Escolar Documentos
Profissional Documentos
Cultura Documentos
thermal technology
separation
refining
gas processing
petrochemicals
PETROLEUM TECHNOLOGY QUARTERLY
ptq
Q2 2013
cover and spine copy 15.indd 1 11/03/2013 13:28
1
2
0
9
_
e
w w w. a irliq u id e . c o m
Whatever the impurity, whatever the composition, Air Liquide Global E&C Solutions
has the right treatment.
Does your raw natural gas contain hydrogen sul de,
carbon dioxide, mercaptans or more?
The composition of natural gas varies
tremendously: almost every source
contains a dierent blend of impurities.
The options for treatment are almost
as diverse. Thats why oering a solution
specically designed for your gas eld
is crucial. We as your partner of choice
provide solutions for all types of natural gas,
including associated and unconventional
gas, from a single source. Customised
and e cient.
air liquide.indd 1 07/03/2013 13:27
2013. The entire content of this publication is protected by copyright full details of which are available from the publishers. All rights
reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic,
mechanical, photocopying, recording or otherwise without the prior permission of the copyright owner.
The opinions and views expressed by the authors in this publication are not necessarily those of the editor or publisher and while every care
has been taken in the preparation of all material included in Petroleum Technology Quarterly the publisher cannot be held responsible for any
statements, opinions or views or for any inaccuracies.
3 The heavy side of light crude
ChrisCunningham
5 Technology in Action
13 Tackling the gasoline/middle distillate imbalance
MarielleGagnire,AnnickPucciandEmilieRousseauAxens
23 Strategies to control sediment and coke in a hydrocracker
MarcoRespini,SilviaEkresandBruceWrightBaker Hughes
Robertajdlk Slovnaft
37 Simulation of bitumen upgrading processes
AntonAlvarez-MajmutovandJinwenChenCanmetenergy, natural resources Canada
MugurelMunteanuWorleyParsons
45 Insulating tank terminals
PeterBoonPentair Thermal Management
49 Understanding thermal oxidisers
PeterPickardandKevinMcQuiggZeeco
57 Diagnostic tools and control parameters for refnery heaters
YahyaAktas ,HakanOg
uz,GrayAksu,Dog
usUzunandMuammerSever
Tpras Izmir refnery
65 Minimising the cost of hydrogen production
LorenaMosca,PalmaContaldo,MenicaAntonelliandPaolaVolpi
Tecnimont KT
75 Options for biojet fuel
JitendraKumarSatyarthi,ChiranjeeviThota,GokakDattatrayaTammannaShastry
andPoyyamaniSwaminathaViswanathan
Bharat Petroleum Corporation
81 HCN emissions in fuid catalytic cracking
XunhuaMo,BartDeGraafandPaulDiddamsJohnson Matthey
87 Oil/water separation technologies
BertholdOtziskKurita europe
91 Customised tower design
DariusRemesatandMichaelKrelaKoch-glitsch Canada
99 Sour water strippers: design and operation
NormanLiebermanProcess Improvement engineering
105 Online cleaning: a mindshift for improving operational excellence
MarcelloFerraraandJustinWeatherford ITW
113 Lessons learned from MpC revamp of two naphtha crackers
VanessaConz,SuzanaFuchsandPriscilaMacedoMouraBraskem
MarkDarbyCMiD Solutions
DougNicholsonIPCOS (UK)
125 Energy-effcient vacuum systems
CChandraSekharaReddyandSVNaiduAndhra University
GPRangaiahnational University of Singapore
Phillips66becameanindependentdownstreamenergycompanywithaglobalrefningcapacityof2.2millionb/dduring
2012. Photo: Phillips 66
Q2 (Apr, May, Jun) 2013
www.eptq.com
p
t
q
Y L R E T R A U Q Y G O L O N H C E T M U E L O R T E P
ed com.indd 1 11/03/2013 16:27
GLOBAL SOFTWARE AND CONSULTING SOLUTIONS FOR
UPSTREAM | GAS PROCESSING | REFINING | PETROCHEMICAL
Linking cut t ing edge t echnology wit h consult ing expert ise for over 30 years
AMER: +1 281 293 8200 ASIA: +65 6735 5488 EMEA: +44 1932 242424
answers@kbcat .com | www.kbcat .com | blog.kbcat .com
KBC Advanced Technologies plc
SOLUTIONS ACROSS THE VALUE CHAIN
| St rat egic Invest ment Support
| Asset Opt imisat ion
|
Sust ainable Workforce Development
C
M
Y
CM
MY
CY
CMY
K
KBCAdv (A4).pdf 1 3/7/2013 7:56:13 AM
kbc.indd 1 07/03/2013 14:18
T
he European Union has arguably
been the global leader in biodiesel
production and use, with overall
biodiesel production increasing from 1.9
million tonnes in 2004 to nearly 10.3 million
tonnes in 2007. Biodiesel production in the
US has also increased dramatically in the
past few years from 2 million gallons in
2000 to approximately 450 million gallons
in 2007. According to the National Biodiesel
Board, 171 companies own biodiesel
manufacturing plants and are actively
marketing biodiesel.1. The global biodiesel
market is estimated to reach 37 billion
gallons by 2016, with an average annual
growth rate of 42%. Europe will continue to
be the major biodiesel market for the next
decade, followed closely by the US market.
Although high energy prices,
increasing global demand, drought
and other factors are the primary
drivers for higher food prices, food
competitive feedstocks have long
been and will continue to be a major
concern for the development of biofu-
els. To compete, the industry has
responded by developing methods to
increase process effciency, utilise or
upgrade by-products and operate
with lower quality lipids as
feedstocks.
Feedstocks
Biodiesel refers to a diesel-equivalent
fuel consisting of short-chain alkyl
(methyl or ethyl) esters, made by the
transesterifcation of triglycerides,
commonly known as vegetable oils or
animal fats. The most common form
uses methanol, the cheapest alcohol
available, to produce methyl esters.
The molecules in biodiesel are pri-
marily fatty acid methyl esters
(FAME), usually created by trans-
esterifcation between fats and metha-
nol. Currently, biodiesel is produced
from various vegetable and plant oils.
First-generation food-based feedstocks
are straight vegetable oils such as
soybean oil and animal fats such as
tallow, lard, yellow grease, chicken fat
and the by-products of the production
of Omega-3 fatty acids from fsh oil.
Soybean oil and rapeseeds oil are the
common source for biodiesel produc-
tion in the US and Europe in quanti-
ties that can produce enough biodie-
sel to be used in a commercial market
with currently applicable
PTQ Q2 2013 3
Editor
Chris Cunningham
editor@petroleumtechnology.com
Production Editor
Rachel Storry
production@petroleumtechnology.com
Graphics Editor
Rob Fris
graphics@petroleumtechnology.com
Editorial
tel +44 844 5888 773
fax +44 844 5888 667
Business Development Director
Paul Mason
sales@petroleumtechnology.com
Advertising Sales
Bob Aldridge
sales@petroleumtechnology.com
Advertising Sales Offce
tel +44 844 5888 771
fax +44 844 5888 662
Publisher
Nic Allen
publisher@petroleumtechnology.com
Circulation
Jacki Watts
circulation@petroleumtechnology.com
Crambeth Allen Publishing Ltd
Hopesay, Craven Arms SY7 8HD, UK
tel +44 844 5888 776
fax +44 844 5888 667
ptq (petroleum technology quarterly) (ISSN
No: 1632-363X, USPS No: 014-781) is published
quarterly plus annual Catalysis edition by Crambeth
Allen Publishing Ltd and is distributed in the US
by SP/Asendia, 17B South Middlesex Avenue,
Monroe NJ 08831. Periodicals postage paid at New
Brunswick, NJ. Postmaster: send address changes to
ptq (petroleum technology quarterly), 17B South
Middlesex Avenue, Monroe NJ 08831.
Back numbers available from the Publisher
at $30 per copy inc postage.
Vol 18 No 4
Q2 (Apr, May, Jun) 2013
The heavy side of
light crude
W
ith its air permit from the state environment department recently in
place, a joint venture of MDU Resources Group and Calumet Specialty
Products Partners is set to begin construction of the USs frst new
greenfeld site refnery since the 1970s. The JV, Dakota Prairie Refning, will
build and operate a $300 million diesel refnery near Dickinson, North Dakota.
The project is in the fnal stages of engineering, and design and construction
was expected to begin in early April, with frst oil to the plant possibly by the
end of 2014. The new refnery will process 20 000 b/d of Bakken crude oil and
will employ approximately 100 people. Around a third of the crude arriving at
the plant will be processed to supply local transport with diesel fuel.
Just a month earlier, Delta Airlines received its frst shipment of Bakken
crude from North Dakota to make jet fuel at its refnery in Pennsylvania.
Last year, the airline-refner decided to replace more expensive foreign crude
supplies with light, home-produced shale oil. Monroe Energy, a Delta subsid-
iary, had previously slowed production at its 185 000 b/d refnery. The
arrival of Bakken crude is expected to improve the sites margin and retain
jobs for the refnerys 400 workers.
In yet more positive news as the northern winter entered its declining
days, the US Energy Information Association (EIA) announced that US crude
production hit 7 million b/d in the fnal months of 2012. This is the highest
output since December 1992 and underlines the reversing trend of US crude
oil imports.
Meanwhile, Australian interests are looking to push ahead with their plans
to develop shale oil production. At frst glance, it appears that the swathe of
fracking sites springing up across the central US is leading a light, sweet era
for the petroleum processing industry across the globe.
If only the energy business were so straightforward. Remember peak oil
the reason refners invested so many billions of dollars in technology to be able
to process sour, heavy crudes? All set up to handle heavy crudes from Mexico
and Venezuela, or oil sands bitumen from Canada, refners may not welcome a
sudden food of high-quality crude as warmly as you might imagine.
Bakken and other light, tight sources of crude could well prove to be the
saving of beleaguered refneries in the eastern US that never made the transi-
tion to heavy feedstock processing. But without a growing feet of Dakota
Prairie-type operations, it looks set to cause a few headaches both within and
beyond the US border. Will the XL pipeline to carry supplies from Canada to
the US Gulf ever be built? Will the complications to crude price differentials
brought on by shale oil supplies make plans to expand Canadian upgrading
a hopeless option? And how will West African producers of light, sweet
crude deal with the loss of a major export market?
With a heavy regime of cuts about to bite the US economy, you might think
that the least troublesome and most productive option would be to export
shale-derived crudes that are not needed for those domestic refneries set up
for lighter feeds. Not so: uniquely among major producers, the US has
deemed crude exports illegal for the best part of a century, a view reinforced
by the oil crises of the 1970s. The remaining option is to process lighter
crudes to a minimum acceptable level for export to markets with, for the
time being, less stringent environmental regulations.
CHRIS CUNNINGHAM
p
t
q
Y L R E T R A U Q Y G O L O N H C E T M U E L O R T E P
ed com.indd 2 08/03/2013 11:05
Theres only one way to stay ahead of
the competition: always be evolving.
Meet Max Ovchinnikov:
Industry Game-Changer.
When you re leading the world in catalyst development, your work isnever done and thatsjust ne
by M ax O vchinnikov, C RITERIO N sSenior Research C hemist and co-developer of such advanced
catalyst platformsasSEN TRY and C EN TERA
a
t
io
n
:
w
w
w
.
g
r
a
t
y
.
f
r
Leader in Process Viscometry
Since 1972
PIVI
viscometer
Compact, hand-held
viscometer for
laboratory or
on-field use
Thermoset
MIVI
Powerful online viscosity
analyzer at reference
temperature
Offer the
best value to
your products
MIVI
viscometer
Process inline solution
for a reliable
measurement
and control
www.sofraser.com
T. +33 (0)2 38 85 77 12 F. +33 (0)2 38 85 99 65 instruments@sofraser.com
case studies copy 7.indd 3 08/03/2013 11:34
For more information, visit www.uop.com/unifex.
2012 Honeywell I nternational, I nc. All rights reserved.
refinement redefined
UOP Unihox rosiduo-upgrading toonnology yiolds 25% moro
oloan uol, turning "bottom o tno barrol" into "top o tno lino."
Got tno most rom ovory barrol. Tno UOP Unihox Proooss oan doublo tno diosol yiold
o otnor rosiduo-upgrading toonnologios, turning tno "bottom o tno barrol" into moro
blaok on your bottom lino. Tno Unihox Proooss is a nign-oonvorsion, oommoroially
provon toonnology tnat prooossos low-quality rosiduo stroams liko vaouum rosiduo into
nign-quality distillato produots loading to rohnory margin inoroasos o up to 100%.
Simply put, witn tno Unihox Proooss rom UOP, you'll maximizo produotion and prohts
rom ovory barrol.
SPM-UOP-30 Uniflex Ad_Comp_v6.indd 1 3/8/12 2:35 PM
uop.indd 1 9/3/12 12:31:44
syngas from the feedstock hydrocarbon source into a
slurry-bubble column with a Cobalt-based catalyst,
and FT product upgrading using hydrocracking and
hydroisomerisation. Products of the process include:
25% pure paraffnic naphtha for onward petrochemical
processing; 25% kerosene, which is compatible with
refnery Jet A1 fuel; and 50% high-cetane (>75),
zero-sulphur diesel, with cold fow properties that are
adjustable in the range -20C to -40C. Alternatively,
says Axens, a larger diesel cut can be produced by
mixing the kerosene and diesel products.
Feedstocks for Forest BtL include wood from the
thinning of young forests, which would otherwise be
left in the forest. This is trimmed for ease of transpor-
tation. Other sources include tree stumps, which are
left to dry in the forest before processing, and logging
residues the crowns and branches otherwise left in
the forest after felling. The biomass is allowed to dry
in the forest, where it may be chipped before transport
to the plant.
Due to the low density of biomass and its relatively
high moisture content, the average economic transport
distance for the raw material is below 200 km. The raw
material is mostly stored in production areas, and only
the amount of biomass required for continuous
production is stored at the plant site.
For Forest BtL, Axens will deliver a process design
package and technical assistance during start-up and
operation, and will manufacture and supply the associ-
ated catalysts.
10 PTQ Q2 2013 www.eptq.com
Sulzer Pumps production sites in Suzhou and
Dalian, China, will produce 34 pumps. The remaining
pumps will be manufactured in Bruchsal, Germany,
and packaged in Suzhou.
Gas processing technology that will supply feedstock
and products for petrochemicals production,
and for refnery diesel and jet fuel pools, is being
installed at a second-generation biofuels project in
Finland.
Forest BtL has signed a licence for the use of the Gasel
technology suite from Axens in a biomass-to-liquid
(BtL) plant to be constructed at Ajos, near Kemi. Ajos
BtL will convert sustainable forest residue into 3700 b/d
of renewable liquid fuels. The biomass will be gasifed
to produce syngas, feeding the Gasel technology.
Now licensed by Axens, Gasel was originally co-de-
veloped as a process and catalyst development
programme started in 1996 by IFP Energies nouvelles,
Eni and Axens.
Axens has been responsible for pilot plant design
and catalyst preparation and production, starting with
the 20 b/d Sannazzaro Fischer-Tropsch (FT) pilot plant
in Italy. The pilot plant was operated in campaigns
between 2001 and 2010, testing, proving and improv-
ing the technology and the catalyst.
The package includes FT conversion of purifed
From forest to petrochems and distillates
Today s Appli cati on :
LN G -P R O C ESSI N G , H AN D LI N G AN D STO R AG E
H ER METI C desi gn features
Lowlife-cycle-costs
Lownoise level
High reliability
C ustomer s tech n i cal speci cati on
Capacity: 120 m /h
Head: 1400 m
H E R M E T I C A L L Y S E A L E D C E N T R I F U G A L P U M P S
HERMETIC-Pumpen GmbH
hermetic@ hermetic-pumpen.com www.hermetic-pumpen.com
H ER METI C-P ump Type C AMTV 5 2 /6 +6
case studies copy 7.indd 4 08/03/2013 11:21
Reliable SwiSS Quality
API 618
Rod load up to 1'500 kN/335'000 Ibs
Power up to 31'000 kW/42'100 HP
Full Range:
You get moRe than Just a
PRocess gas comPRessoR
Lubricated up to 1'000 bar, non-
lubricated up to 300 bar
For highest availability: We recom-
mend our own designed, in-house
engineered compressor valves and
key compressor components
Designed for easy maintenance
We are the competent partner
with the full range of services
worldwide
www.recip.com/api618
YouR BeneFit:
lowest liFe cYcle costs
burckhardt.indd 1 7/3/12 21:39:05
pcs myth.indd 1 8/3/12 09:55:39
Tackling the gasoline/middle distillate
imbalance
T
he varying constraints on the
refning industry result in a
market that is in constant
evolution. Regulations on low-sul-
phur gasoline and diesel fuel, curbs
on refnery emissions and the
increase in heavy crudes over the
past years have infuenced the refn-
ers market. Today, new forces have
emerged, with improvements in
light vehicle fuel economy, the prev-
alence of emerging markets, and the
rise of shale oil and gas shifting the
outlook. In this article, we present
the market trends and outlook
ahead to 2020 and 2030, and foresee
that the current thirst for middle
distillates will be accentuated in the
US and Europe, but also in emerg-
ing markets. In the second section,
the oligomerisation technology,
PolyFuel, which can adjust refnery
output towards middle distillates, is
presented. Finally, an economic
evaluation of the performance of
PolyFuel assesses its value to the
refnery based on 2012 prices and
costs.
Market outlook
In 2012, global oil demand was
around 89.7 million b/d according
to the latest IEA estimates, driven
by growth in emerging countries.
Broken down by product, global
growth in demand for oil is driven
by middle distillates, comprising
kerosene and diesel, unbalancing
the product slates in some regions.
Worlds appetite for middle distillates
From 2000 to 2010, growth in world
consumption of middle distillates
exceeded growth in demand for
gasoline and total oil products: the
annual average growth rates
An oligomerisation technology can produce a signifcant increase in middle
distillate production
MARIELLE GAGNIRE, ANNICK PUCCI and EMILIE ROUSSEAU
Axens
(AAGR) were 1.6%/year and 1.3%/
year, respectively. In 2020, contin-
ued robust global demand for
distillate is expected, with a growth
rate still at 1.6%/year, above total
growth in demand for oil. By
contrast, growth in demand for
gasoline is forecast to be weak,
with an increase of 0.7%/year.
From 2020 to 2030, while global
growth in demand for oil is
expected to be restrained at 0.6%/
year, middle distillates still stay a
primary driver with an annual
growth above that of oil (see
Figure 1).
Atlantic basin gasoline/distillates
balance
Since 1998, Europe has experienced
excess volumes of gasoline produc-
tion. Much of the excess European
gasoline has been exported to
North America, a light vehicle
www.eptq.com PTQ Q2 2013 13
1 . 2
1 . 8
1 . 6
1 . 4
1 . 0
0 . 8
0 . 6
0 . 4
0 . 2
2 0 0 0 -2 0 1 0 2 0 1 0 -2 0 2 0 2 0 2 0 -2 0 3 0
A
v
e
r
a
g
e
a
n
n
u
a
l
g
r
o
w
t
h
r
a
t
e
,
%
0
D ie se l & k e ro se n e
To ta l o il p ro d u c ts
G a so lin e
Figure 1 Growth in world demand for oil products
5 0 0
1 0 0 0
0
5 0 0
2 0 0 2 2 0 0 3 2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7
E
x
p
o
r
t
s
,
k
b
o
e
/
d
1 0 0 0
2 0 0 8 2 0 0 9 2 0 1 0 2 0 1 1
K e ro se n e
D ie se l
G a so lin e
S u rp lu s
D e fic it
Figure 2 European (EU-27) net exports of gasoline, kerosene and diesel Source: JODI
axens.indd 1 08/03/2013 11:48
14 PTQ Q2 2013 www.eptq.com
effciency for new light duty vehi-
cles (CAFE standards) and ethanol
blending requirements.
On the crude supply side, a game
changer is also modifying US refn-
ing dynamics: rapid growth in
shale oil production. US crude
production is surging in states
where shale oil is available. The
Bakken feld in North Dakota and
Eagle Ford feld in South Texas are
the largest, but many other fnds
are spread across the country. This
development has transformed the
US refning industry. Refners in
the US Gulf Coast, where 44% of
the US refning capacity is located,
beneft from this cheap crude.
Yet, there are some disparities:
refners on the East Coast struggle
and experience closures. In PADD
(Petroleum Administration for
Defense District) 1, refners cannot
easily access the new shale oil
supplies; they remain dependent on
crude imports and face competition
from European gasoline imports.
Nevertheless, imports of crude and
oil products are collapsing and the
US has become a net exporter of oil
products. Furthermore, US Gulf
Coast refneries that run shale oil
will produce more light products
due to this light sweet crude
supply, and natural gas liquids
(NGLs) associated with shale
production will increase in the
coming years. All of these factors
will weigh on the whole light ends
complex and may lead to a naphtha
surplus in the US.
Non-OECD economies dominate
global oil demand growth
Predictions point to the non-OECD
share in demand for oil rising from
48% to 60% in 2030, most of this
concentrated in Asia, the Middle
East and Africa (see Figure 4).
Emerging countries, particularly
non-OECD Asia, are forecast to
dominate growth in demand for
middle distillate, with a signifcant
shift towards diesel in the passen-
ger vehicle market in India and a
demand growth driven by commer-
cial freight in China.
In conclusion, with a global
surplus of gasoline reducing export
opportunities and a growing defcit
in middle distillates regionally,
gasoline market. Today, however,
gasoline production within the US
has increased, while demand for
gasoline from petroleum has
reduced. Consequently, European
gasoline exports to the US have
decreased. With a gasoline surplus
exceeding 800 000 boe/d (see
Figure 2), European refners have
found other export destinations in
Latin America, Africa and the
Middle East.
Accounting for half of the total
oil demand, diesel remains the
principal product consumed in
Europe (EU-27) mainly due to the
transportation sector. The mismatch
between gasoline and diesel supply
and demand increases the pressure
on European refneries that cannot
satisfy their market. As a result, net
diesel imports remain signifcant,
reaching a maximum of
560 000 boe/d in 2010, originating
mostly from the CIS and the US.
Furthermore, Europe is short of
kerosene, with a strong and stable
defcit of around 300 000 boe/d,
heightening the European thirst
for middle distillates (see Figure 2).
The North American situation is
completely different (see Figure 3):
US diesel net exports have
increased considerably since 2008,
reaching more than 1 million boe/d
in 2012
Kerosene imports and exports are
balanced
North America gasoline imports
have reduced by 75% when
compared to 2006 and should be
about 230 000 boe/d for 2012.
With world growth in middle
distillates and reduced domestic
demand for gasoline, US refners
have developed diesel exports over
the past few years, increasing distil-
late yields by changing distillation
cut-points or adapting their refn-
ery units operational mode with
small-scale capital investment.
US petroleum-based gasoline
imports have collapsed since 2011,
due to a reduced demand for
fnished motor gasoline, which has
been declining by 1.6%/year since
2007 because of economic condi-
tions and improved engine energy
5 0 0
1 0 0 0
0
5 0 0
2 0 0 2 2 0 0 3 2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7
E
x
p
o
r
t
s
,
k
b
o
e
/
d
1 0 0 0
2 0 0 8 2 0 0 9 2 0 1 0 2 0 1 1 2 0 1 2
K e ro se n e
D ie se l
G a so lin e
S u rp lu s
D e fic it
Figure 3NorthAmericanetexportsofgasoline,keroseneanddiesel
6 0
8 0
7 0
5 0
6 0 %
4 8 %
4 0
3 0
2 0
1 0
2 0 1 2 2 0 2 0 2 0 3 0
O
i
l
d
e
m
a
n
d
,
M
b
/
d
0
N o n -O E C D
O E C D
Figure 4 OECDandnon-OECDtotaloildemandfrom2012to2030
axens.indd 2 08/03/2013 11:48
pcs 1.indd 1 8/3/12 09:50:20
either in OECD countries or emerg-
ing economies, reners are led to
rebalance their product slates. New
technologies are needed to respond
to these trends, producing more
middle distillates and compensat-
ing for a reduction in demand for
gasoline in OECD countries and a
potential over-supply of light prod-
ucts and gasoline in the US.
PolyFuel technology
Optimisation of the products slate
around the FCC unit
Since demand for middle distillates
is growing faster than that of gaso-
line, increased interest in process
options to shift the ratio of diesel to
gasoline production has been regis-
tered. This trend has been seen in
FCC units, along with an increased
demand in propylene from FCC to
help satisfy the much faster grow-
ing demand for petrochemicals. An
underlying theme has been
increased regulation to sharply
16 PTQ Q2 2013 www.eptq.com
reduce the sulphur content of both
gasoline and diesel, and limit the
olen content in gasoline.
An increasingly popular technol-
ogy option to address the fuels
imbalance and product exibility
requirements is olen oligomerisa-
tion. Reactive olens can be
oligomerised, mainly into dimers
and trimers, to increase the size of
the molecules and shift the product
slate towards heavier components.
A well-proven technology to
convert C
3
-C
4
olens into gasoline
and diesel range components has
been in commercial operation since
1986 under the trade name of
PolyNaphtha. Unlike older polym-
erisation technology oriented to
gasoline production and based on
solid phosphoric acid catalyst, this
process uses a regenerable catalyst
with the ability to produce gasoline
and distillate fuels. It offers a
low-cost alternative to alkylation
and an alternative outlet for
propylene, particularly when recov-
ery is not economic.
When distillate production is
favoured, an extension of this
concept can be used to convert
C
5
-C
9
gasoline olens to distillate in
a simple, low-cost PolyFuel process.
A generic block ow diagram of
the process is shown in Figure 5.
In the process, light olens are
oligomerised catalytically in a
series of two xed-bed reactors.
Conversion and selectivity are
controlled by reactor temperature
adjustment, while the heat of reac-
tion is simply removed by feed
efuent heat exchange (see Figure
5). The reactor section efuent is
fractionated, producing gasoline
depleted in olens and middle
distillate fractions. The gasoline
fraction is partly recycled to the
reaction section to enhance middle
distillate production. The distillate
fraction is typically sent to existing
kerosene and diesel hydrotreating
directly.
The reaction section uses swing
reactors to allow for continuous
operation with on-stream catalyst
regeneration or replacement. The
management of the reactors is opti-
mised to maximise catalyst run
length. With the objective to
produce high-quality distillate, the
process utilises the IP 811 catalyst,
which can be operated at high
severity to maximise the middle
distillate fraction as a result of its
high activity and stability.
Typically, the middle distillate
yield exceeds 70% as compared to
feed olens. The IP 811 catalyst can
undergo multiple regeneration
cycles, with over 95% activity
recovery demonstrated. Catalyst
regeneration can be performed in
situ or ex situ.
Like PolyNaphtha, PolyFuel
produces a high-quality Jet A1 frac-
tion after hydrotreatment to
saturate olens. The smoke point
exceeds 35 mm, which is far above
the Euro V specication.
Feedstock selection for PolyFuel
depends on the availability of
olen-rich streams and the ren-
ery-specic product requirements.
When distillate is more favoured
than gasoline, portions of the FCC
gasoline product are logical feeds.
O le fin ic
fe e d
R e c yc le fo r
m a x im u m d istilla te s
M id d le
d istilla te
G a so lin e
Reaction section
F ix e d -b e d
swin g re a c to rs
I P 8 1 1 : in -situ
o r e x -situ
re g e n e ra tio n
Splitter
Figure 5 PolyFuel process general ow diagram
C
5
C
5
/C
6
fe e d
to p o lyfu e l
C
6
C
7
C
1 0
C
1 1 C
1 2
O
l
e
f
i
n
s
,
%
C
u
m
u
l
a
t
i
v
e
s
u
l
p
h
u
r
,
%
Increasing cut point
M e rc a p ta n s
T h io p h e n e
C
1
-T h io p h e n e
C
2
-T h io p h e n e
C
3
-T h io p h e n e
B e n zo -
T h io p h e n e
A lk yl-b e n zo -
T h io p h e n e s
C
9
C
8
O le fin s
C u m u la tive su lp h u r
S
S
Figure 6 Typical FCC gasoline olens and sulphur prole
axens.indd 3 08/03/2013 11:49
www.eptq.com PTQ Q2 2013 17
any cracking process such as FCC,
coking, steam cracking or any other
olen sources, such as efuent from
parafn dehydrogenation or
alcohol dehydration. Specic
pretreatment options are adapted,
depending on feed origin and
contaminants level.
An attractive FCC-based renery
conguration to maximise middle
distillates consists of mixing availa-
ble LPG with a C
5
-C
6
LCN cut and
feeding the blend to a PolyFuel
unit. Available sources of LPGs in a
renery may be:
C
3
s when not sent to
petrochemicals
C
4
s when no C
4
cut upgrading
processes are present
Or excess C
4
s in the case of an
existing C
4
cut upgrading processes
bottleneck (alkylation unit and so
on).
The relative amounts of LPG and
splitter top, thereby minimising the
pretreatment cost of the process
(see Figure 7).
To maximise renery protabil-
ity, the feedstock can also include
olenic C
3
and/or C
4
fractions,
depending on the renerys objec-
tives and economics. Olen-rich
feed components can come from
In principle, the process can accept
the full range of C
5
-C
9
gasoline cut
produced from the FCC unit. The
C
5
-C
6
portion contains the highest
concentration of reactive olens.
