installed in Saudi Aramcos Berri

PTQ Q1 2014 93

13/12/2013 12:05

Knockout
drum

Steam
drum

LP
steam

Steam
separator

Acid
gas

MP/HP
steam

Blowdown
system

Auxiliary
burner

Reaction
furnace

Auxiliary
burner

Clean gas to
atmosphere

Auxiliary
burner

Stack
SO2

Catalytic
converter

Air
blower

Catalytic
converter

Catalytic
converter

Thermal
oxidiser
Tail gas

No. 1
condenser

No. 2
condenser

No. 3
condenser

No. 4
condenser

Liquid sulphur

Liquid
sulphur
storage

Sulphur
pit

Figure 1 Modified Claus process employed by Saudi Aramco

Gas Plant, in its SRU200. This

portion of the article is excerpted
from a paper written by Mossaed
Al-Awwad, Heat Transfer Engineer
for the Aramco Consulting Services
Department (CSD). Blasch gratefully acknowledges Saudi Aramco
for the following data.

Saudi Aramcos Berri Gas Plant

The current refractory system used

on the internal reaction furnace
lining in the Saudi Aramco sulphur
recovery unit (SRU) is a dual layer
of fire bricks as a hot face layer and
a castable layer as back-up with
stainless steel anchors to support
the castable refractory. Part of this
refractory systems internal features
is a checker wall that must be
incorporated into the refractory hot
face lining design. The existing
checker wall system that was used
for over 20 years frequently failed
due to improper design and installation techniques. This article
addresses the experience of a
checker wall refractory system, the
modified system and its impact on
the reaction furnace operation.1
Refractory lining description

The

normal

94 PTQ Q1 2014

blasch.indd 2

furnace

operating

temperature at Saudi Aramco SRU

units is 1800-2000F (9801090C);
however, the hot face material must
be capable of withstanding temperatures of 2500-3000F (13701650C)
which can occur during start-up
firing of natural gas. All SRU reaction furnace linings should utilise
at least 90% alumina hot face
material.
Brick linings at the hot face layer
are more durable than castable or
plastic ram materials. The initial
installation of brick may be slightly
more expensive and require a
highly skilled installer, but brick
normally provides a longer lining
life and requires less maintenance.
Part of this refractory systems
internal lining is a checker wall that
is made of a high density alumina
brick system that has good high
temperature strength to sustain the
harsh environment.
Checker wall system main function

The main function of the checker

wall is to have proper gas mixing
due to the residence time increasing
as a result of improved destruction
of H2S. In addition to that, the
checker wall works as a protector to
the tube sheet from radiant heat

and other effects of process upsets.

The exiting conventional checker
walls installed in Saudi Aramco
reaction furnaces either have a
cylindrical shape that resists high
stress due to the geometry, or a
matrix box shaped brick system
jointed with mortar that normally
failed due to thermal shock cycling,
mortar failure and vibration.
Modified Claus process

All Saudi Aramco SRUs use the

Modified Claus process shown in
Figure 1, where acid gas is initially
treated in an acid gas scrubber and
knock out drum to remove water,
hydrocarbon liquids and diglycolamine. Then the acid gas is heated to
500F (260C) in a pre-heater and
introduced to a reaction furnace
where a third of the H2S is burned
in the presence of heated air to
produce SO2. The generated SO2
reacts with the remaining H2S in
three in-series catalytic stages to
produce sulphur. Each stage
contains an auxiliary burner
(reheater),
converter
and
a
condenser. Liquid sulphur is
collected from all condensers and
stored in a sulphur pit where it is
degassed and then exported. The

www.eptq.com

11/12/2013 12:52

unreacted H2S and other sulphur

components, in the tail gas leaving
the last condenser, are oxidised to
SO2 in a thermal oxidiser.
Modified Claus reactions:
3 H2 S + 3/2O2 H2O + SO2 + 2H2S

(1)

2 H2S + SO2 2 H2O + 3/x Sx

(2)

Where x is = 1, 2, 3...
In a split flow configuration of
the Modified Claus process, a third
of the acid gas passes through the
reaction furnace and waste heat
exchanger and the remaining fraction is mixed with the waste heat
exchanger exit gases just before the
first converter.
Hot process air comes from the
air preheater at 700F (370C) and
9 psig pressure to meet the acid gas
line at the burner nozzle. The air is
rationed to provide enough air to
burn a third of the total H2S in the
acid gas with no excess oxygen.
The exact amount of air needed to
maintain a H2S/SO2 ratio of 2:1 is
controlled. The ratio of 2:1 H2S/SO2
must be obtained in the combustion
chamber to progress all the way
through the train. If there is more
than 2:1 H2S/SO2 in the reaction
furnace, the SO2 content of the gas
will gradually lessen all the way
through the train leaving an excess
of H2S. If the ratio is less than 2:1
H2S/SO2 in the reaction furnace,
the reverse happens with more SO2
remaining than H2S in the tail gas.
On the initial start-up of the plant,
fuel gas is used to raise the temperature in the reaction furnace because
the burning characteristics of the
cold acid gas are poor. When the
main burner is lit on fuel gas, the
gas/air ratio is adjusted so that the
fuel gas is burned stoichiometrically.
Too much gas or substoichiometric
fuel gas firing will deposit carbon
on the catalyst and too much air or
above stoichiometric fuel gas firing
will cause a temperature runaway in
the converter bed if there is sulphur
present, due to the oxidation of the
sulphur.
Fuel gas addition is normally
used to warm up the train during
start-up prior to acid gas admission
and to clean up the converter beds
before a planned shutdown. The

www.eptq.com

blasch.indd 3

Figure 2 Cylindrical shape conventional

checker wall

Figure 3 Matrix box shape checker wall

hot process air and acid gas lines

are provided with 75 psig steam
injection. Steam is injected to maintain a stable flame and to keep the
flame temperature below 2400F
(1315C).

external ambient conditions, the

premature failure of the existing
refractory system was due to
several factors including the design,
bricks quality, installation, and so
on.

Investigative results

Investigative results and

recommendations

The exiting conventional checker

walls installed in Saudi Aramcos
reaction furnaces either have a
cylindrical shape that resists high
stress due to their geometry, or
matrix box shape bricks jointed
with mortar. Investigation revealed
that the existing checker wall has a
clear design deficiency and should
be redesigned. The main cause of
the failure was due to thermal
cycling during the unit start-up and
shut-down. Liquid carry-over to the
combustion chamber loosens the
brick system and causes complete
failure. In addition to that, running
the unit above its design capacity
causes damage to the conventional
checker wall as it cannot sustain
high vibration due to over-design
flow rate.
Figures 2 and 3 show damage
caused to both checker wall
designs.
Additional investigation revealed
that, considering there were no
drastic changes in the internal
firing, gas composition and the

Saudi Aramco had bad experiences

with the current checker designs
due to: repeated failures, short
service life, prolonged repair shutdown, and extensive repair costs.
It is recommended to use an
interlocking hexagonal ceramic
block system which is made of
ceramic (VectorWall, see Figure 4).
Because of its special design, the
modified system which has hexagonal shape increases capacity and
lowers the operation and maintenance costs of medium and large
diameter high temperature reaction
chambers such as those used in
Claus sulphur recovery units. This
is accomplished by altering the
reactant flow path creating significant mixing within the chamber
thereby optimising use of the available volume within the chamber.

Field experience with the

VectorWall checker wall: process
CPC Corporation - Taoyuan refinery

CPC Corporation, Taiwan (CPC) is

Figure 4 VectorWall at Berri Gas Plant

PTQ Q1 2014 95

11/12/2013 12:52

H2S
acid gas
Insulation
Air
Main burner
H2S + NH3
acid gas

Zone (I)

Zone (II)

Boiler
inlet

Figure 6 A partially completed VectorWall

Start-up: open
Normal: close

Fuel
gas
Checker wall
H2S
acid gas

Figure 5 Schematic of Taoyuan refinerys F6301 reaction furnace

/4
/
10 201
/6 2
/
30 201
/7 2
/
18 201
/9 2
17 /20
/1 12
1
27 /20
/1 12
2/
2
15 01
/2 2
/2
6/ 013
4/
26 201
/5 3
/
15 201
/7 3
/2
3/ 013
9/
20
13

12

01

20

21

3/

2/

/2

/1

20

After revamp

Zone (I) flue gas

Zone (II) flue gas

12

1/
/1
23

featuring a complex production

process and a complete range of
equipment. It has a capacity of
220 000 b/d of crude oil. The Dalin
renery, which split off from
Kaohsiung to become independent
in 1996, has four offshore mooring
buoys as well as both large and
small docks for the unloading and
loading of crude oil and petroleum
products. It has a topping capacity
of 300 000 b/d. The Taoyuan
Renery was established in 1976
and, following some debottlenecking renovations and the addition of
a second distillation plant, currently
has a daily capacity of 200 000 bbl
of crude oil.4
There are 15 sulphur recovery
units within CPC. As a part of the

Before revamp

1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0

11

Flue gas temperature, C

the foremost energy enterprise in

the Republic of China with nearly
15 000 employees and a variety of
functional units around the island.
As a state run enterprise, the
company is responsible for the
development and supply of petroleum and natural gas, and is the
core of Taiwans petrochemical
industry.2
CPC of Taiwan owns and operates three oil reneries, located at
Kaohsiung, Dalin and Taoyuan,
respectively. The three reneries
have a combined capacity of
720 000 b/d.3
The Kaohsiung renery, which
has the longest history of the three,
is a large integrated oil rening and
petrochemical production facility

Figure 7 Flue gas temperatures before and after revamping

96 PTQ Q1 2014

blasch.indd 4

Claus sulphur recovery process, the

F6301 reaction furnace is the heart
of No. 3 SRU at the Taoyuan renery. It was designed to process the
acid gas from an HDS (hydrodesulphurisation) plant in the renery.
The acid gas comes from two
sources, the amine regeneration
unit (ARU) clean sour gas stream
containing
mostly
hydrogen
sulphide and the sour water stripper (SWS) foul sour gas stream
with roughly equal amounts of
ammonia, hydrogen sulphide and
water vapour.5
Figure 5 is a schematic of the
F6301 reaction furnace. It is a
refractory lined reactor with an
inner diameter of 2.1m and is about
9m long. A standard brick checker
wall formerly divided the furnace
into two zones. Acid gas from the
HDS was introduced both at the
burner and immediately downstream of the checker wall. The
hydrogen sulphide and ammonia
in the acid gas were converted to
elemental sulphur and SO2 and N2
respectively, in the furnace before
exiting the waste heat boiler. The
elemental sulphur condensed in the
cooling tubes was removed at the
liquid outlet. The leftover H2S and
the SO2 formed in the furnace
reacted further downstream in a
catalytic reactor to form more
elemental sulphur.
The furnace was originally
equipped with a conventional 9in
brick checker wall. Prior to the
retrot, severe furnace vibration
and noise were observed, especially
at high furnace throughput. This
was believed to have been caused
by instability of the damaged
checker wall. Thus it was not possible to operate the furnace at high
temperatures, in particular in the
rst zone. In addition, deposits of

www.eptq.com

11/12/2013 12:52

Cooperative agreement for

sulphuric acid applications

Following the first presentation of

this
technology,
Blasch
was
approached by DuPont Clean

www.eptq.com

blasch.indd 5

2000
1750
1500
1250
1000
750
500
250
/4
/
1 0 201
/6 2
/
3 0 201
/7 2
/
1 8 20
/9 12
17 /20
/1 12
1
27 /20
/1 12
2/
2
1 5 01
/2 2
/2
6 / 013
4/
2 6 201
/5 3
/
1 5 201
/7 3
/2
3 / 013
9/
20
13

12

01

20

21

3/

2/

/2

/1

12

/1

1/

20

11

Figure 8 Comparison of acid gas processed

21 201
/4 2
/
10 201
/6 2
/
30 201
/7 2
/
18 201
/9 2
17 /20
/1 12
1
27 /20
/1 12
2/
2
15 01
/2 2
/2
6/ 013
4/
26 201
/5 3
/
15 201
/7 3
/2
3/ 013
9/
20
13

01

3/

/2

/1

1/

20

11

Sulphur yield, tons/day

/1
23

After revamp

Before revamp

150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
0

2/

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 beneficial 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 N2
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 m3/hr.
In addition, no more furnace vibrations were observed.
Figure 9 compares the sulphur
yield. After the revamp, the average 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.

After revamp

2250

12

In light of the above problems, CPC

decided to revamp the furnace in
May 2012. The revamp included
installing a new insulation refractory 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 backfilled with tight
castable.

Foul sour gas processing rate,

m3/hr

Retrofit

Before revamp
2500

23

ammonium
bisulphate/sulphate
were observed in the downstream
equipment due to insufficient
destructions of the ammonia in the
furnace.

Figure 9 Comparison of sulphur yield

Technologies to explore application

of the VectorWall technology in
sulphuric acid applications with the
objective of developing a more efficient, compact furnace design
utilising this technology.
DuPont was already using Blasch
HexWalls in two of its own plants.
The first 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 first was constructed in a
choke ring configuration, and the
second as a standard checker wall
a configuration we would see
again in Claus plants later on. The
first wall was located a mere 8ft 5in

Figure 10 VectorWall after 18 months

service.

from the burner, and the second

21ft 5in.
Figure 11 shows the walls during
an inspection in 2012, some five
years after installation. They are
still in service.
Also, in 2007, a pair of HexWalls

PTQ Q1 2014 97

11/12/2013 12:52

oped designs that are being

marketed.
Interestingly, the Borderland Plant
SAR was experiencing issues relative 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 configuration and

VectorWall performance

Figure 11 Five years service at James River

Figure 12 New HexWall (left) and after five years service (right) in the Borderland plant

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.
Residence
time
0

0.5

1.0

Mixing
0% 100%

0%

Figure 13 Straight through furnace design

98 PTQ Q1 2014

blasch.indd 6

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 casting 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 installation, 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-

So, what has been learned from all

this? Access to software and hardware that can create a realistic,
accurate, multiphysic (heat, mass,
ow and kinetic) reactor simulation, 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 conguration 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 optimal 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 performance, 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 solution 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 characteristics is desired. For example, in

www.eptq.com

11/12/2013 12:53

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 unreacted 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 temperature, or any of the reaction kinetics.

www.eptq.com

blasch.indd 7

Velocity
ft/s
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

Figure 14 Choke ring on top, VectorWall on bottom

Residence
time
1.0
0.9
0.9
0.8
0.7
0.6
0.6
0.5
0.4
0.4
0.3
0.2
0.1
7.1e0.02
0

Figure 15 Choke ring on top, VectorWall on bottom

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 w w w. c p c . c o m . t w / e n g l i s h / c o n t e n t /
index01.asp?sno=715&pno=64
4 w w w. c p c . c o m . t w / e n g l i s h / c o n t e n t /
index01.asp?sno=715&pno=64

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 fired equipment
specialist, Saudi Aramco Oil Company,
responsible
for
providing
technical
consultations related to fired 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 Refining & 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.

PTQ Q1 2014 99

11/12/2013 12:53

REFINING CHINA 2014

Regent Hotel, Beijing, 2-3 April
REGIST
ER
NOW

ional
logy.
ess
ry
om

ted
om

Join Us at the Top International Refining Conference in China

Into its 8th year, REFINING CHINA was the first refining technology conference in China and was hailed by many
past delegates as the best of its kind in China. The distinguished speakers that are confirmed thus far is an
affirmation that the conference will be yet another resounding success.

PRINCIPAL CHINESE SPEAKERS IN 2014

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 refining enterprises
Hua Wei, Vice President and Chief Engineer, Sinopec

REGISTER EARLY TO ENJOY SAVINGS

Yanshan Petrochemical The green low-carbon path of

petroleum refining
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 Pacific Petrochemical
Company, Dalan, Safe use of petrochemical equipment

SPONSORS

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)

For more information, visit www.refiningchina.com

rc ad.indd 1

13/12/2013 12:51

Characterising and tracking contaminants

in opportunity crudes
The specific 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

rude oil characterisation and

tracking of specific properties
and contaminants are critical
aspects in the design and operation
of the modern refinery. Given the
influx of new sources for oil
(synthetic crude from upgraders,
oil shale, etc), understanding the
effect of these crudes on the refinery 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 methodologies 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 refiners. Those facilities
that are configured for light/sweet
crude now have access to significant volumes of tight oil that match
nicely with the assets. Over the last
five to 10 years, many refiners have
gone through reconfigurations to be
able to handle heavy Canadian and
Venezuelan crudes.
As these assets come on-line, those
refiners 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 balancing their new asset capability
against changing availabilities of
both light and heavy crudes slates.

www.eptq.com

kbc.indd 1

Key stream properties and impacts

Property
API

Comment
Impact
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 H2

and produce more H2S

Corrosion

Nitrogen

High nitrogen levels require H2

and produce more NH3

Corrosion

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

guard bed catalysts

Corrosion/catalyst
deactivation

Concarbon

Requires carbon rejection

mechanism

Catalyst deactivation and

yields

Asphaltenes

Increases potential for fouling that

requires shutdowns to resolve

Fouling

Naphthenic acids

High levels cause corrosion

Corrosion/fouling

Compatibility

Certain crude and blends are

incompatible

Affects allowable crude

blend and fouling

Chlorides

Typically associated with alkali

metals

Corrosion

Methanol
Helps prevent hydrate formation

Water/oil separation,
catalyst deactivation

Viscosity

High transportation costs

If too high to pump, requires

diluents or redesign

Table 1

Ultimately, this dramatic shift in

crude availability is opening up the
crude slate that most refiners
can process, and is influencing a
facilitys need to track and understand how different crudes will
and are impacting a facilitys
performance. KBC has been recommending for some time that a
premium should increasingly be
placed on a refinerys 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 refinery is beyond the
scope of this article. However, a
high level summary is warranted to
provide context on the need for
refiners to track and monitor these
properties and contaminants.
Tables 1 and 2 summarise key
properties and how they impact
refining operation.

PTQ Q1 2014 101

11/12/2013 13:01

Crude metals and sources

Conventional metals
Nickel, vanadium
In asphaltenes

Source
Natural organics

Concerns
Catalyst poison

Iron
Iron oxides

Corrosion products,
sulphides

Catalyst poison,
foulant

Silicon

Polydimethlysiloxane

Defoamer

Catalyst poison

Arsenic

As organics

Natural organics

Catalyst poison

Exotic metals
Phosphorous

Source
Pigging gel,
acidising gel

Concerns
Catalyst poison

Titanium

Naturally occurring

Catalyst poison

With bitumen solids

Alkali metals
Ca, Mg, Na
Naturally occurring

Calcium stearate
Flow improver

Catalyst poison, foulant

Mercury

Naturally occurring

Catalyst poison

Selenium

Naturally occurring

Environmental

Crude fouling

The following are typical examples of how these properties could

impact refining 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 sufficient metallurgy to
avoid naphthenic acid corrosion?
Are the crudes themselves compatible with one another when mixed,
or will asphaltene precipitation
occur? Can the desalter sufficiently
remove basic sediment and water
(BS&W) and salts to avoid downstream 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

sufficiently 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 C3+
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

kbc.indd 2

Property data sources

Several data sources exist to understand and estimate the properties

of a given crude or refining stream.

Laboratory analysis

Table 2

102 PTQ Q1 2014

Will resid processing capacity limit

refining 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 refinery will perform and how the
assets generate sustainable and reliable
profitability.
Therefore,
understanding these properties and
contaminants, not only in crudes
but also in intermediate streams
and final products, is critical to
selecting a crude slate and preparing and executing an operating
plan.

requirements? Can hydrotreated

naphtha be routed to gasoline
blending to help manage pool
octane? What is the proper disposition for C5s and C6s 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,

fluid catalytic crackers,
hydrocrackers)

How will the gasoil sulphur and

nitrogen impact these units? Can
the downstream amine, sour water,
and sulphur units manage the H2S
and NH3 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 sufficient hydrogen available to meet contaminant
removal and/or conversion targets?

Resid units (coker, ebullated bed

hydrocracker, fixed bed resid
hydrotreater)

How will the resid concarbon level

impact coker yields and utilisation?

The best way to understand a

streams properties is to measure
them in a laboratory. While this
option may seem intuitively obvious, accurate measurement of a
given property is more difficult
than it may initially seem.
First, multiple methods exist for
measuring a specific property for a
given stream type. Most refiners
utilise standard methods according
to internationally accepted measurement techniques, such as those from
the American Society for Testing
and Materials (ASTM). However,
the refiner must decide on the
method that is utilised in their facility 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 various properties and testing methods.
Therefore, when the laboratory
reports a given value for a property, it really should be used within
the context of the values confidence level rather than as an
absolute value. Also, the ability of
the laboratory to accurately measure a property is strongly
influenced by the following factors:
Availability and accuracy of the
measurement equipment
Training of laboratory staff and

www.eptq.com

11/12/2013 13:01

Process 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
FDSDFLW\ IRU DEVRUELQJ DFLG JDVHV LQFUHDVHV VLJQLFDQWO\ DV WKH WHPSHUDWXUH
decreases, resulting in reduced circulation rate and associated operating costs.

Typical Physical Solvent Process

Comparing Physical
Solvents for Acid Gas
Removal

PC (Propylene Carbonate)

The Fluor Solvent process uses JEFFSOL PC and is by Fluor

Daniel, Inc. The light hydrocarbons in natural gas and hydrogen in synthesis
gas are less soluble in PC than in the other solvents. PC cannot be used for
selective H2S treating because it is unstable at the high temperature required
to completely strip H2S from the rich solvent. The FLUOR Solvent process is
generally limited to treating feed gases containing less than 20 ppmv; however,
LPSURYHGVWULSSLQJZLWKPHGLXPSUHVVXUHDVKJDVLQDYDFXXPVWULSSHUDOORZV
treatment to 4 ppmv for gases containing up to 200 ppmv H2S. 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.

DEPG
at 25C

PC
at 25C

NMP
at 25C

MeOH
at
-25C

H2

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

Gas Component

DEPG (Dimethyl Ether of Polyethylene Glycol)

DEPG is a mixture of dimethyl ethers of polyethylene glycol.

Solvents containing DEPG are marketed by several companies including
Coastal Chemical Company (as Coastal AGR), Dow (Selexol), and UOP
(Selexol). DEPG can be used for selective H26UHPRYDODQGFDQEHFRQJXUHG
to yield both a rich H2S feed to the Claus unit as well as bulk CO2 removal.
DEPG is suitable for operation at temperatures up to 347F (175C). The
minimum operating temperature is usually 0F (-18C).

MeOH (Methanol)

The most common Methanol processes for acid gas removal are
the Rectisol process (by Lurgi AG) and Ifpexol process (by Prosernat). The
PDLQ DSSOLFDWLRQ IRU WKH 5HFWLVRO SURFHVV LV SXULFDWLRQ RI V\QWKHVLV JDVHV
GHULYHG IURP WKH JDVLFDWLRQ RI KHDY\ RLO DQG FRDO UDWKHU WKDQ QDWXUDO JDV
treating applications. The two-stage Ifpexol process can be used for natural gas
applications. Methanol has a relatively high vapor pressure at normal process
conditions, so deep refrigeration or special recovery methods are required to
prevent high solvent losses. The process usually operates between -40F and
-80F (-40C and -62C).

NMP (N-Methyl-2-Pyrrolidone)

The Purisol Process uses NMP and is marketed by Lurgi AG.

7KH RZ VFKHPHV XVHG IRU WKLV VROYHQW DUH VLPLODU WR WKRVH IRU '(3* 7KH
process can be operated either at ambient temperature or with refrigeration
down to about 5F (-15C). The Purisol process is particularly well suited to the
SXULFDWLRQRIKLJKSUHVVXUHKLJK&22 synthesis gas for gas turbine integrated
JDVLFDWLRQFRPELQHGF\FOH,*&&V\VWHPVEHFDXVHRIWKHKLJKVHOHFWLYLW\IRU
H2S.

CO2

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

H2S

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
DYDLODEOHDOWHUQDWLYHVXVLQJDYHULHGSURFHVVVLPXODWRUVXFKDV3UR0D[ which has
EHHQYHULHGZLWKSODQWRSHUDWLQJGDWD
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.

WDZ/
^K'Z/
h^

bre.indd 1

7/3/12 10:45:54

Repeatability and reproducibility examples

Property
Sulphur -~2 wt%
Sulphur -~500 ppmw
Concarbon -~ 2 wt%
Concarbon -~ 0.5 wt%
Nitrogen -~3000 ppmw
Nitrogen -~30 ppmw
Vanadium -~100 ppmw
Vanadium -~100 ppmw

Method
ASTM D4294
ASTM D4294
ASTM D4530
ASTM D4530
ASTM D5762
ASTM D4629
ASTM 5708
ASTM 5863

Repeatability
0.06
63
0.12
0.05
261
1.1
4
11

Reproducibility
0.25
128
0.39
0.15
798
4.7
19
21

Table 3

adherence to a given procedure

Sampling,
preparation,
and
handling techniques of a given
stream sample.
Examples abound of refiners 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 conditions 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 properties 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 predictive basis. Hence, refiners 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

kbc.indd 3

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 separations 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 field. 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 influenced by the age of the production
field. In general, as fields age the
amount of heavy material increases,
thereby concentrating contaminant
levels.
Content of trace metals and
contaminants is influenced 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

field, the properties and boiling

point content of various crude shipments can vary substantially.
Crude assays can be powerful
information in deciding a crude
slate for a given facility. The information 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, refiners and crude traders can be lulled into a false sense
of security and knowledge as they
assume that a given crude will
have a fixed set of properties and
yields continuously.
In reality, the assay is a snapshot in time. Validations of crude
assays and backcasting of actual
versus predicted crude properties
and volumes are critical components of effective crude selection
and management processes.

Plant data

Most refiners regularly sample

plant streams and complete laboratory analyses, either as needed for
monitoring or dictating unit performance or as required for selling
intermediate or final products. Due
to the influx of new crude sources,
a few specific areas should be
reviewed and assessed within the
facility.
First, validation of crude assays is
becoming more critical. As an
example, refiners processing tight
oils are observing swings in crude
API gravity of five to 10 numbers,
or more, for a given shipment or
batch. Since API gravity is an indicator of the amount of light and
heavy material, these levels of
swings can dramatically impact
how the refinery manages and
responds to these quality changes.
Therefore, the following recommendations 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

www.eptq.com

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.
burners

flares

thermal oxidisers

aftermarket parts, service & solutions

Zeeco, Inc. 22151 E. 91st St., Broken Arrow, OK 74014 USA

+1-918-258-8551 sales@zeeco.com zeeco.com

zeeco .indd 1
Zeeco_PTQ_Dec.indd
1

Zeeco, Inc. 2014

13/12/2013
15:44
12/13/13 8:42
AM

transfer meters are properly and

regularly calibrated and account for
changes in crude specific 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 specifications
for crude quality and understand
the incremental economic and
performance impact for relaxing
those specifications
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 property testing methods for the
laboratory, and ensure the laboratory staff are properly trained and
have the proper equipment
Consider periodic round-robin or
independent laboratory testing to
validate and certify that the internal 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 engineering 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 training and situational/what-if drills.