The heavier C
7
+ portion contains
fewer olens and can have a high
content of contaminants (sulphur,
nitrogen and oxygenates), which
reduce catalyst cycle length. To
illustrate, a typical gasoline prole
of olens and sulphur as a function
of boiling point is shown in
Figure 6.
Based on these observations,
Axens has identied the C
5
-C
6
olen-rich cut as the preferred feed
to be processed in the PolyFuel unit
to reach the highest protability by:
Maximising the amount of
oligomer product
Maximising catalyst run length
Minimising feed pretreatment
Signicantly reducing the olen
content in the remaining gasoline
product.
The previous point on olens
reduction in the remaining gasoline
pool can be signicant in regions
where olens content is tightly regu-
lated. Typically, the olens in the
C
5
-C
6
cut are reduced by over 80%
as compared to the C
5
-C
6
PolyFuel
feed. When the FCC unit is operated
at high severity to maximise propyl-
ene, the concentration of C
5
-C
6
olens is very high in the gasoline
and may result in an off-spec gaso-
line pool. As a result, the new
process can help to satisfy the need
for increased distillate and improve
the gasoline pools properties.
In order to prepare the feed to
the PolyFuel unit, one can take
advantage of existing FCC gasoline
selective desulphurisation unit
design. The most widely selected
technology is the Prime-G+ process
with 214 licensed units as of
January 2013. The Prime-G+ rst
step, which consists of a selective
hydrogenation unit (SHU) on the
full-range cracked naphtha (FRCN)
with a downstream splitter,
provides an ideal feed pretreatment
by removing the diolens and
reducing the sulphur content of the
light C
5
-C
6
cut. In reneries that
include Prime-G+, a PolyFuel unit
can easily be installed on the light
cracked naphtha (LCN) from the
F R C N
O p tio n a l C
3
/C
4
L C N
H C N
M a k e -u p
U L S g a so lin e
a n d d istilla te
Prime-G+
first step
P rim e -G +S H U
se le c tive
h yd ro g e n a tio n
Splitter
Prime-G+
selective
HDS
U L S g a so lin e
PolyFuel
H
2
Figure 7 Prime-G+ integration with PolyFuel for contaminant control
Motor fuel
PolyFuel FCC
Max distillate
Products
G a so lin e D istilla te to D H T
R a ffin a te
C
3
-C
6
o le fin s
Figure 8 PolyFuel in a FlexEne conguration oriented to distillates production
To maximise renery
protability, the
feedstock can also
include olenic C
3
and/or C
4
fractions,
depending on the
renerys objectives
axens.indd 4 08/03/2013 11:49
18 PTQ Q2 2013 www.eptq.com
scheme of PolyFuel integrated in
the FlexEne concept is shown in
Figure 8. Operation of the FCC
unit, selection of feedstock cuts,
operating severity and oligomer
product cut recycled to the FCC
unit are all parameters subject to
optimisation, depending on prod-
uct demand and pricing.
Rening/petrochemical integration
When the renery is integrated with
a petrochemicals complex, there are
further opportunities to share feed
and product streams to improve
protability. The oligomers from the
PolyFuel unit can be entirely recy-
cled to the FCC unit to maximise
propylene production and the
parafnic rafnate sent to the steam
cracker to produce more petrochem-
icals or gasoline blend stock.
As is indicated in the name, the
FlexEne concept utilising FCC with
PolyFuel provides great exibility
to optimise the production of
propylene, gasoline and distillates
according to market demand.
In the context of faster worldwide
growth in demand for distillates
and propylene than for gasoline,
but with regional disparities, Axens
has developed a new process with
the exiblility to:
Process lower-valued olenic
feeds in the C
3
-C
6
range from a
variety of sources
Maximise distillate production
and reduce gasoline production
according to market demand
Further optimise distillate or
propylene production when inte-
grated with FCC in a FlexEne
conguration
Improve the overall renery
economics as feed and product
prices uctuate.
Economic appraisal
The following economic appraisal
evaluates the difference in return
between the sale of products from a
classic FCC conguration and
the cost of installation of a PolyFuel
unit downstream of the FCC unit,
and the revenue associated with the
increased middle distillates and
upgraded gasoline.
The present case study assumes
the following feedstocks:
C
5
-C
6
light cracked naphtha
In addition to a stand-alone
conguration, the process can be
applied within the FlexEne
concept, which is Axens combina-
tion of FCC and oligomerisation
processes. This combination
enhances product exibility by
controlling the balance of propyl-
ene, gasoline and middle distillates
production with a low additional
capital investment to the FCC
complex. This exibility is achieved
via oligomerisation of light FCC
olens into heavier fractions and
recycle of undesired oligomer cuts
for further cracking in the FCC
unit, thus increasing the desired
product slate often an increase in
propylene and distillate.
A typical distillates-oriented ow
C
5
-C
6
cuts and the operating sever-
ity of the PolyFuel unit can be
optimised to adjust the distillate
product rates in accordance with
market demand while reducing the
gasoline surplus.
In summary, the new technology
addresses the gasoline/middle
distillates imbalances by selec-
tively converting unwanted
olenic streams in the C
3
-C
6
range
into distillate fuels while meeting
increasingly tight product speci-
cations. A PolyFuel unit can be
designed to operate over a wide
range of LPG and/or C
5
-C
6
olenic
feed compositions and with the
exibility to adjust the gasoline/
middle distillates product ratio to
regional requirements.
FCC
R2R
DCC
HS-FCC
VG O
re sid
Prime-G+
Steam
cracker
Aromatics
complex
ParamaX
PRU
O lig o m e r
re c yc le
M ix e d C
4
P ro p a n e
M ix e d C
3
P a ra ffin ic C
4
-C
6
C
3
G a so lin e
L C N
H C N
D istilla te /
fu e l o il
P o lym e r
g ra d e
e th yle n e
P o lym e r
g ra d e
p ro p yle n e
B T X
PolyFuel
Catalytic cracking
unit with FlexEne
Petrochemical
integration
Figure 9 Rening and petrochemicals integration with FlexEne and PolyFuel
M ix e d
L P G /C
5
-C
6
fro m F C C
M id d le
d istilla te
L P G
G a so lin e
Reaction section Splitter
Figure 10 PolyFuel case study
axens.indd 5 08/03/2013 11:49
How Can KBR Answer Your
Refining Challenges?
Owners of refineries continue to confront
many challenges rising feedstock prices,
shrinking margins, varying global demands
and a changing regulatory landscape that
includes ever-more stringent specifications
on sulfur and carbon footprints. As refinery
owners debottleneck and enhance existing
facilities, they call on KBR to deliver.
To learn how KBR can address your refining
challenges, go to:
refining.KBR.com/PTQ
refining.kbr.com/PTQ
K13009 2013 K BR, I nc. All Rights Reserved.
kbr.indd 1 08/03/2013 10:45
20 PTQ Q2 2013 www.eptq.com
prices, to achieve an internal rate of
return (IRR) of 15%, the price of
middle distillate has to be $38/t
higher than the gasoline price.
When compared with the 2008
middle distillate/gasoline price
difference, which was $154/t for
kerosene and $102/t for diesel, this
break-even differential is not
excessive.
Due to the excess of gasoline
production in Europe, all the gaso-
line produced is not sold at the
European gasoline market price.
Part of the production has to be
exported, possibly at lower prices
than the European market value.
The local difference between gaso-
line and middle distillates prices at
the outlet of a refnery is thus most
probably lower than the market
price difference, leading locally to
economics that are already profta-
ble for the implementation of
PolyFuel technology.
Today, the price differentials
observed between gasoline and
middle distillates do not seem to
correspond with market conditions.
In the near future, with even more
pressure on middle distillates on the
demand side, we can expect that
middle distillate prices are likely to
be higher than those of gasoline,
reversing the current price structure.
US economics
Prices
In the US in 2012, because of the
shale gas effect, the average price
for LPG was $636/t, while the
gasoline price reached $1129/t. The
prices of feedstock and products in
2012 were as shown in Table 3.
Results
In 2012 in the US, the IRR for
PolyFuel based on mixed LPG-LCN
feed is 13% with a net present
value of $12 million.
Investment and economics
hypothesis
The PolyFuel unit inside battery
limit (ISBL) capital investment cost
is $29 million for a capacity of 16
500 b/d. The depreciation duration
is 15 years, with a discount rate of
10% and a proft tax rate of 30%.
Utilities and catalyst cost
The utilities costs used in the
economic evaluation are shown in
Table 1. Overall, utilities and cata-
lyst costs for this case study were
calculated to be $19/t of feed or
$1.8/bbl of feed.
European economics
Prices
In Europe in 2012, the Brent crude
price was $112/bbl, while the
spread between gasoline and diesel
was $57/t and for gasoline/Jet A1
was $23/t. The prices of feedstock
and products in 2012 were as
shown in Table 2.
Results
Considering 2012 LPG and gasoline
(LCN) from an existing Prime G+
unit, currently sent to the gasoline
pool
Available LPG olefnic cut
currently sent to the LPG pool
with a ratio of 50%LPG:50%LCN.
The capacity chosen is 16 500 b/d
to illustrate a median case
for economic evaluation of
PolyFuel.
Feeds and products
The light cracked naphtha and LPG
cut from FCC are highly olefnic
streams. Olefns in the feed are
converted by oligomerisation in the
PolyFuel unit into middle distil-
lates. After hydrogenation, the
middle distillates produced can be
sold either as diesel or Jet A1
kerosene.
The quality of the remaining
gasoline is improved by the
oligomerisation process, olefn and
sulphur content being reduced,
while maintaining good octane
numbers. The gasoline product is
thus sent directly to the gasoline
pool.
3 0
3 5
2 5
2 0
1 5
1 0
5
1 2 0 1 0 0 8 0 6 0 4 0 2 0 0 2 0 4 0
I
R
R
,
%
Spread kerosene-gasoline, $/t
0
1 5 % I R R
9 6 $ /t
Figure 11 IRR sensitivity for PolyFuel based on mixed LPG-LCN feed according to middle
distillate gasoline price spread with 2012 USGC-based price
Utilities Cost Unit
Electricity 75 $/Mwh
HP steam 30 $/ton
MP steam 25 $/ton
LP steam 20 $/ton
Cooling water 0.20 $/ton
Utilities costs
Table 1
2012 - Europe $/t $/bbl
Brent 843 112
Premium 95 10 ppm 1037 124
Propane 861 74
Butane 859 74
LPG 860 74
Kerosene Jet A1 1014 129
Diesel 979 136
Feedstock and products prices in Europe
in 2012
2012 - US Gulf Coast $/t $/bbl
WTI 718 94
Premium gasoline 1129 136
Propane 523 45
Butane 750 64
LPG 636 55
Kerosene Jet A1 1018 130
Diesel 936 130
Feedstock and products prices in the
US in 2012
Table 2 Table 3
axens.indd 6 08/03/2013 11:49
www.eptq.com PTQ Q2 2013 21
rener and particularly the ability
to anticipate market needs in differ-
ent regions as constraints evolve.
PolyFuel and PolyNaphtha are trademarks of
Axens.
Marielle Gagnire is Technology Manager
for hydroprocessing and olens technologies
downstream FCC, especially oligomerisation
and etherication technologies, in Axens
Marketing, Technology and Technical
Assistance Department. She is an engineering
graduate from the Ecole Nationale Suprieure
de Chimie de Paris, and holds a post-graduate
engineering degree from the IFP School.
Annick Pucci is Deputy Product Line Manager
in the eld of light ends hydrotreatment and
a specialist in rening olens processing,
particularly for FCC efuent upgrading.
She holds a bachelors degree in chemical
engineering from Ecole Nationale Suprieure
des Industries Chimiques de Nancy, France.
Emilie Rousseau is a Strategic Marketing
Engineer in Axens Marketing Department. She
holds a chemical engineering degree from the
Ecole Nationale Suprieure des Ingnieurs en
Arts Chimiques et Technologiques de Toulouse,
a masters in chemical engineering from
Imperial College in London and a masters in
energy economics and corporate management
from IFP School.
It is in Europe where the
difference between renery yield
structure and market demand is
critical, especially since conven-
tional rening tools do not have the
exibility to reduce excess gasoline
production and to increase the
amount of middle distillates.
Moreover, with European reneries
facing increasing difculty in nd-
ing export markets for their excess
gasoline and given the tensions in
middle distillate supply, PolyFuel
should full a primordial role in
adjusting the gasoline-distillate
production to better t market
demand.
In other regions, new tendencies
such as shale oil and shale gas are
revolutionising the US market,
providing additional light products
and consequently inuencing
market balance and prices. Today
in the US, as a result of the impact
of shale gas on the cost of LPG,
PolyFuel is already protable for a
mixed feed of LPG and C
5
/C
6
cut.
The exibility of the new process
offers many advantages to the
The new technology is protable
taking into account todays US
market prices even if the middle
distillate price is lower than the
gasoline price. Indeed, as a result of
shale gas production, LPG prices are
low. Adding LPG (C
3
and/or C
4
) cut
in a PolyFuel unit lowers feedstock
costs and contributes to increased
protability, while maximising
middle distillates production in the
renery.
To reach 15% IRR for PolyFuel
with prices based in 2012 in the US
Gulf Coast, the middle distillate
price can be $96/t lower than gaso-
line. If the middle distillate price
were equal to the gasoline price
($1129/t) and the LPG price kept at
$636/t, the IRR would reach 28%.
Conclusion
With the world market for middle
distillates growing and a reduced
demand for gasoline in certain
regions, the new process for olenic
gasoline oligomerisation allows the
renery scheme to be adapted to a
maximum distillate mode.
Taking the Guesswork Out of CO
2
Removal
ProTreat Simulation Software
Get Real
Dont Guess!
SyngaS, Lng and Carbon Capture
ProTreat accurately predicts column performance.
100% Mass-Transfer-Rate based no ideal stages.
Grassroots? Revamp? Designing for Packing?
Using MDEA with piperazine? HotPot?
Build a virtual plant with the industrys
most advanced gas treating simulator.
www.ogt rt .com +1 281 970 2700 info@ogt rt .com
Learn how state-of-the-art simulation science can beneft you.
OGT-0117 PTQ Half Horizontal Ad.indd 1 11/13/12 10:40 PM
axens.indd 7 08/03/2013 11:49
Sulzer Chemtech
Legal Notice: The information contained in this publication is believed to be accurate and reliable, but is not to be construed as implying any warranty or guarantee of performance.
Sulzer Chemtech waives any liability and indemnity for effects resulting from its application.
Sulzer Chemtech, USA, Inc.
8505 E. North Belt Drive | Humble, TX 77396
Phone: (281) 604-4100 | Fax: (281) 540-2777
TowerTech.CTUS@sulzer.com
www.sulzer.com
sulzer.indd 1 11/03/2013 09:57
Strategies to control sediment and coke
in a hydrocracker
G
lobal supply of good-quality
crude is decreasing, while
at the same time the
requirement for clean and ultra-low-
sulphur-containing products is
increasing. Ebullated-bed resid
hydrocracking processes, known as
LC Finer or H-Oil, have gained
increasing interest during the last
decade due to their capability of
producing light and middle distil-
lates that are high in quality and
low in sulphur. Major economic
drivers for ebullated-bed processes
are run length, maintenance cost
and, most of all, conversion. Often, a
compromise has to be made
between operations and conversion
due to fouling and the limitations/
specifcations of sediments in the
heavy fuel oil produced from the
hydrocracking process. Controlling
fouling at optimal conversion can
provide signifcant operational and
performance beneft.
The limiting factor is the fouling
of the fractionation section, espe-
cially the bottom stream areas, and
specifcally the atmospheric and
vacuum column bottoms and
vacuum column furnace, as
observed by Putek, Sherwood and
McNamara.
1-3
The bottom stream
heat exchanger run lengths and
maintenance requirements set limits
to the overall unit conversion and
economic effciency. Fouling also
appears in the high- and mid-
pressure separators downstream of
the ebullated-bed reactors, in the
vacuum fractionator heater, and in
the atmospheric and vacuum frac-
tionators. In many cases, fouling
limits unit operation.
The deposit generation mecha-
nism from hydrocracked residues
Fouling control techniques enabled an ebullated-bed resid hydrocracker to
operate at high severity while the rate of fouling of critical areas was controlled
MARCO RESPINI, SILVIA EKRES and BRUCE WRIGHT Baker Hughes
ROBERT AJDLK Slovnaft
is mainly due to lost solubility of
asphaltenic fragments generated
by thermal cracking at reactor
temperatures. These cracked
asphaltenes may recombine easily
to more insoluble coke precursors.
When these molecules exceed their
maximum solubility in the resid-
uum matrix, they phase-separate
into non-solvated particles and
grow rapidly via polymerisation
reactions into coke particles of
suffcient size and agglomerating
potential to become a fouling
problem.
This mechanism is slowed and
partly controlled by catalytic
hydrogenation, which saturates the
free radical fragments of the
cracked asphaltenes at least at
the reaction section where thermal
cracking reactions are faster and
catalyst is present together with
hydrogen.
3
These reactions enable
much higher yields compared to
thermal processing, such as in
visbreaking or delayed coking.
Conversion via thermal cracking
leads to increased asphaltene
decomposition, which is moderated
due to hydrogen saturation reac-
tions that inhibit sediment
formation, as observed by Bartholdy
and Andersen.
4
However, typically,
the top 10% volume of the reactor
has no catalyst present, and in that
area reactions are purely thermal,
leading to uncontrolled conversion
of coke precursors to coke, as there
is no moderating or inhibiting effect
from the hydrogenation reactions
that occur in the lower reactor
section.
As a result, the increase in temper-
ature necessary to increase
conversion above certain limits
leads to increased sediments
and coke generation, creating
diffcult-to-control foulants that
deposit in critical plant areas and/or
www.eptq.com PTQ Q2 2013 23
I
n
c
r
e
a
s
i
n
g
f
o
u
l
i
n
g
r
a
t
e
Increasi ng conversi on
C o n ve rsio n lo ss
O p tim a l
o p e ra tin g
win d o w
F o u lin g
I n c re a se d m a in te n a n c e c o st
S h o rte r ru n le n g th s
Figure 1 Exponential fouling tendency increase with conversion increase
baker hughes.indd 1 08/03/2013 12:10
www.eptq.com
cause sediment formation in the heavy fuel oil. The
fouling tendency can be shown to increase exponen-
tially with conversion increase (see Figure 1). Therefore,
the value beneft of the conversion increase can be lost
to fouling and sediments. However, this fouling-to-con-
version relationship can also show that there is no major
advantage in decreasing severity whenever fouling rates
are acceptable or controllable. To do so would simply
result in lost conversion without a corresponding value
improvement in terms of sediment deposition control
and run lengths. This relationship brings about the
concept of optimising conversion as a function of the
rate of fouling and fuel oil sediment formation.
Fouling problems, their monitoring and control
Problem areas
The sections most prone to fouling are the atmospheric
column, vacuum column and preheat exchangers. At
very high conversion, the reactor and separator may
also suffer from high coke generation. Extensive foul-
ing of the separators and columns can lead to
unplanned shutdowns, downtime and lost production.
The same trends for conversion severity are valid for
sediment generation in the fuel oil. Below certain limits
of conversion the fuel will be stable, while above
certain limits the tendency to generate sediment with
time cannot be controlled.
From the above considerations, it is clear how setting
the proper operating conditions is important, as this
enables the best trade-off between conversion maximi-
sation and production of stable fuel and acceptable
rates of fouling. Optimal plant management requires
continuous control for resid product stability. The
stability is related to the tendency to produce fouling
deposits and generate sediment.
Optimal severity depends on the properties of the
feed being processed. The feed changes whenever the
refnery feed quality, residual feed make-up or the
plant feed rate changes.
1
Feed composition-related
factors that may infuence severity/conversion are the
stability reserve of asphaltenes in the vacuum resid
(often reported as p-value), the content of asphaltenes
and the intrinsic solubility of these asphaltenes. Low
stability reserve, high asphaltene content and poor
solubility will all contribute to an increased tendency
to generate coke and unstable residuum product and
fouling deposits.
The amount of metallic contaminants (especially
sodium and iron) is another factor that can impact
process performance by affecting catalyst performance
and, in some cases, by increasing coking tendency,
favouring dehydrogenation and conversion of coke
precursors those less soluble thermally cracked
asphaltenes into coke.
To avoid deposit generation, the LC Finer catalyst
plays a major role.
3
The type of catalyst utilised in the
process can have a great effect. In ebullated-bed reac-
tors, the catalyst is changed continuously to maintain
catalyst activity and to remove metals from residue oil.
The effect of the catalyst operations on the process and
fouling can be monitored easily with sophisticated
effciency trays must be used to reduce refux ratio
and lower the overhead cooling load. Flue gas
turbines, energy recovery hydraulic turbines, energy-
saving motors, frequency conversion motors and air
fow regulation systems of compressors should be
employed to recover pressure energy and reduce elec-
tric energy consumption. High-effciency intensifed
burners need to be used in furnaces to improve eff-
ciency. The exhaust temperature of furnaces should be
reduced to improve thermal effciency by 2-3%. New
insulation material needs to be employed to reduce
the heat loss of equipment and pipelines.
Low-temperature heat should be utilised by apply-
ing low-pressure steam generation technology,
low-temperature Organic Rankine Cycle (ORC)
systems and Kalina Cycle systems.
Energy optimisation of the area should be carried
out. Heat integration between the refnery and local
co-generation power plant needs to be realised. A
large quantity of low-temperature heat in a refnery
cannot be recovered because a heat sink is not availa-
ble, and this portion of low-temperature heat may be
used as a heat source for demineralised water and
boiler feed water in a cogeneration power plant. Also,
integration with the local chemical plant is important
to achieve material exchange and optimisation, as well
as optimum energy use. Integrated energy optimisa-
tion of the area should be carried out rather than
energy optimisation of a single refnery.
Conclusions
The priority of refning technology development and
the confguration of refning units in China have
distinct characteristics. During a long period in the
future, the FCC unit will still be the main secondary
conversion unit for gasoline and diesel production.
Therefore, improving the quality of FCC gasoline
and diesel is very important for oil product quality
upgrading technology in the future.
With the trend towards poor-quality crude oils,
future refneries should further optimise process fows
and change unit confguration so as to improve the
comprehensive utilisation rate of resources, and to
meet the requirements of energy savings and emis-
sions reduction.
More detailed classifcation of crude oil fractions
and the consideration of various narrow-cut process-
ing technologies will actively promote the low carbon
emission of refneries. Narrow-cut processing will
become the development trend in the overall process
fow optimisation of refneries in the 21st century.
Sun Lili is Vice President of Sinopec Engineering Incorporation in
charge of processing solution studies, engineering, construction and
startup of grassroots refneries and refnery modifcation projects. She
has over 20 years experience in engineering design and startup of
hydroprocessing units, has won several Prizes of National Scientifc
and Technological Progress, and holds a BS in petroleum refning from
China University of Petroleum.
Email: sunlili.sei@sinopec.com
www.eptq.com
www.dpharp.com
High Performance
Unmatched precision and real world
accuracy
Unconditional long term stability with lowest
Total Cost of Ownership (TCOj
Multi-sensing functionality
Safety
lnherently fail-safe sensor with no
undiscovered failure modes.
Certifed to lEC61508 as standard for single
use SlL2 and dual use SlL3
Robustness
Rugged, yet compact construction
Multiple material options for harsh industrial
environments
ln-built overpressure mechanism for
increased reliability
A New Standard
in Pressure
Measurement
EJA-E-270x86.indd 1 07/09/12 13:50
sinopec.indd 8 10/12/12 13:02:15
baker hughes.indd 2 08/03/2013 12:10
monitoring techniques described
below. The effect of the catalyst-
changing sequence and the
catalysts age are directly observed
with these techniques.
Catalytic hydrogenation has two
additional effects that are related to
an increase in the hydrogen-to-
carbon ratio from the hydrogena-
tion reactions. These reactions
increase conversion due to the
decrease in boiling point of light-
ends fuids, and they reduce the
capability of the residuum to
disperse/solvate asphaltenes, as
some aromatic carbon content is
lost. Aromatics are excellent
solvents for asphaltenes. The above
reactions reduce the stability
reserve of the produced LC Finer
residuum compared to the feed
residuum due to lost aromaticity.
Asphaltene thermal cracking
produces less soluble, more
condensed, aromatic, carbon-rich
molecules, often called polynuclear
aromatic (PNA) compounds.
Hydro-cracking decreases the capa-
bility of the main oil phase to
solvate asphaltenes due to hydro-
genation of the aromatics, further
decreasing asphaltene stability.
This reduction in stability reserve
leads to rapid fouling of the down-
stream reaction section, starting
with the separators and progress-
ing to the columns and preheat
exchangers. Finally, reduced stabil-
ity leads to precipitation of the
most insoluble fraction of
asphaltenes, resulting in a high
concentration of sediment, as meas-
ured by the Hot Filtration Test
(HFT). There is also a tendency to
experience increased sediment
formation over time when LC Finer
resid is stored in tanks. In fact, a
lack of suffcient solvation enables
the asphaltene aggregates to come
into contact and grow in size to
levels that are above the one
micron limit, as measured by the
HFT procedure.
Tests used to control operations to
minimise fouling
Control of process severity is tradi-
tionally done by measurement of
the sediments by the HFT on the
produced vacuum bottom residue
and the heavy fuel oil derived from
www.eptq.com PTQ Q2 2013 25
fuxing this residue. When fuel is
submitted to ageing procedures
that can reproduce long-term stor-
age behaviour for sediment
deposition, the HFT is usually
reported as Total Sediment
Potential and Total Sediments by
Accelerated Aging.
A toluene insolubility test is often
used to detect the amount of coke
formed in the process by dissolving
the vacuum residuum and fltering
it through a 0.45 micron porosity
flter.
Other common tests measure the
asphaltene stability in the residue
streams by adding an asphaltene
precipitant (paraffn) to the cracked
resid or fuel, and determining the
quantity of the precipitant needed
to cause focculation. The detection
methods for asphaltenes precipita-
tion may be based on paper
chromatography (spot test) or on
microscopy methods, known gener-
ally as p-value. Apart from
detecting the so-called stability
reserve by dosing with paraffnic
solvents, the sediments can also be
measured by fltration methods.
These methods suffer from exten-
sive problems due to inherent
errors and the low accuracy of such
analytical techniques. In practice,
when these methods are applied,
the problems lead to a reduced
capability to operate the unit at
proper severity, with the following
consequences:
Lost conversion
Ineffective fouling control, lead-
ing to reduced run lengths
Problems with sediment forma-
tion in the produced heavy fuel oil.
Several parameters are usually
monitored to track the impact of
fouling on unit operations. In the
case of the Slovnaft LC Fining unit,
these include:
Decrease in atmospheric and
vacuum bottom temperature read-
ing due to coking
Decreased temperature at the
column bottom and skin of the
bottom suction line; this drops
when a coke layer grows inside the
pipe in both fractionators
Skin temperature increases in the
vacuum column preheater (mainly
in the convection zone)
Pressure drop increase and/or
outlet temperature increase for heat
exchangers downstream of the
vacuum column.
Baker Hughes techniques for
conversion unit optimisation
To overcome the diffculties of the
traditional test and control meth-
ods, Baker Hughes has developed
proprietary technologies for analys-
ing the operational impacts of
conversion units on fouling
tendency and fuel stability. The
company has run successful appli-
cations for more than 10 years in
visbreakers and delayed cokers.
These monitoring programmes
enable refners to optimally manage
plant severity and fouling control.
The objective is to achieve maxi-
mum conversion with the best
balance between conversion and
run length, while avoiding uncon-
trollable deposition of foulants. The
fnal target is to determine the best
economic scope for operating the
LC Finer.
Slovnaft was interested in devel-
oping more advanced monitoring
methods in co-operation with Baker
Hughes. The methods described
below were implemented along
along with routine refnery process
monitoring and are being used to
help maintain the process at the
optimal process window. The
methods can also identify sudden
changes in the process that might
affect the fouling control.
Continuous reporting and informa-
tion exchange take place with
refnery operations as well as with
the refnerys technology and
process department.
Baker Hughes methods are based
on the control of two aspects
related to controlling the impacts of
severity/conversion on fouling and
fuel oil sediment formation:
Stability measurement for asphaltene-
containing residuals
This test is based on the addition of
paraffn precipitants. Asphaltene
focculation is detected by accurate
laser turbidimetric measurement in
the near-infrared region. This tech-
nique provides good accuracy,
reliability and low error (less than
2% compared to the traditional
p-value), rapid results (less than 30
baker hughes.indd 3 08/03/2013 12:10
www.eptq.com
minutes for a measurement), and has a wide database
of correlations with fouling and sediment formation.
The technique can be used effectively to optimise unit
conversion.
Coke and coke precursors particle measurement
High severities in thermal cracking reactions cause
coke generation, with an exponential trend at high
conversion. For LC Finer processes, the generation of
coke particulates must be controlled or rapid fouling
takes place. Coke particulates cause problems with the
catalyst and generate excessive deposits in the unit
downstream of the reactors. This impact can be seen as
fouling and sediments in the fuel. Baker Hughes has
developed a patented method based on sizing and
counting coke particles, and highly insoluble
asphaltenes. The technique uses a light obscura-
tion-based instrument, and is very accurate in the
trace-level range typical of coke and sediments genera-
tion in LC Finer operations. The particle analysis
technique enables clear improvement on unit manage-
ment compared to traditional methods.
Conversion and run length maximisation
The above methods enable the optimisation of the
units operational parameters to maximise severity and
conversion, while at the same time controlling the rate
of fouling and formation of fuel sediments.
Fouling control for atmospheric and vacuum column section
Fouling in the atmospheric and vacuum column is
related to reaction severity and the fash zone tempera-
ture in the column. Raising severity increases
conversion, while an increase in fash zone tempera-
ture enables more distillate recovery. Both aspects have
an exponential impact on column fouling, causing
column bottom coking and fouling above the fash
zone. Fouling in the column can lead to black gas oil
and sometimes excessive pressure drop across the
column. Baker Hughes uses antifoulant additives and
the monitoring approach to control fouling in this
section and to monitor column operations.
Control of skin temperature increase due to coking in the
vacuum heater
Vacuum heater feed is the thermally cracked, highly
unstable atmospheric bottom residuum that has a high
tendency to foul/coke. This section is a critical area of
the unit, where uncontrolled fouling will result in
decreased run length. Antifoulant additives enable
control of the rate of fouling in the vacuum heater,
while helping the plant to maximise conversion.
Fluxing of ebullated reactor unit streams
A common practice in ebullated- bed conversion units is
to add aromatic hydrocarbons, such as FCC heavy cycle
oil, as fux oil to the residue stream and/or to the heavy
cracked fuel oil products. The fux oil will increase the
stability of the asphaltenes and decrease the tendency of
these streams to form coke. Addition of the fux oil can
be performed in various locations of an ebullated
SAMSON AG MESS- UND REGELTECHNIK
Weismllerstrae 3
60314 Frankfurt amMain Germany
Phone: +49 69 4009-0 Fax: +49 69 4009-1507
E-mail: samson@samson.de
Internet: www.samson.de
SAMSON GROUP www.samsongroup.net
A
0
1
1
2
0
E
N
Partner with
the Best
With over 50 independent subsidiar-
ies and more than 220 engineering
and sales ofces spread across the
world, SAMSON ensures the safety
and environmental compatibility of
your plants on any continent.
To offer the full range of high-quality
control equipment used in industrial
processes, SAMSON has brought
together highly specialized compa-
nies to formthe SAMSON GROUP.
baker hughes.indd 4 11/03/2013 12:56
www.eptq.com PTQ Q2 2013 27
polynuclear aromatic-type struc-
tures with an increasingly higher
percentage of aromatic carbon,
which decreases their solubility in
the resid matrix.
The solvating power/solubility
blending number (SBn) determines
the capability of the resid oil to
hold asphaltenes in suspension.
This information can be applied to
the effects of distillate fux oil on
residue fuel oil stability. Catalytic
hydrogenation decreases the resi-
dues ability to solvate asphaltenes
by reducing the aromaticity of the
effective dose rate of antifoulant
chemistry will enable the unit to be
run at high severity/conversion
levels to maximise run lengths and
therefore profts.
Several stability-related indices
can be calculated from the ISI
measurement. The asphaltene insol-
ubility number (In) is a measure of
the tendency of the asphaltenes to
precipitate. This tendency is greatly
increased by thermal cracking (see
Figure 3). In the LC Finer process,
thermal cracking causes dealkyla-
tion of asphaltenes, resulting in
reactor process. This practice should
allow an increase in conversion. The
impact of this operation depends,
apart from quantity, on the resid
feed quality and the fuxant quality.
In some situations, the fuxant is not
able to stabilise the residual feed.
One can determine the capability of
the fuxants to stabilise the streams
using the stability test and the coke
particle measurements, and the
probable stabilisation that might be
achieved as a function of the
percentage of fuxant. This data
enables optimal choice of the type
and amount of fux oil for various
streams.
Ebullated reactor unit bottom stream
exchangers fouling control
The stability of the asphaltenic
compounds in the cracked residue
will impact fouling in the bottom
stream exchangers. These exchang-
ers typically see extensive fouling.
Fouling can be controlled by reduc-
ing severity until the limit at which
excessive fouling potential is
reached, and by injecting antifoulant
additives. The amount of antifoulant
additive can be balanced with
reduced severity to achieve an opti-
mal conversion limit.
Results of analysis methods at
Slovnaft
Stability analysis
The stability analysis, called
Infrared Stability Index (ISI), meas-
ures the stability reserve of
asphaltenes in LC Finer unit-
produced resid by the addition of
an asphaltenes precipitant. Onset of
precipitation is determined by
turbidimetry. Laser signal attenua-
tion in the near-infrared region
detects the asphaltenes aggregation
point.
Typically, units are operated
close to the solubility limit of
asphaltenes (see Figure 2), yielding
low ISI stability reserve values. To
produce more stable, high-
quality bottom products for heavy
fuel oil blending and to avoid a
critical fouling situation, the unit
severity must typically be
decreased so much that conversion
and overall unit economics suffer.
Understanding this operations
envelope and the application of an
S e ve re fo u lin g
M e d iu m fo u lin g
L o w fo u lin g
L C F in e r va c u u m b o tto m
L C F in e r a tm o sp h e ric b o tto m
L C F in e r va c u u m re sid u a l fe e d
ISI stability reserve range
Figure 2 Average ISI stability reserve values at Slovnaft
L C F in e r
fe e d
L C F in e r
m e d iu m re a c to r
te m p e ra tu re
L C F in e r
h ig h re a c to r
te m p e ra tu re
I
n
Figure 3 Average asphaltene insolubility numbers
L C F in e r
fe e d re sid
L C F in e r
a tm o sp h e ric
re sid
L C F in e r
va c u u m re sid
S
B
n
Figure 4 Differences in residuum solvating power (SBn)
baker hughes.indd 5 11/03/2013 12:56
3225 Gallows Road, Fairfax, Virginia 22037-0001, USA
www.exxonmobil.com/tsl +1-703-846-2568 fax +1-703-846-3872 tsl.email@exxonmobil.com
ExxonMobil Global Leader in
Fuels and Lubes Process Technologies
ExxonMobil Technologies are applied across our corporation and by licensees worldwide in a growing
list of process industry applications. Take advantage of our vast experience and ongoing commitment
to continuously improve our industry leading technologies.
Advanced Sol ut i ons
for t he Wor l ds Toughest
Energy Chal l enges
T ECHNOL OGI ES:
exxon.indd 1 07/03/2013 13:36
www.eptq.com PTQ Q2 2013 29
content and size distribution of
solids/coke precursors and coke
particles in a certain range of
particles.
The advantages of this method in
comparison to the traditional tolu-
ene insoluble method are:
The quantitative accuracy of the
method
The particle size range and its
changes can be detected
Very sensitive to process changes
Very quick test method.
The amount and size can be
related to fouling and unit run
length.
Typically, coke particle formation
exhibits exponential growth with
increased reactor temperature and
severity. Figure 7 illustrates the
increase in coke particle index as a
function of unit severity (conver-
sion). The particle index considers
both particle size and amount.
Particle distribution is a distribu-
tion of solids, coke precursors and
coke particles at micron scale. The
changes in coke particle size and
progressively extracted, starting
from more soluble ones, such as
heptol solvent with a high amount
of heptane, to less soluble ones,
requiring higher toluene to be
solvated. The results are shown in
Figure 6. There is a clear shift
towards less soluble asphaltenes;
the total amount of asphaltenes in
the vacuum resid is less when
compared to vacuum residue due
to hydrocracking of asphaltenes.
However, the remaining
asphaltenes are shifted towards
more insoluble species. In pure
toluene, the LC Finer feed
produced only trace quantity of
asphaltenes, while a relatively high
amount of asphaltenes was found
in the thermally cracked bottoms.
These asphaltenes are coke precur-
sors in nature (they are structurally
and in terms of solubility very close
to coke), which confrms the CSI
data shown in Figure 6.
Particle analysis
The particle index measures the
residue phase. The cracked product
residue has a lower SBn index
when compared to the parent feed
resid and therefore will impart a
negative impact on asphaltene or
polynuclear aromatics stability.
Some of the asphaltene solvating
capability is recovered when the
cracked atmospheric resid is
distilled through the vacuum
column, as paraffns are preferen-
tially removed as light and heavy
vacuum gas oil. Aromatics are not
thoroughly distilled in the vacuum
tower, leaving them in the vacuum
residue and increasing its capability
to disperse asphaltenes.
There is a clear decreasing trend
of SBn due to catalytic hydrogena-
tion going downstream from the
feed to the atmospheric resid, and
then a partial recovery when atmos-
pheric resid is distilled through the
vacuum column (see Figure 4).
Paraffnic compounds distill up the
vacuum tower, leaving behind an
increased concentration of aromatic
compounds in the vacuum residue,
which increases dispersion of the
asphaltenes. Typically, this is seen
as improved stability of vacuum
column bottom products over what
is in the atmospheric tower bottom
products.
The coking stability index (CSI)
measures the most insoluble
asphaltenes and coke precursors,
which have the greatest precipita-
tion potential. Above 400C, coke
precursors are quickly converted to
coke, leading to fouling deposits.
The thermal cracking component of
the process decreases the degree of
alkylation of asphaltenes, making
them less and less soluble in the
resid at higher temperatures. These
asphaltenes become PNA
compounds, which form particles.
TheCSIdataareshowninFigure5.
The generation of more insoluble
asphaltenes in the Slovnaft LC
Finer process was also measured by
an asphaltene fractionation tech-
nique. Resid feed and vacuum resid
asphaltenes were precipitated with
an excess of heptane and redis-
solved in various heptane/toluene
mixtures (heptol) with increasing
toluene, which provides an increas-
ing capability to solvate
asphaltenes. The fuids were
L C F in e r
fe e d
L C F in e r
m e d iu m re a c to r
te m p e ra tu re
L C F in e r
h ig h re a c to r
te m p e ra tu re
Figure 5 Average coke precursor stability
h /t 7 0 /3 0 h /t 5 0 /5 0 h /t 3 0 /7 0 h /t 2 0 /8 0 h /t 1 0 /9 0 h /t 0 /1 0 0
A
s
p
h
a
l
t
e
n
e
s
c
o
n
t
e
n
t
,
w
t
%
Heptol solvent with increasing aromaticity
L C F in e r va c u u m re sid
F e e d va c u u m re sid
Figure 6 Asphaltenes fractionation results
baker hughes.indd 6 08/03/2013 12:11
30 PTQ Q2 2013 www.eptq.com
Particle size also affects fouling of
the furnace tubes and heat exchang-
ers, as greater-sized particles easily
deposit, even when velocities are
high. Coke and solid particles
create hard deposits on heat
exchanger surfaces, requiring
mechanical cleaning.
Finally, the formation of particles
of 1.6 microns or more from submi-
cron particles by agglomeration
and growth leads to an increase in
heavy fuel oil sediments.
The average particle size distribu-
tion histogram for the Slovnaft LC
Finer unit is presented in Figure 8.
It is interesting to note how, even
below thermal cracking tempera-
tures (below 400C), the reactions
leading to fouling related to aggre-
gation of particles are not stopped.
Clearly, there is an increase in
particle concentration above 2.5
microns radius in the vacuum
residuum compared with the
atmospheric resid. These larger
particles give a high contribution to
the total mass of coke particles and
they have much more impact on
fouling compared to the smaller
ones.
Antifoulant programme
Slovnaft decided to perform a trial
with the Baker Hughes antifoulant
programme to improve unit run
length. The dosing points of the
antifoulant are indicated in Figure
9. Before the trial, a baseline for
process and monitoring values was
created. (No antifoulant was used.)
In addition, the antifoulant
programme was compared to other
tested antifoulants, referred to as
ADD1 and ADD2 (see Figures 10
and 11).
Severity of fouling for an LC
Finer depends on unit severity,
catalyst type and feedstock qual-
ity. Therefore, one needs to
consider process conditions when
performing a comparison between
different cases. The reported
results are from runs with feed
quality in line with reference base
case data and with unit severity
(reactor temperatures at a given
processed feed rate) that was, on
average, higher.
The programme was adopted and
modifed for an LC Finer from
particles may lead to severe foul-
ing. Bigger particles easily
accumulate in various places in the
fractionation section, such as
separators and column bottoms.
number can be easily detected and
correlated to changes in operation
and antifoulant treatment. Growth
in particle size and an increase in
the concentration of various size
Figure 8 Coke particle distribution in atmospheric and vacuum residuum samples
VT 5
C 3 0 1
Wa sh
o il
C 3 0 1
B o tto m
H 3 0 1
C 2 0 1
E309A E309B
E302A
Vac. heater
H301
Vac.
C301
ATM
C201
LP/HT
V103
P305
P304
E302B
E303
Figure 9 LC Finer: Baker Hughes antifoulant dosing points
0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 3 . 0 3 . 5 5 . 0 6 . 0 7 . 5 1 0 . 0 1 2 . 5 1 5 . 0
Size, microns
L C F in e r va c u u m re sid
L C F in e r a tm o sp h e ric re sid
C
o
k
e
p
a
r
t
i
c
l
e
s
i
n
d
e
x
(
m
a
s
s
c
o
n
c
e
n
t
r
a
t
i
o
n
c
o
k
e
i
n
r
e
s
i
d
.
)
S lo vn a st
lo w
se ve rity
S lo vn a st
m e d iu m
se ve rity
S lo vn a st
h ig h
se ve rity
L C F in e r A
lo w
se ve rity
L C F in e r B
m e d iu m
se ve rity
L C F in e r C
ve ry h ig h
se ve rity
Figure 7 Coke particles as a function of severity (Slovnaft data compared to other units)
baker hughes.indd 7 08/03/2013 12:11
Flexible H
2
S Removal
reliability. Over 200 licensees in 29 countries are meeting todays compliance requirements by
turning nasty H
2
and environmentally sound, LO CAT is the technology of choice for H
2
S removal / recovery.
LO CAT
www.merichem.com
Sweet Solutions
-
-
-
-
C
M
Y
CM
MY
CY
CMY
K
Merichem Company LO-CAT.pdf 1 3/4/2013 7:36:22 AM
merichem.indd 1 08/03/2013 10:47
Y ou r s e a r c h f or e x p e r t i s e e n d s
w i t h u s . We b u i l d b e t t e r b oi l e r s .
Satisfed customers discover the power of our practical knowledge the ability to design and build
boilers that operate effciently, safely and cleanly in a variety of industrial applications, including
refning, petro-chemical and power generation. The know-how of our engineers and technicians
combined with our expanded facilities and equipment, including a new membrane panel welding
machine results in economic value and competitive advantage for you. We ve been designing
and building boilers for people who know and care since 1996.
www.RENTECHBOILERS.COM
rentech.indd 1 12/03/2013 09:42
Vacuum heater optimisation
An analysis of unit process data and
monitoring results indicated that
another area of improvement to
further control fouling was the possi-
ble optimisation of vacuum column
heater operating parameters.
A process simulation was
performed by Baker Hughes to
evaluate several process variables
that seemed likely to impact the
downstream fouling-related param-
eters. As a result of that modelling,
one of the controllable process
parameters was identied as
having a signicant impact on
downstream fouling.
Based on this model, a plan for a
test run was agreed upon between
Slovnaft and Baker Hughes. The
impact from the change in process
Figure 12. The fouling control
additive was able to increase the
ability of the atmospheric bottoms
stream to hold asphaltenes, thereby
allowing increased operational
severity.
Slovnaft also decided to test the
impact of terminating the antifou-
lant programme during a treated
run. The antifoulant programme
was stopped for about one week to
determine its impact on operations.
On the one hand, the test period
was rather short to clearly indicate
degradation of signicant process
parameters. On the other hand, the
particle index measurements
showed a signicant increase in
total particle count during the time
when no antifoulant programme
was in place (Figure 13).
high-severity and high-temperature
applications (such as visbreaker,
coker). The programme improves
operations through the control of
several related mechanisms:
Thermal degradation of oils
Asphaltene destabilisation
Catalytic effects of metals in feed-
stocks and on equipment surfaces
Precipitation of organic and inor-
ganic solids.
On average, application of the
antifoulant programme, together
with monitoring and optimisation
of renery unit operations, have
shown good improvements in
performance at zones critical for
fouling in this unit. The most criti-
cal aspects were coke accumulation
in the atmospheric and vacuum
bottom columns, coking of the
vacuum bottom column suction
line, and increased skin tempera-
tures in the vacuum heater
convective section. Good results
were also obtained due to an opti-
misation of the dose rates between
the three injection points available:
at the LC Finer atmospheric
bottom, in front of the vacuum
heater and at the vacuum bottom
(see Figure 9). Optimisation of the
programme was possible through
the use of monitoring data.
Comparison of the decay of ther-
mocouple readings at the vacuum
column bottom suction line clearly
showed a marked improvement
compared to the baseline run
before implementing the Baker
Hughes programme (see Figure 10).
Fouling of the vacuum heater
showed very good performance. A
comparison of skin temperature
measurements for the vacuum
heater convection zone (where foul-
ing had been most signicant) for
different runs is shown in Figure
11. Taking control over vacuum
furnace convection zone fouling
enabled an increase in the furnace
outlet temperature from 375C
(707F) to 388C (730F), while
maintaining acceptable fouling
rates (applied in a middle of run
BAKER_3).
The impact of antifoulant on the
asphaltene stability reserve in
Slovnafts LC Finer atmospheric
bottom stream during medium-
severity operations is presented in
www.eptq.com PTQ Q2 2013 33
T
e
m
p
e
r
a
t
u
r
e
,
C
Operat i ng days
C 3 0 1 T b o tto m lin e B A K E R 3
C 3 0 1 T b o tto m lin e A D D 1
C 3 0 1 T b o tto m lin e A D D 2
C 3 0 1 T b o tto m lin e B A K E R 2
C 3 0 1 T b o tto m lin e B A K E R 1
C 3 0 1 T b o tto m lin e B L A N K
* T h e in le t te m p e ra tu re s to th e su c tio n lin e a re sim ila r fo r e a c h ru n .
* * T h e B A K E R _2 ru n wa s in te rru p te d d u e to th e lo ss o f u tility su p p ly a n d
c o n se q u e n t lo ss o f k e y e q u ip m e n t ( m e c h a n ic a l issu e s) , a n d n o t d u e to fo u lin g .
Figure 10 LC Finer vacuum bottom suction line temperature decay due to fouling
Operat i ng days
T
e
m
p
e
r
a
t
u
r
e
,
C
H 3 0 1 T c o n v1 B A K E R 3
H 3 0 1 T c o n v1 A D D 1
H 3 0 1 T c o n v1 A D D 2
H 3 0 1 T c o n v1 B A K E R 2
H 3 0 1 T c o n v1 B A K E R 1
H 3 0 1 T c o n v1 B L A N K
Figure 11 LC Finer vacuum heater skin temperature increase due to fouling
baker hughes.indd 8 08/03/2013 12:11
34 PTQ Q2 2013 www.eptq.com
With o u t d isp e rsa n t With B a k e r H u g h e s d isp e rsa n t
I
S
I
s
t
a
b
i
l
i
t
y
r
e
s
e
r
v
e
H F T se d im e n ts c o n tro l lim it
Figure 12 Antifoulant impact on typical Slovnaft LC Finer atmospheric bottom stability
C
o
k
e
p
a
r
t
i
c
l
e
s
i
n
d
e
x
Operat i ng days
Va c . re sid .
A tm . re sid .
A n tifo u la n t
tre a tm e n t
re sta rte d
A n tifo u la n t
p ro g ra m
p a u se d
Figure 13 Impact on the coke particle index when antifoulant programme stopped
Operat i ng days
Va c . b tm . p a rtic le s
A tm . b tm . wB P I
P ro c e ss p a ra m e te r
Figure 14 Correlation of the coke particles index and control process parameter
Progressive changes in this parame-
ter continued to lower the vacuum
bottom coke particle index. All the
while, the atmospheric bottom coke
particle index remained relatively
consistent, indicating a reduction
in downstream fouling impact.
However, there was a point at
which severity/conversion could be
negatively impacted. Further analy-
sis and testing determined the
optimum range of operation for
this parameter to maintain or even
maximise severity/conversion with
respect to the relative reduction in
fouling potential.
Conclusions
In ebullated-bed hydrocracking
units economics are greatly affected
by fouling. Well-managed fouling
control can allow the plant to oper-
ate at high severity while
controlling the rate of fouling in
critical areas. Baker Hughes, in
co-operation with Slovnaft, demon-
strated methods for studying and
controlling fouling phenomena. The
stability method detects tendency
towards asphaltene destabilisation
and precipitation. The particle
method detects the amount and
size of coke particles generated in
the reactors. Reduction in fouling
was demonstrated with the use of
antifoulant, and a change in fouling
was also demonstrated with a
change in antifoulant dosage.
Lastly, operational optimisation
was shown to further enhance foul-
ing control. The critical fouling
threshold was identied and able
to be monitored.
Antifoulant injection delivered a
good performance in controlling
fouling in critical unit sections and
in increasing run length, while
keeping the units performance
within a high-severity/high-
conversion operational window.
Satisfactory performance was
achieved even in cases of deterio-
rating feed quality and during unit
upsets. The performance of the
antifoulant was additionally
enhanced with monitoring capabili-
ties and attention to optimum
operational conditions.
Target values were established
for critical parameters. The LC
Finer operation used these target
Figure 14 demonstrates that the
primary fouling indicator, the
vacuum bottom coke particle index,
reduced signicantly as the heater
process parameter was changed.
parameter would be monitored by
measuring the fouling indicators,
and by normalising them at
constant process parameters with a
proprietary procedure.
baker hughes.indd 9 08/03/2013 12:11
www.eptq.com PTQ Q2 2013 35
MBA from the University of Houston. He is a
registered professional engineer in the state
of Texas, and is an inventor or co-inventor
on eight US patents and has published nine
technical papers.
Robert ajdlk is a Slovnaft Bratislava Refnery
Technologist with the MOL Group. During the
past two years, he has been a technologist
for the refnerys LC Fining unit and FCC
pretreater unit. He is a graduate of the Slovak
Technical University, Bratislava, Faculty of
Chemical Technology, and holds a PhD in
chemical engineering and process control.
Email: robert.zajdlik@slovnaft.sk
she holds a masters in chemistry from the
University of Novi Sad, Serbia, and a PhD in
chemistry from the University of Vienna.
Email: silvia.ekres@bakerhughes.com
Bruce Wright is a Baker Hughes Senior
Technical Support Engineer in the Industrial
Technology Department in Sugar Land, Texas,
specialising in the hydrocarbon process
industries. He has more than 30 years
industry experience and is currently involved
in technical support and troubleshooting
of refnery fouling problems. He holds a BS
in chemical engineering from Rensselaer
Polytechnic Institute, Troy, New York, and an
control values to drive the process
close to the fouling threshold where
rapid exponential fouling starts,
thus maximising severity/conver-
sion with respect to an acceptable
and controllable rate of fouling. The
approach can be used to determine
the best economical operating
window, even with changes in
feedstock and operations, because
the demonstrated methods can
rapidly detect the impact of
fouling.
LC Fining is a mark of Chevron Lummus Global.
This article is based on a paper presented at the
International Bottom of the Barrel Technology
Conference & Exhibition, Rome, May 2012.
References
1 Putek S, Gragnani A, First resid hydrocracker
to produce stable, low-sulphur diesel fuel
from ural vacuum residue, ERTC 10th Annual
Meeting, Vienna, 2005.
2 Sherwood D E Jr, Barriers to high conversion
operations in an ebullated-bed unit relationship
between sedimentation and operability, NCUT
Workshop, Edmonton, 2000.
3 McNamara D J, Sherwood D E Jr, Bhan
Opinder K, Getting more out of your resid
upgrading unit, 6th BBTC, Barcelona, 2008.
4 Bartholdy J, Andersen S I, Changes in
asphaltene stability during hydrotreating,
Energy and Fuels, 2000.
Marco Respini is a Baker Hughes Technology
Development Specialist in the Industrial
Technology Group in Europe, specialising in
refnery process fouling control. He has 14
years refning experience and is currently
involved in developing new technologies for
monitoring fouling and severity control in
resid conversion processes. He holds a degree
in industrial chemistry from Milan University
and is a registered professional chemist in
Italy. An inventor of two US patents, he has
published fve technical papers and four
conference papers on visbreakers and heavy
fuel oil stability problems.
Email: marco.respini@bakerhughes.com
Silvia Ekres is a Baker Hughes Account
Manager responsible for Downstream
Chemicals, Industrial Portfolio product
line sales, and specialises in refnery and
petrochemical applications, including
mitigating fouling-related issues in LC Fining
units. The author of several publications,
baker hughes.indd 10 08/03/2013 12:12
dupont.indd 1 23/2/12 15:04:26
Simulation of bitumen upgrading
processes
T
he Canadian oil sands indus-
try currently produces about
1.6 million barrels of
combined mining and in-situ-based
bitumen. About 55% of it is
upgraded to synthetic crude oil
(SCO) and then processed in
Canadian and US refneries.
1
The
production of transportation fuels
from Canadian bitumen requires
either new, integrated upgrading
and refning facilities, or the updat-
ing of existing refneries that use
conventional crudes to allow more
input of bitumen feedstocks.
2
In
either case, it is important and
useful to model and simulate the
entire upgrading and refning
scheme under different process
confgurations and product scenar-
ios, to minimise process-related
energy intensity while achieving
the best economic benefts. This can
provide guidelines to existing
upgrading and refning operations
as well as to process design for new
upgraders and refneries.
There are two major primary
bitumen upgrading technologies:
coking and hydroconversion.
3,4
Historically, coking has predomi-
nated as the choice for primary
upgrading.
5
As the frst step to
producing a bottomless (zero resi-
due) SCO, it handles easily the
higher solids and water contents
in mining-based bitumen. The
byproduct coke helps to trap solids,
as well as concentrate and remove
metals, and some of the sulphur
and nitrogen contained in bitumen.
However, the total liquid product
yield is relatively low due to coke
formation.
6
In comparison, ebullat-
ed-bed catalytic hydroconversion is
more selective towards a total
Two bitumen upgrading processes currently used in the Canadian oil sands
industry were investigated using modelling and simulation techniques
ANTON ALVAREZ-MAJMUTOV and JINWEN CHEN CanmetENERGY, Natural Resources Canada
MUGUREL MUNTEANU WorleyParsons
liquid yield due to hydrogen addi-
tion.
7
In commercial operations, the
conversion of bitumen is not 100%
in the hydroconversion unit, gener-
ating a small portion of residue that
is either further processed in a
coking unit or withdrawn as
byproduct. Therefore, bitumen
upgraders currently use either
coking or ebullated-bed hydrocon-
version, or a combination of both.
8
Both of the two primary upgrad-
ing processes produce liquid
products with boiling ranges simi-
lar to those of conventional crudes.
However, they have high concen-
trations of impurities, such as
sulphur and nitrogen.
9
The down-
stream hydrotreating units remove
these impurities to produce sweet
blending feedstocks for SCO with-
out changing the boiling range of
the liquid too much. The SCO boil-
ing range is essentially controlled
by the primary upgrading step. In
reviewing some of the major chal-
lenges that the oil sands industry is
facing, bitumen upgraders need to
capitalise on, or address, the
following:
Take advantage of some
relatively minor upgrading at the
recovery stage
Take advantage of the necessity
to move to alternative energy and
hydrogen sources, particularly
internally generated residues,
which may greatly infuence the
selection of the main upgrading
process
Address major environmental
and greenhouse gas (GHG) emis-
sion concerns in an integrated way
Improve SCO quality with exist-
ing facilities.
The oil sands industry is, by its
very nature, in a position to infu-
ence technology development for its
relatively unique needs. It is impor-
tant to identify the possible avenues
for better upgrading technology for
current and future projects.
10
While
some new primary upgrading tech-
nologies are being developed and
implemented, such as slurry-phase
hydrocracking and supercritical
solvent extraction/deasphalting, the
existing commercial processes and
technologies namely, coking,
ebullated-bed hydroconversion,
hydrotreating and hydrocracking
will still dominate for the
foreseeable future.
4
Therefore,
modelling, simulation and optimisa-
tion of these integrated processes
will provide valuable and helpful
knowledge towards innovative
process design and operation.
The objective of this study is to
evaluate commercial bitumen
upgrading schemes by performing
process modelling and simulation,
which will help the industry to
improve process effciency, reduce
GHGs emissions and other related
environmental impacts (for
instance, optimised H
2
and water
www.eptq.com PTQ Q2 2013 37
The oil sands industry
is, by its very nature,
in a position to
infuence technology
development for its
relatively unique
needs
canmet.indd 1 08/03/2013 13:04
Value adding catalysts, absorbents, additives and
process technology for oil refning processes.
www.jmcatalysts.com/refneries
UK
Tel +44 (0)1642 553601
USA
Tel +1 630 268 6300
Tel +1 732 223 4644
Naphtha HDS
Diesel HDS
Hydrotreating catalysts
Vacuum tower
Hydrogen catalysts
FCC unit
FCC additives
SOx NOx removal
Light olefn production
Bottoms conversion
Metals traps
Activity boosters
CO oxidation
Gas / liquid
purifcation
S, Hg, Cl removal
Crude unit
VGO HDS
Hydrogen plant
Process
diagnostics
Untitled-1 1 31/07/2012 17:18 jm.indd 1 26/02/2013 17:04
www.eptq.com PTQ Q2 2013 39
units were carefully calibrated
based on in-house experimental
data and published data.