106 PTQ Q1 2014

kbc.indd 4

The bottom line is that refiners

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 refiners 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 refineries, to monitor and track key
proprieties. Of key interest currently
are measurements for vapour pressure and H2S for environmental
compliance and personnel protection purposes, particularly given
the usage of trucks and rail cars for
transporting tight oils.
Production entities, pipelines,
and refiners are jointly developing
standard specifications for crude oil
qualities. As an example, a specification standard for West Texas
Intermediate crude has been developed 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 development 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 promoting fundamental research projects
on crude oil quality and testing
methods.
Websites, such as www.crudemonitor.ca, can provide data on
property variability of select crude
types and agglomerated crude
sources.
Crude compatibility has drawn
renewed interest in the last five
years, given the desire by many
refiners to process highly paraffinic
tight oils with heavy Canadian

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 expanding the usage of Infochems
Multiflash technology in predicting
phase equilibria, including hydrates,
waxes, and asphaltenes. This technology is currently used extensively
in the upstream production side for
flow assurance modelling The
intention is to apply this same technology to downstream operations to
supplement and enhance KBCs
existing capabilities in predicting
incompatibility.

Methods and tools for tracking

contaminants

Refiners need ways to predict

stream qualities throughout a facility to make decisions on crude
purchasing and operating strategy.
Ultimately, refiners 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 refiners rely on some crude

assay management system (CAMS)
to convert paper assay data into
usable information for making decisions. Essentially, these software
programs allow the user to mix
multiple crudes together and generate 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 generate the cut properties. Instead,

www.eptq.com

11/12/2013 13:01

crude and vacuum fractionation

sectional efficiencies are assumed
based on typical unit performance
(or anticipated generic performance) and are essentially fixed.
Though reasonably accurate for
most users and tasks, this simplified distillation method may not
fully capture how a given crude
mixture and rate will respond in
the crude or vacuum unit. In addition, the CAMS tools cannot
capture limitations on heat recovery, 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 refiners rely on Excel spreadsheets to estimate stream properties

and the impact on unit performance and yields. Often, these
property calculations are based on
open literature equations or simplified 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 operational 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 estimate critical input properties to
complete the calculations.

Linear programs

Most refiners rely on a linear

program (LP) to make crude, feedstock, and high level operating
strategy decisions. The LP has the

www.eptq.com

kbc.indd 5

capability to model the entire facility from crude through to finished

products. Each process unit is
represented by a linear sub-model,
which takes feed flows and key
properties and generates product
rates and key properties. Each
sub-model is a linear representation. 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 mitigate this limitation to some degree,
but the LP is dependent on good
delta shift vectors which are generated by other tools and technology
licensors.
LPs can be useful tools for evaluating different crudes and the
impact on refinery performance and
economics, but certain limitations
will impact the number of properties a refiner can predictively track.

Process simulations

Over the last 10 years, process

simulation packages, such as KBCs
Petro-SIM, have evolved and
improved substantially in modelling refineries. Process simulation
packages not only include detailed
thermodynamic packages and the
ability to simulate pumps, compressors, exchangers, separators, and
fractionators, but also include the
capability to model kinetic reactors,
track and predict stream properties
from crude through finished product, and simulate an entire facility
on a rigorous, non-linear basis.
Hence these models can help a
refiner understand how the entire
facility will respond to given feedstock quality change and determine
how the distribution of contaminants will change throughout the
facility.

Use of multiple components and

narrow cuts allows the process simulation 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 property 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 representation. The reactor model for an
individual process unit can be used
to run a series of cases, where individual feed property and unit
operating conditions are perturbated 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 refinery
responds to a given change, which
an LP cannot easily replicate. The
process simulation technology is
proprietary, which requires licensing from a software vendor,
although many refiners already
license unit simulators and flowsheets for traditional uses and
require only incremental licences to
develop a whole plant simulation.
Table 4 summarises the advantages
and disadvantages of each tool.

PTQ Q1 2014 107

12/12/2013 12:17

Tool advantages and disadvantages

Tool
Advantages
Disadvantages
Create standard crude cut estimates Estimates fractionation efficiency
Crude assay
No downstream impacts
management system Ease of use

Spreadsheet
Ease of use
Need estimate of input properties
Open literature based
for changing conditions
correlations
Linear programs

Commonly used
Solution speed
Facility wide prediction capability

Linearisation of non-linear processes

Limited number of cuts tracked

Process simulations

Facility wide, non-linear

Solution time
representation
Proprietary methods
Predict changes in properties from
separations/reactions
Detailed component property curves

Table 4

Property distributions

To help understand how property

and contaminant distributions can
be different for various crudes and
the resulting impact on refining
operation, KBCs Petro-SIM was
utilised to generate property curves
for some of the key properties
discussed above.

Sulphur

Sulphur content, wt%

One of the primary crude properties

refiners focus on is sulphur content.
Higher sulphur crudes require
additional assets to remove and
process the sulphur, such as hydrotreaters, hydrogen plants, amine
systems, sour water systems and
sulphur plants. Figure 1 illustrates
the sulphur distribution for several

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 significantly 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 boiling range.
Tight oil crudes can have a
sulphur advantage over sister
conventional
crudes,
thereby
making these feeds attractive to
sweet crude configured refineries.

Albian heavy synthetic (unconventional)

Athabasca bitumen (unconventional)
Athabasca dillbit (unconventional)
Bakken (unconventional)
Brent (conventional)
Maya (conventional)
Syncrude sweet (unconventional)
WTI (conventional)

Cutpoint temperature
Figure 1 Sulphur for conventional and unconventional crudes

108 PTQ Q1 2014

kbc.indd 6

There are several considerations

and options to manage crude
sulphur content and distribution.
Accurate assay information is critical 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
refinery performance monitoring.
For those facilities processing
purchased intermediates (gasoil,
diesel, naphtha), consider periodic
sulphur distribution testing to properly 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 significant impact on catalytic units. In
particular, FCC and gasoil hydroprocessing (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 hydrotreated 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 unconventional 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

www.eptq.com

11/12/2013 13:01

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 formation
from
hydroprocessing
reactions dictate overall cycle
length. In the FCC, concarbon has a
significant effect on the reactor heat
balance, thereby influencing catalyst 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

www.eptq.com

kbc.indd 7

Nickel content, ppm wt

Albian heavy synthetic (unconventional)

Athabasca dillbit (unconventional)
Bakken (unconventional)
Brent (conventional)
WTI (conventional)
Zuata medium (unconventional)

Cutpoint temperature
Figure 2 Nickel for conventional and unconventional crudes

units, feedstock concarbon content

determines the coke yield, thereby
influencing 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 unconventional crude does not necessarily
mean a crude will have high
concarbon levels However, several
of the example conventional crudes
have concarbon levels that appear

Vanadium content, ppm wt

of metals, even in the gasoil range.

The type of diluent will have a
significant impact on this effect.
Tight oil crude (Bakken) has
substantially
lower
vanadium
content than even a sister conventional 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 (distributors, wash rates, grid/packing),
flash zone (tower entry, vapour
disengaging), and transfer line
(velocity and flow regime) will
influence entrainment, which can
impact gasoil metals as much as the
natural distribution of metals.
Regular metals balances among
HVGO, overflash (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 monitoring and monthly production
planning backcasting, utilise a
process simulator and actual plant
data to help identify inaccuracies in
crude assay metals content and
distribution profile.

at much lower distillation temperatures than most conventional and

unconventional crudes
Brent and Syncrude Sweet have
similar concarbon distributions,
though they are produced by
different methods. Syncrude, by
definition, is a synthetic crude
that is created by coking and
severe hydrotreating of Canadian
bitumen. This process significantly
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

Arabian heavy (conventional)

Athabasca bitumen (unconventional)
Athabasca dillbit (unconventional)
Bakken (unconventional)
Maya (conventional)
WTI (conventional)
Zuata medium (unconventional)

Cutpoint temperature
Figure 3 Vanadium for conventional and unconventional crudes

PTQ Q1 2014 109

11/12/2013 13:02

Concarbon content, wt%

facility impact for these crude

modications, as long as the tool
has representative assay information and can properly account for
downstream unit impacts on varying concarbon levels.

Arabian heavy (conventional)

Athabasca dillbit (unconventional)
Bakken (unconventional)
Brent (conventional)
Maya (conventional)
Syncrude sweet (unconventional)
Zuata medium (unconventional)

Carbon to hydrogen ratio

Cutpoint temperature
Figure 4 Concarbon for conventional and unconventional crudes

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 operations monitoring information.
Also as with metals, specialised
concarbon destruction catalyst can
be installed in the gasoil hydrotreater to help manage concarbon
and ensure the unit meets cycle
length targets. In addition, monitoring of feedstock concarbon levels
should be part of the catalyst activity 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 conventional crudes. Therefore, the unit
may be able to increase cutpoint if
the vacuum tower is hydraulically

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 maximum 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 simulators can help analyse the overall

Crude and refinery product C/H ranges

Crudes
Dilbit and bitumen
Synthetic crude
Conventional heavy
Conventional medium
Tight oils

Table 5

110 PTQ Q1 2014

kbc.indd 8

C/H range
7.3 -8.1
7.0 -7.2
6.9 -7.3
6.3 -6.9
6.3 -6.5

Products
LPG
Gasoline
Jet
Diesel
Fuel oil

C/H range
4.5 -4.9
5.7 -6.2
6.2 -6.7
6.4 -6.9
>7.0

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 product to meet quality specications,
such as specic gravity, cetane,
aromatics, and so on. These quality
specications can vary substantially, 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 summarises some representative ranges for
C/H for various crude grades and
typical rening products.
Although C/H ratio can be measured, most reners do not regularly
test for this property, and most
crude assays do not carry it,
although they do carry other proxies 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 accurate process simulation, as the
kinetic reactors need to predict and
utilise this information to determine the amount of hydrogen that

www.eptq.com

11/12/2013 13:02

 www.digitalrefining.com is an
easy-to-use website for the refining, gas
and petrochemical processing industries.

and other information from over 40

categories.
Find relevant industry events worldwide.

A simple search facility enables you to

find and display relevant information
across the whole spectrum of articles,
literature, news, events and companies.
Readers can download technical articles

dig ref copy 2.indd 1

Read news items focusing on contracts,

company news and new products.
Download articles from our new White
Papers section

16/09/2013 10:02

Carbon/hydrogen ratio

Case studies
Albian heavy synthetic (unconventional)
Arabian heavy (conventional)
Athabasca dillbit (unconventional)
Bakken (unconventional)
Brent (conventional)
Maya (conventional)
Syncrude sweet (unconventional)
Zuata medium (unconventional)

Cutpoint temperature
Figure 5 C/H ratio for conventional and unconventional crudes

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 hydrogen 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 hydroprocessing assets, as well as hydrogen
purchase and production options,
are critical to ensure a new crude
will meet the necessary product

112 PTQ Q1 2014

kbc.indd 9

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 reneries, the resulting liquid products
have poor C/H ratios. Therefore,
additional hydroprocessing hydraulic 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 distillate 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 kerosene range material to the diesel
pool, which impacts cetane index
and hydroprocessing unit severity
and hydraulic requirements.

To illustrate the observations, principles,

and
recommendations
discussed within this article, two
case studies are presented.

Refinery reconfiguration

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 products to multiple regions. Therefore,
to select the correct renery conguration 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 streamline the design process, thereby
reducing overall project timetable.
In addition, the selected conguration has been stress tested against
alternate
crude
slates
and
operational
scenarios,
thereby
demonstrating the robust congurations ability to meet future
processing objectives.

www.eptq.com

11/12/2013 13:02

Tight oil evaluation

A North American refiner was in

the process of examining the
impact of processing tight oil in
their facility. The refiner wanted to
understand not only the amount of
tight oil the existing facility and
configuration could process, but
also how strategic investment
options would impact the amount
processed.
A Petro-SIM simulation of the
existing configuration 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 refiner 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 opportunities that were not immediately
apparent, due to Petro-SIMs ability
to track key stream qualities and

understand the second and third

order effects of a crude slate change
on the refinery 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 specific properties and
contaminants is critical in modern
refineries. 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 challenge is converting this raw
data into useful information. The
key consideration is applying the
tools and methodologies that are
fit for purpose and allow the
refiner to meet economic, reliability,
environmental,
and
safe
operating targets.

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, CC01-153.
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.

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

www.eptq.com

kbc.indd 10

09/12/2013 14:49:02

PTQ Q1 2014 113

11/12/2013 13:02

sabin.indd 1

27/02/2013 15:01

New process arrangements for upgrading

heavy oils and residua
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

n this article, we consider

the results of integrating residue
fluid catalytic cracking (RFCC),
hydrocracking (HCR) and hydroconversion (HRH). A target refinery
was simulated and calibrated, then
the HRH unit patented by the
Research Institute of Petroleum
Industry (RIPI) was added to these
units. After validating HysysRefinery 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 flow 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 refined
products indicate that middle distillates (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

www.eptq.com

ripi.indd 1

gasoline is a crucial subject, so any

improvement in the refinery
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 RIPI2 to convert residue and heavy cuts to lighter
products.
Hysys-Refinery is a simulation
tool commercialised by KBC
Advanced Technologies and AEA
Technology-Hyprotech. This simulator has made significant advances
in detailed representation of reactor
sections. Mohaddecy and Sadighi3
have demonstrated the ability of
this software to simulate the HCR
and catalytic reforming units. Lee4
used Hysys-Refinery 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 residue 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 maximise
ethylene
and
propylene
production.6
Initially, the target refinery with
HCR and RFCC units was simulated using the HCR-Sim and
FCC-Sim reactor modules available
in the Hysys-Refinery 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-Refinery as
a yield reactor, to satisfy the yields
reported by the licensor. Finally,
while keeping constant the feed
quality and flow 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 upgrading extra heavy oil. The HRH
process recently developed by RIPI
produces higher valued products,
which alternatively could be used
as feedstock for refineries.7 The
main objective of this process is the
break-up of high molecular weight
hydrocarbons
to
light
and
medium molecular weight products. 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

PTQ Q1 2014 115

11/12/2013 13:08

hydrogenation
and
operating
conditions allow higher conversion
without coking and excess polymerisation. The HRH process has
high flexibility with regards to the
type of feed and the amounts of
sulphur and heavy metal contents
in the feed. The process can eliminate all the heavy metals and
almost 50% of the sulphur components 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.

HC offgas
HC LPG

Hydrocracker

HC light naphtha
HC heavy naphtha
HC kerosene

HC yield HC yield
(mass) (vol)

Splitter

HC diesel
HC off test

HCR

HC

HFeedM combined feed residue

HC
fresh
feed

Recycle
adjust
Recycle
feed

Recycle
input

Recycle
LVGO

Base case

FCC yield FCC yield

(mass) (vol)

HVS

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).

FCC
offgas
FRCG

Treated
residue

FCC
Mixer 1

HGO

FCC
Mixer 2

OCT
LPG
CDU
LPG

Raw FCC feed

LCO

Unstabilised LPG

CSO
RFCC
coke

Residue fluid
catalytic cracker

Figure1 Arrangement of units and streams in the base case

HC offgas
HC LPG

Hydrocracker

HC light naphtha
HC heavy naphtha
HC kerosene

HC yield HC yield
(mass) (vol)
HCR

HC
fresh
feed

HC diesel

Splitter

HC off test

HC

HFeedM combined feed residue

Recycle
adjust
Recycle
feed

Recycle
input

FCC yield FCC yield

(mass) (vol)

HVS
Treated
residue
HGO
OCT
LPG
CDU
LPG

HRH offgas
HRH gasoline

A
HRH feed
mixer

Recycle

LVGO

HRH yield HRH yield

(mass) (vol)

HRH diesel

HRH feed

FCC
offgas

HRH FCC feed

Hydroconverter

FRCG
FCC
Mixer 1
FCC
Mixer 2

Raw FCC feed

Unstabilised LPG

Residue fluid
catalytic cracker

LCO
CSO

RFCC
coke

Figure 2 Arrangement of units and streams in the simple series strategy

116 PTQ Q1 2014

ripi.indd 2

www.eptq.com

11/12/2013 13:08

Feed flow rates of HRH and RFCC in

simple series strategy

Feed flow rates in the base case

HCR feed flow rate
Fresh feed rate, BPD
HCR recycle, BPD

24 090
13 150

HRH feed flow rate

HCR off-test, BPD
CSO, BPD

674.3
8951

RFCC feed flow rate

LVGO, BPD
HVS, BPD
Treated residue, BPD
HGO, BPD

10 430
9568
60 860
5344

RFCC feed flow rate

LVGO, BPD
HVS, BPD
Treated residue, BPD
HGO, BPD

5344
10 430
60 860
9568

Feed flow rates of HRH and RFCC in

series strategy

Table 1

RFCC feed flow rate

LVGO, BPD
HVS, BPD
TR, BPD
HGO, BPD
HCR off-test, BPD

10 430
9568
60 100
5344
758.7

HRH feed flow rate

TR, BPD
CSO, BPD

754.3
8927

Table 2

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 feedstock has been kept the same as in
the base case. The block diagram
and feed flow rates of the base case

Table 3

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 refinery.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

specifications of the RFCC and

HCR units the same as in the base
case. The HRH unit patented by
RIPI is flexible towards a variation
in feed quality, so the variation in
the feed specification is deemed not
important for this unit.

Simple series integration

The block diagram and feed flow

rates of a simple series integration
strategy are shown in Figure 2 and
Table 2, respectively. The feed and

HC offgas
HC LPG

Hydrocracker

HC light naphtha
HC heavy naphtha
HC kerosene

HC yield HC yield
(mass) (vol)
HCR

HC
fresh
feed

HC diesel

Splitter

HC off test

HC

HFeedM combined feed residue

Recycle
adjust
Recycle
feed

Recycle
input

HRH yield HRH yield

(mass) (vol)

HRH gasoline

A
HRH feed
mixer

Recycle

HRH diesel

HRH feed

HVS
FCC split
Treated
residue

F treated
residue

FCC
offgas

HRH FCC feed

Hydroconverter

FRCG
LVGO
HGO

FCC-ADJ A

FCC yield FCC yield

(mass) (vol)

HRH offgas

OCT
LPG
CDU
LPG

FCC
Mixer 1
FCC
Mixer 2

Raw FCC feed

Unstabilised LPG

Residue fluid
catalytic cracker

LCO
CSO

RFCC
coke

H treated residue

Figure 3 Arrangement of units and streams in the series strategy

www.eptq.com

ripi.indd 3

PTQ Q1 2014 117

11/12/2013 13:08

HC offgas
HC LPG

Hydrocracker

HC light naphtha
HC heavy naphtha
HC kerosene

HC yield HC yield
(mass) (vol)

HC diesel

Splitter

HC off test

HFeedM

HC
fresh
feed
LVGO-HCR

HCR
HC
combined feed residue

Recycle
adjust

REC-HCR

LVGO-HCR

Recycle
feed

LVGO

REC-S

HRH yield HRH yield

(mass) (vol)

Recycle
input

HRH gasoline
HRH feed
mixer

residue

HRH diesel

HRH feed

Recycle

LVGO-FCC LVGO-ADJ REC-FCC

FCC yield FCC yield
(mass) (vol)
Treated

HRH offgas

FCC
offgas

HRH FCC feed

Hydroconverter

FRCG

HVS
HGO
OCT
LPG
CDU
LPG

FCC
Mixer 1

Raw FCC feed

FCC
Mixer 2

Unstabilised LPG

Residue fluid
catalytic cracker

LCO
CSO

RFCC coke

Figure 4 Arrangement of units and streams in the parallel strategy

product properties for HCR were

the same as in the base case. The
HRH unit takes the HCR off-test
stream and the RFCCs clarified
slurry oil (CSO) as feed. The RFCC
takes the LVGO, HVS, TR and
HGO as feed.

Series integration

The block diagram, feed and product flow rates of a series integration
strategy are shown in Figure 3 and
Table 3, respectively. The feed and
product properties for HCR were
Feed flow rates of HCR, RFCC and HRH
in parallel strategy
HCR feed flow rate
HCR recycle, BPD
LVGO, BPD
RFCC feed flow rate
LVGO, BPD
HVS, BPD
TR, BPD
HGO, BPD
HCR off-test, BPD
HRH feed flow rate
TR, BPD
CSO, BPD

11 530
1690
8741
9568
60 860
5344
1688

0
9021

Table 4

118 PTQ Q1 2014

ripi.indd 4

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 advantage 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 flow of untreated residue
can be redirected to the HRH unit,
and this unit is capable of removing heavy sulphur and metallic
contamination.
Feed flow rates of RFCC and HRH in
residue upgrading strategy
RFCC feed flow rate
LVGO, BPD
HRH FCC feed, BPD
HVS, BPD
TR, BPD
HGO, BPD
HCR off-test, BPD

8741
4660
9568
56 170
5344
1688

HRH feed flow rate

TR, BPD
CSO, BPD

4686
8983

Table 5

Parallel integration

The block diagram and feed flow

rates of a parallel integration strategy 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 flow

rates of a residue upgrading strategy are shown in Figure 5 and
Table 5, respectively. The feed for
HCR was the same as for the parallel 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-Refinery
simulator.

www.eptq.com

11/12/2013 13:08

There is more
to Metso than
meets the eye.

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 field 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.

Discover more at www.metso.com/oilandgas/flowcontrol

Neles Jamesbury Mapag

metso.indd 1

10/12/2013 12:43

HC offgas
HC LPG

Hydrocracker

HC light naphtha
HC heavy naphtha
HC kerosene

HC yield HC yield
(mass) (vol)

HC diesel

Splitter

HC off test

HFeedM

HC
fresh
feed
LVGO-HCR

HCR
HC
combined feed residue

HRH yield HRH yield

(mass) (vol)

Recycle
adjust

REC-HCR

LVGO-SP

Recycle
feed

LVGO

REC-S

Recycle
input

HRH gasoline
HRH feed
mixer

Treated
residue

F treated
residue

Recycle

HRH FCC feed

FCC
offgas

FCC yield FCC yield

(mass) (vol)

Hydroconverter

FRCG

HVS
FCC
Mixer 1

HGO

H treated residue

LCO

Raw FCC feed

FCC
Mixer 2

OCT
LPG
CDU
LPG

HRH diesel

HRH feed

LVGO-FCC LVGO-ADJ REC-FCC

TR-SP

HRH offgas

CSO

Unstabilised
LPG

RFCC
coke

Residue fluid
catalytic cracker

FCC feed recycled

HRH-REC

Figure 5 Arrangement of units and streams in the residue upgrading strategy

Gasoline increase,
%

Figures 6 and 7 show the percentage 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

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 operating 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

6
5
4

4.96%

4.88%

6.98%

5.51%

3
2
1
0

Simple series

Series

Parallel

Residue
upgrading

Integration cases
Figure 6 Comparison to base case of growth in gasoline production

120 PTQ Q1 2014

ripi.indd 5

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 considerably
in
different
integration
strategies. As was mentioned
before, the product quality of RFCC
unit is closely tied to its feed quality, 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.

www.eptq.com

11/12/2013 13:08

Feed quality of HCR unit in base case and integration strategies

HCR feed
Base
Simple series
Series
Parallel
Residue upgrading

Ni+V, ppmwt
2.98
2.98
2.98
3.18
3.18

Sulphur, wt%
1.222
1.222
1.222
1.323
1.323

IBP, C
309
309
309
285.5
285.5

FBP, C
516.4
516.4
516.4
516.3
516.3

Table 6

Diesel increase,
%

60
50
40
30

53.69%
39.89%

41.27%

37.53%

20
10
0

Simple
series
Figure 7 HGO crude
exchanger

Series

Parallel

Residue
upgrading

Integration cases

market considerations and by the processes might increase in the

Figure
7 Comparison
to base
of growth infuture,
diesel production
and the fouling of units
downtime
required
for case
a mechanical cleaning turnaround. With a might become a more serious
downtime of 15-20 days, it is much concern, potentially impacting a
Feed quality
RFCC
in baselarge
case and
integration
number
of strategies
production days.
more economical
to of
run
theunit
units
under non-optimised conditions By regular application of online
RFCC feed
ppmwt This,
Sulphur, cleaning,
wt%
IBP,the
C unit can
FBP,always
C
be
rather
than lose Ni+V,
production.
Base
14.723
1
264.6
592.6
operated
under
clean
conditions.
In
however,
results
in
energy
losses,
14.723
1
264.6
592.6
Simple series
addition,
the
vast
majority
of
giveaway
and
capacity
reduction,
Series
14.615
1.006
264.6
599.5
267.3
592.6
Parallel negatively 14.874
mechanical
work, which
takes up a
which
impact unit0.9667
Residue upgrading
13.734
1.088
267.3
significant
number of 592.1
days during
economics.
The introduction of online clean- unit shutdowns, can be replaced
Table
ing, 7by cleaning the unit in 24 with an online cleaning process that
hours on an oil-to-oil basis, allows requires no opening of equipment
References
Heavyentry,
Hydrocarbonaceous
Feedstock,
and can be carried
out
for the recovery of losses and the ofor man
1 Bhaskar,
G, Balarman
K S, European
Patent
Application,
EP 1754770A1,
in as little
as 24
hours oil-to-oil.
operationM,ofValavarasu
units under
improved
Advantages
of mildreliable
hydrocracking
FCC feed Research Institute of Petroleum Industry.
and more
conditions.
-Furthermore,
a pilot plant study,
Petroleum Science
M, Sadighi S, Mohaddecy S R,
turnarounds
can and
be 8 Bahmani
Acknowledgement
Technology, 21, 9 & 10, 1439-1451, 2003.
Mashayekhi
M, Hydrocracker parametric
avoided or rescheduled with Special thanks to all the Lotos team for the
2 European
patent
application,
EP sensitivity study, PTQ, Q2 2009.
reduced downtime.
support and
valuable assistance
provided
1754770A1, Process for Hydroconverting of 9 Sadighi
S, Mohaddecy
S R, Ghabouli
O,
during
the
entire
time
that
ITW
In
the
case
of
turnaround
a Hydrocarbonaceous Feedstock, Research Rashidzadeh M, Optimisation of productOnline
yield
Cleaning
been applied
on site.
improvement,
anNTI company,
additional
Institute
of Petroleum and
2007. and
coke has
formation
in a RFCC
unit, PTQ, Q2
reduction
in
downtime
can
be
3 Mohaddecy S R, Zahedi S, Sadighi S, Bonyad 2010.
Mariusz Hoowacz has been Crude Oil
achieved
by applying
ITWs
H,
Reactor modeling
and simulation
of Sepehr
Sadighi
is Project
Manager,
Catalysis
Distillation
Complex
Manager
with
Grupa
catalytic
reforming,
Petroleum & Coal, 48 (3), and
Division,
Catalytic
improved
degassing/decontaminaLotos Nanotechnology
in Gdansk, Poland, since
2008 and,
since
28-35,
2006.
Reaction
Engineering
Department,
tion technology.
2010, Manager
of the new
distillationResearch
unit 120.
4 Lee R, Leunenberger E, Powell R, Optimizing
the
cat feed hydrotreater/FCCU complex with
Conclusion
detailed
simulation
tools, Online
Process Technology
The results
of ITW
Cleaning
Update, Desulphurization Process, Word
have opened up new possibilities
Refining, Jul/Aug 2001.
for Lotos, whereby online cleaning
5 Dean R R, Mauleon J L, Combined Fluid
can be Cracking
appliedandduring
a plant
run,
Catalytic
Hydrocracking
Process,
to
recover
a
units
performance
and
US patent No: 4426276, 17 Jan 1984.
improve
the
level
of
operational
6 Talman J, Jonsgma B, Thamprajmachit B,
excellence,
and
in preparation
Cackett
S, Wijk R,
Synergistic
integration offor
FCCa
and
hydroprocessing
upgrading
turnaround,
to facilities
reduce fordowntime
bottom
the barrel, Asiascope
PacificofRefining
and theofturnarounds
work.
Technology
Lumper.
These Conference,
options 2001,
are Kuala
increasingly
7 Kadiev
K M,asMezhidov
V K, of
Zarkesh
J,
important,
the amount
opporMasoudian
S
K,
Process
for
Hydroconverting
tunity
crudes
the
company

www.eptq.com
www.eptq.com

ripi.indd
itw.indd 65

Institute
of Petroleum
Tehran,
Rafa Zaprawa
was Industry
VDU (RIPI),
and propane
Iran.
He holds unit
a PhD
in chemical
engineering
deasphalting
Shift
Team Manager
with
from
Universiti
Grupa
Lotos inTeknologi
Gdansk,Malaysia.
Poland, from 2001Email:
@ripi.irManager of the crude oil
2008, Sadighis
then Deputy
Reza
Seif complex.
Mohaddecy is Project Manager,
distillation
Catalysis
and is
Nanotechnology
Division,
Marcello Ferrara
the Chairman of ITW.
With
Catalytic
Reaction Engineering
Department,
27 years experience
in the petroleum
business,
RIPI.
He holds
MS in chemical
including
oil aexploration
and engineering
production,
from
Sharif University
of Technology.
refining,
petrochemicals,
transportation,
Email:
Seifsr @ripi.ir
and energy
production, he holds a PhD in
Kamal
Masoudian
is Project
Manager, Catalysis
industrial
chemistry
and international
patents
and
Nanotechnology
Division,
Catalyst
for new processes and additive compositions
Characterization
andcontrol
Evaluation
Department,
for environmental
and for
improving
RIPI.
He holds a BSc in chemical
engineering.
petroleum/petrochemical
processes.
Email:
@ripi.ir
Email:masoudiansk
mferrara@itwtechnologies.com