6-18
Coking and hydroconversion
kinetics were formulated based on
simple reaction networks involving
the following lumps: gas (C
1
-C
4
,
H
2
S and NH
3
), naphtha (IBP-204C),
LGO (204-343C), HGO (343-
524C), vacuum residue (524C+)
and coke (for the coker unit
only). The hydroconversion model
also included hydrodesulphurisa-
tion (HDS), hydrodenitrogenation
(HDN), hydrodemetallisation
(HDM) and hydrogen consump-
tion. The rate equations accounted
for the effect of temperature and
pressure in the case of hydrocon-
version, whereas they were only
temperature dependent in the case
of coking. The models were prop-
erly tuned and implemented in a
Hysys spreadsheet. This feature
enables user-dened calculations to
be performed using any variable
from the simulation owsheet and
then the results exported to any
use) in future bitumen upgrading
and rening schemes.
Bitumen upgrading process
Figure 1 presents a generic ow
diagram of the bitumen upgrading
process. The diluted bitumen from
extraction and froth treatment plant
rst passes through the diluent
recovery unit (DRU), where the
diluent is recovered and recycled to
the extraction plant. The atmos-
pheric-topped bitumen (ATB) is fed
to a vacuum distillation unit (VDU)
to obtain naphtha, light gas oil
(LGO), heavy gas oil (HGO) and
vacuum-topped bitumen (VTB).
The VTB is processed either in a
coker unit or in an ebullated-bed
hydroconversion reactor, or even in
both. The total liquid product (TLP)
from the coking and/or hydrocon-
version stage is fractionated, and
the resulting streams are combined
with the corresponding fractions
coming from the VDU and routed
separately to hydrotreating (HDT)
units. After hydrotreating, these
streams are blended to form SCO,
which is transported by pipeline to
reneries in Canada and the US.
Modelling and simulation approach
The simulation was conducted in
the Aspen Hysys environment.
Two typical bitumen upgrading
schemes were analysed: coking and
hydroconversion based. Oil proper-
ties of interest (API gravity,
sulphur, nitrogen and metals,
among others) and product yield
shifts at every conversion step were
handled by the Petroleum Rening
tool in Hysys. Since Hysys does not
have modules for certain units of
the bitumen upgrading scheme
namely, coker, ebullated-bed
hydroconverter and naphtha
hydrotreater kinetic models
were developed and implemented
into Hysys to predict the perfor-
mance of those processes, as will be
discussed below. Other units, such
as distillation columns and gas oil
hydrotreaters, were simulated with
built-in Hysys modules. All of the
C o k e
D ilu te d
b itu m e n
A tm o sp h e ric
to p p e d
b itu m e n
U n c o n ve rte d
re sid u e
D ilu e n t
Naphtha hydrotreating
S o u r
g a s
G a so lin e
H
2
LGO hydrotreating
S o u r
g a s
D ie se l
H
2
HGO hydrotreating
S o u r
g a s
F C C fe e d
H
2
S yn th e tic
c ru d e o il
Coking/hydroconversion
F e e d
+H
2
P ro d u c ts
C a t.
C a t.
F e e d
P ro d u c ts
S o u r
g a s
Diluent
recovery
Vacuum
distillation
N a p h th a ( I B P 2 0 4 C )
L G O ( 2 0 4 3 4 3 C )
H G O ( 3 4 3 5 2 4 C )
Fi gure 1 Generic diagram of bitumen upgrading process
canmet.indd 2 08/03/2013 13:04
40 PTQ Q2 2013 www.eptq.com
stream or unit. In this case, the
spreadsheet imported the operating
conditions from the feed stream to
calculate the product distribution.
The yield shift was then performed
by the Petroleum Shift Reactor tool
using the product yield estimates
as input. The effect of hydrotreat-
ing on the chemical composition
and physical properties of the feed
was handled by the Assay
Manipulator. This tool adjusts the
distribution of assay properties
(such as sulphur content) along the
entire boiling range by specifying a
target bulk value, which is deter-
mined by hydrotreating kinetics.
Naphtha HDT was modelled
according to a kinetic model for
coker naphtha hydrotreating
reported in the literature.
19
The
model accounts for HDS, HDN and
aromatics saturation. Simulation of
LGO and HGO hydrotreating was
performed using the gas oil hydro-
cracking-hydrotreating module.
Rate constants were properly cali-
brated in order to produce typical
operational parameters such as
reactor temperature profles, hydro-
gen consumption and product
quality.
20
Results and discussion
Simulation fowsheets were
constructed for the two upgrading
schemes by assuming a processing
capacity of 100 000 b/d diluted
bitumen. The bitumen-to-diluent
ratio was assumed to be 75/25 vol/
vol. The properties of the bitumen
feed and its products after diluent
recovery (atmospheric distillation)
and vacuum distillation are shown
in Table 1. The ATB, which is
equivalent to the raw bitumen, is
mainly comprised of VTB and
HGO, with a small amount of LGO
and almost no naphtha. It is charac-
terised by having a low API gravity
(10.7API) and high contents of
undesired contaminants such as
sulphur, nitrogen, metals and
Conradson carbon residue (CCR),
all of which make bitumen upgrad-
ing operations very diffcult, costly
and environmentally detrimental.
Practically all metals and carbona-
ceous material are concentrated in
the VTB fraction, which constitutes
the coker/hydroconverter feed. The
naphtha and gas oil fractions also
require further processing due to
their high concentrations of sulphur
and nitrogen.
The details of coker and hydro-
conversion reactor performance are
shown in Table 2. It is clear from
the product distribution that hydro-
conversion is more selective
towards valuable liquid products,
without generating solids. This is
the main advantage of hydrogen
addition technologies over carbon
rejection (thermal cracking) ones.
The residue conversion using H
2
addition upgrading is about 78%,
and there is a substantial catalytic
removal of sulphur, nitrogen and
metals. Hydroconversion also
generates fewer C
1
-C
4
compounds,
but more H
2
S and NH
3
due to
sulphur and nitrogen conversion.
Conversely, coking offers almost
complete residue conversion but
with signifcant coke formation
(about 31 wt% based on coking
feed). There is also nearly 100%
rejection of metals and partial
sulphur and nitrogen elimination,
all of which are concentrated in the
coke. Table 3 shows the properties
of the liquid streams after fractiona-
tion of the total liquid product for
the two schemes.
Generally and relative to coking-
based schemes, hydroconversion
produces better-quality distillates
due to the catalytic hydrogenation
effect. Also noted is that for the
hydroconversion-based scheme
there is a portion of unconverted
vacuum residue with much lower
quality than the original VTB feed.
Usually, this stream is recycled to
Diluent recovery Vacuum distillation
a
Property Diluted bitumen Diluent ATB Naphtha LGO HGO VTB
Yield, wt% 100.0 20.4 79.6 0.8 12.1 30.7 56.4
SG 60/60F 0.9381 0.7659 0.9953 0.7730 0.8691 0.9565 1.0556
API gravity 19.3 53.2 10.7 51.6 31.3 16.4 2.5
Sulphur, wt% 3.72 0.81 4.46 0.89 1.63 3.20 5.80
Nitrogen, wppm 3299 10 4139 11 114 1301 6602
Aromatics, wt% 61.7 14.7 73.8 15.8 35.2 57.6 91.6
Nickel, wppm 59.1 - 74.1 - - 0.1 131.3
Vanadium, wppm 196.8 - 247.1 - - 0.2 437.8
CCR, wt% 10.5 - 13.2 - - - 23.4
a) Yield values are based on total ATB feed.
Properties of the bitumen feedstock and its distillation products
Table 1
Coker Hydroconverter
Product yields, wt%
H
2
S 1.5 4.3
NH
3
0.1 0.5
Gas (C
1
-C
4
) 6.1 4.0
Naphtha 15.5 12.6
LGO 20.8 34.0
HGO 23.5 27.2
Vacuum residue 1.5 19.2
Coke 31.0 -
Total 100.0 101.7
H
2
consumption, scf/bbl - 1218
Conversion ,wt%
Sulphur 39.1 67.3
Nitrogen 68.4 55.2
Metals (Ni+V) 98.5 76.6
Vacuum residue 98.5 78.0
a) Product yield values are based on total VTB feed.
Coker and hydroconverter performance
a
Table 2
canmet.indd 3 08/03/2013 13:04
www.eptq.com PTQ Q2 2013 41
the ebullated-bed reactor or fed to a
coker unit. For this study, it was
considered that this stream is recy-
cled back to the hydroconversion
reactor, meaning that there is
complete overall residue conver-
sion even when the once-through
conversion reported in Table 2 is
78%.
The fractionated products from
the VTB conversion step, along
with the corresponding VDU distil-
lates, require further hydrotreating
to reduce sulphur, nitrogen and
aromatic contents for producing
suitable blending components for
SCO. The properties of the hydro-
treated products and fnal SCO are
provided in Table 4. The SCO has
no residue and very low sulphur
and nitrogen contents. The
coking-derived SCO is slightly
more aromatic than that derived
from the hydroconversion-based
scheme. Table 5 summarises the
details of the HDT units. It is
observed that the hydrotreating of
coker products requires higher
Coker Hydroconverter
Property Naphtha LGO HGO Naphtha LGO HGO Vacuumresidue
Yield, wt% 26.8 34.1 39.1 14.0 37.9 27.8 20.4
SG 60/60F 0.7363 0.8715 0.9736 0.7365 0.8710 0.9715 1.1240
API gravity 60.7 30.9 13.8 60.6 31.0 14.2 -5.6
Sulphur, wt% 1.79 3.67 4.43 0.25 0.71 1.37 6.90
Nitrogen, wppm 316 1694 3973 393 1582 3063 8905
Aromatics, wt% 27.1 58.0 66.0 20.5 42.5 53.0 99.8
Nickel, wppm - - 8.2 - - 3.0 262.9
Vanadium, wppm - - 27.4 - - 6.2 458.3
CCR, wt% - - 1.5 - - 0.2 14.3
a) Yield values are based on total liquid product.
Propertiesofcokerandhydroconverterliquidproducts
a
Table3
Coker-basedscheme Hydroconverter-basedscheme
Property Naphtha LGO HGO SCO Naphtha LGO HGO SCO
SG 60/60F 0.7375 0.8597 0.9131 0.8680 0.7375 0.8573 0.9115 0.8665
API gravity 60.4 33.1 23.5 31.5 60.4 33.6 23.7 31.8
Sulphur, wt% 0.018 0.112 0.16 0.13 0.014 0.044 0.15 0.10
Nitrogen, wppm 12 204 740 482 21 264 455 336
Aromatics, wt% 19.9 41.8 54.2 45.9 15.8 36.5 50.9 41.7
PropertiesofhydrotreatedproductsandSCO
Table4
test runs results, it could be
concluded that the revamp targets
for the CDU-1 main fractionator
(C-150) were achieved. No hydrau-
lic constraint was experienced in
achieving the design intake of
13 000 t/d and the required prod-
uct quality was achieved.
Conclusions
The performance of Shell ConSep
trays in the HGO pumparound
section of the CDU-1 main fractiona-
tor met the target of capacity
enhancement without any drawback
compared to the pre-revamp condi-
tions. During the test run, the trays
were operating at 10-15% lower than
the design capacity even at the
design intake of 13 000 t/d due to
heavier crude feed and lower feed
temperature. However, the built-in
capacity margin enabled stable oper-
ation for the trays at much above
the capacity limit of the frst genera-
tion of high-capacity trays.
The options to debottleneck
columns already equipped with the
frst generation of high-capacity
trays are limited. ConSep trays
provide an attractive solution for
www.eptq.com PTQ Q1 2013 77
such cases. In this revamp project,
use of only three of these trays in
the most capacity-constrained
section of the column made it possi-
ble to retroft the existing column
and made the capex option more
attractive over the other debottle-
necking options.
* Shell ConSep, Shell CS and Shell HiFi are Shell
trademarks. ** Mellapak Plus 252Y is a Sulzer
Chemtech trademark.
References
1 Refnery expansion means NZ more self
reliant, media release by NZRC, 16 July 2010.
2 Wilkinson P M, De Villiers W E, Mosca
G, Tonon L, Achieve challenging targets in
propylene yield using ultra system fractionation
trays, ERTC 2006.
3 De Villiers W E, Bravo J L, Wilkinson P M,
Summers D R, Further advances in light
hydrocarbon fractionation, PTQ Q3 2004.
KaushikMajumder is Distillation Team Lead of
Shell Projects & Technology in Bangalore, India.
He holds a bachelors degree from Jadavpur
University, India, and a masters and doctorate
from Indian Institute of Technology, Delhi.
Email: Kaushik.Majumder@shell.com
Giuseppe Mosca is the Global Refnery
Technology Manager of Sulzer Chemtech.
He holds BS and MS degrees in chemical
engineering from the University La Sapienza
Rome, Italy.
Email: Giuseppe.Mosca@sulzer.com
Kent Mahon is a Process Engineer at Refning
NZ. He was the Senior Process Engineer and
Commissioning Process Engineer during the
Point Forward Project.
Email: Kent.Mahon@refningnz.com
Parameters Design Testrun
Froth backup/CS height, % 68 60
Tray pressure drop, mbar 12.3 9.2
Tube food , % 73 60
Flow parameter 0.17 0.19
Overall column load factor, m/s 0.12 0.10
Flooding (CS tray), % 133 112
KeyperformanceindicatorsforConSeptrays
Table3
sulzer.indd 5 11/12/12 18:05:58
canmet.indd 4 08/03/2013 13:04
42 PTQ Q2 2013 www.eptq.com
due to the fact that the
ebullated-bed hydroconversion unit
completely converts the vacuum
bottoms into gaseous and liquid
products without forming any solid
byproducts. However, this comes
with the cost of almost three times
more hydrogen consumption (1269
scf H
2
/bbl bitumen vs 426 scf H
2
/
bbl bitumen), which in return
makes this scheme more energy
intensive.
Summary
This article presents a detailed
simulation of two typical bitumen
upgrading schemes using Aspen
Hysys. Process models for every
major unit were constructed and
calibrated based on representative
data. The simulations predicted
typical upgrader performance and
allowed tracking of the transforma-
tion of the feedstock at every stage
of the process. It was concluded
that the hydroconversion-based
scheme is more selective towards
valuable products such as LGO and
HGO and therefore exhibits a
higher SCO yield (about 94 wt%)
when compared with that obtained
in the coking-based scheme
(76 wt%).
However, hydrogen consumption
in hydroconversion-based upgrad-
ing is much higher than that in the
coker-based scheme (1269 scf H
2
/
bbl bitumen vs 426 scf H
2
/bbl bitu-
men), which makes the former
option more energy intensive. In
general, the simulation fowsheet is
versatile and can be used for
exploring different operating
scenarios. Therefore, it can be
considered as an important and
useful tool to guide bitumen
upgrading process design, opera-
tion and optimisation.
Acknowledgments
Partial funding for this study was provided
by the Canadian Interdepartmental Program
of Energy Research and Development (PERD
1.1.3). Comments and suggestions from Dr
Edward Little and Dr Mustafa Al-Sabawi on
revising the manuscript are greatly appreciated.
The authors are grateful for the support and
help from IT staff at CanmetENERGY.
References
1 LENEF Consulting Limited, Bitumen &
the two process schemes is
presented in Table 6. The hydro-
conversion-based upgrading
scheme exhibits a substantially
higher yield of SCO than the coker-
based scheme (94 wt% vs 76 wt%
SCO). The difference in the SCO
yield between the two schemes is
average reactor temperatures than
those of hydroconverter products
(naphtha HDT: 285C vs 275C;
LGO HDT: 340C vs 335C; HGO
HDT: 370C vs 363C) in order to
achieve similar conversion levels.
This also explains why hydrogen
consumption is higher for coker
products. In the case of HGO
hydrotreating, the product distribu-
tion shows that there is mild
hydrocracking of HGO into LGO
and naphtha. Total conversion of
aromatics is relatively low during
LGO and HGO hydrotreating, since
most of the aromatics hydrogena-
tion is partial saturation of
polyaromatics to form di- and
mono-aromatics.
Finally, an overall comparison of
Coker-basedscheme Hydroconverter-basedscheme
Naphtha LGO HGO Naphtha LGO HGO
Operatingconditions
Liquid hourly space velocity (LHSV), h
-1
2.5 2 1 2.5 2 1
Average temperature, C 285 340 370 275 335 363
Pressure, MPa 4 5 9 4 5 9
H
2
/oil, scf/bbl 1680 1680 3370 1680 1680 3370
Productyields,wt%
H
2
S 1.8 2.7 3.6 0.3 1.0 2.5
NH
3
0.03 0.1 0.2 0.04 0.1 0.2
Gas (C
1
-C
4
) 0.0 0.2 0.1 0.0 0.1 0.1
Naphtha 91.7 1.5 1.6 92.8 1.4 2.3
LGO 6.8 95.9 15.2 7.1 97.8 12.7
HGO - - 80.4 - - 83.2
Total 100.3 100.4 101.1 100.2 100.4 101.1
H
2
consumption, scf/bbl 142 262 736 100 259 690
Conversion,wt%
Sulphur 99.0 95.9 95.6 95.4 95.7 94.0
Nitrogen 95.8 77.7 66.0 94.3 76.0 77.1
Aromatics 25.7 12.2 12.4 21.9 9.7 10.5
a) Product yield values are based on total hydrotreater feed.
Hydrotreaterperformance
a
Table5
Coker-basedscheme Hydroconverter-basedscheme
Products,wt%
Gases 7.0 8.0
Naphtha 10.3 11.0
LGO 30.3 42.5
HGO 35.5 40.4
Coke 17.6 -
Total 100.6 101.9
SCO, wt% 76.1 93.9
SCO, vol% 87.1 107.3
H
2
consumption, scf/bbl 427 1269
a ) Product yield and H
2
consumption values are based on total ATB feed.
Overallperformanceofthetwoschemes
a
Table6
Thesimulation
fowsheetisversatile
andcanbeusedfor
exploringdifferent
operatingscenarios
canmet.indd 5 08/03/2013 13:04
www.eptq.com PTQ Q2 2013 43
Very Heavy Crude Upgrading Long Term R&D
Opportunities, 1994.
2 Yui S, Chung K H, Syncrude upgrader revamp
improves product quality, Oil Gas J, 2007, Vol.
105, 46, 52.
3 Chrones J, Germain R R, Bitumen and heavy
oil upgrading in Canada, Fuel Sci Tech Int, 1989,
7, 783.
4 Rana M S, Samano V, Ancheyta J, Diaz J
A I, A review of recent advances on process
technologies for upgrading of heavy oils and
residua, Fuel, 2007, 86, 1216.
5 Speight J G, The Chemistry and Technology
of Petroleum, 2007, 4
th
ed, CRC Press/Taylor &
Francis, Boca Raton, FL.
6 Sayles S, Romero S, Understand differences
between thermal and hydrocracking,
Hydrocarbon Process, 2011, Sept, 37.
7 Martinez J, Sanchez J L, Ancheyta J, Ruiz R
S, A review of process aspects and modeling
of ebullated bed reactors for hydrocracking of
heavy oils, Catal Rev Sci Eng, 2010, 52, 60.
8 Yui S, Producing quality synthetic crude oil
from Canadian oil sands bitumen, J Jpn Petrol
Inst, 2008, 51, 1.
9 Yui S, Athabasca oil sands produce quality
diesel and jet fuels, Oil Gas J, 2000, Vol. 98, 47,
58.
10 Yui S, Chung K H, Processing oil sands
bitumen is syncrudes R&D focus, Oil Gas J,
2001, Vol. 99, 17, 46.
Variante 4a
A Member of Unitherm-Cemcon Group
CS Combustion Solutions GmbH
Simmeringer Hauptstrasse 24
A-1110 Vienna
Austria
T+43 1 740 40 2812
F+43 1 740 40 2815
sales@comb-sol.com
www.comb-sol.com
your experts
n ACID/TAIL GAS
n LIQUID SULFUR
n SPENT ACID
n CHEMICAL/PETROCHEMICAL RESIDUES
The team of CS Combustion
Solutions is experienced in
engineering, supply and
commissioning of vertical
and hori zontal burners
and combustors for
more than 20 years.
11 Wadsworth D, LC-Fining options for heavy
oil upgrading, Proceedings of the NPRA Annual
Meeting, San Diego, CA, 9-11 March 2008.
12 Ordorica-Garcia G, Croiset E, Douglas P,
Elkamel A, Gupta M, Modeling the energy
demands and greenhouse gas emissions of the
Canadian oil sands industry, Energy Fuels, 2007,
21, 2098.
13 Morawski I, Mosio-Mosiewski J, Effects of
parameters in Ni-Mo catalysed hydrocracking
of vacuum residue on composition and quality
of obtained products, Fuel Process Technol,
2006, 87, 659.
14 Danial-Fortain P, Gauthier T, Merdrignac
I, Budzinski H, Reactivity study of Athabasca
vacuum residue in hydroconversion conditions,
Catal Today, 2010, 150, 255.
15 Ding F, Ng S H, Xu C, Yui S, Reduction of
light cycle oil in catalytic cracking of bitumen-
derived crude HGOs through catalyst selection,
Fuel Process Technol, 2007, 88, 833.
16 Botchwey C, Dalai A K, Adjaye J, Kinetics
of bitumen-derived gas oil upgrading using a
commercial NiMo/Al
2
O
3
catalyst, Can J Chem
Eng, 2004, 82, 478.
17 Yui S, Sanford E, Kinetics of aromatics
hydrogenation of bitumen-derived gas oils, Can
J Chem Eng, 1991, 69, 1087.
18 Yui S, Sanford E, Mild hydrocracking of
bitumen-derived coker and hydrocracker heavy
gas oils: kinetics, product yields, and product
properties, Ind Eng Chem Res, 1989, 28, 1278.
19 Yui S, Removing diolefns from coker
naphtha necessary before hydrotreating, Oil
Gas J, 1999, 97, 36.
20 Chang A-F, Liu Y A, Predictive modeling of
large-scale integrated refnery reaction and
fractionation systems from plant data. Part 1:
hydrocracking processes, Energy Fuels, 2011,
25, 5264.
Anton Alvarez-Majmutov is an NSERC Visiting
Fellow at CanmetENERGY working on bitumen
upgrading process modelling and simulation.
He holds a PhD from Mexican Institute of
Petroleum (IMP).
Jinwen Chen is a Senior Research Scientist and
Group Leader at CanmetENERGY. He holds
a PhD in chemical engineering from Tianjin
University.
Mugurel Munteanu is a Lead Process
Engineer at CoSyn Technology, a division of
WorleyParsons, in Edmonton, Canada. He holds
a PhD in chemical engineering from Laval
University, Canada.
canmet.indd 6 08/03/2013 13:04
Weir Minerals Lewis Pumps is the world
standard for manufacturing pumps
and valves in the sulphur chemicals
industry. O fferin g a fam ily o f Lewis steam -
jacketed su lp h u r p u m p s, o u tstan d in g
reliab ility in h ig h -tem p eratu re su lp h u ric
acid ap p licatio n s, an d a co m p lete fam ily
o f valves, Weir M in erals Lewis P u m p s
co n tin u es a lo n g trad itio n o f o fferin g
su p erio r p ro d u cts an d services.
With its stro n g co m m itm en t to service, Weir
M in erals Lewis P u m p s can sh ip stan d ard
rep lacem en t wear p arts in 7 2 h o u rs to
m o st in tern atio n al airp o rts in em erg en cy
situ atio n s. With a co m p lete lin e o f p u m p s
an d valves, Lewis P u m p s p ro d u cts are rig h t
o n targ et. C u sto m ers in m o re th an 1 0 0
co u n tries can t b e wro n g .
Lewis
Pumps
products
are right
o
n
t
a
r
g
e
t
Excellen t
M in erals
S o lu tio n s
L E W I S
P U M P S
Vertical C h em ical P u m p s
8 6 2 5 G ran t R d .
S t. Lo u is, M O 6 3 1 2 3
T: +1 3 1 4 8 4 3 -4 4 3 7
F: +1 3 1 4 8 4 3 -7 9 6 4
Em ail: sales@ lewisp u m p s. co m
www. weirm in erals. co m
C o p yrig h t 2 0 1 1 , En viro Tech P u m p system s, I n c. All rig h ts reserved . LEWI S an d LEWI S P U M P S are reg istered trad em arks o f En viro tech P u m p system s, I n c. ; WEI R is a reg istered trad em ark o f Weir En g in eerin g S ervices Ltd .
LewisTarget Ad Valves.indd 1 2/28/2012 8:46:55 AM
weir pumps.indd 1 07/03/2013 13:41
Insulating tank terminals
S
torage tank insulations are
becoming more important than
ever. Global CO
2
emissions
have risen dramatically over the
past few years and are set to
increase by up to 90% by 2050
should industrial emissions remain
unchecked.
1
Under pressure from
the EUs ambitious 20-20-20
target, tank terminal operators are
increasingly looking for effective
insulation systems to limit costs
and prepare for new legislation.
This article discusses the different
insulation solutions available and
establishes how tank operators can
effectively reduce energy costs and
cut CO
2
emissions.
Thermal insulations with long-term
benefts
Although thermal insulation
systems are easily installed and
highly cost effective, studies have
revealed that industrial insulations
are often either poorly maintained
or non-existent, resulting in exces-
sive heat losses.
2
However, with the
EUs ambitious plans to cut green-
house gas emissions by 20%
compared to its 1990 levels by 2020
3
and a carbon tax more likely,
industrial insulation solutions can
play a key role in achieving energy
targets. Primarily used for products
that require storage at high temper-
atures, tank insulations provide a
number of benefts.
One of the most obvious advan-
tages of effective thermal insulation
is the reduced loss of heat. By wrap-
ping tanks in materials with low
thermal conductivity, heat within
the tank is retained more effectively.
As a result, tank operators need
to use considerably less heat-
Energy-effcient tank insulation systems result in lower energy costs, less
maintenance, and a prolonged lifespan of contents and equipment
PETER BOON
Pentair Thermal Management
generating equipment to produce
the same amount of heat. According
to a study commissioned by the
European industrial insulation
Foundation (EiiF), improving insula-
tion to a cost-effective level would
avoid 66% of current heat loss,
which amounts approximately to
the annual energy consumption of
10 million households.
2
This, in turn,
would also have an impact on
carbon emissions, resulting in a
potential reduction of 49 Mt/y of
CO
2
. As prices for energy rise, these
cost-saving potentials will become
even more obvious, making
energy-effcient insulations an
attractive investment for the future.
Although, thermal insulation is
primarily used for tanks that store
temperature-sensitive materials, it
can, in fact, conserve the quality of
many substances and offers an
effective protection against frost
and fre. Furthermore, it can signif-
cantly reduce the effects of thermal
stresses on the tank itself, protect-
ing its structure from external
forces, such as rain, wind and
abrupt temperature fuctuations.
The results are lower energy costs,
less maintenance and a prolonged
lifespan of contents and equipment.
Finding the solution
When it comes to choosing a
system, tank operators have a
number of insulation solutions
available. Depending on the
contents, external environment and
required temperature level, some
solutions might be less or more
suitable than others. Whatever the
project, involving insulation engi-
neers in the planning process is
vital and will ensure that the
energy-saving potential of the tech-
nology is maximised.
One of the established insulation
materials available is mineral wool.
Often used to insulate buildings,
pipes and storage tanks, mineral or
stone wool has become one of the
most common insulation materials.
While demonstrating good sound
absorption, fre resistance and
reasonable thermal behaviour,
mineral wool insulations require a
complicated and lengthy installation
process. In particular for larger
tanks, a complex scaffolding system
is required to set up the outer insu-
lation structure. Conventional tank
insulation systems with mineral
wool may also require extensive
welding of insulation support rings
and insulation pins, next to the need
for installing banding bars and
screws that cause increased heat loss
and increased tank builders cost.
While this might not cause major
concerns at the early stages of a
project, it does, however, result in
signifcant downtime once the tank
is in service, making it less suitable
for retroft insulations.
To minimise the installation
impact on active tanks, some tank
terminal operators are turning to
polystyrene or polyurethane foam
insulations. Although these foams
can be easily sprayed onto the walls
of a tank and will provide instant
thermal insulation, the lifespan of
the material is comparatively low.
Easily accessible to rodents and
birds, spray foam insulation needs
to be inspected regularly to ensure
the system remains fully sealed.
Furthermore, spray foam can only
be installed in completely dry condi-
tions and it can encourage mould
www.eptq.com PTQ Q2 2013 45
pentair.indd 1 08/03/2013 13:08
46 PTQ Q2 2013 www.eptq.com
the use of scaffolding. As a result,
the installation process is safer and
can be completed much quicker
than conventional methods. Time
spent on-site will be a deciding
factor for many operators, in
particular those looking to retrot
terminals. Insulation panels can be
easily tted during scheduled
maintenance work or when tank
contents are being replaced, avoid-
ing costly disruptions. As the
panels are customised and can have
lengths up to 12 metres, the amount
of horizontal joints is reduced to
the minimum. The system allows
for individual panels to be easily
replaced should they become
damaged. The fully locked seam
system also eliminates water pene-
tration, reducing the occurrence of
under-insulation corrosion and
offering consistent performance
levels throughout its life (see Figure
3). As a result, the system hardly
requires any maintenance and has a
much higher lifespan than conven-
tional systems.