Non-intrusive flow
measurement

up to 400C
Trouble free operation at
extreme pipe temperatures
No clogging, no abrasion,
no pressure losses
Installation and maintenance
without process interruption
Independent of fluid or pressure
Hazardous area approved

Field-Proven at Refineries
Heavy crude Oil
Atmospheric Distillation
Vacuum Distillation
Coker & Visbreaker Feed
Fluidized Catalytic Cracker
Bitumen

refinery.flexim.com

PTQ Q1 2014
Revamps
2013 121
63

11/12/2013
10/09/2013 20:18
17:06

enersul.indd 1

6/12/12 13:25:03

Increasing conversion and run length

in a visbreaker
Combining anticoke/antifoulant treatments with monitoring technology enabled
a refinery to keep its visbreaker unit at the best process severity for any feed
MATTEO VIRZI ISAB
MARCO RESPINI Baker Hughes

isbreaker
economics
are
mainly based on achievement
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 fouling control and conversion.
This article presents several
advances that were put in place for
the ISAB Priolo refinery visbreaker,
resulting in improved performance
after
successful
results
were
obtained in the past1 and met all of
the desired targets. This was possible due to Baker Hughes VisTec
technology and the capability of the
ISAB refinery to use this technology in a very effective way.

Visbreaking (thermal cracking)

limitations due to fouling

The main impact as a result of thermal 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

www.eptq.com

baker hughes.indd 1

Figure 1 Thermal cracking of asphaltenes during visbreaking

Saturates

Lighter
saturates

Resins
Asphaltenes

Low MW
asphaltenes

Carboids
coke

Figure 2 Thermal cracking reactions in

visbreaking

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
paraffins by thermal cracking of
heavier molecules contributes to
the decrease in stability. Thermal
cracking reactions that occur in
visbreaking can be schematically
simplified (see Figure 2).

The asphaltene cores (low molecular 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 naphthenoaromatics 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 polymerise, 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 temperatures. As a result of the process
temperatures used, visbreaking
tends to produce high levels of
coke particles.
A very interesting insight and
confirmation on the impact of thermal cracking comes from the direct
measurement of changes in the

PTQ Q1 2014 123

12/12/2013 21:14

Sbn and In

Sbn >> In
High stability reserve

Sbn
In

Feed residuum

Vistar

Process severity/conversion
Figure 3 Solubility blending number (SBn) and insolubility number (In) changes from
visbreaker feed resid to vistar

Conversion loss

Fouling
Increased maintenance cost
Shorter run lengths

Fouling increases

Optimal operational
window

Conversion increases

Figure 4 Rate of fouling vs conversion

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 classification captures
a broad range of molecular structures. 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 fouling. 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).

Coke

Limitations in visbreaker conversion

Furnace outlet temperature

Figure 5 Coke generation vs furnace outlet temperature (from experimental data on
industrial visbreaker)

124 PTQ Q1 2014

baker hughes.indd 2

For any blend of processed crudes

there is an optimal conversion for
the downstream visbreaker.
Fouling rates have an exponentially increasing trend with severity
and conversion. At low severities,
the impact of the conversion
increase is very limited. As conversion limits are reached, any further
minor increase in severity has drastic consequences on fouling rates,
resulting in severe reduction in unit
run length. This is shown schematically in Figure 4.

www.eptq.com

12/12/2013 21:15

www.eptq.com

baker hughes.indd 3

Fouling factors increase rate,

h.m2.C/Kcal day

4.0e6
3.5e6
3.0e6
2.5e6
2.0e6
1.5e6
1.0e6
5.0e7
0.0e+0

Vistar stability reserve, VSI

Figure 6 Heat exchanger fouling rates vs stability reserve of vistar

800

probe signal derivative

Fouling can severely impact the

heater run length due to coke deposition. 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 stability 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% without a reduction in severity or
conversion; however, they give the
best results in high fouling situations 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 treatment 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 acceptable vistar stability reserve limit is
necessary to run the unit at maximum conversion, together with an
effective method to measure the
stability.
This was achieved at ISAB by
following Baker Hughes proprietary guidelines and by continuous
and prolonged experience and optimisation on this unit.
Another very important aspect is
the refiners technology and the ability of the operations staff to adjust
process parameters related to severity 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.

700

Destabilisation point

600
500
400
300
200
100
0
+100

0.5

0.7

0.9

1.1

1.3

1.5

1.7

1.9

VisTec stability index

Figure 7 Asphaltene destabilisation point

ISAB refinery 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.

For any blend of

processed crudes
there is an optimal
conversion for
the downstream
visbreaker
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 microscopy based p-value. The method
uses light absorbance in the
near-infrared range to detect
asphaltene destabilisation.
The VSI has four main advantages over other methods:

It is extremely accurate and
repeatable
It is designed specifically for the
determination of visbreaker feed
and tar stability
It gives rapid results, allowing
refiners to react quickly and maintain optimum operating conditions

It has been developed and
proven in use on many visbreaker
units throughout the world.

PTQ Q1 2014 125

12/12/2013 21:15

10

Number of particles,
thousands

9
8
7
6
5
4
3
2
1
0

10

Particle size, m
Figure 8 Example of the VCI particle size population analysis

Conversion,%

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 measurement 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

Visbreaker tar and heavy fuel oil

stability.

ISAB visbreaking unit targets

The rst run at the ISAB visbreaker

with
the
VisTec
programme

The achievement of
targets was based on
more severe control
limits compared to
fouling indicators
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:

Target

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 turnaround 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 conversion 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.

11
14
/9
/2
01
1
23
/1
2/
20
11
1/
4/
20
12
10
/7
/2
01
2
18
/1
0/
20
12
26
/1
/2
01
3
6/
5/
20
13

6/
6/
20

26
/2
/2
0

11

ISAB visbreaking unit constraints

Figure 9 Conversion vs target

126 PTQ Q1 2014

baker hughes.indd 4

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.

www.eptq.com

12/12/2013 21:15

www.eptq.com

baker hughes.indd 5

5/
6/

/1
26

20

01
/2

20
0/
/1

18

13

12

2
01
/7

10

1/

4/

/2

20

20
2/
/1

/2
/9

12

11

1
01
23

20
6/

14

26

6/

/2

/2

01

11

Conversion,%

Atmospheric distillates
Vacuum distillates

Figure 10 Atmospheric and vacuum conversions

VSI stability reserve

Lost conversion

3
25

/6

/2

01

2
7/

12

/2

01

2
21

/5

/2

01

1
3/

11

/2

01

1
01
/2
/4
17

/9

/2

01

Excessive fouling

29

6/
5/
20
13

VCI, wBPI unit volume

concentration coke

Figure 11 VSI stability of visbreaker residue

18
/1
1/
20
10
17
/4
/2
01
1
14
/9
/2
01
1
11
/2
/2
01
2
10
/7
/2
01
2
7/
12
/2
01
2

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 fouling was, as observed in the previous
run, a concern, resulting in limitations 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 management of the heater burners, allowing
the avoidance of excessive uctuations 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

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 stability reserve (VSI) and coke particle
generation (VCI), the throughput
was on average rather high and

PTQ Q1 2014 127

12/12/2013 21:15

350
300

Unit feed rate

250
200
150
100
50

13

2
1/

2/

20

01

/2

01

12

/2

1/

10
1/

8/

20

12

12
20

1/

1/

6/

20

12

12
20

1/

2/
1/

4/

1
01

1
/2
12
1/

1/

10

/2

01

11
20

11

8/

20

1/

6/
1/

1/

4/

20

11

Figure 13 Unit feed rate

y = 0.132x 4878.4
R2 = 0.3851

VKTIMS.PV heating normalised

VKTIMS.PV heating

700

Temperature, C

650
600
550
500
450

1/
4/
20
12
10
/7
/2
01
2
18
/1
0/
20
12
26
/1
/2
01
3
6/
5/
20
13

01
/2
12

/2

01
23
/

11

14
/9

20
6/
6/

26
/

2/

20

11

400

Figure 14 Optimisation of burners and impact on skin temperature

Skin temperature model results

VKTR1_9M.PV VKTIM5.PV
Skin modelling result
soaking
heating
Average coking rates, C/day
0.179
0.135
Impact of feed rate, C/(ton/h)
0.134
0.094
Heater outlet temperature impact, C/C 2.214
2.157

VKTIM1.PV
heating
0.169
0.103
2.008

Table 1
Skin temperature model from previous run

Skin modelling result

Average coking rates, C/day
Impact of feed rate, C/(ton/h)
Heater outlet temperature impact, C/C

Table 2

128 PTQ Q1 2014

baker hughes.indd 6

VKTIM6.PV
soaking
0.467
0.206
3.324

VKTIM5.PV
heating
0.237
0.111
3.791

VKTIM3.PV
heating
0.216
0.191
2.089

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 typically 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, indicating 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 temperature control in this run.
The good results obtained in the
latest run are also due to excellent
optimisation of burner operation
throughout the run.

www.eptq.com

12/12/2013 21:16

KALDAIR

j zink.indd 1

12/12/2013 09:39

13

7/

6/

3/

20
1/

20

13

2
/2

7/

8/

11

9/

/2

20

01

12

2
01

2
/7
10

/5

/2

01

2
01
11

12

12

/3

/2
/1

/2

01

Delta P trend, mmHg

11
20
1/
/1
13

Figure 15 Wash grid delta P trend

70

HVGO, ton/h

60
50
40
30
20
10

13
20

2
1/

2/

01
/2

12

1/

/2

01

12
20

10
1/

1/

8/

12
20

6/

4/
1/

1/

20

12

12
20

2/
1/

12

/2

01

1
1/

1/

10

/2

01

11
20

11

8/
1/

20
6/

1/

1/

4/

20

11

Figure 16 HVGO yield

40

Wash oil rate,

m3/h

35
30
25
20
15
10

13
/1
1/
20
11
12
/1
/2
01
2
12
/3
/2
01
2
11
/5
/2
01
2
10
/7
/2
01
2
8/
9/
20
12
7/
11
/2
01
2
6/
1/
20
13
7/
3/
20
13

Figure 17 Wash oil rates

Some critical skins showed an

increased trend in the rst period
and the renery was able to properly adjust burner ring and ame

130 PTQ Q1 2014

baker hughes.indd 7

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 properly adjust burner ring and ame
distribution in order to minimise
the impact of bad heat ux distribution on skin temperature increase
rates.
Another important aspect of the
unit is the vacuum wash grid. This
is a critical section in terms of fouling. 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 deposits, increasing delta P and possibly
allowing deposits to degrade on the
grid. In these situations, a
controlled decrease in wash oil can

www.eptq.com

12/12/2013 21:16

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 technique 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.

which provided refinery personnel

with the right information to be
able to constantly keep the unit at
the best process severity for any
processed feed.

A controlled decrease
in wash oil can result
in better handling of
fouling, and at the
same time increase
HVGO directly

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 treatments with monitoring technology,

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.

3 Respini, Jones, Spanu, Sesselego, Avoiding foul

play, Hydrocarbon Engineering, Nov 2006.
Matteo Virzi is Senior Technology Manager
with ISAB Priolo refinery 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 refinery and petrochemical
process improvements in fouling control.
He has 15 years of refining experience and
is currently involved in developing new
technologies for improving refinery conversion
processes. With extensive experience in
asphaltene related problems in oil production
and refining, he is an inventor of five 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 field 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

alves
Best V 67
8
since 1

www.eptq.com

baker hughes.indd 8

PTQ Q1 2014 131

13/12/2013 11:38

HydroCOM
the 1000th!
Long-time customer and partner JX Holdings Inc.
from Japan orders anniversary HydroCOM
Experience real capacity control and energy savings
HydroCOM is an efficient, 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.