Case study
In 2012, Pentair Thermal Manage-
ment was asked to complete a
major tank insulation project for
the Zeeland Renery in
Nieuwdorp, The Netherlands. A
joint venture between French oil
giant Total and Russias Lukoil, the
Zeeland renery processes a total
of 9.5 million tonnes of crude and
heavy oil per year, making it one of
the largest oil and product trading
centres in the world. It is also
considered to be one of the worlds
most energy-efcient reneries. As
such, it requires an insulation
system that allows it to signicantly
reduce its energy consumption and
cut CO
2
emissions to an absolute
minimum.
Pentair Thermal Management
was commissioned for the design,
fabrication and installation of insu-
lation systems on two of the
renerys tanks. On one of the
tanks, a 60m-diameter oil tank, the
company was also required to
remove the existing mineral wool
insulation, which had become
badly damaged by heavy winds,
while water ingress had also caused
corrosion of the tanks carbon steel
growth if not applied properly.
Recent ndings also suggest that
polystyrene foam does not demon-
strate the same stability as other
materials when exposed to re.
4
Locking in the heat
An effective and improved alterna-
tive to conventional systems and
spray foam systems consists of
prefabricated panels constructed by
laminating insulation material to a
metal jacket. These are robust,
double-locking, vertical-standing
seam insulation panels that can be
easily tted around a tank. Pentairs
insulation system is called Trac-Loc.
Due to the use of insulation mate-
rials with a much lower thermal
conductivity value compared to
mineral wool or polystyrene foam,
the heat loss is highly reduced,
resulting in substantial energy
savings and, in direct relation to
that, a reduction in the carbon foot-
print. The preferred insulation
material is polyisocyanurate (PIR),
a non-brous closed cell structure
material resulting in minimised
water/uids absorption and degen-
eration, a superior low lambda
value, and a long and stable life-
time (see Figure 2).
Trac-Loc panels are usually
prefabricated off-site in a controlled
workshop environment, making the
installation less dependent on
weather conditions. They are then
assembled on-site using a hanging
basket or cherry picker, eliminating
Tra c -L o c
c lip
P a n e l N o . 1 P a n e l N o . 2
Trac-Loc clip and panels to assembly
Alignment of Trac-Loc clip and panels
Double locking of panels creates seal
Final seam
Figure 1 Trac-Loc assembly
pentair.indd 2 08/03/2013 13:08
shell. The second part of the project
involved the renovation of a
36m-diameter storage tank that
required an energy-efcient insula-
tion that could maintain the
temperature of its contents at
180C.
Completion of the project was
further challenged by severe
weather conditions in the area,
including high winds and heavy
rain. The Trac-Loc system,
however, provided the solution, as
it did not require any scaffolding
and minimised welding, reducing
the working hours on-site and
limiting potential safety risks to an
absolute minimum. As a result, it
took the installation site team only
four weeks to install a new insula-
tion system on the larger oil tank,
less than a third of the time conven-
tional methods would have taken.
Overall, Pentair Thermal Manage-
ment installed a total of 5633m
2
of
Trac-Loc panels to insulate both
tanks. Each of the panels had
specially shaped edges that were
mechanically seamed together,
creating a rigid and secure struc-
ture that is resistant to heavy winds
and rain. Zeeland Renery selected
PIR as insulation material based on
its thermal behaviour and to offer
additional protection against
under-insulation corrosion. The
result is an energy-efcient, solid
insulation system that is not only
lowering CO
2
emissions, but will
also, over its lifetime, provide
Zeeland Renery with substantial
reductions in total cost of
ownership.
Conclusion
Energy-efcient tank insulation
systems play an important role in
the global strategy to reduce indus-
trial CO
2
emissions. While panel
systems, such as Trac-Loc, may
require a higher initial investment,
they will amount to considerable
cost savings over the systems life-
time. Their high energy savings
potential, short installation time
(fewer eld man-hours reduces the
potential for safety incidents), low
operational costs and durability in
challenging conditions make them
well suited for most storage tank
insulation projects. In particular for
www.eptq.com PTQ Q2 2013 47
large tank terminals with high
temperature requirements, the
combination of minimised mainte-
nance costs and high energy
savings will be signicant and have
a real impact on future
protability.
Trac-Loc is a mark of Pentair Thermal
Management.
References
1 Energy technology perspectives, IEA, 2010.
2010: scenarios and strategies to 2050, Paris.
2 Climate protection with rapid payback
Energy and CO
2
savings potential in industrial
insulation in EU27, Ecofys, 19 June 2012.
3 http://ec.europa.eu/clima/policies/g-gas/
index_en.htm
4 Fire performance of LNG carriers insulated
with polystyrene foam, Journal of Hazardous
Materials, 2008.
Peter Boon is Business Development Manager
Tank Farms at Pentair Thermal Management
and is responsible for supporting the companys
Trac-Loc projects in the tank terminal market.
He holds a degree in mechanical engineering
from the Technical University, Amsterdam.
0 . 0 8
0 . 0 6
0 . 0 4
0 . 0 3
0 5 1 0 1 5
T
h
e
r
m
a
l
c
o
n
d
u
c
t
i
v
i
t
y
,
W
/
m
K
Vol ume rel at ed moi st ure cont ent , %
0 . 0 2
E x tru d e d p o lystyre n e fo a m , 8 0 k g /m
3
P o lyu re th a n e , 8 0 k g /m
3
E x p a n d e d p o lystyre n e fo a m , 8 0 k g /m
3
M in e ra l wo o l, 8 0 k g /m
3
C o rk
Figure 2 Characteristics of insulation materials
Figure 3 Trac-Loc is mounted using stainless steel clips
pentair.indd 3 12/03/2013 11:36
enersul.indd 1 12/03/2013 09:45
Understanding thermal oxidisers
T
here are many similarities
between thermal oxidisers and
fred heaters. A thermal
oxidiser and a fred heater both
include one or more burners, an
insulated/refractory-lined combus-
tion chamber, and a fuel
measurement and control system.
Both can be designed to operate
under either forced draft or natural
draft conditions. However, the
similarities between them quickly
begin to diverge upon a detailed
examination of their design and
intended function (see Figure 1).
The purpose of a fred heater is to
heat a process fuid using fuel gas
or oil. Examples of common fred
heater applications are:
Ethylene cracking furnaces
Thermal fuid heaters
Crude oil heaters
Regeneration gas heaters
Delayed coker heaters.
The primary purpose of a ther-
mal oxidiser is effcient waste
stream destruction. Common ther-
mal oxidiser applications include:
Tail gas from sulphur recovery
systems in natural gas plants and
refneries
Chlorinated hydrocarbons
from vinyl chloride monomer
manufacturing
Vent streams from batch reactors
Vents from tank farms
Waste solvents from pharmaceu-
tical manufacturing
Tail gas from carbon black
manufacturing
Carbon monoxide-laden regener-
ation fue gas from catalytic
cracking furnaces.
The function of a thermal
oxidiser is to effciently destroy
hydrocarbons and other
Dealing with common misconceptions on the design of vapour and liquid waste
thermal oxidiser systems
PETER PICKARD and KEVIN MCQUIGG
Zeeco
environmental contaminants;
therefore, it must be designed with
suffcient mixing, reaction time,
reaction temperature and excess
fue gas oxygen content to achieve
the goal. On the other hand, the
function of a fred heater is to eff-
ciently heat process fuids, so it is
designed to conserve fuel. This
primary functional difference
results in the primary design
difference. A thermal oxidisers
combustion chamber is lined with
refractory and sized for one to two
seconds of fue gas residence time
before contact with heat transfer
surfaces. In contrast, a fred heater
typically has heat transfer surfaces
located very close to the fame.
Thus, a fred heater has no
refractory-lined chamber separat-
ing the fame from the heat
transfer tubes. If a fred heater was
used as a thermal oxidiser, the
cold heat transfer surfaces would
inhibit effcient destruction of
hydrocarbons. If a thermal oxidiser
was used as a fred heater, the
NOx emissions would be higher
than expected; the cost to heat the
fuids would be unnecessarily high
due to the purposeless refracto-
ry-lined chamber upstream of the
heat recovery section; and the fuel
usage would be higher.
API Standard 560
A thermal oxidiser is not a fred
heater. In fact, the two are, at best,
only distant cousins for reasons
already discussed. However, in the
absence of a clearly defned design
standard, many engineers attempt
to adopt API Standard 560 (API
560), entitled Fired Heaters for
General Refnery Service, as a
primary thermal oxidiser design
specifcation.
The frst two paragraphs of API
560 are:
Paragraph 1.1.1 This standard
covers the minimum requirements
for the design, materials, fabrica-
tion, inspection, testing, preparation
for shipment, and erection of fred
heaters, air preheaters, fans and
burners for general refnery service.
Paragraph 1.1.2 A fred heater is an
exchanger that transfers heat from
the combustion of fuel to fuids
contained in tubular coils within an
internally insulated enclosure.
Notice that thermal oxidiser (or
incinerator) is not mentioned in
either of these statements. Nor is
thermal oxidiser or incinerator
listed in the Defnition of Terms,
mentioned or illustrated anywhere
else in the entire document.
There are 14 sections of API 560.
Of these, only two Section 7,
Refractory and Insulation, and
Section 11, Centrifugal Fans and
Drivers for Fired Heater Systems
are primarily related to the design
of a thermal oxidiser.
The frst paragraph of Section 7,
Refractory and Insulation, is quite
often incorrectly applied to a
www.eptq.com PTQ Q2 2013 49
The function of a
thermal oxidiser is
to effciently destroy
hydrocarbons and
other environmental
contaminants
zeeco.indd 1 11/03/2013 12:58
50 PTQ Q2 2013 www.eptq.com
thermal oxidiser. This paragraph,
labelled as optional, limits the
shell temperature of the outside
casing of the radiant and convec-
tion sections and hot ductwork to
82C (180F). This may be accept-
able for red heaters, where the
main purpose is the efcient use of
energy, but it is frequently unac-
ceptable for a thermal oxidiser.
Limiting the shell temperature can
create many problems with the
design of a thermal oxidiser. These
problems include:
Most thermal oxidisers are inter-
nally lined with some combination
of a rebrick and castable lining.
The inclusion of an 82C shell
requirement maximises the thick-
ness of this lining, resulting in a
larger thermal oxidiser vessel and
higher capital cost
If we assume that the reason for
the low-temperature casing require-
ment is the conservation of energy,
it makes very little sense for a ther-
mal oxidiser burning an exothermic
waste to incorporate this feature,
unless there is some type of down-
stream heat recovery involved and
energy conservation is very impor-
tant. Even for a thermal oxidiser
using fuel gas to burn an endother-
mic waste, the amount of heat lost
from the shell is usually a minor
fraction of the total heat release. In
these cases, an in-depth analysis
should be performed to determine
if the low-temperature shell justies
the requirement for larger equip-
ment and higher capital cost
Many thermal oxidisers burn
wastes containing elements (such
as sulphur, halogens and some-
times phosphorus) that create acid
gases with dew points well above
the API 560s recommended shell
temperature. The long-applied
solution for handling high sulphu-
ric acid dewpoint problems in the
sulphur industry has been to
design thermal oxidisers with
casing temperatures at least 28C
(50F) above any anticipated
sulphuric acid dewpoint and then
enclose the thermal oxidiser shell
with a rain shield (either alumin-
ium, galvanised steel or stainless
steel). The rain shield maintains the
thermal oxidiser shell at a fairly
constant operating temperature,
Thermal oxidiser
reaction chamber
Waste heat
recovery unit
Vent
stack
Burner
Refractory
Fired heater
chamber
Heat transfer
tubes
Burner
Refractory
Figure 1 Arrangements of a red heater (top) and a thermal oxidiser (bottom)
zeeco.indd 2 11/03/2013 13:11
z4413.indd 1 3/11/2013 9:03:23 AM linde.indd 1 11/03/2013 10:15
regardless of the ambient weather
conditions. This outer rain shield
also serves as personnel protection
by providing an external surface
that is below 120F.
NOx emissions
Fired heater NOx emission
requirements are often mistakenly
applied to thermal oxidisers. Fired
heaters typically fre fuel gas and
air into a chamber containing a
heat transfer surface. Even though
the fue gas from a fred heater
may contain only 2% to 3%
oxygen, the fame temperatures in
the heater are minimised and
produce less NOx. This is because
recently developed burner technol-
ogy such as that employed by the
patented Zeeco GLSF Free Jet
Ultra-Low NOx burner circulates
low oxygen fue gas that has been
cooled by the internal heat transfer
surface. The circulating fue gas
decreases burner fame tempera-
tures and reduces NOx emissions.
In thermal oxidisers, waste gases
of various compositions are fred
in refractory-lined adiabatic
chambers. In high excess air appli-
cations, there are no low-oxygen,
cooled fue gases to circulate. The
oxygen content and temperatures
in these adiabatic chambers are
often higher than in fred heaters.
This means thermal oxidisers
frequently generate more NOx
than fred heaters designed for a
52 PTQ Q2 2013 www.eptq.com
similar heat release.
Regulations regularly specify
fred heater NOx emissions in units
of mass divided by heat release
based on a lower heating value (for
example, lb/MMBtu LHV). For a
system fring fuel gas and air, the
conversion of lb/MMBtu to fue
gas concentration is almost univer-
sally repeatable. For instance, an
emission requirement of 0.2 lb/
MMBtu when fring natural gas
and air will correspond to a fue
gas NOx concentration of 150 parts
per million by volume (ppmv)
corrected to 3% oxygen, dry basis.
This relationship between lb/
MMBtu and fue gas concentration
will hold true as long as the fuel
gas composition contains mostly
organic compounds. The relation-
ship does not hold if the fuel gas
contains signifcant amounts of
inerts (nitrogen or carbon dioxide)
or hydrogen.
The compositions of the wastes
and fuels fred in thermal oxidisers
will vary signifcantly. In fact,
seldom are any two thermal
oxidiser applications identical. For
example, consider two different
thermal oxidiser applications using
natural gas as fuel. The waste gas
to thermal oxidiser A is essentially
100% nitrogen; the waste gas to
thermal oxidiser B is 100% CO
2
.
With both thermal oxidisers at the
same fuel fring rate, thermal
oxidiser A with the N
2
waste can
handle more waste than thermal
oxidiser B with CO
2
waste. Assume
the NOx requirement is specifed at
0.2 lb/MMBtu for each thermal
oxidiser. This will translate to fue
gas NOx concentrations of 28 ppmv
and 35 ppmv, respectively (values
corrected to 3% oxygen, dry basis)
for A and B, with both thermal
oxidisers operating at 1600F
(1800C). If the thermal oxidisers
should have to operate at 1800F
(982C) to achieve a higher waste
destruction effciency, the 0.2 lb/
MMBtu value now corresponds to
32 ppmv and 40 ppmv, respec-
tively, for A and B.
The above example, plus the fact
that it is usually more diffcult to
control NOx in a thermal oxidiser
than in a fred heater, demonstrates
that NOx regulations for fred heat-
ers should not be generally applied
to thermal oxidisers (see Figure 2).
Sulphur oxide (SOx) emissions
SOx is produced in fue gas by the
presence of sulphur-containing
combustibles in the feed streams.
These are commonly elemental
sulphur, H
2
S, COS, CS
2
and various
mercaptans. The most common
reactions are:
S + O
2
SO
2
H
2
S + O
2
H
2
O + SO
2
COS + O
2
CO
2
+ SO
2
CS
2
+ O
2
CO
2
+ SO
2
Carbon monoxide can be further
oxidised to CO
2
by effcient burner
and oxidiser design:
CO + O
2
CO
2
NOx production can be inhibited,
Figure 2 Two vertical low-NOx thermal oxidisers used for NOx reduction
The compositions
of the wastes and
fuels fred in thermal
oxidisers will vary
signifcantly
zeeco.indd 3 11/03/2013 12:58
www.eptq.com PTQ Q2 2013 53
design cannot inhibit the presence
of these elements in fue gas.
Burner testing
The following examples are
This same discussion applies to
other inorganic substances in typi-
cal waste streams, such as chlorine,
sodium, potassium, phosphorus
and silica in the sense that burner
as discussed above. However,
while it is possible to design a
burner that will minimise NOx
and CO, it is not possible to design
a combustion system to minimise
SO
2
formation. CO
2
is the normal,
expected and intended product of
combustion for any compound
containing carbon. Similarly, SO
2
is
the normal, expected and intended
product of combustion for sulphur
compounds. Therefore, sulphur
content in the feed will produce a
predictable quantity of SOx (SO
2
and SO
3
) in the fue gas, independ-
ent of burner or combustion
system design (see Figure 3).
The only way to lower SOx
emissions is to remove sulphur
from the feed stream or remove
SOx from the fue gas with SOx
scrubbing equipment such as a
spray tower, packed column or a
dry scrubber. However, due to the
high cost of such systems and high
water usage, the typical solution is
a tall vent stack (45-65 m) to
disperse the SOx over a broad
area.
Figure 3 Sulphur recovery unit (SRU) tail gas thermal oxidiser with a fre tube-style
waste heat boiler. This unit is designed for a total 75 MW heat release
To keep pace w it h t he demanding
qualit y requirement s of modern
f uels, advanced, precise and easy t o
use analyt ical t echnology is required.
Wit h a complet e range of XRF and
ICP spect romet ers, SPECTROs
unique solut ions f or at -line and
laborat or y element al analysis
are capable of meet ing t he most
demanding product specif icat ion
t est ing requirement s.
Visit us at
Pit t con 2013,
17-21 March,
Philadelphia, PA,
USA
Element al Analysis
of Fuels and Oils
Discover more excit ing det ails at
w w w.spect ro.com/ fuels
Det erminat ion of Sulfur and ot her
element s at -line and in t he laborat ory
zeeco.indd 4 11/03/2013 12:59
54 PTQ Q2 2013 www.eptq.com
module will accurately calculate the
required supplemental fuel. For
example, it is common in the
carbon black industry to burn tail
gas that is roughly:
Component Vol%
Nitrogen 40
Water vapour 40
Carbon monoxide 10
Hydrogen 10
Temperature 450F
Lower heating value 60 Btu/scf
Now, let us compare that to the
following hypothetical vent gas
stream:
Component Vol%
Methane 10
Carbon monoxide 10
Carbon dioxide 80
Temperature 100F
Lower heating value 123 Btu/scf
If the combustion of each of these
streams is simulated using 25%
excess air at 982C (1800F) operat-
ing temperature, neither requires
supplemental fuel. However, expe-
rience shows that the hypothetical
vent gas stream requires supple-
mental fuel, whereas carbon black
tail gas does not. Please note that
this is the case even though the
hypothetical waste is twice the
heating value of carbon black tail
gas. Intuitively, tail gas combustion
is self-sustaining due to high feed
temperature and hydrogen content
(see Figure 3). It is important to
note that there is not a mathemati-
cal formula available that will
reliably predict the required fuel
for many exothermic waste streams.
Simply put, there is no substitute
for experience in thermal oxidiser
design.
Mechanicaldesignpressure
Many engineers are uncomfortable
with the concept of a high-pressure
stream fowing to a system that
operates at relatively low pressure.
This apprehension is quite under-
standable because thermal oxidisers
(and fares) are often the only
process in the plant that has an
unobstructed fow path to atmos-
phere. Typically, there are no
block valves to obstruct the fow
the test. However, this type of
limited test represents signifcant
additional cost and provides very
little useful information.
Calculationofrequired
supplementalfuel
The proliferation of chemical
process simulation software has
provided an increase in productiv-
ity for engineers around the globe.
However, like all software, it has its
limitations. It is common among
combustion engineers to speak of
exothermic and endothermic
streams. These are rather loose def-
nitions: exothermic streams are
generally considered to be ones
that require no supplemental fuel
to reach the oxidiser operating
temperature when burned.
Conversely, endothermic streams
require supplemental fuel. The
danger is assuming the result from
the process simulator reactor
primary reasons why factory
testing of thermal oxidiser burners
prior to shipment is not a routine
practice:
While a mixture of gases can
easily be made to produce a stream
that will be very close to the
expected fuel for a fred heater
burner test, it is rarely feasible to
produce a waste stream that
approximates what the thermal
oxidiser will typically burn. This is
usually due to the high volume
and/or unusual nature of the waste
components
If it is possible to produce a
stream that is comparable to the
expected waste, it is likely that the
components of the test waste will
require a special environmental
permit. Most projects do not allow
for the time and expense required
to obtain such a permit and most
test facilities are not willing or able
to pursue this type of permit
If it is possible to prepare a
comparable waste stream and a
permit can be obtained, it may not
be feasible to heat the stream to the
expected operating temperature
Thermal oxidiser burners are
frequently too large to ft into an
existing burner test furnace
Thermal oxidiser burners are
frequently too large to be fred at
the available test facility fuel rate
Most available test furnaces are
designed to simulate the conditions
of a fred heater. The behaviour of
a thermal oxidiser burner in a fred
heater will be quite different from
the same burner in a refractory-lined
thermal oxidiser chamber. As
discussed previously, the NOx
emissions will certainly be differ-
ent, as will the fame length and
almost any other combustion
performance measurement that
may be of interest
It is usually not feasible to test
the thermal oxidiser burner in its
combustion chamber due to the
cost of a feld simulation test setup.
Despite the above limitations,
many thermal oxidiser manufactur-
ers will provide an optional burner
fame stability test using fuel oil,
natural gas or propane fuel gas
(subject to fuel capacity and test
furnace size constraints), but only if
the customer specifcally requires
Equipment Pressuredrop(mbar)
Oxidiser vessel 1 to 3
Water tube boiler 2.5 to 20
Fire tube boiler 5 to 20
Superheater 1 to 7
Economiser 1 to 7
Combustion air preheater 7 to 40
Waste gas preheater 7 to 50
Baghouse (particulate flter) 5 to 20
Water spray quench 5 to 10
Submerged quench 40 to 150
Venturi scrubber 20 to 70
Packed column scrubber 10 to 25
Acid gas absorber 10 to 25
Dry electrostatic precipitator 1 to 7
Wet electrostatic precipitator 1 to 7
Vent stack Usually 0
Typicalpressuredropinathermal
oxidisersystem
Table1
Misconceptions
regardingthermal
oxidiserdesign
areoftenfound
tobepervasive
throughoutthe
processindustries
zeeco.indd 5 12/03/2013 11:37
Peter Pickard is an Applications
Engineer for Zeeco, Inc, a combustion
equipment manufacturer based in Broken
Arrow, Oklahoma, USA. His experience includes
the design, construction and commissioning
of thermal oxidiser systems in locations
around the world. He holds a BS in chemical
engineering from the University of Tulsa.
Kevin McQuigg is a Senior Process Engineer for
Zeeco, Inc. He has over 30 years of experience
in the design, construction and commissioning
of all types of combustion-related systems in
locations around the world. He holds a MS in
chemical engineering.
www.eptq.com PTQ Q2 2013 55
Conclusion
The misconceptions discussed
regarding thermal oxidiser design
are often found to be pervasive
throughout the process industries.
Thermal oxidisers and red heaters
differ from each other in terms of
purpose, designed function, NOx/
SOx generation, emissions control
and designed operating pressure. It
is important to understand these
differences to ensure the design of
a robust system that meets project
and operational requirements.
downstream of the burner. Pressure
that can exist downstream of the
valve separating high pressure from
the incineration system is the sum of
the pressure drop of the equipment/
piping downstream of the valve. To
get a sense of the pressure drop of
typical thermal oxidiser system
equipment, see Table 1.
Waste streams are often injected
into the burner. Typical required
oxidiser injection point pressure
drops for various streams are
shown in Table 2.
The reason these pressures are
listed in a separate table is that the
system we are describing actually
begins downstream of these injec-
tion points. The pressure in the
pipe feeding these injection points
must be greater than the pressure
drop of the injection point.
Downstream of this point, the pres-
sure is dictated by the downstream
equipment pressure drops. Thus,
the burner housing and combustion
chamber in which a hydrocarbon
liquid is being burned does not
need to be designed for 10 barg,
although the pipe that feeds the
burner does need to be.
Theoretically, it is possible to
exceed the maximum expected
operating pressure in cases of
explosion (a rare and usually
avoidable event), vastly higher than
expected waste vapour stream ow
or blockage of the ue gas path
(due to fouling, for example). A
safety and control system designed
by any reasonable standard should
be in place to prevent these scenar-
ios. However, in the unhappy event
that ue gas pressure drop does
increase without the intervention of
instrumentation, it is much more
likely that excess pressure will be
relieved through the vent stack exit
or through the combustion air fan
housing than oxidiser system
over-pressurisation occurring.
Did you know that compressor operators report oil losses
as high as 1 gallon per day through their existing oil wiper rings?
Packings featuring metallic ring designs cause piston rod wear
and, paired with oil losses, result in large costs over time.
If you want to improve effciency of your oil wiper rings,
contact us and learn more about our latest revolutionary innovation:
ringsandpacking@hoerbiger.com
www.hoerbiger.com
Have you got used to the
of your metallic oil wiper rings?
oil leakage
Stream Pressure drop (bar)
Aqueous liquid 10 to 12
Hydrocarbon liquid 10 to 12
Exothermic vapour 1 to 0.1
Endothermic vapour 0.01 to 0.3
Required injection point pressure drop
Table 2
zeeco.indd 6 11/03/2013 13:11
GET A RETURN ON INVESTMENT
WITHIN 24 HOURS
Using ITW Online Cleaning on a proactive way will help improve
TURNAROUND OPERATIONS and OPERATIONAL EXCELLENCE
An entire Process Unit can be ITW Online Cleaned in as low as 24 hours on an oil-to-oil basis. Our cleaning technology will
clean all of the Units equipment quickly, safely and environmental friendly.
To recover a Units performance, ITW Online Cleaning can be applied in a Turnaround during shutdown operations, to
reduce/eliminate the need for mechanical cleaning, or during the run.
Furthermore, use of our cleaning technology eliminates the need to open or enter hazardous process equipment.
ITW Online Cleaning is the only technology which has an immediate Return On Investment (ROI).
By using ITW Online Cleaning during the run, to recover Units performance, at Unit start-up (just after 24h, ITW
Technology has been applied), all the benefits are immediately monetized in terms of:
- improved throughput
- improved energy efficiency
- improved product quality
- reduced emission taxes
Regular application of ITW Online Cleaning will target an increased run length under clean conditions, with a far higher
ROI.
For Turnaround applications, a further ROI increase can be added by applying ITW Improved Degassing/Decontamination
to achieve quick and effective safe entry conditions.
EVALUATE ITW ONLINE CLEANING TODAY TO IMPROVE YOUR PLANTS PROFITABILITY !
For more informations contact : ITW S.r.l.- C.da S.Cusumano - 96011 Augusta - Italy
Tel. +39 (0931) 766011
E-mail: info@itwtechnologies.com
www.itwtechnologies.com
Join ITW Team worldwide and send your Curriculum Vitae to : jobs@itwtechnologies.com
ITW
Innovative
Technologies
Worldwide
Now hiring
Professionals
Worldwide
itw.indd 1 11/9/12 11:45:38
Diagnostic tools and control parameters
for refnery heaters
T
he basic principle of good
combustion is to fully oxidise
the carbon, hydrogen and
sulphur in the fuel, leading to the
formation of carbon dioxide, water
and SOx. Stoichiometric levels of
oxygen should normally be suff-
cient to complete combustion.
1
Improvement of furnace opera-
tion is very important to achieve
effcient or continuous production,
and safety issues have always been
the priority for chief engineers in
refneries.
Generally, for combustion in
refnery heaters, there are two main
parameters that affect the thermal
effciency of the entire combustion
process:
Lowering the value of the oxygen
content of the fue gas
Lowering the value of the
temperature of the fue gas.
The thermal effciency of furnaces
can be improved by increasing heat
recovery from the fue gases.
Usually, a 20C reduction in stack
fue gas temperature corresponds
to a 1% gain in effciency. In addi-
tion, a 1% reduction in the total
oxygen level leads to a 0.5%
improvement in effciency.
Reducing CO
2
emissions is an
absolute necessity and an expensive
challenge for refneries to meet
environmental targets according to
legal requirements. In Izmir refn-
ery, we are trying to decrease the
production of greenhouse gases by
making some operational changes
and investments.
In order to reduce emissions of
CO
2
, CH
4
and O
3
, as well as energy
demand, we undertook the follow-
ing tasks:
Heat integration projects, for the
To optimise or manage furnace operation more effectively and safely, engineers
and operators need to understand and control the entire process
YAHYA AKTAS, HAKAN OG
Catalysts
Co-Catalysts
Additives
6000 OIL is an
innovative hightech version in ContiTech's extremely
successful steam hose family. It is especially suitable
for conveying steam in the chemical and petrochemical
industry, in reneries, the building industry, in shipyards
and for various other industrial applications under the
most grueling conditions in all places, where existing
steam hoses fail at an early stage due to oil contamina-
tion from outside.
q heat resistant EPDM inner layer material
q temperature range for saturated steam up to
max. +210C, short term +220C at 23 bar
(saturated steam)
q working pressure up to 18 bar
q burst pressure: min. 180 bar, safety factor 10:1
q reinforcements: 2 steel wire braids, excellent
resistance to corrosion
q oil resistant special cover, resistant to ozone,
weather, UV and abrasion
q electrically conductive, R < 10
6
q meets all requirements as per EN ISO 6134 2B
ContiTech Schlauch GmbH
Postfach 1120
D-34481 Korbach
industrial.hoses@uid.contitech.de
DAM PF TRIX
6000 O IL
highest quality for conveying
steam according to
EN ISO 6134 - 2B -
w w w. c o n tite c h . d e /a lw
Conti Anzeige Dampf Trix OEL GB 86x270 01-2013.indd 1 04.02.13 14:18
tecnimont.indd 7 08/03/2013 13:39
No Harm,
NO
FOUL
Sai nt -Gobai n NorPro
www.norpro.saint-gobain.com
Ohio, U.S.A.