For more information:

compressor-mechatronics@hoerbiger.com

www.hoerbiger.com

HydroCOM_CE_Buyers_guide.indd
1
hoerbiger.indd
1

11.12.2013
14:45:40
11/12/2013
14:59

Troubleshooting refinery equipment with

multiphase CFD modelling
Computational fluid 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

uch of the common process

equipment
in
refineries
today was designed and
built according to traditional empirical design methodologies that were
developed decades ago. Without an
intimate understanding of the
complex flow patterns present
within a given system, designers
had to rely on conservative assumptions and trial and error to ensure
that equipment met design requirements. Modern computational fluid
dynamics (CFD) tools allow designers to pull back the veil on complex
internal flows, but their use has
been limited by available computing
power. As computing power continues to increase, CFD is becoming a
practical tool for industrial scale
problems. Through a deeper understanding 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 modifications can eliminate the need to
design and fabricate new equipment, resulting in significant cost
savings without compromising
performance. This article will
discuss several cases demonstrating
the application of CFD to traditional process equipment. Each
case presented will discuss the
motivation for the use of CFD, the
assumptions required to yield a
practical and robust CFD simulation, 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 problems is an ability to simplify a

www.eptq.com

pcs.indd 1

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 fidelity
simulation is often unjustifiable or
unattainable for many industrial
problems. In most cases, a simplified
modelling
approach
specifically developed to examine
the variable(s) of interest is the
most efficient practice. By carefully
considering all of the independent
and dependent variables relevant to

A simplified
modelling approach
developed to examine
the variable(s) of
interest is the most
efficient practice
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 physics
complicates
the
setup
of a simulation, significantly
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 discovered
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
significant
expense,
replacement of the new drum
would have been difficult due to
space restrictions. It was hypothesised that internals could be added
to the drum to adequately increase
the vapour-liquid separation. For
verification, CFD could be used to
confirm the effectiveness of any
design changes that would see the
drum operate while significantly
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 modifications 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
flow at the inlet, breakup/coalescence of the liquid droplets,
formation of a liquid film 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 simplified,
however,
by
carefully
considering the variables of interest. The functions of the liquid

PTQ Q1 2014 133

11/12/2013 14:15

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 variables 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 film
thickness are not significant to the
overall function of the drum and
were not modelled, as their omission did not significantly affect the
variables of interest.
Modelling was further simplified
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 counterintuitive that two models would be
more efficient 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 first 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
flow was low. Small, light particles
follow vapour flow streamlines
more closely than larger particles
because they have less momentum
relative to the drag caused by the
bulk vapour flow (lower Stokes
number). With this principle in
mind, a conservative particle size
should be smaller than the maximum allowable droplet size. These
conservatively sized particles were
injected with the vapour at the
separator
inlet
and
tracked

134 PTQ Q1 2014

pcs.indd 2

throughout the domain as the

vapour travelled from the separator
inlet to the outlet. A simplifying
assumption was that the particles
underwent partially elastic collisions (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 flow 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 therefore not consequential to the
design. The design was modified
and simulated iteratively until no
particles escaped through the drum
outlet.

Use of computational
fluid dynamics
allowed many design
iterations to be
evaluated within a
few days to arrive at
an optimal solution
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 fluid
(VOF) model to track a well defined
vapour-liquid
interface.
Liquid
volumetric flow rates were significantly lower than gas volumetric
flow rates, and liquid entering and
exiting the drum was not significant
to the problem of maintaining a
stable liquid layer. The model was
therefore built with no liquid flow
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 flow. If
the flow agitated the liquid layer to
the point where liquid mass
escaped through the outlet, the
design was modified. Stabilising the
free liquid layer with high vapour

flow rates proved more difficult

than trapping the initially entrained
particles. The high velocity vapour
flow tended to re-entrain significant
quantities of liquid. However, a
design was developed that could
satisfy both requirements.
The use of these two complementary CFD simulations led to a robust
design and confidence 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 comparison, 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 diameter 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 flawed. The overhead
receiver was designed with the
inlet at the bottom of the drum,
presumably to reduce the complexity 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 fluid in the
drum would interrupt the separation 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
flow velocities and fluid residence
time. Traditional methods assume
plug flow through the vessel, meaning that each phase has a uniform
cross-sectional velocity equal to the
volumetric flow rate divided by the

www.eptq.com

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 residence 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-section of the drum fails to hold. If
the plug ow assumption is not
met, some fraction of the liquid in
the separator may experience residence 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 variables
of
interest
(anomalous
cross-sectional velocities), so this
was omitted from the simulation.
Any regions with signicant movement 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

pcs.indd 3

Raw gasoline
volume fraction
1.00e+000
9.47e001
8.95e001
8.42e001
7.89e001
7.37e001
6.64e001
6.32e001
5.79e001
5.26e001
4.74e001
4.21e001
3.68e001
3.16e001
2.63e001
2.11e001
1.58e001
1.05e001
5.26e002
0.00e+000

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

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 hydrocarbon 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
Bottom inlet
agitates liquid
in drum

Water in reflux
carries salts
Salt deposition
plugs trays

Figure 2 Poor separation in the overhead

receiver leaves water in the reflux, 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

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 mechanisms by which this device works
and to look for simple improvements 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 placement of the pipes. By fully
understanding the mechanisms that
allow this inlet device to achieve
good results, better design decisions can be made in future
installations.

PTQ Q1 2014 135

11/12/2013 14:16

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 differences 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 essentially 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 sufciently large (high Stokes number),
the particles tended to detach from
the vapour ow and follow a trajectory 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

Inlet device

Quench

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

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 distribution (see Figure 4). The simulation
nonetheless provided evidence to
explain the decreased coke formation 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 redirected 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

Velocity, ft/s
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
0.00e+000

Figure 4 Velocity profile on a vertical plane within the tower. The case with the inlet
device (left) is similar to the base case (right). Significant quantities of vapour are not
directed upward by the inlet device

136 PTQ Q1 2014

pcs.indd 4

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 distributor 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 scatter 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 formation for two reasons: 1) the liquid
impacting the back of the tower
did so over a larger area, concentrating 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

www.eptq.com

11/12/2013 14:16

maximum, rather than in the stagnation region at the

point of jet impingement, coke formation is minimised. 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.

Partner with
the Best

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 scattering are the two key principles that allow this device
to prevent excessive coke formation will lead to
future designs that specifically maximise these
phenomena rather than designs that attempt to
improve the device through misguided attempts to
affect the distribution of the flow within the tower.
CFD is an extremely useful and increasingly practical tool for improving the design of and increasing
the understanding of common process equipment. By
carefully defining a problem and the variables of
interest, simulations can be built that give good information where necessary without incorporating
unnecessary complexity. Finally, it is worth mentioning here that CFD inputs, assumptions and results
should be carefully scrutinised. Commercial CFD software 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 computational 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.

With over 50 independent subsidiaries and more than 220 engineering

and sales offices 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 companies to form the SAMSON GROUP.

A01120EN

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 refinery industry
worldwide.
Steve White is a Chemical Engineer with Process Consulting
Services. He has more than 37 years of process design experience in
refinery 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

www.eptq.com

pcs.indd 5

11/12/2013 14:16

SOGAT2014 Ad1 Full Page Bleed 303mmX216mm for PTQ.pdf 1 12/3/2013 12:06:10 PM

CM

MY

CY

MY

sogat.indd 1

09/12/2013 11:32

High-fidelity operator training simulators

Integrating a high-fidelity simulation model with a real plant control system
requires some finessing to produce a true high-fidelity operator training simulator
MARTIN SNEESBY
APESS

he current trend in operator

training simulators (OTS) is to
connect
a
high-fidelity
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-fidelity OTS that
will maximise training value.
However, success is not guaranteed
by the project model, and there are
often some complex and difficult
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 congratulate themselves on their decisions
and look forward to an OTS that
meets
all
their
expectations.
Thereafter, the project implementation team will be put to work
following a fairly standard project
model (see Figure 1).
The first major implementation
task is model development. A new
dynamic process model will require
an enormous amount of data to be
processed and configured within a
model framework. This needs the
attention of simulation experts in
order to ensure robustness, accuracy and adequate simulation

www.eptq.com

Build high-fidelity
dynamic
simulation model

Modify existing
high-fidelity dynamic
simulation model
Configure plant
DCS for OTS

OR

Configure or
emulate plant
ESD

Validate
model

Integrate
components

Test
integrated
system

Delivery
(FAT and SAT)

Figure 1 Standard project model for an operator training simulator

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 significant work to adapt it to
the specific OTS requirements,
current plant configuration 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-fidelity model.

Process understanding is almost

always crucial to any problem solving on an OTS, including issues
with the DCS and ESD functionality, 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 fits all
execution model
Ring-fencing parts of an existing
process model against any change.
These approaches should be
avoided, even if they appear superficially cost effective.
Pre-existing models often come
with baggage, too. Rose-tinted

PTQ Q1 2014 139

glasses can make old models unrealistically 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 floundered
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 testing the models dynamic behaviour.
The model responses should be
reasonable and realistic for a wide
range of transients. Judging reasonableness and realism is difficult.
Typically, experienced operating
staff are asked to make this assessment and, while this is usually the
best option, it is not quantifiable
and is subject to preconceptions
and other human factors. For example, it is very common to have
experienced operators reporting
that a model reacts more slowly
than the real plant. This misconception 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 difficult 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 sufficiently confident
of the models fidelity, or they feel
the schedule is more pressing than
quality control. However, skipping
model validation (or failing to take
sufficient notes during the process)
is usually the prelude to a projects
disaster.
In parallel with model development (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) mechanism to ensure the project
hardware costs are minimised so
that redundancy (unnecessary in an
OTS) is not built in accidentally.

140 PTQ Q1 2014

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 highfidelity dynamic simulation model
of the process and an OTS version
of the real plant DCS. The integration 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 difficulty 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 finish 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 first
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

will conclude that the tuning is

right so that the model is wrong.
The simulation engineer will then
be asked to fix the high-fidelity
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
efficient (although it is often possible to get good, noise-free data
from a high-fidelity 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 introspection. 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

www.eptq.com

system response. Exactly how good

is the high-fidelity dynamic simulation model?

Misleading
plant data

Sources of model error

High-fidelity 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 configure these
models are required to accept and
make many assumptions and estimates.
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 sufficient 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 (flooding, side reactions, and so on)
Stability
and
robustness
constraints imposed by the simulation mathematics
Misleading plant data that

suggest
correlations
between

www.eptq.com

Unskilled
simulation
engineer

Inadequate unit
operation models

Model
error

Inaccurate
VLE

Estimated
data

Equipment
design margins

Hidden
assumptions

Poor engineering
assumptions

Figure 2 Sources of plant model mismatch with high-fidelity OTS models

actions and effects where no direct

connection exists or responses
influenced 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 application
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 efficiency 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 highfidelity dynamic model. For example, variation in throughput with
stage efficiency (or weir height) is
often seen on process plants, but is
not predicted. Similarly, flooding
(like surge on a compressor)
involves pseudo-random phenomena
that are beyond most models. Even

the hold-up on a tray can be difficult 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 thermodynamic models used in a
simulation as a potential source of
uncertainty. Hopefully, the modelling engineering will be sufficiently
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 (homogeneous) volumes even though we
know that perfect mixing is
unlikely to be realised. Pure deadtime 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

PTQ Q1 2014 141

Tp = 20
td = 4
Kp= 1

Response

Kc = 5

Ti = 4

1
0.5
0
0.5
0

Ti = 11

0.4

0.4

0.3

0.3

0.2

0.2

0.1

0.1

0.1

0.1

0.2

50

Response

Time

Kc = 3

150

0.2

0.4

0.4

0.3

0.3

0.3

0.2

0.2

0.2

0.1

0.1

0.1

0.1

0.1

0.1

50

100

150

0.2

50

100

150

0.2

0.4

0.4

0.3

0.3

0.3

0.2

0.2

0.2

0.1

0.1

0.1

0.1

0.1

0.1

0.2

0.2

100

Time

150

50

100

100

150

50

100

150

Time

50

Time

0.4

50

Time

Time

Time

Response

100

0.4

0.2

Kc = 1

Ti = 20

150

Time

0.2

50

100

150

Time

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)

than definitive different valve

manufacturers will use different
equations to calculate flow, 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 engineering books have been written on
troubleshooting plant problems.
Books by Norman Lieberman and

142 PTQ Q1 2014

Henry Kister are especially noteworthy. 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 conditions? Of course, we do not. We
just occasionally get seduced by the
accuracy of high-fidelity models
and forget that all model predictions are subject to uncertainty.

Controller tuning parameters

Process control engineers understand that all control loops have to

be tuned for their specific characteristics. Some loops are simple and
need little attention. For example, a
simple flow controller is almost
guaranteed to work well with a low

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
specific set of controller tuning
parameters (responses are indicative 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

www.eptq.com

Kc = 3
Ti = 1

Response

Kp = 1.5

Tp = 10

1
0.5
0
0.5
0

Tp = 20

0.4

0.4

0.3

0.3

0.2

0.2

0.1

0.1

0.1

0.1

0.2

50

Response

Time

Kp = 1.0

150

0.2

0.4

0.4

0.3

0.3

0.3

0.2

0.2

0.2

0.1

0.1

0.1

0.1

0.1

0.1

50

100

150

0.2

50

100

150

0.2

0.4

0.4

0.3

0.3

0.3

0.2

0.2

0.2

0.1

0.1

0.1

0.1

0.1

0.1

0.2

0.2

100

Time

150

50

100

100

150

50

100

150

Time

50

Time

0.4

50

Time

Time

Time

Response

100

0.4

0.2

Kp = 0.5

Tp = 30

150

0.2

Time

50

100

150

Time

Figure 4 Predicted control responses with fixed controller parameters and variable process responses

more robust against model controller 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-fidelity process simulation

is built from unit operation models
that are based on accepted theory
and configured with design performance 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

www.eptq.com

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 immediate need. These actions, even if
superficially effective, are unlikely
to address the real problem and
often create further issues in themselves because a high-fidelity 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 fixes are excluded, the
modelling engineer must review
the process model with a finetoothed comb. This involves
checking and rechecking all the
input data (thousands and thousands of parameters) and assessing
all the implicit assumptions. A
solution may or may not be available via this route.

In short, tuning the model to

match a desired performance is
difficult to impossible.
The pragmatic engineering solution is to use a modified 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 simulated DCS, but that process is
already sufficiently 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

PTQ Q1 2014 143

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 controller can be recalibrated to match the
model dynamics, the overall
response can be significantly
improved.
It is important to realise the limitations of retuning controllers. If
the model is significantly 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 unstable in the OTS (with plant tuning
constants, see Figure 5) then retuning 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 fix 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 physics and the first law of chemical
engineering (the mass balance).
The
operating
plant
often
provides clues: an exchanger seemingly operating at better than
design performance, a compressor
not delivering the expected head,
possible flooding in a column, and
so on. These issues can often be
addressed by tuning the equipment
parameters, for instance, (increasing the specified 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

144 PTQ Q1 2014

Plant behaviour
0.4
0.3
0.2

0.1
0
0.1
0.2

50

100

150

Time
Response of high-fidelity model
with plant tuning constants

0.5
0
0.5
0

Time
Response of high-fidelity model
with calibrated tuning constants
0.4
0.3
0.2
0.1
0

Conclusions

0.1
0.2

might make the model unstable

under some conditions!
Transmitter time constants are
typically estimated. In fast loops,
these time constants may be significant in the overall loop response.
The approach to equilibrium (for
example, efficiency factors) can also
affect responses. Again, these
values are typically estimated and
some tuning can be justified.
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 dynamics and has negligible impact on
the model mismatch at normal
operating conditions.
Tuning should always be sensible. 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 engineering in the model construction phase
or that the engineer is tuning the
wrong input altogether.

50

100

150

Time
Figure 5 Potential benefits of retuning
controllers to work with a high-fidelity
simulation instead of the actual plant
dynamics

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 estimated inaccurately because they do
not affect the steady state solution
and larger volumes help model
stability. It is tempting to overestimate volumes in order to make
the model robust. A corollary of
this is that reducing model volumes

Model building is a skilled exercise

and should be treated as such. A
high-fidelity model requires deep
process understanding and experience. It is not an off the shelf item.
Modelling engineers must recognise 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 beneficial. 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 specific data
set. Plant tuning constants will

www.eptq.com

often need adjustment to work well

with even the most accurate simulation models. Failure to allow for
controller tuning on the simulator
system may be detrimental to the
overall projects success.

Prognostications

High-fidelity 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 timesteps, more detailed unit operation
models), although there are diminishing returns.
The increase in fidelity of DCS
emulation that is achieved by using
real hardware and control configurations is laudable, but it should
be remembered that a control

Green-field 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
benefit of a high-fidelity OTS for
training is small much can be
achieved with relatively simple, stable
and robust models.

system is complementary to a plant

or simulation. A low-fidelity
process model with a complementary (well tuned) control layer
might provide better overall

responses that a high-fidelity model

with a non-complementary control
layer imported from the operating
plant. Use the high-fidelity control
emulation, but be tolerant of the
process model and accept that
some controller tuning is likely to
be beneficial.
OTSs will continue to become
more accurate, but this will not be
achieved through software developments alone. The more detailed
the simulation, the more data and
understanding that it holds, and,
consequently, the more engineering
effort required. As with any engineering 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

HigH standard valves

for non-standard conditions.
www.zwick-armaturen.de

ww.eptq.com

PTQ Q1 2014 145

EXTEND
YOUR REACH
ATTEND THE 2014
AFPM ANNUAL MEETING
Dont miss out
on the industrys
most important
event of the year!

afpm.org

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
refining industrys most
important players gather to
network for three days of insight,
innovation and solutions. Explore
the challenges and opportunities
that are on the refining industrys
horizon at AFPMs Annual
Meeting in Orlando!
Registration is now open.

AFPM_2014_AM_PTQ_PRINTER.indd
1
000_afpm.indd
1

11/18/13 2:47
PM
09/12/2013
11:25

What your transmitter may not be

telling you
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

n process instrumentation applications measuring pressure, ow

or level, the engineer or technician 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 conditions 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.

Transmitter
Manifold
Impulse
line
PIV x 3

Process line

Figure 1 The process instrumentation loop

consists mainly of a process interface valve
(PIV), impulse line and fittings, a manifold
and the transmitter

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.

Standardisation: a step towards

simplified maintenance

Close coupling

Before we review the process instrumentation loop in detail, let us look

rst at a recently developed alternative. It is an elegantly simple
solution, and if your application
allows you to employ it you should.

www.eptq.com

swagelok.indd 1

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 temperature 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 accessible, 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.

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

If your goal is an optimal design,

there are a limited number of ways
to set up the process instrumentation loop. And yet, in most plants,
there is a multitude of variations,
and many of them are not optimal.
These variations have been developed over time, by different
engineers and/or contractors for

PTQ Q1 2014 147

11/12/2013 14:52

different projects and reasons. Such

a situation can be a drain on your
time and energy. Each system has
different needs in terms of maintenance, 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, maintenance and accuracy. The result is a
high degree of standardisation. For
example, before standardisation, a
rening plant may have 30 different congurations for process
instrumentation loops. After standardisation, the same plant may
have only ve, with each containing the same basic components: a
transmitter mount, manifold system
and redundant pressure measurement. 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 applications. 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 inaccurate transmitter readings (as we
discuss in more detail below). If you

148 PTQ Q1 2014

swagelok.indd 2

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, selfcontained unit, reducing size and weight

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

a single, self-contained unit, which

has these advantages:
Fewer leak points
Flexibility to congure the product using different sized orices
and anged connections
Reduced size and weight, reducing the need for and expense of
structural support of the instrumentation 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 appropriate for all uids, but especially
for those with higher viscosity
when using ball valves. In fact, a
good DBB design should be roddable 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 instrumentation 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

www.eptq.com

11/12/2013 14:51

Normalised HDS temperature,

F

grip-type
tube fittings
not backthe initial investment and on-going from
corrosionof ARDS
(as mentioned
yield selectivity,
in will
particular
configuration
+ RFCC, the
offpropylene
with thermal
cycling
vibra-the catalystTo replacement To
other debris,
yield.
The or
higher
cost will be earlier)
LVGOorfraction
of the such
crudeas is
transmitter
transmitter
tion,
unlike content,
the threaded
fittingsthe much
Burrsto area a VGO
byproduct
of the
hydrogen
the higher
higher than for most hydro- burrs.
routed
hydrocracker
used
with
carbon
steel.
machining
process
during
manufacpropylene
yield.
processing
units.
(HCU)
to
shift
the
yields
ina the
in the
UK, I value
designed
communicate where the module construction of refineries are being Company
promise for
deriving
from
The
third
objective

maintaining
turing.
A
critical
step
during
For
the
types
of
crude
oil
that
The
other
key
features
of
this
desired
direction.
Figure
2
shows
a
water
stripper that
Ammonia
will be installed on the plot
plan.gas applied to the construction of sour
smaller
accumulations
of eliminated
unconvenEqualisation
torange
sulphur
plantconfiguration
temperature
within
a
certain
manufacturing
and
installation
have
been
considered
here,
the
to
maximise
the
schematic
of
the
refinery.
the
unnecessary
features
of
the
unit
Connections between the modules distributed GTL
valveplants. The GTL tional gas that would otherwise be
typical
is
heating
the
should
thefeatures
removal
all gas,
burrs
requires
process
unit production
Figure
2.such as
CCRincreased
content
of
atmospheric
of
petrochemicals
Theinbekey
ofof
this
refinery
are achieved
designed
toby
be similar
in configprocess involves
two operations:are:
the shown
left
underground,
shale
Isolation
impulse
lines.
You
insulate
your
from
all
surfaces.
Otherwise,
capabilities
inthat
acan
number
of13 PSIG
refining
Figure
3wetted
shows
the essentials
of
a
valves
uration
construction
is
conversion
of
natural
gas
to
a
tight
gas,
coal
bed
methane
and
residues
is so
about
12-14
wt%,
and

Process
straight-run
light
naphconfiguration
are:
190F
Temp. 1
Temp. 2 Temp. 3 Temp.
Temp.
1dislodge
Temp.
Temp.
3
process
particularity
hydroimpulse
lines
manually
Ventech
field
they
can
when
manicorrect
sour
water
stripper
design.
relatively
straightforward.
mixture
of carbon
monoxide
(CO) checkstranded
gas
(gas2 tower
fields
located
too to
metals
areunits,
80-90
wtppm.
Thus, tha
along
with
light
catalytic
1 A
vacuum
is the
installed
Sour The
90F
processing
assets.
Feed
in
atcatch
ambient
tracing
or that,
purchase
tubing
that
fold
isis inbrought
service
and
intoorbe
), steam
knowncracker
as back
estimates
withfeed
modularisation,
and hydrogen
far
from
existing
pipeline
infrapretreatment
of
the
ismagnitude
required.
naphtha
(LCN)
in(Hthe
produce
a diesel-type
cut
2
water
60
(15)
of
impact,
ofunit
course,
on
(70-100F,
feed
syngas,
by LPa Fischerapproximately
70%
a factory
project
is
structure).
A in
small,
modularised
has The
been
insulated
in isof
theadepends
scratch
the
valve
stem
orfrom
seat,
HPfollowed
ARDS
specially
Process
heavy straight-run
naph- conditions
processed
the 21-38C)
distillate
hydrothe
particular
feedeven
processed,
sour
feedcut
tank.
Toprocessed
Tropsch
process reforming
to convert the
already
complete
GTL
plantwater
the
flexibility
toheat
be
and
encased
in a polymeric
jacket.the tha
preventing
aLVGO
positive
shut-off.
Burrs
50catalytic
designed
hydrotreater
thatbefore
pretreats
in a(FT)
unit the
treater,
ahas
to be
inherent flexibility
of the
facilitys
the
feed from
(32C)
to
250F
syngas
into
paraffinic
hydrocarbons
are RFCC
shipped
their Figure
installed
close
toalmost
themonoflanges
trapped
4 A manifold,
consisting
of three
Pre-insulated
bundled
tubing
may
be
an90F
issue
in
themodules
feed to the
unitfrom
tocomes
reduce
(CRU)
to
produce
reformate
in also
a two-stage,
full-converoperations
and
the
ability
to
(120C)
requires
about
16
wt%
that
can
be
further
refined
to
facility.
This
greatly
decreases
field
resource
and
then
used
to
process
valves, 40
enables
techniciannaphtha
to
ready
to install inincluding
coiled lengths.
It needle
and
globe pattern
contaminants
sulphur,
Process
heavythecatalytic
sionother
hydrocracker
unit needle
oriented
offset
this
increased
processing
steam

ow,
or
about
1.3-1.4lb
of
produce
a
wide
range
of
construction
time
to
deliver
an
that
resource
locally.
Associated
gas
isolate
and
calibrate
the
transmitter
in
cannitrogen,
be heated
with steammetals
or left
valves.
organometallic
and
(HCN),
pygas
and
reformate
towards
the
production
of
middle
requirement.facility (see Figure 1).
steam
per gallon
of
stripper
hydrocarbon-based
products,
includoperational
(gas
produced
along
with
oil)
is
30
(8)
the
field
unheated,
as
the it application
content
CCR.In Additionally,
increases
the throughMedium
the aromatics
complex to distillates
making
this
move,
proper
which
is close
to a typical
ing
clean-burning,
sulphur-free bottoms,
These
methods
also
facilitate
easy
another
area
of
opportunity
for
NH3 water to
Seat
sealsHVGO
in needleisvalves
requires.
It is content
importantofwith
hydrogen
thepre-inatmosparaxylene
and benzene
The
combined
hydrotreaters
20jet fuel. Speciality
selection
requires
good
understripping
forwith
diesel and
prod- design
disassembly
and a relocation,
if produce
modularised
GTL steam
plants. ratio
This gas
transmitter
in
a
standard
differenAnother
reason
why
valves
leak
is
sulated
tubing
bundles
to
follow
the
s
with
relatively
high
pheric
residue,
thus
improving
the