Tel: +13306773552
norpro.stow@saint-gobain.com
Steinefrenz, Germany
Tel: +49 6435 96570
norpro.steinefrenz@saint-gobain.com
Protect your catalyst bed from being
fouled with new MacroTrap
XPore 80 guard bed media
MacroTrap XPore 80 guard bed media offers
increased macroporosity which ultimately increases
fltration capacity. MacroTrap XPore 80 medias
proven technology traps particulate
contaminants for pressure drop
stability and improved reactor
cycle performance so catalyst
bed fouling is sidelined.
Dont let contaminants foul
your reactor. Protect your catalyst
investment with MacroTrap XPore 80 guard bed
media from Saint-Gobain NorPro, the leading name
in guard bed media.
Magnifcation
of Macropores
Norpro MacroTrap XPore 80 Ad_newpub.indd 1 2/12/13 3:46 PM
norpro.indd 1 26/02/2013 16:19
Options for biojet fuel
A
viation turbine fuel (ATF), or
jet fuel, is a specialised type
of petroleumbased fuel
used to power aircraft. It is gener-
ally of a higher quality than fuels
used in less critical applications
such as heating or road transport. It
is colourless to straw coloured and
is a complex blend of hydrocar-
bons, such as normal paraffns,
isoparaffns, cyclic and aromatic
hydrocarbons. It is obtained by the
direct distillation of petroleum
crude, as well as from the hydroc-
racking of heavier crude fractions.
Tables 1 and 2 list the chemical
composition and various properties
of jet fuel used for commercial
airlines. A product of petroleum
refning, which belongs to the
middle distillate group, it has a
distillation range between 150C
and 300C, but not generally above
250C.
About 25-40% of the operating
costs of an airline are on fuel
expenditure. Annual jet fuel
demand is over fve million barrels
per day. Airlines alone spent $140
billion on jet fuel in 2010. That cost
was expected to reach $200 billion
in 2012. Therefore, airline compa-
nies, as the largest users of jet fuel,
are exposed to price risks owing to
extreme volatility in prices world-
wide. Besides these, there is
environmental concern about the
air pollution due to greenhouse gas
emissions. The high cost of ATF,
along with volatility in petroleum
prices and pressure to reduce
carbon emissions, are the main
factors propelling researchers and
industry to look for alternative
renewable fuel options for the avia-
tion industry.
A refner reviews the available technologies for introducing drop-in
jet fuel production
JITENDRA KUMAR SATyARThI, ChIRANJEEVI ThoTA, GoKAK DATTATRAyA TAMMANNA ShASTRy and
PoyyAMANI SwAMINAThA VISwANAThAN
Bharat Petroleum Corporation
Due to the stringent regulation of
jet fuel and high investment in
current jet engine technology as well
as the supply infrastructure, the
aviation industry wants no solution
other than a drop-in fuel alternative
to petroleum jet fuel. Therefore,
biojet fuel technology should fulfl
the following requirements to have
any signifcant impact:
Flexibility in feedstock process-
ing, because even the same type of
feedstock from renewable sources
can vary in composition; for
instance, vegetable oils can have
different levels of unsaturation and
acid content
Should meet jet fuel specifcations,
especially freezing point, thermal
stability and energy content
Should be a drop-in replacement
for petroleum-based jet fuel
Should be environmentally
acceptable
Production, operating and main-
tenance costs should be competitive
with petroleum jet fuel.
In the last few years, signifcant
developments have been observed
towards the use of biojet fuel.
Besides providing an alternative and
sustainable option, it greatly helps
in reducing carbon emissions
because the use of biojet fuel recy-
cles CO
2
and gives a carbon-neutral
alternative. Signifcantly, there has
been approval of the specifcations
for Fischer-Tropsch-based and
hydrotreated renewal jet fuels by
ASTM and the commercial trials of
biojet fuel by several airlines, includ-
ing Lufthansa, Finnair, Interjet,
Aeromxico, Thomson Airways, Air
France, Air China, Alaska Airlines,
Thai Airways and Jetstar.
Biojet fuels are nearly chemically
identical to conventional petroleum
jet fuels and they match petroleum
jet fuel for performance. There are
various routes and approaches by
which biojet fuels can be prepared,
and some important routes are
discussed here.
Biomass to jet fuel
(Fischer-Tropsch process)
In this process, a synthetic gas
(syngas) consisting of carbon
monoxide and hydrogen is frst
produced by gasifying the biomass
using heat, oxygen and steam. This
syngas is converted into a paraffnic
wax by the Fischer-Tropsch (FT)
process. Finally, the paraffnic wax
www.eptq.com PTQ Q2 2013 75
F
5 0
I m m e d ia te sta b le
N O x fo rm a tio n
C O o x id a tio n a c tivity
d e a c tiva te s ra p id ly
N O x
C O
Figure 4 CO promotion activity of platinum CO promoter has a shorter half-life than
NOx production
w/o COP Pt-COP Non-Pt COP
N in coke, wt% 1.16 1.17 1.17
NOx in FG, ppmw 112 388 145
N in coke to NOx 24% 79% 29%
Balance N in coke to N
2
76% 21% 71%
Impact of CO promoter type on FCC NOx formation: commercial FCC data
Table 2
CO/CO
2
Vol/Vol NO Vol% HCN Vol% N
2
O Vol% Total N rel. to w/o COP
w/o COP 0.52 39 12 0.3 1.0
Pt-COP 0.30 92 8 - 2.0
Non-Pt COP 0.31 67 6 - 1.4
Impact of deactivated COP on N-species selectivity in TPO of spent catalyst
(IR detector)
Table 3
j matthey.indd 4 12/03/2013 09:50
www.eptq.com PTQ Q2 2013 85
promoters prefer to do this by
creating mainly NO, whereas
non-platinum promoters prefer to
convert HCN to N
2
.
Xunhua Mo is a Catalyst Development
Scientist with Johnson Matthey Technology
Center in Savannah, USA. She holds a PhD in
chemistry from Clemson University.
Bart de Graaf leads the FCC Additives R&D
group in Johnson Matthey Technology Center
in Savannah, USA. Prior to this, he worked on
biofuels for Johnson Matthey in Billingham,
UK, and for AkzoNobel/Albemarle on FCC in
various roles. He holds a PhD in heterogeneous
catalysis and chemical engineering.
Paul Diddams is Vice President of Johnson
Mattheys FCC Additives business (formerly
Intercat). Prior to Johnson Matthey, he held
senior positions with Total, Grace Davison and
BP. He holds a PhD in catalysis from Cambridge
University and has been involved in rening for
over 25 years.
formed and its chances of leaving
the regenerator before it is
converted into N
2
or NO depend
on its subsequent residence time
and contact and how quickly its
temperature is increased. Platinum
and non-platinum combustion
promoters both effectively reduce
HCN by catalysing its conversion
to N
2
or NO. Platinum combustion
hence in the regenerator. During
coke combustion, HCN is readily
formed, especially at low tempera-
tures in the presence of oxygen (for
instance, where spent catalyst
enters the regenerator). HCN is
thermodynamically unstable under
FCC conditions; it wants to be
converted to N
2
or NO, which are
much more stable. The fate of HCN
3 0 0
2 5 0
2 0 0
1 5 0
1 0 0
5 0
0 2 0 4 0 6 0 8 0 1 0 0 1 2 0
N
O
x
,
p
p
m
Days
0
C O P -N P a d d itio n s b e g in
C o m m e rc ia l F C C U tra n sitio n s
fro m P t-C O P to C O P -N P.
N O x d ro p s b y 7 0 %
P t-C O P
Figure 5 NOx emissions remain for a long time following discontinued use of platinum
CO promoter
ABB C onsulting.
Safe, reliable, efficient
operations.
O il and G as operators aiming to achieve operational excellence in the face of
ageing assets, harder to reach reserves, cost pressures and sk ills shortages,
often require expert assistance. ABB C onsulting offer the oil and gas industry
expertise in; industrial energy efficiency, inspection, specialist functional
engineering, integrity management, operations improvement, process safety,
projects services, technical software and technical training. We benefit customers
by reducing risk, optimising cost and improving manufacturing efficiency.
w w w . a b b . c o m /c o n s u lti n g
A B B C o n s u lti n g
Tel: +44 (0)1925 741111
E-mail: contact@ gb. abb. com
PTQ Q2 2013 (125x178) Mar 2013.indd 1 04/03/2013 11:00:46
j matthey.indd 5 12/03/2013 09:49
PROFLUX
INNOVATION
severe service
grid packing
YOU CAN RELY ON US.
Find relevant industry events worldwide.
Read news items focusing on contracts,
company news and new products.
dig ref copy 2.indd 1 11/03/2013 16:36
Online cleaning: a mindshift for improving
operational excellence
O
nline cleaning of renery
and petrochemical produc-
tion units has been
positively proven in many severe
applications such as the bottom
section of visbreaker and vacuum
columns with both trays and struc-
tured packing. In turnaround
applications, visual inspection has
conrmed the results. An entire
unit can be cleaned online in as
little as 24 hours. Furthermore,
cleaning can be implemented
during unit operations to recover
unit performance, operate under
clean conditions and improve oper-
ational performance.
In one recent case, ITW applied
its Online Cleaning technology to
the tar stripper section of an atmos-
pheric vacuum distillation unit in a
major US oil renery. The crude
unit was scheduled to operate at
reduced rates for three to four days
due to other maintenance activities.
While at a reduced rate, the ren-
ery performed online cleaning of
the heavily fouled tar stripper
section to improve operational
performance and to validate the
technology for strategic application
in future turnarounds/slowdowns
in lieu of mechanical equipment
cleaning.
ITWs asphaltene stabiliser was
injected into a closed loop circula-
tion that utilised 2-inch temporary
piping to route liquids into the
crude side of four preheat exchang-
ers. Even withstanding the low
circulation rate (normal circulation
utilises 10-inch lines), the results
achieved were impressive.
Due to the short duration of the
unit slowdown, there was no time
available to visually inspect the
A patented method of chemical cleaning can reduce downtime arising from
severe fouling in process units
MARCELLO FERRARA and JUSTIN WEATHERFORD
ITW
heat exchangers for cleanliness.
However, the renery isolated and
visually checked one of the two tar
stripper bottom coke pots as a rst
check of the results of cleaning.
The tar stripper coke pots suffer
severe fouling of jack-hammer-type
coke, and the coke screens are very
difcult to remove after even very
short operating times. According to
the renerys maintenance superin-
tendent, the previous time the tar
stripper coke pots were dealt with,
there was a need to chip the coke
out and more than eight hours
were required just to remove the
basket.
After being cleaned online, the
coke pot was clean. The coke basket
was removed in approximately 10
seconds without any hindrance.
Any remaining deposit was dry,
soft and friable, and could be
broken by hand. In addition, after
the unit was returned to service,
process data were compared to
immediately prior to the cleaning.
Heat exchanger duties were
improved and the pressure drop
across four heat exchangers in the
loop decreased by approximately
15-17 psi, relieving the hydraulic
limitation of the crude unit and
allowing for an increase in
throughput.
Safety and environmental factors
The hydrocarbon processing indus-
try continues to try and limit
airborne emissions through techni-
cal efforts and investments. For
example, consider the investments
and revamps required (valve
substitutions, insulation of anges,
double seals on pumps and double
seals on tanks, for instance) to limit
volatile organic compounds
(VOCs). Then imagine the effects of
hydroblasting on these assets.
Until recently, hydroblasting was
the only effective technology for
cleaning heat exchangers and
process equipment. However, by
performing cleaning operations in a
circulation loop, online cleaning
eliminates the problems associated
with hydroblasting. Moreover,
sludge is converted to a fully reusa-
ble/reprocessable product and the
washing uids are compatible with
the hydrocarbon processing indus-
try (when reused as feed in a
process unit) and combustion
processes (when reused as a blend-
ing component in fuels). As a result,
the equipment can be put on-stream
immediately after cleaning.
Another problem encountered
during turnaround is the duration
of steam-out operations and their
associated environmental problems,
which directly impact equipment
downtime.
Current operations
Mechanical cleaning is currently
performed only where there is a
stringent need for it. This suggests
www.eptq.com PTQ Q2 2013 105
Figure 1 Coke pot showing clean screen
upon opening the vessel head
itw.indd 1 08/03/2013 16:35
106 PTQ Q2 2013 www.eptq.com
unit can be immediately put back
on-stream without extracting the
bundles. However, if visual inspec-
tion is to be performed, a clean
bundle can be extracted. If any
residual hard coke remains, it will
be dry, friable and very easily
removed, as the technology elimi-
nates all the coke binders.
Increasing revenues
The following case histories may be
considered unusual by some, while
others may see them as typical of
the hydrocarbon processing indus-
try. Paradoxically, the industry is
not energy effcient. As energy is so
readily available, costs are not
always taken fully into account.
Another problem is that cleaning
is a lengthy and unsafe operation
that can lead to a loss in produc-
tion, which can overwhelm the
energy-saving benefts. For this
reason, cleaning is usually post-
poned until the very last minute.
Online cleaning helps to address
production, energy-saving, environ-
mental, safety and reliability issues,
as it can achieve effective equip-
ment cleaning via closed loop
circulation for the entire production
unit in one day or less, thereby
making equipment immediately
available for production.
In very diffcult-to-solve cases,
online cleaning can be performed
during different periods of the run
to avoid certain problems.
Depending upon the magnitude of
the problem and the particular
application, even coke formation
inside equipment (for example,
columns, packings and exchangers)
or delta P build-up can be
prevented by regular application of
online cleaning.
At current high crude oil costs
and low operating margins, being
able to address all these issues can
lead to signifcant improvements in
revenue. For example, by reducing
the amount of fuel burned, there
will be a reduction in SOx, NOx,
CO, CO
2
and other airborne pollut-
ants at no extra cost. When
considering all the investments the
hydrocarbon processing industry is
implementing to reduce waste/
airborne pollutants, online cleaning
can provide these advantages:
that equipment is being allowed to
degrade to the minimum acceptable
performance before cleaning. This
situation arises from a number of
peculiarities in the hydrocarbon
processing industry.
First, the industry sees cleaning
as a troublesome, lengthy operation
that leads to a loss in production,
so it is avoided. Also, the mechani-
cal cleaning process involves at
least 20 operations, each of which
has an associated hazard.
When the production/mainte-
nance department faces a cleaning
situation, it normally reacts in one
of two ways:
The situation is fairly typical and
managed as a routine operation
The situation is problematic and
postponed until it can no longer be
tolerated.
In the frst case, the management
team is unlikely to consider some-
thing that is not routine. Or, in the
best-case scenario, it will be excited
about trying something new.
However, unless the management
team is strongly committed to the
necessary testing of the new tech-
nology, there is a tendency to
procrastinate.
In the second case, the manage-
ment team may not be inclined to
look for a new technology. Instead,
it addresses the problem in a
manner that is ineffcient and costly.
In both cases, process unit econom-
ics and environmental performance
can suffer signifcantly.
Advantages of online cleaning
An array of technologies is availa-
ble for solving problems related to
equipment cleaning, including
ITWs patented technology for
cleaning heat exchangers and
process equipment in a closed
circulation loop.
Equipment can be cleaned online
with this technology without
having to extract the heat
exchanger bundles. This is particu-
larly benefcial in severe
applications such as visbreakers
and columns with structured pack-
ing internals. Crude distillation
units and vacuum units have also
been successfully cleaned online.
An entire unit can be cleaned in as
little as 24 hours. The advantages
online cleaning have over hydro-
blasting include:
Reduced downtime
Closed loop operation, without
the need for extracting the bundles
Simultaneous cleaning of multi-
ple process units
No waste generation
No emissions
Improved safety.
Online cleaning involves a
patented method and the injection
of a patented oil-based chemical
that transforms sludge and coke-
like material into a fully reusable
product, thereby removing any
sludge and deposits from metal
surfaces. The washing solution can
be reused/reprocessed because the
stabilising properties of the
chemical avoid re-aggregation of
deposits, thereby eliminating any
potential precipitation during stor-
age and/or fouling of the process
equipment. Moreover, the chemi-
cals used in ITWs method are
compatible with hydrocarbon
processing industry/combustion
processes because they:
Do not contain any metals
Do not contain any compounds
based on P, B, S, As, Bi, Si or Pb
Do not contain any halogens
Do not contain any compounds
that might be harmful to any plant
metallurgy
Do not contain any carcinogenic
compounds
Do not contain any compound
that in operating dosages might
interfere with biological waste
treatment processes
Do not negatively impair fnished
products specifcations.
After being cleaned online, the
heat exchangers or entire process
After being cleaned
online, the heat
exchangers or entire
process unit can be
immediately put back
on-stream without
extracting the bundles
itw.indd 2 08/03/2013 16:35
Reduced waste generation
Reduced VOC emissions
Reduction in airborne pollutants
(following energy savings and
reduction in fuel burned)
Reduced CO
2
emissions (follow-
ing energy savings and reduction
in fuel burned)
Improved safety.
It is important to note that the
health, safety and environment
benets are achieved without any
extra investment or cost. The cost
of installing spare equipment is
also avoided. The additional bene-
ts of online cleaning are increased
run length, operation under cleaner
conditions and potential avoidance
of a turnaround.
Turnaround operations
When applied in preparation of a
turnaround, online cleaning greatly
reduces downtime, besides
improving health, safety and envi-
ronmental performance. In this
application, a technology is also
available that uses another patented
chemical, added during steam-out,
to quickly and effectively achieve
safe entry conditions. The chemical
also eliminates any noxious odours
during steam-out, as well as any
pyrophorics-related problems.
Additionally, the chemical does not
create emulsions. By signicantly
reducing steam-out time, steam
requirements will be greatly
reduced too, which is important
during turnaround operations
when steam production is at a
minimum and its consumption is at
stake.
By applying online cleaning
followed by improved degassing/
decontamination, there is little, if
any, need for mechanical cleaning,
and safe entry conditions are
achieved more quickly. This results
in a signicant reduction in down-
time, which, in turn, leads to an
increase in production.
Alternatives
There are instances where online
cleaning technology was consid-
ered from the onset as an
alternative to mechanical cleaning
or to improve turnaround opera-
tions. This was due to the fact that
the operator required visual
www.eptq.com PTQ Q2 2013 107
inspection to validate the technol-
ogy. After such positive validation,
ITW Online Cleaning technology
was implemented during the
actual run to immediately recover
unit performance, to run under
cleaner conditions and to increase
run length. It is therefore used as a
preventive tool for unit optimisa-
tion and to improve operational
excellence. Achievable revenues
from these optimisations depend
upon unit conguration and
production layout. They will
certainly reach several million
dollars if applied proactively.
Climate conditions and cleaning
schedules
Very cold or very hot climates often
dictate the time required for
process unit cleaning. As a result,
reneries may be forced to tolerate
inefciencies until conditions
permit. However, climate condi-
tions only affect those mechanical
cleaning techniques such as hydro-
blasting and manual cleaning that
are lengthy and labour/equipment
intensive. Climate-related factors
do not affect online cleaning to the
same degree. ITW technology only
requires 24 hours on an oil-to-oil
basis, with normally only one or
two engineers to monitor the clean-
ing and no heavy hardware, just
the chemical and the injection
pump.
Case study: online cleaning of a
visbreaker unit
Online cleaning was performed
during a scheduled decoking of the
furnaces after one years operation
of a European renery visbreaker
unit. Cleaning of the heat exchang-
ers was originally not scheduled, as
the unit can bypass the critical
preheat exchangers, allowing for
mechanical cleaning during the
run. The following is a summary of
the results:
In 36 hours, 28 bottom exchang-
ers of a 310 t/h visbreaker were
cleaned
Figure 2 Visbreaker bottom exchanger following online cleaning
Figure 3 Visbreaker main fractionator
bottom following online cleaning
itw.indd 3 08/03/2013 16:36
bete.indd 1 26/02/2013 11:17
www.eptq.com PTQ Q2 2013 109
was facing an unscheduled
shutdown to mechanically clean
the entire vacuum bottom section.
To solve the problem, the
vacuum section bottom was online
cleaned in 48 hours. The visbreaker
was then started up immediately.
The temperature recovery on
TI1826 was on average 45C,
which remained stable during the
run (see Figure 5). The unit could
easily reach its scheduled turna-
round and run fve more months.
(The full run is not reported in
Figure 5.)
layout of the refnery with this
visbreaker is shown in Figure 4.
The refnerys problem was
related to an increase in the
vacuum section bottom train outlet
temperature (reading TI1826). The
refnery had to shut down the unit
for cleaning when the temperature
approached 280C. This event
occurred long before the conven-
tional run length, and the refnery
had to reduce throughput to
the minimum rate. However, this
did not recover the temperature,
so, after one month, the visbreaker
The bottom of the main fractiona-
tor was clean. Normally, more than
6ft of sludge and coke is found,
and at least four days are sched-
uled for chipping and removing the
deposits via buckets through the
nearest man-way. Following online
cleaning, only two hours was
required for removing the deposit
The bottom packing of the main
fractionator was clean
Furnace decoking time was
reduced by about 20%
The visbreaker started up 12
hours before the scheduled time,
with all the preheat, packing and
lines clean.
Visual inspection of the most
severe service bottom exchanger
was performed to validate the
results. However, the other
bottom/preheat train equipment
was not opened. Figure 2 shows the
visual inspection of the frst bottom
exchanger cleaned after one year in
operation.
Figure 3 shows the bottom of the
main fractionator at the man-way
opening, wherein almost no coke
was found. The unit started up
with no mechanical cleaning of heat
exchangers.
Case study: online cleaning of
visbreaker vacuum section
bottom train
Online cleaning was applied to the
vacuum section of a visbreaker. The
Topping
Vacuum
Vacuum
section
Bottom
train
Tank
TI 1826 Visbreaker
IGCC
Figure 4 Online cleaning of visbreaker vacuum bottom section refnery layout
2 1 0
2 5 0
2 4 0
2 3 0
2 2 0
2 6 0
3 0 0
2 9 0
2 8 0
2 7 0
2 0 0
1 9 0
1 7 0
1 8 0
1 6 0
0 2 0 4 0 6 0 8 0 1 0 0 1 2 0 1 4 0 1 6 0 1 8 0 2 0 0 2 2 0
T
e
m
p
e
r
a
t
u
r
e
,
C
Days on st ream
1 5 0
I T W o n lin e c le a n in g
M e c h a n ic a l c le a n in g
o f two s h e lls
A v e ra g e re c o v e ry
fro m I T W
L im it fo r u n it s h u td o w n
Figure 5 Visbreaker vacuum section bottom train outlet temperature recovery after
online cleaning
itw.indd 4 08/03/2013 16:36
110 PTQ Q2 2013 www.eptq.com
hours. The recovery of the quench
temperature was about 50C
(see Figure 6) and, as the units are
integrated, a recovery of 18C on
the vacuum furnace inlet tempera-
ture was also achieved (see
Figure 7).
Case study: online cleaning of a slop
distillation unit
A slop distillation unit processes,
on a campaign basis, all the slop
collected on-site at a 280 000 b/d
renery. The facility has four
topping/vacuum units, one FCC
unit, one residue hydrocracker and
two ethylene units. After a
one-month campaign, the slop
distillation unit is shut down for
cleaning, as it suffers very heavy
fouling and coke accumulation.
To demonstrate the effectiveness
of online cleaning, the unit was run
as usual. It was then shut down
and opened to check the fouling
conditions in the main fractionator
and two bottom exchangers. The
unit had about one metre of coke-
like sludge in the bottom of the
main fractionator. Moreover, the
main fractionators trays and
bottom exchangers were heavily
fouled.
The online cleaning lasted 24
hours. Degassing/decontamination
was then implemented to reduce
steam-out time, eliminate any
noxious emissions and pyrophorics
ignition. The main fractionator
bottom and trays were free of any
deposits, and practically no depos-
its (only 12 kg) were found in the
bottom exchangers.
Furthermore, these deposits were
loose and easily removed using the
units service water at 23 bar. No
mechanical cleaning was carried
out on the unit and it was hence-
forth started up. Figures 8 and 9
show the situation before online
cleaning, and Figure 10 shows the
improvements.
Case study: online cleaning a
260 000 b/d crude distillation unit
A renery wanted to validate our
technologies for future turna-
rounds/slowdowns and during the
run.
Online cleaning was performed
on the entire CDU (which has two
which was limiting throughput.
Mechanical cleaning of four heat
exchanger shells (with the hottest at
the bottom) had not helped to
recover the temperature, so there
was a steady increase in quench
temperature.
After being cleaned online, the
two units were started up in 36
Case study: simultaneous online
cleaning of a vacuum unit and
visbreaker
Online cleaning was applied to a
vacuum unit integrated with a
visbreaker at a European renery.
The reason for considering online
cleaning was the increase in the
visbreakers quench temperature,
2 7 0
2 9 0
2 5 0
2 3 0
2 1 0
1 9 0
1 7 0
1
J
a
n
1
4
J
a
n
2
7
J
a
n
9
F
e
b
2
2
F
e
b
7
M
a
r
2
0
M
a
r
2
A
p
r
i
l
1
5
A
p
r
i
l
2
8
A
p
r
i
l
1
1
M
a
y
2
4
M
a
y
6
J
u
n
1
9
J
u
n
2
J
u
l
1
5
J
u
l
2
8
J
u
l
1
0
A
u
g
2
3
A
u
g
5
S
e
p
1
8
S
e
p
1
O
c
t
1
4
O
c
t
2
7
O
c
t
9
N
o
v
5
D
e
c
T
e
m
p
e
r
a
t
u
r
e
,
C
1 5 0
A fte r I T W
L in e a r ( b e fo re I T W)
B e fo re I T W
Ave ra g e re c o ve ry fro m I T W
I T W o n lin e c le a n in g
VB -E 1 m e c h a n ic a l
c le a n in g
2 9 0
3 0 0
2 8 0
3 1 0
3 2 0
2 7 0
2 6 0
2 5 0
2 4 0
1
J
a
n
1
4
J
a
n
2
7
J
a
n
9
F
e
b
2
2
F
e
b
7
M
a
r
2
0
M
a
r
2
A
p
r
i
l
1
5
A
p
r
i
l
2
8
A
p
r
i
l
1
1
M
a
y
2
4
M
a
y
6
J
u
n
1
9
J
u
n
2
J
u
l
1
5
J
u
l
2
8
J
u
l
1
0
A
u
g
2
3
A
u
g
5
S
e
p
1
8
S
e
p
1
O
c
t
1
4
O
c
t
2
7
O
c
t
9
N
o
v
5
D
e
c
T
e
m
p
e
r
a
t
u
r
e
,
C
2 3 0
A fte r I T W
L in e a r ( b e fo re I T W)
B e fo re I T W
I T W o n lin e c le a n in g
Ave ra g e
re c o ve ry
fro m I T W
Figure 6 Simultaneous online cleaning of vacuum and visbreaker: quench temperature
recovery
Figure 7 Simultaneous online cleaning of vacuum and visbreaker: furnace inlet
temperature recovery
Figure 8 Main fractionator bottom before
online cleaning
Figure 9 Main fractionator bottom before
online cleaning
itw.indd 5 08/03/2013 16:36
www.eptq.com PTQ Q2 2013 111
Thus, using online cleaning and
polymer modication technology,
all of the problems related to
quench water loop fouling can be
eliminated. The same approach can
be applied to cleaning a quench oil
loop.
Marcello Ferrara is the Chairman of ITW. He
holds a PhD in industrial chemistry and has 27
years experience in the petroleum business. He
holds international patents for new processes
and additive compositions for environmental
control, and for improving petroleum/
petrochemical processes.
Email: mferrara@itwtechnologies.com
Justin Weatherford is the Sales Manager North
America for ITW. He holds a BS in chemical
engineering from Kansas State University.
He has 12 years experience in the rening/
petrochemical business.
Email: jweatherford@itwtechnologies.com
some cases, this is not sufcient to
restore the cleanliness of the
bundle. Often, the bundle must be
replaced with a new one, as
mechanical cleaning is almost
impossible to achieve. Figure 12
shows a quench water exchanger
with its typical fouling.
ITW applied its Polymer
Modication technology to this
problem. By making tests on some
deposits taken from a severely
fouled exchanger, it was possible
to validate the technology.
Apparently, there were different
types of fouling layers in the bundle:
Coke
Coke and polymer
Inorganic (a white powder-type
deposit, looking like calcium
carbonate).
All of these were hard, rock-like
materials. The samples were all of
the polymer type and it was possi-
ble to dissolve/modify them. After
treatment, only a fully broken, very
soft and friable polymer was left.
preheat trains) and lasted 24 hours.
Subsequent to that, as the renery
wanted to further recover down-
time by reducing steam-out
requirements, improved degassing/
decontamination technology was
applied. This also lasted 24 hours.
At the man-ways opening, the
following results were achieved:
LEL = 0%; HC = 0 ppm; H
2
S = 0
ppm; benzene = 0 ppm; ammonia =
0 ppm; pyrophorics = nil.
Moreover, for the rst time in the
units history, there was no need
for columns wetting, as pyrophor-
ics were completely eliminated. All
the maintenance operations were
conducted under dry conditions.
Visual inspection of the bundles
conrmed the results of online
cleaning, shown in Figure 11.
The renery decided, therefore,
not to extract any bundles during
turnaround, but rather to check
their performance at unit startup.
At CDU startup, the results of
online cleaning were comparable to
mechanical cleaning. The outlet
temperatures of the trains were the
same and the fouling factor of the
heat exchangers was comparable to
the mechanically cleaned ones.
No waste was generated during
online cleaning. All of the washing
uids were sent to a slop oil tank
and were fully reprocessed at unit
startup. There were no quality or
operational issues during repro-
cessing. As a result, the renery has
implemented the following objec-
tives into its operating plan:
Avoid CDU slowdown to clean
the preheat trains (one day of
online cleaning vs about 25-30 days
at half capacity for slowdown)
Rescheduling of future turnaround
downtime by taking into considera-
tion both the reduced time for safe
entry and the reduced need for
cleaning exchangers.
Case study: online cleaning of
ethylene unit quench water loop
exchanger
The quench water loop exchangers
in an ethylene unit normally suffer
severe fouling problems. It is
common to require two weeks to
extract a single bundle. On top of
that, hydroblasting might take an
additional three to ve days. In
S lu d g e le ve l
b e fo re I T W
Figure 10 Main fractionator bottom after
online cleaning
Figure 11 CDU bottom exchangers after
online cleaning
Figure 12 An ethylene quench water loop exchanger might take two weeks to be
extracted
itw.indd 6 08/03/2013 16:36
epc.indd 1 07/03/2013 13:39
Lessons learned from Mpc revamp of two
naphtha crackers
B
raskem is the eighth largest
petrochemical company in the
world and the largest polymer
producer in the Americas. The
Braskem Triunfo site includes two
naphtha crackers with a combined
olefns capacity of 1250 KTA ethyl-
ene and 685 KTA propylene.
AspenTechs proprietary DMCplus
controllers were built for both
olefns plants over the past 10
years, but were not fully
operational.
In 2010, a project was initiated to
review the performance of the
advanced control systems and
improve their operability, service
factor and economic impact. A
further objective was to increase
the site process control engineers
knowledge. The 2010 project
covered two Braskem olefns units
located at the same site in the south
of Brazil. The site actually has three
olefns plants and one aromatics
unit. Two plants, referred to as
Olefns 1 and Olefns 2, process
naphtha. The third olefns plant
uses ethanol as feed (green ethyl-
ene). The multivariable predictive
control (MPC) project discussed
within this context is based on the
two naphtha-fed olefns plants.
Leveraging advanced controls
Braskem was formed in 2002
through the merger of several
petrochemical and chemical compa-
nies. The company has a focused
growth strategy and has acquired
companies in Brazil and, more
recently, in the US, Germany and
other countries. Braskem has indus-
trial units in the four largest
petrochemical centres in Brazil. It
utilises advanced control as a
A project overview includes changes made at the regulatory and advanced control
level, and results obtained. Techniques and tools used in the project are highlighted
VAnessA Conz, suzAnA FuChs and PrisCiLA MACedo MourA Braskem
MArk dArBy CMiD Solutions doug niChoLson IPCOS (UK)
means to reduce variable costs,
increase energy savings and
improve productivity.
Olefns 1 started production in
1982. It is the largest ethylene
producer in the complex and has
the capacity to process 312 t/h of
naphtha. The plant has 12 naphtha
furnaces and two recycle ethane
furnaces. Olefns 1 normally oper-
ates with high feed rates typically
limited by raw gas compressor
constraints. It is responsible for the
pressure control of the site ethylene
supply grid and is therefore
affected by consumer demand vari-
ation. The base instrumentation is
the oldest on site, with too many
control loops in manual mode.
These characteristics make Olefns 1
challenging for operators with a
high number of interventions from
the control panel.
Olefns 2 started production in
1999 and has the capacity to
process 180 tph naphtha. The plant
has seven naphtha furnaces, one
being a swing furnace that can be
used on either olefns plant, and
one ethane furnace. It is more diff-
cult for this plant to handle feed
composition variability, as this unit
processes more LGP and C
4
s along
with naphtha.
Braskem owns several sites in
Brazil. A particular challenge the
company faces is job rotation of its
people, such as production and
control engineers. New plant
projects require experienced senior
engineers. To succeed with MPC, it
is well known that site engineers
require the process and tools
knowledge to maintain the installed
base of applications. The use of
projects to accelerate engineer train-
ing and learning is strategic to the
company.
Both olefns plants already have
MPC controllers installed. Olefns
1 has had controllers since
2004 and Olefns 2 since 2001.
Table 1 lists the controllers that
have been installed. After plant
shutdowns and minor revamps,
some of the controllers were
www.eptq.com PTQ Q2 2013 113
MPCs implementation year
Olefns 1 Furnaces with O
2
optimisation using draft pressure (14) 2007
De-ethaniser 2004
C
2
splitter 2006
MAPD reactor 2004
C
3
splitter (2 columns) 2004
Debutaniser 2004
Olefns 2 Furnaces with O
2
optimisation using dampers (8) 2001
Primary fractionator and quench tower (2) 2001
High- and low-pressure depropaniser (2) 2001
De-ethaniser (1) 2001
C
2
splitter (1) 2001
MAPD reactor (1) 2001
C
3
splitter (1) 2001
HPG hydrogenation reactor (1) 2001
installed MPC controllers at the start of the project
Table 1
braskem.indd 1 12/03/2013 11:38
114 PTQ Q2 2013 www.eptq.com
project scope (for example, in iden-
tifying benets opportunities
during the project). This approach
requires discipline to calculate
benets and determine priorities.
Lessons learned and relearned
A multi-skilled project team was
formed with control engineers,
process engineers, production engi-
neers, instrument technicians and
operators. MPC consultants partici-
pated in all steps. DCS
modications were executed by the
site control engineers. In addition,
the site supply chain group
provided important information
about company optimisation objec-
tives and economic scenarios. The
skills diversity of the group was
essential to the success of the
project and reinforced the conclu-
sion that teamwork is very
important for MPC projects. The
scope of the project initially
involved regulatory control stabili-
sation to avoid ooding of the
distillation columns, to allow oper-
ation closer to products
not operational and others were
not working well.
Revamp project
The online factor in Olens 1 was
poor and the benets in the existing
MPCs were not being captured.
The objective of the project was to
improve the economic and qualita-
tive benets, as well as to increase
the online factor. In Olens 1, for
example, the variability of C
2
split-
ter impurities was high and the C
3
splitters had a large propylene loss
(>20% mass). In Olens 2, the goal
was to remodel the parts of the
controllers that were not working
properly from the last shutdown in
2005. In addition, a parallel project
installed new actuators on the
secondary air dampers of the
furnaces in order to control excess
oxygen more efciently. The project
was an opportunity to review the
furnace models and to add the
damper positions as additional
MVs to the furnace controllers.
The site had just one engineer
responsible for MPC for a number
of years. In 2010, the company
hired a trainee to become a control
engineer, and a production engi-
neer was assigned to the team to
help the control engineers maintain
the controller online factor. The
project was a good opportunity for
engineers from the site to learn
about MPC project methodology
and to develop expertise on
advanced process control in order
to handle future projects by
themselves.
The scope of the project was chal-
lenging. There were many columns
and furnaces involved, and few
engineers to execute the work. To
speed things up, and to empower a
team environment, the control
group decided to hire advanced
control consultants. The project was
justied on economic return.
Although some of the equipment in
the scope had MPC controllers,
certain controllers were selected to
be remodelled due to their poor
performance. The contract was
made based on consultant hours in
order to have exibility in the
F C
L C
T C
F e e d
P G C
C 2 S
C 3 V
P G C
F C
C 3 L
F C
R e c yc le e th a n e
Side reboiler
13T03
P C T
M
V
F C
P C
L C
F D C M V
M V
L C
V13 ethylene
product vessel
Ethylene tank
Internal reflux
(ext. reflux draw)
Reflux drum
F C
P C
T C
C 3 R
C 3 R
L C
F C
Ve n t
Vaporisation via
C2 compressor
(fixed rate)
Plant 2
ethylene vapour
C 3 L
C 3 V
E th yle n e va p o u r
to c u sto m e rs
F C M V
Ethylene
vaporisation
L I
Figure 1 Revised C
2
splitter regulatory controls
braskem.indd 2 08/03/2013 16:53
Don t mi ss t hi s i mpor t ant
i ndust r y event !
Orlando World Center Marriot t
Orlando, Florida
May 21-24, 2013
Within the next ten years, the
industr y will be run by todays
up and coming engineers.
BUILDING
THE NEXT
GENERATION
RELIABILITy & MAINTENANcE
cONfERENcE AND EXHIBITION
This years conference emphasizes
how this new generation of
engineers can take the torch
and improve reliabilit y in the
refning and petrochemical
industries. Dont miss our
great speakers and professional
development sessions!
Register at www.afpm.org.
af pm.or g
afpm.indd 1 08/03/2013 12:14
116 PTQ Q2 2013 www.eptq.com
specications and, particularly for
Olens 2, better energy perfor-
mance on the cracking furnaces.
These were the obvious opportuni-
ties in MPC. However, working
with the consultants gave a better
understanding of the entire process
and led to other opportunities. In
order to prioritise these new oppor-
tunities against the initial scope, it
was important to measure the
benets on an ongoing basis. This
also helped project promotion and
acceptance.
To ensure the success of the
project, it was crucial to monitor
the results and performance in
order to help the operators and
engineers use the new capabilities
in the controllers.
Realising the full benets of the
MPC applications required solving
a control problem associated with
the ethylene product header. A
challenge to the Olens 1 plant is
the disturbances caused by disrup-
tions in the demand of the ethylene
customers, which historically had
required overriding DMCplus
targets or switching off the
DMCplus controller. Solving this
specic control problem reduced
the impact of demand disruptions
on both operations and the other
MPC controls, resulting in less
operator interaction, smoother
operation and higher online service
factors.
Another important control prob-
lem is the DCS regulatory controls.
Some MPC projects take the regula-
tory layer as they nd it. Changing
a DCS structure is hard work, oper-
ations must be convinced of the
benets, safety reviews must be
carried out, drawings must be
updated and console technicians
trained. Nonetheless, change is
often the right thing to do. Despite
both Olens 1 and 2 being well-es-
tablished ethylene units with
functional multivariable controllers,
several of the improvements in this
discussion are achieved by chang-
ing the regulatory control system.
The correct regulatory structure can
make plant operation and multivar-
iable control better. As an example,
using a pressure-compensated
temperature to control the tray
temperature reduces the impact of
pressure changes on the column.
This simplies the controller
design, making day-to-day perfor-
mance better and long-term
support easier. Another example is
reux subcooling compensation in
the DCS. In C
3
splitters, this is a
signicant effect. If not corrected
for, any independent variable that
changes subcooling (ie, pressure,
cooling water temperature) will
have a model to the overheads
composition and column pressure
drop.
Specic improvements
Regulatory design, or redesign in
this instance, can have a signicant
impact on MPC control. This
includes linearisation and improved
rejection of disturbances due to the
faster execution of DCS loops.
Proper regulatory design can also
increase the robustness of the over-
all system to model errors, simplify
the controller design and better
accommodate plant constraints. A
summary of the redesigns follows:
C
2
splitter In Figure 1, the tray 13
temperature correlates very well
with the bottoms ethylene composi-
tion. However, incorporating a
temperature controller with the
existing strategy of accumulator
level control to reux would result
in an unstable control arrangement
(ie, there would be a snowball
effect when the TC adds heat, the
overhead accumulator LC adds
reux, etc). Therefore, an internal
reux controller was added, which
manipulates the side-draw FC.
Internal reux is calculated as
reux ow minus side-draw ow,
based on setpoint values. As a
result, incremental reux changes
translate into incremental side-
draw changes (at a constant
internal reux). This control modi-
cation permits pressure compensate
temperature (PCT) control of tray
13 via reboiler ow. The PCT
controller has stabilised bottoms
ethylene and the tower material
balance (as seen in a more stable
bottoms level)
Green oil tower The original strat-
egy had the overhead accumulator
level in cascade to the reux and
the bottoms level in cascade to the
reboiler quench water ow. The
side draw was manipulated by a
DCS resident internal reux ow
F C
D P !
F C
Q
T
I N
T
O U T
Q WS
B e fo re
A fte r
Q WR
L C
F e e d
113T07
113T06
Figure 2 Plant 2 C
3
splitter revised regulatory control
The correct
regulatory structure
can make plant
operation and
multivariable control
better
braskem.indd 3 08/03/2013 16:53
www.eptq.com PTQ Q2 2013 117
Column DP is a constraint, result-
ing from tray downcomer liquid
backup. Internal reux control,
calculated based on the difference
between overhead and reux
temperature, was implemented in
the DCS. A duty controller was
incorporated into the bottoms LC
to reboiler ow cascade. Although
disturbances in quench water inlet
temperature (due to furnace
swaps) are realised relatively
quickly in the bottoms level, the
addition of the duty controller
showed denite improvement in
providing tighter level control
with less control action (and less
interaction with the overhead
system)
C
3
splitter plant 2 This large
column (see Figure 2) is made up
of an upper column and a lower
controller (calculated as reux
minus side draw); with this strat-
egy, the level controllers would
ght when the FDC was saturated,
leading to a violation in the column
DP limit. The strategy was modi-
ed to have the level controller
manipulate the C
3
splitter feed ow
and the internal reux controller
manipulate the reux ow
C
3
splitter plant 1 The original
strategy was a traditional design
for high-purity, small bottom
product columns: bottoms level to
reboiler quench water ow, with
DMC manipulating reux ow
and bottoms ow. The overhead
condenser is water cooled and,
therefore, subject to ambient
effects, leading to disturbances in
reux temperature and, therefore,
changes to the internal reux ow.
column. There is a side reboiler at
the bottom of the upper column
heated by hot water. This
exchanger reboils approximately
40% of the reux. The original
strategy was to x the reboiler duty
and control the level in the upper
column bottoms by adjusting the
ow to the lower column.
However, even with a Q calcula-
tion on the reboiler duty, the duty
of the reboiler was not constant
when the water supply temperature
changed. These errors in turn
caused ow changes to the lower
column and duty changes to the
upper column, upsetting both
columns unnecessarily. The lower
column had a strict ooding limit
so accurate control was important.
The cure was simple: the base regu-
latory control strategy was altered
No blue background.
Always translate into UK
English (except images).
Leading cap on rst word
only.
Pumps and valves as below.
Cylinders to have graduated
ll, always in colour
Block diags: 2pt lines
Schematic diags: 0.75pt lines
Arrows all type 4, 90%
FULL PAGE: 175 x 250 mm
Helvetica Neue
Bold 9pt
H e lve tic a N e u e
M e d iu m 8 p t
F C M V
F C
F C M V
F C M V
F e e d
S te a m
R a tio
R a tio
M V
T C A VG M V
4 1
F C
F u e l
g a s
M a ss
H L
D AM D AM
P C L P C H
M V M V D AM M V
F C M V
F C
F C M V
F C M V
F e e d
S te a m
E fflu e n t
R a tio
R a tio
M V
T C A VG M V
2 3
F C
F u e l
g a s
H L
D AM D AM
P C L P C H
M V M V D AM M V
P C
M
V
Figure 3 Plant 2 furnace with revised inlet dampers
braskem.indd 4 08/03/2013 16:53
the furnace operating range. The
plant 2 furnaces have four damp-
ers, two on each side, as shown in
Figure 3.
Most of the current furnace
controller models were judged
satisfactory, so response testing
was limited to testing only varia-
bles that signicantly changed the
furnace air ow rate through the
furnace. These are the four
revamped inlet dampers and the
furnace draft pressure. Testing of
these variables was done manually
over several days.
A challenge with all cracking
furnace MPC controllers is ensuring
the nal control matrix is correctly
collinear. This analysis is compli-
cated by the physical arrangement
of the furnace, which means
dependent variables, such as fuel
gas pressure and valve position, in
the radiant section will have a
different gain ratio (by a factor of
two) to variables such as radiant
coil inlet temperature in the convec-
tion section.
The identied model gains were
modied using a custom Excel
spreadsheet designed to nd the
minimum necessary changes to the
identied gains to achieve co-
linearity. Using Excel allowed indi-
vidual curves to be weighted
according to how accurate each
gain was. It also allowed the calcu-
lation process to be documented for
further use if the furnaces are
retested.
The original furnace controller
made use of a number of custom
code calculations designed to
inhibit the controller during certain
conditions. For example, the
controller would suppress move-
ment of the draft pressure
whenever the draft pressure was
being reduced, but not when it was
being increased. At times, this
would cause oscillatory movement
in draft pressure, moving slowly
down and quickly up. Such calcula-
tions are usually applied after the
initial controller is built, in an effort
to mitigate poor performance with-
out retesting. These calculations are
confusing and can only be imple-
mented using a trial and error
approach. For the revised furnace
controller, all of these calculations
to control the level by adjusting the
reboiler duty and the accurate ow
to the lower column was held xed.
This simple change corrected the
inaccurate reboiler duty and
allowed staff to easily keep both
columns steady during hot water
temperature transients. This
resulted in simpler models and a
stronger controller.
118 PTQ Q2 2013 www.eptq.com
3 5 0 0
4 5 0 0
4 0 0 0
3 0 0 0
2 5 0 0
2 . 5 2 . 7 2 . 9 3 . 1 3 . 3 3 . 5 3 . 7 3 . 9 4 . 1
P
P
M
d
r
a
w
Feed propane, %
2 0 0 0
L in e a r ( to p )
L in e a r ( b o tto m s)
To p
B o tto m s
y = 1 7 5 0 . 8 x 2 6 9 2 . 2
R
2
= 0 . 9 9 4 9
y = 1 3 . 0 5 3 x +8 6 . 1 6 7
R
2
= 0 . 9 9 9 8
3 0
5 0
4 0
2 0
1 0
0
6 0
C
3
i
n
b
o
t
t
o
m
s
,
%
Figure 4 Simulated effect of feed composition on overhead and bottoms product
compositions
3 0 . 0
3 2 . 5
2 7 . 5
2 5 . 0
2 2 . 5
2 0 . 0
1 7 . 5
1 8 1 6 1 6 3 2 2 4 4 8 3 2 6 4 4 0 8 0 4 8 9 6 5 7 1 2 6 5 2 8 7 3 4 4 8 1 6 0
F
e
e
d
,
t
o
n
s
/
h
r
Ti me, 12-mi nut e i nt erval s
1 5 . 0
3 0 . 0
3 2 . 5
2 7 . 5
2 5 . 0
2 2 . 5
2 0 . 0
1 7 . 5
1 5 5 9 1 1 1 8 1 6 7 7 2 2 3 6 2 7 9 5 3 3 5 4 3 9 1 3 4 4 7 2 5 0 3 1 5 5 9 0
F
e
e
d
,
t
o
n
s
/
h
r
Ti me, 12-mi nut e i nt erval s
1 5 . 0
Figure 5 C
3
splitter 1 feed
Plant 2 furnace upgrade
The primary purpose of the plant 2
furnace MPC upgrade was to
include the revamped secondary air
dampers to the existing furnace
controllers. Inlet air dampers adjust
the burner secondary air ow.
These dampers are very important
for effective air rate control; draft
pressure alone is not adequate over
braskem.indd 5 08/03/2013 16:53
revised MPC controller was built to
exploit capacity of both vessels so
that the ow of ethylene to storage
www.eptq.com PTQ Q2 2013 119
the size of the surge capacity.
Previously, the reux drum had
been under tight level control. The
were removed, as they were not
required with the revised controller
model. Once the rst furnace was
completed, all identical furnaces
were changed to use the same
controller model.
All composition variables were
modelled as steady-state variables,
as opposed to ramp variables (as is
often done to get a shorter control-
ler horizon), to provide a consistent
set of steady-state targets to the
controller. The C
3
splitters can be
constrained due to column DP or
condenser limitations, both of
which are relieved by reducing the
internal reux. Note that a devia-
tion above the upper propane
impurity limit of the propylene
product invariably results when the
reux is decreased to control at one
of these constraints due to the
slower dynamic effect of the
bottoms draw compared to the
reux. In Olens 1, the DMCplus
controller manipulated the feed
rates to both C
3
splitters (and
included the green oil tower). This
provides the future capability to
adjust the feed split to balance
propylene losses in the two splitters
when the models for the second
splitter (heat pump column) are
added.
On Olens 1, the side draw from
the C
2
splitter goes to a separation
vessel (V13) under ow control.
From this vessel, the majority of the
ethylene is vaporised to supply the
site vapour customers. The
pressure of the customer pipeline
is controlled by adjusting the
ow of ethylene from this vessel
through two vaporisation routes.
The remaining ethylene is refriger-
ated for cryogenic storage.
This ow to storage consumes a
signicant amount of refrigeration
duty.
Level control of this relatively
small vessel was a constant chal-
lenge for the Olens 1 operators
due to customer demand changes,
plant production changes and
Olens 2 production rate changes.
This situation was improved by
changing the C
2
splitter control
scheme, as described earlier, and
controlling both the C
2
splitter
reux drum level and the V13
product level, effectively doubling
6 0 0 0
5 0 0 0
4 0 0 0
3 0 0 0
1 8 1 6 1 6 3 2 2 4 4 8 3 2 6 4 4 0 8 0 4 8 9 6 5 7 1 2 6 5 2 8 7 3 4 4 8 1 6 0
C
o
n
t
a
m
i
n
a
n
t
s
i
n
p
r
o
p
y
l
e
n
e
,
p
p
m
Ti me, 12-mi nut e i nt erval s
2 0 0 0
6 0 0 0
5 0 0 0
4 0 0 0
3 0 0 0
1 5 8 1 1 1 6 2 1 7 4 3 2 3 2 4 2 9 0 5 3 4 8 6 4 0 6 7 4 6 4 8 5 2 2 9
C
o
n
t
a
m
i
n
a
n
t
s
i
n
p
r
o
p
y
l
e
n
e
,
p
p
m
Ti me, 12-mi nut e i nt erval s
2 0 0 0
Figure 7 Contaminants in propylene product of C
3
3 0
2 5
2 0
1 5
1 8 1 6 1 6 3 2 2 4 4 8 3 2 6 4 4 0 8 0 4 8 9 6 5 7 1 2 6 5 2 8 7 3 4 4 8 1 6 0
F
e
e
d
,
t
o
n
s
/
h
r
Ti me, 12-mi nut e i nt erval s
1 0
3 0
2 5
2 0
1 5
1 5 5 9 1 1 1 8 1 6 7 7 2 2 3 6 2 7 9 5 3 3 5 4 3 9 1 3 4 4 7 2 5 0 3 1 5 5 9 0
F
e
e
d
,
t
o
n
s
/
h
r
Ti me, 12-mi nut e i nt erval s
1 0
Figure 6 C
3
splitter 2 feed
braskem.indd 6 08/03/2013 16:54
was changed as slowly as possible,
thereby minimising the impact of
refrigeration capacity. At the start
of the project, several of the opera-
tors believed that this level was
impossible to control automatically
120 PTQ Q2 2013 www.eptq.com
Average Standard deviation
Variable Before Before Before After
OVHD contaminants, % 0.350 0.408 0.067 0.045
Bottoms propylene, % 15.3 6.46 11.80 4.28
Delta Press T06, kgf/cm
2
0.308 0.313 0.018 0.005
Delta Press T07, kgf/cm
2
0.732 0.723 0.019 0.016
Olens 2 C
3
splitter results
Table 6
Column Variable Before After
C
3
splitter 1 Propylene in bottoms ow, % 6.6 5.5
Overhead contaminants, ppm 570 200
Delta pressure, kgf/cm
2
0.11 0.04
Feed, ton/h 1.86 1.03
C
3
splitter 2 Feed C
3
splitter 2, ton/h 2.16 0.73
C
3
splitter standard deviations before and after results
Table 4
Variable Before After
Overhead contaminants, ppm 255 76
Ethylene in bottoms product, ppm 1653 652
C
2
splitter standard deviations before and after results
Table 5
as everything affects it. The addi-
tion of the reux drum capacity
allowed the ow to storage to be
automatically controlled during all
but the most severe upsets. It is
very popular with the operators.
Recommended best practices
The DCS loops associated with the
DMCplus controllers were tuned
for disturbance rejection. For cases
where this tuning resulted in signif-
icant overshoot for setpoint
changes, the available DCS option
of setpoint ramping was used to
keep overshoot below 20%. For
slower responding loops and levels,
a tuning package
1
was used.
Models for the tuner were devel-
oped either using DMCplus or by
tting data to simple transfer func-
tions (for example, dead-time plus
ramp). The benets of using a
tuning package included:
Ability to tune new loops with
condence (important for operator
acceptance)
Ability to achieve critically
damped responses for level loops
which were retained (ie, not
opened for control in DMCplus);
for example, the four cold box
separator levels of the demetha-
niser feeds.
The loss of downstream custom-
ers creates a signicant disturbance
to the plant and the refrigeration
systems in particular. Many loops
within the refrigeration system
(compressor SIC, suction drum
PICs and LICs of the various users)
were retuned. It was important to
verify tuning (and retune as neces-
sary) during the hottest months of
the year, when the system was
most loaded.
Steady-state simulations of
high-purity distillation columns
such as C
2
and C
3
splitters are quick
to build and calibrate to observed
data. Models of the four high-purity
columns were constructed using the
AspenTech AspenPlus steady-state
simulator. These models were used
primarily to conrm the experimen-
tal steady-state gains for the
controller response models. High-
purity distillation processes such as
C
3
splitters have a long settling time
that can make the steady-state gains
difcult to estimate experimentally.
Knowledge of the theoretical
process gain is an advantage, espe-
cially for feed forward independent
variables, such as feed composition,
that are not easy to move. Figure 4
shows the estimated effect of feed
propane composition on the
Independent variable Overheads propane in propylene
Theoretical gain Experimental gain
Reux, t/h 24.5 ppm 21 ppm
Feed, t/h 555 ppm 496 ppm
Feed propane, mol% 1750 ppm 1400 ppm
Bottoms ow, t/hr* -4670 ppm -1572 ppm
* The error in the bottoms ow is likely due to truncation of the process response at the controller settling
time interval of 20 hours. The actual response is about 48 hours.
Simulation vs experimental gains
Table 2
Variable/transform type Non-linearity
Various controller OPs/CV OP vs ow non-linearity
Demethaniser methane/CV Gain increase with higher impurity
Demethaniser chiller level/MV Dimensioning effect of chiller level on reux ow with
increasing chiller level*
C
2
splitter product ethane/CV Gain increase with higher impurity
C
2
splitter bottoms ethylene/CV Gain increase with higher impurity
Controller transforms
Table 3
braskem.indd 7 11/03/2013 13:05
www.eptq.com PTQ Q2 2013 121
overheads and bottoms composition
of the plant 2 C
3
splitter. Table 2
compares theoretical to experimen-
tal process gains.
The steady-state models were
also used to estimate several DCS
constants used in the regulatory
control system, such as the effect of
reux subcooling changes on calcu-
lated internal reux and pressure
compensated temperatures.
Transforms were used in several
situations where signicant non-
linearities were observed. A key
reason for including the transforms
was to provide stable, non-
oscillatory responses for situations
where current operation might be
far away from the steady-state
targets, such as when a DMC is
tuned on, and to provide for the
possibility of future operation at
signicantly different steady-state
targets. All transforms used on the
project were based on piece-wise
linear transforms on either MVs or
CVs based on the data. CV trans-
forms were developed based on
plots of predicted versus actual,
and two iterations (typically) were
used to develop the transform.
Table 3 lists the key transforms
developed on the project.
Results
In the following discussion,
selected trends and statistics are
presented to illustrate the improve-
ments obtained with the revamped
controllers compared to previous
operation. The data include both
normal, including constrained situ-
ations, and upset situations; no
data have been excluded. Most
plots show approximately 68 days
of operation prior to the project
and 47 days afterwards.
Prior to the project, Olens 1 C
3
splitters operators would manipu-
late C
3
splitters feeds by changing
PID setpoints. Including these feeds
as manipulated variables in the
APC controller, the parallel
columns better absorbed feed and
severity variations in furnaces. The
number of interventions per day in
the feed was reduced from 39 to
practically zero. Figures 5 and 6
show the feed of both splitters
before and after APC remodelling.
C
3
splitter 1 is a conventional
4 0
3 0
2 0
1 5 8 1 1 1 6 2 1 7 4 3 2 3 2 4 2 9 0 5 3 4 8 6 4 0 6 7 4 6 4 8 5 2 2 9
B
o
t
t
o
m
p
r
o
p
y
l
e
n
e
,
%
m
a
s
s
Ti me, 12-mi nut e i nt erval s
1 0
0
4 0
5 0
3 0
2 0
1 0
1 8 1 6 1 6 3 2 2 4 4 8 3 2 6 4 4 0 8 0 4 8 9 6 5 7 1 2 6 5 2 8 7 3 4 4 8 1 6 0
B
o
t
t
o
m
p
r
o
p
y
l
e
n
e
,
%
m
a
s
s
Ti me, 12-mi nut e i nt erval s
0
1 5 8 1 1 1 6 2 1 7 4 3 2 3 2 4 2 9 0 5 3 4 8 6 4 0 6 7 4 6 4 8 5 2 2 9 5 8 1 0
Ti me, 12-mi nut e i nt erval s
1 . 4
1 . 5
1 . 2
1 . 3
1 . 0
1 . 1
0 . 9
p
r
e
s
s
u
r
e
,
k
g
f
/
c
m
2
0 . 8
1 . 4
1 . 5
1 . 2
1 . 3
1 . 0
1 . 1
0 . 9
1 8 1 6 1 6 3 2 2 4 4 8 3 2 6 4 4 0 8 0 4 8 9 6 5 7 1 2 6 5 2 8 7 3 4 4 8 1 6 0
p
r
e
s
s
u
r
e
,
k
g
f
/
c
m
2
Ti me, 12-mi nut e i nt erval s
0 . 8
Figure 8 Propylene loss in C
3
splitter
Figure 9 Delta pressure in C
3
splitter
braskem.indd 8 11/03/2013 13:05
122 PTQ Q2 2013 www.eptq.com
shown in Figure 7. The most
important benet in C
3
splitter 1
was the reduction in propylene
loss in the column bottoms. The
reduction in the average propylene
losses was 5.2% compared to the
baseline average (see Figure 8).