The
C
vacuum
residue
to
be
pretreated
standing and
modelling
water disposed
strippers.ofThe
E-1byfeed
4
food-grade waxes, sour
necessary,
at some
point of
in (1)the ucts including
is typically
either
re- in
tial
pressure
setup,
the
technician
poor
design.
Most
needle
valves
manufacturers
instructions
for
sealolefins
content
are
either
sold
or
crackability
and
selectivity
of
the
the
vacuum
residue
desulphurisaDN-3651
alternative
feed
feed
10 lubricants can also be
facility as
as a strong
knowl- solvents and
preheater,
uxalternative
pump (P-2)
and
future.
Forwell
example,
a remotely
injection, DN-3551
atre
considerable
expense,
the two
isolation
valves
andthe
employ
aDN-3551
metal
to(E-2)
metal
seal.
The
ingresidue.
the
insulation
when
into
DN-3651
current
GOthe
current
GOor
recycled
to from
the
second
riser
of
asources
typicalsplicing
ARDS
HDS closes
tion
(VRDS)
unit
before
processing
edge
ofFor
the processing
and
processing
re
ux
cooler
shown
in
produced
the
paraffinic
located
gas
facility
could
back
into
the
reservoir
by
the
Base
opens
the
equalisation
valve
(see
metal
tip
of
the
stem
moves
downor rate
cutting
into
the
pre-insulated
E-1
FCC
(directly
or
through
oligomeriof
90%,
the
CCR
reduction
is
it
in
the
RFCC
unit
requirements
of the
available
Figure
all eliminated.
How,
be
easily taken apart
and moved
to hydrocarbons.
wasteful2 areand
environmentally
Steam
4). This
procedure
results
in then,
ward
to
seal
against
a metal
seat
tubing
bundle.
crudes.
Withand
the
continued
develdoes
one
know
thatwhich
the
sation)
tocommercial-scale
maximise
propylene
about
65-70%
metals
Depending
the
quality
ofofthe
Large,
GTL
a new
natural
gasthe
source
if anreducexist- Figure
damaging
practice
flaring,
Alternative
feed
Current
feedonof
transmitter
being
zeroed
matching
shape.
The
tip
of
the
stem
opment
andrate)
increased
availability
shown
in
Figure
3
will
tion
(HDM
isdepleted
about
85-90
combined
vacuum
residue
plants,
including
the Sasol
Oryx out
and design
ing
supply
was
in
its theproduction
is subject to increasing regulation. and
Catalyst age
Manifold
Ifplants
one RFCC
of the
two
may
beBecause
shaped
like
abuilt
ball
or way
athis
vee. a
of such
thisRFCC
knowledge
it to
was
this
thecalibration.
Shell Pearl
(both
located
current
location.
Modularised
GTL
plants
enable
Products
from
the
unit
and work?
wt%.
Thefeeds,
HDT
feed needs
will for
HVGO
feed
the
VRDS
unit,
to
be Low
continuously
updated
to
(at
the
Amoco
re
nery
in
Milford
The
manifold
consists
of
a
set
of
isolation
valves
is
leaking,
however,
In
either
case,
it
is
important
that
in
Qatar),
have
been
built
at
enorApplying
modularisation
to
refinotherwise
wasted
gas
to
be
250F
steam
cracker
are
combined
in
a
then
have
aNH
CCR
content
of
less
slip
stream
of
vacuum
residue
water
3
tobodies
desalters
Figure
15
DN-3651
and
DN-3551
normalised
temperature
ensure
minimal
operational
Haven,
UK)
in
1970,
wherethe
valves
whose
are
calibration
will
be
unsuccessful
thecould
tip Wales,
does
not and
rotate
with
mous
capital
cost.
The
Oryx
plant, to HDS
ery
has machined
advantages
converted
intotaken
additional
revenue.
light
ends
recovery
section
than
6 construction
wt%
and metals
content of thesingle
be
blended
with
surprises.
Reviewing
curves
worked
justeconomic
wellpicture,
as the
designed
for
production
levels
of it
with
toblock
productivity,
prodP-1 and
In
the
a oil
from
a regard
single
ofTBPmetal,
all subsequent
readings
will be
stem.
If larger
it such
does,
it RFCC
may
grind
into
produce
ethylene
and propylene,
less
than
10 wtppm.
diluents
asas
decanted
and
bulk
whole
crude
conventional
design
shown
in a to
34Dekkers
000 b/d,
cost
around
$1.5
billion
uct quality
andsteel.
ensuring
the
safety
modular
can
be(LCO)
the
usually
stainless
manithe(DCO)
seat, GTL
scoring
it cycle
and
creating
Ovchinnikov
iscapability
a Senior
Research
2
C, Daane
Oil & Gas
J., 1999,
97, Max
minimising
the R,capital
investment
Due
to theproperties
feed
rate,ofThe
nature
of
the inaccurate.
and
light
oilChemist
Figure
4
Two-stage
sour
water
stripper
design
without
feed
preheat
or
even
the
individual
cut-point
Figure
2.
to
build.
The
Shell
Pearl
plant,
with
of
construction
personnel.
Since
the
key
factor
that
enables
the
construcCriterion
Catalysts
andoil.
Technologies
145.
fold
mounts
to the side
of the a Further,
afterfrom
calibration,
pathway
forsome
leakage.

Hydrogen
the CRU the
and with
feed,
and processing
objectives,
produce
fuel
The VRDS
ranges and
doesina acritical
not 3anPuri
Houston,
Texas. projects
is primarily
ultimate
design
capacity
of
140 based
B K,will
Irgolicclose
K J, Environ.
Geochem.
Health,
modules
are serves
built
well-lit,
upstream
that
transmitter
equalisation
There
are then
two
main
steam
cracker
is the
supplemented
by tion
relatively
high
pressure
and
low operator
unitinofwould
beHe
able
todesign
reduce
increased
water
in
the
sulphur
Figure
1.
In
other
words,
a
lot
of
Two-stage
sour
water
stripper
necessarily
characterise
the
feed
engaged
in otherwise
the
research
and
development
of
1989,
11, 95.
000
b/d
of
GTL
products
and
120
climate-controlled
environment,
would
be
cancelled
function,
enablingdesign
calibration
or valve
andhydrogen
open the production
two isolation
approaches
to the
achieving
a nonfurther
from
space
velocity
is
required
the
CCR
to
levels
required
catalysts
hydroprocessing
applications
and by
4
NielseofB, Villadsen,
Appl.
Cat.,
1984,
11, 123.
Figure
4forshows
a sour
water
stripplant feed
and
thusaround
more
condensaequipment
is added
to
generate
difficulty
or
indicate
the full
impact
b/d
natural
gas
liquids,
cost
work
can
continue
the
clock
because
of
poor
results
derived
service
of thetotransmitter.
valves,
placing
the transmitter
back
rotating
stem
tip. especially
Ininthe
case
of athe
steam
reforming
be used
in isthe has
compared
other processing
hydroprocessing
the
RFCC
feed,
when
12
years
experience
heterogeneous
5
(a)ux Internal
Criterion
per
with
a of
side
draw-off.
The exampartly
tionthe
in individual
the sulphur
plantconditions,
feed
NH
re
when
no to
fractionation
on
units.
around
$18-19 communication,
billion.
Conventional
regardless
of weather
from
economic
models.
For
Quality
is especially
important
in 3 in hydrotreaters.
service.
If
the
equalisation
valve
vee
tip,
there
may
be
a
knuckle
technologies.
Even
so,
the
cycle
CCR
content
of
the
feed
to
andsour
refining
technologies.
He hasare
co- the
& Technologies;
(b)
R N, catalysis
gas
knock-out
drum (and
thus Catalysts
stripped
is extracted
required
between
theMerryfield
feedeconoand
GTL plant
designs rely
on
for greater
productivity
and easier
ple, some
shale water
gas
discoveries
a length
manifold.
During
calibration
or
does
not
achieve
a
positive
shut-off,
joint
that
enables
the
upper
stem
authored
over
20
technical
publications
and
Gardner
L
E,
Parks
G
D,
Catalyst
Characterization
of
the
ARDS
unit
is
typically
VRDS
unit
exceeds
24
wt%.
References
water
that
has Since
to be module
recycledheight
back overhead
8 andbydirected
to the to
only from
mies of product.
scale to Again,
drive the
positive
quality
control.
being tray
hampered
high developawhile
PhD degree
inlower
organic stem
chemistry
from
Science,
ACS configuration
Symposium
Series 1985,
1. the holds
normal
operation,
at
least
one
of
high
pressure
will
leak
across
turn
the
remains
1
Henke
K
R,
Arsenic,
Environmental
Chemistry,
Complex
with
only
one
year.
For
the
case
evaluto
the sour water
stripper).
for result
use as
make-up
the tower
is to
striponly
out hydrotreaters
financial of
returns
and are
viable
is restricted,
safety
is enhanced, as purpose
ment costs, which
in marginal
State University.
6 Bhan
Othe
K, Arsenic
removal
catalyst
and Iowa
theated
valves
in
the
manifold
is
in
the
seat
to
low
pressure
side,
stationary,
except
for
axial
(up
and
Health
Threats
and
Waste
Treatment,
Wiley,
Complex
configuration
with
HCU
+
VRDS
+
RFCC
and
ARDS
operating
and the
Hlarge
S. supplies of water
removal
The design
in Figure
in thedue
saltto(NH
the
NHthere
where
are
workers
build shown
at limited
heights2 method
economics
gas4HS)
prices
that
3 making
2 US Patent
for
same,
6759364.
2009, 186.
offobjectives,
position.
If
the
shut-off
is
less
rendering
the
differential
reading
down)
movement.
The
intention
of
this
refinery
configSDA
+
DCU
+
HCU
+
RFCC
at
least
two
reactor
step
of
the
reactor
ef
uent.
was
common
in
the
1960s.
low-priced natural gas.
within the fabrication facility.
are often low. These projects can be
than
complete,
result
istoan
inacIn
case
of configurations
a ball
tip,gas
a to
ball
Correct
stripper
designoption
water
from
However,
itthe
also
suffers
from
the inaccurate.
Thetherefinery
illusuration
is to another
increase
the production
trains,
each with
three
five
very
However,
being Completely
enhanced
by stripped
converting
the
curate
reading
from
transmitter.
would
these
floats
in the
stem
tip,
like
in a residue
ballIn
1969,
while
for expense
theleak?
now
sour
water
stripper
bottoms
is
same
heat
balance
and Why
Modularising
GTLthedrawbacks
trated
above,
where
of
middle
distillate
atvalves
the
developed
working
smaller-sized
and of the
large
and
thick-walled
reactors, will
higher-value
clean
fuels
produced
For
example,
when calibrating
cause is
debris,
either
point
pen.
Among
ball tips,
there
vanished
Amoco
International
theprocess.
crude reduce
desalter.
While
seenathe
in One
The
samecomplications
advantages
ofasboth
modular
distributed
GTL
plants

shows
in
thetoGTL
hydrotreaters
the
feed
petrochemicals.
Relative
toscaling
theOil
first sent
beneedless
required.
Consequently,

ITW

Innovative
Technologies
Worldwide

HIRING EXPERIENCED PROFESSIONALS 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:
x
refining/petrochemicals process specialty chemicals sales
x
refining/petrochemicals process technology, operations, maintenance, turnaround
and should be ZLOOLQJWRWUDYHOQDWLRQZLGHDQGZRUOGZLGHDORQJZLWKEHLQJSURQHWRKDUGZRUNLQJDQGVDOHV
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
102 www.eptq.com
PTQ Q1 2013

www.eptq.com
20
PTQ Q4 2013

www.eptq.com
lieberman.indd 3

velocys.indd
swagelok.indd
3 3 2
f wheeler.indd

jobs@itwtechnologies.com

www.itwtechnologies.com
PTQ Q2 2013 101
www.eptq.com

PTQ
Q1 2014 149
www.eptq.com

PTQ Q3 2013 83
08/03/2013 16:02
10/12/12
14:27:5614:52
11/12/2013

12/09/2013 13:40

moscow.indd 1

09/12/2013 11:28

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 without leakage.
In addition, a differential hardness 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 film on the seat. A
low differential hardness between
the stem tip and the seat can cause
premature failure due to the inability of the stem tip to coin out the
seat, or from excess torque when
closing the valve.

Seat seals in ball valves

The ball in a ball valve usually seals

against a fluoropolymer seat. High
pressure in the fluid 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 fluctuating. In valves that do not contain
these components, an interference fit
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

the same level of attention.

Measurement
accuracy
is
as
dependent on the transmitter accuracy 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, efficiency and plant
profitability.
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.

POLYSTINGER
LED HAZ-LO
130 LUMENS
15,000 CANDELA

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.
E-FLOOD
LITEBOX
615 LUMENS
4,000 CANDELA

In fact, you dont have to think

about your flashlight at all.
Because its a STREAMLIGHT.

KNUCKLEHEAD

2013 STREAMLIGHT, INC.

200 LUMENS
1,150 CANDELA

TRIDENT LED
HEADLAMP
80 LUMENS
4,000 CANDELA

www.eptq.com

swagelok.indd 4

CONNECT WITH US WWW.STREAMLIGHT.COM

PTQ Q1 2014 151

11/12/2013 14:52

Alphabetical list of advertisers

ABB Consulting

113

Heurty Petrochem

146

Hoerbiger Compression Technology

Air Liquide Global E&C Solutions

IFC

IDTC & BBTC

Albemarle

IBC

ITW

AFPM Annual Meeting 2014

AMACS Process Tower Internals

Axens

38 & 41

OBC

32
132

86
80 & 149

Johnson Matthey Process Technologies

John Zink Hamworthy

23

129

BASF Corporation, Catalysts Division

12

KBC Advanced Technologies

Bilfinger Water Technologies

65

Kurita Europe

67

Bryan Research & Engineering

103

Linde

25

Merichem Company

54

Burckhardt Compression

35

CB&I

11

Criterion Catalyst & Technologies

CS Combustion Solutions
DigitalRefining.com

4
45

111

DuPont Belco Clean AirTechnologies

DuPont Sustainable Solutions

37
83

Elliott Group
Enersul

79

122

European Fuels Conference

62

Everlasting Valve Company

89

ExxonMobil Research and Engineering Company 15

European Fuels Conference
Flexim

62

121

Metso Automation

119

OHL Gutermuth Industrial Valves

68

Process Consulting Services

Prosim

26
100

Refining China 2014

Sabin Metal Corporation
Samson

137

Sandvik Materials Technology

SOGAT 2014

Spectro Analytical Instruments

61

Streamlight
UOP

151
7, 9 & 31

Yokogawa Europe

Grabner Instruments

91

Zeeco

21

Zwick Armaturen

46
138

73

Grace Catalyst Technologies

GE Oil & Gas, Surface Pumping Systems

114

World Petroleum Congress

16 & 19

74

Four Quest Energy

85

Optimised Gas Treating

Wood Group Mustang

131

Onis International

92

Foster Wheeler

28

150
53

105

145

For more information on these advertisers, go to www.ptqenquiry.com

152

PTQ Q1 2014

ad index copy 2.indd 1

www.eptq.com

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. With a continuing record of success, Albemarles ACTION
is the only commercially proven FCC catalyst to maximize distillate, butylenes, and octane with minimal
gasoline loss. Utilizing its unique zeolite and matrix technologies, ACTION has been successful cracking all
types of feeds, from tight oil to heavy resid. Your success is too important to risk with unproven alternatives.
Achieve success with the undisputed leader demand ACTION.

CONTACT AN ALBEMARLE REPRESENTATIVE AT (281) 480-4747 TO DISCUSS WAYS YOUR CATALYST

OPERATIONS CAN BENEFIT FROM THE ADVANTAGES OF ACTION TECHNOLOGY.

REFINERY CATALYST SOLUTIONS

View our article on ACTION in the latest issue of Catalyst Courier: www.albemarle.com/CatalystCourier82

albemarle.indd 1

11/12/2013 09:32

Stimulate the heart of

your hydroprocessing unit
ImpulseTM, the catalyst technology that combines
the stability you recognize with the activity you need
4JOHMF
TPVSDF
*40

t
*40

t
0)4"4

www.axens.net

Imp_PTQ.indd 1
axens.indd 1

04/12/2012 14:07:08
6/12/12 13:17:48