A signicant part of the variability
in the propylene loss in the
bottoms is due to condenser
limitations.
Flooding was an issue for C
3
split-
ter 1 in Olens 1. Modications in
regulatory control such as includ-
ing internal reux control reduced
the ooding, which can be seen in
the delta pressure (see Figure 9).
Standard deviations are shown in
Table 4.
As discussed previously, imple-
mentation of internal reux was
very important for APC success in
the C
2
splitter. This approach stabi-
lised the overhead and bottoms
compositions, and the result with
the MPC was minor variability in
the product ethylene contaminants
and in the ethylene in bottoms
ethane. These results are shown in
Figures 10 and 11. Standard devia-
tions are shown in Table 5.
The new C
2
splitter controller
signicantly reduced the number of
ow set point changes of ethylene
to storage tank. The ethylene vessel
product level was operating in
manual mode, and the number of
interventions in ow to the tank
per day was around 45. After the
project, this number is around 4.8.
This result means that the operators
now have more time available to
improve the process.
The ratio between ethylene ow
to storage and C
2
splitter feed ow
reduced by 26%. Sending ethylene
to the storage tank is much less
economic than sending it directly
to the product header. Sending
ethylene to tank requires signi-
cant refrigeration duty. The C
2
splitter MPC helped to maximise
ow to the customer header by
better control of the product vessel
level. In addition, the overlling of
the product vessel is avoided.
Figure 12 shows the behaviour
feed ow-to-tank ow ratio
before and after APC remodelling.
Figure 13 shows the product
level.
of the contaminants in propylene
product. The regulatory control
changes along with APC resulted
in the variability reduction, as
column and C
3
splitter 2 is a heat
pumped column.
Another improvement in C
3
split-
ter 1 was the decrease in variability
2 0 0 0
1 5 0 0
1 0 0 0
5 0 0
1 5 3 3 1 0 6 6 1 5 9 9 2 1 3 2 2 6 6 5 3 1 9 8 3 7 3 1 4 2 6 4 4 7 9 7 5 3 3 0
E
t
h
y
l
e
n
e
c
o
n
t
a
m
i
n
a
n
t
s
,
%
m
a
s
s
Ti me, 12-mi nut e i nt erval s
0
2 0 0 0
1 5 0 0
1 0 0 0
5 0 0
1 9 6 0 1 9 2 0 2 8 8 0 3 8 4 0 4 8 0 0 5 7 6 0 6 7 2 0 7 6 8 0 8 6 4 0 9 6 0 0
E
t
h
y
l
e
n
e
c
o
n
t
a
m
i
n
a
n
t
s
,
%
m
a
s
s
Ti me, 12-mi nut e i nt erval s
0
8 0 0 0
6 0 0 0
4 0 0 0
2 0 0 0
7 0 0 0
5 0 0 0
3 0 0 0
1 0 0 0
1 5 3 3 1 0 6 6 1 5 9 9 2 1 3 2 2 6 6 5 3 1 9 8 3 7 3 1 4 2 6 4 4 7 9 7 5 3 3 0
B
o
t
t
o
m
e
t
h
a
n
e
,
p
p
m
Ti me, 12-mi nut e i nt erval s
0
8 0 0 0
6 0 0 0
4 0 0 0
2 0 0 0
7 0 0 0
5 0 0 0
3 0 0 0
1 0 0 0
1 9 6 0 1 9 2 0 2 8 8 0 3 8 4 0 4 8 0 0 5 7 6 0 6 7 2 0 7 6 8 0 8 6 4 0 9 6 0 0
B
o
t
t
o
m
e
t
h
a
n
e
,
p
p
m
Ti me, 12-mi nut e i nt erval s
0
Figure 11 C
2
splitter bottoms ethylene
Figure 10 C
2
splitter overhead contaminants
braskem.indd 9 11/03/2013 13:05
www.eptq.com PTQ Q2 2013 123
Olens 2 C
3
splitter results Due to a
plant routine shutdown, runtime
on the Olens 2 C
3
splitter has been
limited. However, the rst month
of operation showed encouraging
results (see Table 6).
Olens 2 furnace results Table 7
shows the results in one furnace
after remodelling and the addition
of new bottom dampers as manipu-
lated variables to the furnace
controller. The better air control
allowed the minimisation in excess
air and consequently savings in fuel
gas. Similar results have been
obtained for the other six furnaces.
Next steps
There are further opportunities in
both olens plants to improve
process results through improved
process control. A challenge for the
Braskem control team is to main-
tain the revised controllers along
with implementing new projects.
Future work includes integrating
the cold side controllers and imple-
menting furnace severity control.
Braskem is also preparing for the
training of all operators, inde-
pendently of the existence of
projects, as new operators in the
control panel are a frequent reality.
AspenTech DMCplus is a mark of AspenTech.
This article was originally prepared for
presentation at the 2012 AIChE Spring National
Meeting in Houston, TX, 1-5 April 2012 (AIChE
paper 50c; www.aiche.org).
References
1 Dittmar R, Shabroz G, Harpreet S, Darby M,
Robust optimization based multi-loop PID
controller tuning: a new tool and its industrial
application, Control Engineering Practice, April
2012, 356-370.
Vanessa Conz is APC Lead Control Engineer at
Braskem. Email: vanessa.conz@braskem.com
Suzana Fuchs is APC Controller Engineer at
Braskem.
Priscila Macedo Moura is Production Engineer,
Braskem.
Mark Darby is Consultant at CMiD Solutions.
Email: darbymark@sbcglobal.net
Doug Nicholson is Consultant at IPCOS (UK)
Ltd.
0 . 4
0 . 5
0 . 3
0 . 2
0 . 1
1 5 3 3 1 0 6 6 1 5 9 9 2 1 3 2 2 6 6 5 3 1 9 8 3 7 3 1 4 2 6 4 4 7 9 7 5 3 3 0
Ti me, 12-mi nut e i nt erval s
0
C
2
S
t
a
n
k
f
l
o
w
F
e
e
d
f
l
o
w
0 . 4
0 . 5
0 . 3
0 . 2
0 . 1
1 9 6 0 1 9 2 0 2 8 8 0 3 8 4 0 4 8 0 0 5 7 6 0 6 7 2 0 7 6 8 0 8 6 4 0 9 6 0 0
Ti me, 12-mi nut e i nt erval s
0
C
2
S
t
a
n
k
f
l
o
w
F
e
e
d
f
l
o
w
8 0
9 0
6 0
4 0
2 0
7 0
5 0
3 0
1 0
1 5 3 3 1 0 6 6 1 5 9 9 2 1 3 2 2 6 6 5 3 1 9 8 3 7 3 1 4 2 6 4 4 7 9 7 5 3 3 0
Ti me, 12-mi nut e i nt erval s
0
L
e
v
e
l
,
%
8 0
9 0
6 0
4 0
2 0
7 0
5 0
3 0
1 0
1 9 6 0 1 9 2 0 2 8 8 0 3 8 4 0 4 8 0 0 5 7 6 0 6 7 2 0 7 6 8 0 8 6 4 0 9 6 0 0
Ti me, 12-mi nut e i nt erval s
0
L
e
v
e
l
,
%
Figure 13 Ethylene product vessel level
Figure 12 Ethylene ow to consumers and storage ratio
CO
2
reduction 216 ton/month
Fuel gas reduction 29 ton/month
Cracking furnace results
Table 7
braskem.indd 10 08/03/2013 16:54
Untitled-7 1 12/03/2013 11:19
Energy-effcient vacuum systems
V
acuum technology is used in
petroleum refneries to facili-
tate the distillation of heavy
ends at low temperatures, to prevent
coking and degradation of products
and for other applications.
Currently, steam jet ejectors and
steam ejector-liquid ring vacuum
pump (LRVP) combinations are the
most common methods for vacuum
generation in petroleum refneries.
Although steam jet ejectors are very
reliable, they are highly ineffcient.
Due to increasing energy costs and
environmental concerns, it is essen-
tial to reduce the energy required
for vacuum generation. Petroleum
refneries discard a lot of waste heat
to the environment, which could be
used to reduce energy consumption
for vacuum generation. There are
publications illustrating the merits
and limitations of vacuum genera-
tion methods
1,3,4,6
and chilled/
refrigerated water generation.
2
However, there is a need for an inte-
grated approach covering all aspects
of vacuum generation and its energy
reduction possibilities.
The main objective of this study
is to analyse various methods for
developing energy-effcient vacuum
generation systems in petroleum
refneries. Three case studies are
presented to enhance understand-
ing in the selection process.
Vacuum generation in refneries
Steam ejectors and LRVPs are gener-
ally used in petroleum refneries. A
review of various vacuum genera-
tion equipment capacities, operating
ranges and effciencies is available.
6
Steam ejectors may have one or
more stages in series or a series-
parallel combination, with pre- and/
Case analysis of the techniques available to reduce energy consumption in
vacuum systems reveals the potential for cost savings
C Chandra Sekhara reddy and S V naIdU Andhra University
G P ranGaIah National University of Singapore
or interstage condensers, depending
on the level of vacuum required and
the utility optimisation and opera-
tional fexibility sought for various
plant loads. Steam ejectors are
highly reliable, and the availability
of steam in petroleum refneries
makes ejectors the natural choice.
However, they are highly ineffcient
6
(<10%), mainly due to a lack of
moving parts to convert fuid veloc-
ity to pressure effciently.
4
LRVP most commonly uses water
as a seal liquid since it can be sepa-
rated and reused safely. They
are generally more expensive
compared to steam ejectors.
However, they do not require large
heat exchangers to condense the
vapour at their outlet, and the
operating costs of LRVPs are gener-
ally lower than steam jet ejectors.
For better operating cost savings, a
steam ejectorLRVP combination is
sometimes used to replace the last
one or two stages of a multistage
steam ejector system.
design principles and utility
requirements
This section presents useful design
principles and tools for estimating
the utility requirement for steam jet
ejectors and LRVP. Use of pre- and
interstage condensers can reduce
both capital and operating costs for
the vacuum system. The vacuum
produced is limited by the temper-
ature of the cooling water; the
colder the temperature of the cool-
ing water, the lower the vacuum
produced.
Steam requirement for ejectors
can be estimated based on the dry
air equivalent (DAE) of suction
gases (including air, water vapour
and other gases). As per the HEI
(Heat Exchange Institute) proce-
dure for calculating DAE
8
, water
vapour in the suction gases is
considered separately and all other
gases (including air) are treated as
a mixture, in accordance with this
mixtures molecular weight. HEI
has published curves to convert
suction gas streams to DAE using
molecular weight and temperature
entrainment ratios. Molecular
weight entrainment ratio (MW
c
) is
defned as the ratio of the weight of
suction gas to the equivalent
weight of air. Temperature entrain-
ment ratio is defned as the ratio of
the weight of air (or water vapour)
at actual suction temperature to the
weight of air (or water vapour) at
21.1C.
The following equations are
derived from HEI curves
8
for
temperature entrainment ratios
(TC
a
and TC
w
) and MW
c
. These are
convenient for use in computer
programs:
TC
a
= -4 * 10
-10
T
3
+ 3 * 10
-7
T
2
- 0.0005 T +
1.0131 (1)
TC
w
= -1 * 10
-13
T
4
- 7 * 10
-12
T
3
+ 8 * 10
-8
T
2
-
0.0006 T + 1.015 (2)
For M = 0 to 60, MW
c
= 1 * 10
-5
M
3
-
0.00013 M
2
+ 0.0642 M + 0.016 (3.1)
For M = 60 to 150, MW
c
= -2 * 10
-5
M
2
+
0.0077 M + 0.9464 (3.2)
Water vapour and other compo-
nents in the suction gas can be
converted to DAE using the correc-
tion factors from equations 1, 2, 3.1,
3.2 and the following equation:
www.eptq.com PTQ Q2 2013 125
reddy anhdra.indd 1 08/03/2013 17:21
126 PTQ Q2 2013 www.eptq.com
U = 5.678 (220.0417 + 1.6919 ln(NC) -
2.67975 [ln(NC)]
2
- 1.5465 [ln(NC)]
3
) (6)
For a gas vapour mixture with
NC from 50% to 95%:
U = 5.678 (-245 896 + 233 845.3 ln(NC) -
83 300.5 [ln(NC)]
2
+ 13 183.62 [ln(NC)]
3
-
782.58 [ln(NC)]
4
) (7)
The suction pressure of LRVPs is
dictated by the temperature of seal
water. With normal cooling water
temperatures of ~30C, LRVPs are
used to replace steam ejectors oper-
ating at suction pressures >150 torr
(usually the last one or two stages
DAE of suction W
OG
+
W
w
gas or vapour
=
TC
a
* MW
COG
TC
W
* MW
CWV
(4)
The amount of motive steam
required to compress (from suction
pressure to discharge pressure) the
unit DAE mass of suction gas/
vapour in a steam ejector is dened
as R
a
(kg of motive steam/kg of
DAE equivalent of load gas).
Values of R
a
are available
4
as
curves, with suction pressure on
the abscissa and discharge pressure
on the ordinate. For rough estima-
tion of an ejectors steam
consumption, one can simply use
the following equation:
9
R
a
=
>
Pd
(0.434 -
1.338
+ 0.000475 Pa) - 0.187H
Pa Pa
(1.2 -
P
v
- 10.2)
20 (5)
Steam requirement for ejectors
can be estimated by multiplying
DAE and R
a
values.
For condenser calculations,
involving air and water mixtures,
the overall heat transfer coefcient,
U W/(m
2
K), can be estimated using
the following developed equations.
4
For a gas vapour mixture with
non-condensable vapour mole
percentage, NC, from 1% to 50%:
E va lu a te th e typ e o f g a se s
to b e e va c u a te d
M o stly n o n -c o n d e n sa b le
g a se s?
P re -c o n d e n se r u sin g
c o o lin g wa te r
I s c h e a p so u rc e o f m e d iu m
p re ssu re ste a m a va ila b le ?
I n sta ll m u ltista g e ste a m e je c to r
syste m with o p tim u m d e sig n
I n sta ll p re -c o n d e n se r u sin g c h ille d
wa te r ( g e n e ra te d b y m e c h a n ic a l o r
sin g le -sta g e a b so rp tio n c h ille r
I n sta ll two -sta g e ste a m e je c to r
a n d L R VP c o m b in a tio n
N o Ye s
Ye s N o
Ye s N o
Ye s N o
I s c o n d e n sin g te m p e ra tu re
o f va p o u r g re a te r th a n
c o o lin g wa te r re tu rn
te m p e ra tu re ?
I s c h e a p e le c tric p o we r o r
wa ste L L P ste a m a va ila b le ?
Figure 1 Design strategy for a new vacuum system
reddy anhdra.indd 2 08/03/2013 17:22
of ejectors for vacuum distillation
columns). Power required by a
LRVP can be estimated using equa-
tion 8:
3
P =
Pa * V
s
* ln Pd
Pa
27 000 *
e
(8)
Chilled water generation
Chilled water can be generated
economically using absorption heat
pumps and mechanical refrigera-
tion. Absorption heat pumps use
waste heat (such as low-pressure
steam or a hot process stream)
rather than mechanical/shaft
energy for operation. Lithium
bromide absorption pumps are
frequently used due to their lower
cost and application range up to
the freezing point of water.
Compared to mechanical chillers,
absorption chillers have a low
co-effcient of performance (COP).
Nonetheless, their operating costs
can be substantially lower because
they use waste heat, while vapour
compression chillers must be motor
driven. At lower electricity prices, a
mechanical chiller can be attractive
for chilled water generation.
7
www.eptq.com PTQ Q2 2013 127
Optimisation of vacuum system
operating cost
A design strategy and procedure
for minimising the operating cost of
a multistage steam jet ejector
system is discussed in this section
and illustrated with two case stud-
ies. There are several ways to
reduce the operating costs of a
vacuum system. Engineers often
fnd it diffcult to take optimisation
decisions as most of the informa-
tion is vendor specifc. The present
work illustrates the use of simple
techniques to optimise vacuum
system operating costs easily and
quickly. A design strategy for opti-
mising a new vacuum system is
presented in Figure 1. Process
simulators such as Aspen Hysys
can be used to estimate a condens-
ing temperature at which the
majority (~90%) of vapour
condenses, at various stages of the
vacuum system.
Once a vacuum system design is
selected, optimum discharge pres-
sure and steam consumption for
each stage of the multistage steam
ejector system (see Figure 2) can be
obtained by solving the following
optimisation problem; for example,
using the Solver tool in Microsoft
Excel. Defnitions of parameters and
variables in this optimisation prob-
lem are given in the nomenclature.
The following objective function
is to minimise overall steam
consumption for the multistage
steam ejector system:
Wv =
n
i
=1
Wv
i
=
n
i
=1
Pd
i
(0.434 -
1.338
+
Pa
i
Pa
i
0.000475 Pa
i
) - 0.187] (1.2 -
P
v
-10.2
)
20
(W
a
+
W
wi
) (9.1)
MW
CWV
The quantities in the above equa-
tion are as follows.
Suction pressure for (i+1)th stage:
Pa
i+1
= Pd
i
- Dp
i
(9.2)
Water vapour fow rate to the
inlet of (i+1)th stage:
Ww
i+1
= Ww
i
- 18 * Pv
i
29 * (Pa
i+1
- Pv
i
) (9.3)
Saturation pressure of water
vapour corresponding to the vent
temperature of the ith stage
condenser is a function of the ith
stage condenser vent temperature:
P ro c e ss
g a s/va p o u r
F irst sta g e in le t
g a s/va p o u r
S te a m in je c to r S te a m in je c to r S te a m in je c to r
Wa te r-o il
se p a ra tio n
ta n k
Wa te r-o il
se p a ra tio n
ta n k
M P ste a m
C o o lin g wa te r
su p p ly
S lo p o il p u m p
S o u r wa te r p u m p
F irst sta g e
c o n d e n se r
S e c o n d sta g e
c o n d e n se r
n th sta g e
c o n d e n se r o r
a fte r c o n d e n se r
WV
1
P
v
WV
2
P
v
P a
1
P a
2
P a
n
T v
1
P d
1
T vo
1
T v
2
P d
2
T vo
2
T v
n
P d
n
T vo
n
A
1
W
C W1
A
2
W
C W2
A
n
W
C Wn
Ww
1
+W
a
Ww
2
+W
a
Ww
n
+W
a
To fire d
h e a te r
C o o lin g
wa te r
re tu rn
WV
n
P
v
Figure 2 Schematic of a multi-effect steam ejector system
reddy anhdra.indd 3 08/03/2013 17:22
128 PTQ Q2 2013 www.eptq.com
of the optimisation procedure in
Equation 9.
Load/suction gas: 45.36 kg/h of
air + 97.52 kg/h of water vapour at
suction pressure of 15 torr and
suction temperature of 21.1C;
discharge pressure: 815 torr; motive
steam: 10.6 barg.
Cooling water supply and return
temperatures are 32.2C and 37.8C.
Solutions to the optimisation prob-
lem, obtained using Excels Solver,
for a two-, three- and four-stage
ejector system with different
approach temperatures are shown
in Table 1. The vent pressures tally
well with those given in the refer-
ence,
4
with error in the range +14%
to -8%. For three- and four-stage
steam ejector systems, steam fow
rates estimated by the above proce-
dure are greater by up to 25%,
whereas, for a two-stage steam ejec-
tor system, the estimated steam
fow rate was lower by ~20%
compared to the reported values.
4
The main reason is the difference in
Ra values from the reference
graphs
4
and Equation 9.1, used in
the optimisation procedure. For
accurate estimation of steam fow
rates, one can use the optimised
vent pressures and the reference
graphs.
4
Cooling water consump-
tion shown in Table 1 can be
reduced further with some increase
in condenser area by using a series
fow arrangement, where water
from the frst interstage condenser
fows through the other interstage
and after-condensers.
For relatively higher suction pres-
sures (~40 torr), use of a
pre-condenser, with chilled water
A
i
=
L
i
* (Ww
i
+ Wv
i
- Ww
i+1
) * 1000 (9.11)
U
i
* LMTD
i
* 3600
Cooling water required for the ith
stage condenser:
Wcw
i
=
L
i
* (Ww
i
+ Wv
i
- Ww
i+1
) * 0.239 (9.12)
(T
wc
- T
wco
)
Decision variables: Pd
i
for i=1, 2,...
n-1
Bounds and constraints:
Pa
1
< Pd
i
< Pd
n
and Wv
i
> 0
Steam properties in equations 9.4
and 9.9 can be calculated using the
Excel spreadsheet, freely available
at www.x-eng.com.
The above optimisation problem
(Equation 9) is applicable to
systems involving air and water
only. It ignores vapour superheat at
the ejector inlet, and vapour
sub-cooling and liquid sub-cooling
in condenser area calculations.
Also, it assumes simple LMTD
without any correction factor.
Hence, condenser area calculations
are approximate. However, the
optimisation problem in Equation 9
is very useful to arrive at a prelimi-
nary design concept and/or to
verify vendors proposals. For large
systems involving other vapours,
the optimisation problem can be
solved by including simulation data
from process simulators such as
Aspen Hysys, Aspen Plus and Pro/
II.
The following case study (case
study 1), based on the details from
the manual optimisation of a
vacuum system presented by
Power,
4
illustrates the effectiveness
Pv
i
= f(Tvo
i
) (9.4)
Vent temperature for the ith stage
condenser:
Tvo
i
= T
wco
+ TAP
i
(9.5)
Mole fraction of water vapour at
the ith stage condenser inlet:
nw
i
=
Ww
i
+ Wv
i
/ [
Wa
+
Ww
i
+ Wv
i
]
(9.6)
18 29 18
Mole percent of non-condensable
gases at the ith stage condenser:
NC
i
= (1 - nw
i
) * 100 (9.7)
Partial pressure of water vapour
at the ith stage condenser inlet:
Ppw
i
= Pd
i
* nw
i
(9.8)
Saturation temperature of water
vapour corresponding to its partial
pressure at the ith stage condenser
inlet is a function of its partial pres-
sure at the ith stage condenser inlet:
Tv
i
= f(Ppw
i
) (9.9)
Log mean temperature difference
for the ith stage condenser:
LMTD
i
=
(Tv
i
- T
wco
) - (Tvo
i
- T
wc
)
ln
(Tv
i
- T
wco
)
(Tvo
i
- T
wc
) (9.10)
Overall heat transfer coeffcient
for the ith stage condenser, U, is
given by equations 6 and 7.
Required area for the ith stage
condenser:
Approach Valuesfromoptimisation
temperatureUtilities Inter-condensers Total
Number at1st Cooling Vent Condenser heat
Inter- condenser, MPsteam, water, temperature, inletpressure, Inter-andafter- transfer
Case Stages condensers C kg/h kg/h C mmHg condenserarea,m
2
area,m
2
1A 2 1 11.1 878 97 254 43.3 134 23.78/5.27 29.0
1B 3 2 16.7 840 93 300 48.9/54.4 130/284.2 18.1/4.54/2.05 24.7
1C 3 2 11.1 726 82 261 43.3/48.9 106.5/253.6 21.64/4.96/2.31 28.9
10 3 2 5.6 647 74 407 37.8/48.9 88.9/254.6 31.24/4.61/2.3 38.2
1E 3 2 2.8 598 69 558 35/43.3 79.5/229.8 44.84/5.5/2.28 52.6
1F 4 3 5.6 584 68 423 37.8/40.6/48.9 78.4/138.6/263 31.2/7.3/2.5/1.5 42.5
1G 4 3 2.8 557 65 680 35/43.3/48.9 71.8/148/259.3 47.4/5.74/2.6/1.5 57.2
1H 4 3 1.7 539 63 948 33.9/41.7/48.9 68.4/141.3/255 63.2/6.25/2.6/1.5 73.5
Optimisationresultsforcasestudy1:avacuumsysteminvolvingairandwatervapourmixture
Table1
reddy anhdra.indd 4 08/03/2013 17:22
www.eptq.com PTQ Q2 2013 129
VD U o ve rh e a d
va p o u r
1 0 0 7 0 k g /h r
3 5 . 2 to rr
8 0 C
S te a m in je c to r
1 0 0 % c a p a c ity
Wa te r-o il
se p a ra tio n
ta n k
Wa te r-o il
se p a ra tio n
ta n k
4 o ff
4 8 2 m
2
1 2 6 m
2
M P ste a m
4 5 2 5 k g /h r
1 0 . 5 b a rg
2 3 0 C
C o o lin g wa te r
su p p ly
1 0 9 8 8 3 3 k g /h r
5 b a rg
3 2 C
1 4 5 9 6 k g /h r
7 5 to rr
3 8 C
S lo p o il p u m p
S o u r wa te r p u m p
F irst sta g e
c o n d e n se r
S e c o n d sta g e
c o n d e n se r
7 5 4 k g /h r
1 0 . 5 b a rg
2 3 0 C
S te a m in je c to rs
( x 2 )
5 0 % c a p a c ity
7 8 4 . 4 k g /h r
4 7 2 . 3 to rr
4 7 C
1 5 3 8 k g /h r
9 1 5 to rr
4 1 C
1 7 0 0 6 6 k g /h r
3 8 C 4 5 . 6 m
2
F o u rth sta g e
c o n d e n se r
1 4 0 0 k g /h r
1 0 . 5 b a rg
2 3 0 C
S te a m in je c to rs
( x 2 )
5 0 % c a p a c ity
1 3 1 4 k g /h r
8 9 to rr
3 8 C
2 7 1 4 k g /h r
1 9 1 . 3 to rr
4 1 C
3 7 4 3 3 3 k g /h r
3 8 C
1 1 6 0 k g /h r
1 0 . 5 b a rg
2 3 0 C
S te a m in je c to rs
( x 2 )
5 0 % c a p a c ity
9 5 3 . 3 k g /h r
1 7 6 . 3 to rr
4 4 C
2 1 1 3 k g /h r
4 9 1 . 3 to rr
4 1 C
2 8 3 4 0 0 k g /h r
3 8 C 4 6 . 4 m
2
T h ird sta g e
c o n d e n se r
7 4 7 . 8 k g /h r
8 9 4 . 8 to rr
5 0 C
To fire d
h e a te r
C o o lin g
wa te r
re tu rn
Figure 3 Existing vacuum system for a vacuum distillation column (base case case 3A)
C h ille r
VD U o ve rh e a d
va p o u r
1 0 0 7 0 k g /h r
3 5 . 2 to rr
8 0 C
S te a m in je c to rs
( x 2 )
5 0 % c a p a c ity
P re -c o n d e n se r
C h ille d wa te r
re tu rn
1 3 C
Wa te r-o il
se p a ra tio n
ta n k
Wa te r-o il
se p a ra tio n
ta n k
7 1 3 3 3 k g /h r
2 9 2 . 6 m
2
1 2 0 m
2
1 0 0 8 k g /h r
2 4 . 7 5 to rr
1 4 C
1 3 C
M P ste a m
7 9 2 k g /h r
1 0 . 5 b a rg
2 3 0 C
C o o lin g wa te r
su p p ly
5 b a rg
3 2 C
C h ille d wa te r
su p p ly
6 4 2 8 3 3 k g /h r
5 b a rg
7 C
C o o lin g wa te r
re tu rn
1 8 0 0 k g /h r
7 5 to rr
3 8 C
1 4 0 0 m
2
S lo p o il p u m p
S o u r wa te r p u m p
F irst sta g e
c o n d e n se r
S e c o n d sta g e
c o n d e n se r
7 0 5 k g /h r
1 0 . 5 b a rg
2 3 0 C
S te a m in je c to rs
( x 2 )
5 0 % c a p a c ity
7 6 9 k g /h r
4 7 2 . 3 to rr
4 4 C
1 4 7 4 k g /h r
9 1 5 to rr
4 1 C
1 6 0 3 0 0 k g /h r
3 8 C 4 0 m
2
F o u rth sta g e
c o n d e n se r
1 0 9 0 k g /h r
1 0 . 5 b a rg
2 3 0 C
S te a m in je c to rs
( x 2 )
5 0 % c a p a c ity
1 2 7 2 k g /h r
6 9 to rr
3 6 C
2 3 6 2 k g /h r
1 9 1 . 3 to rr
4 1 C
2 9 0 9 0 0 k g /h r
3 8 C
1 0 4 8 k g /h r
1 0 . 5 b a rg
2 3 0 C
S te a m in je c to rs
( x 2 )
5 0 % c a p a c ity
9 4 5 k g /h r
1 7 6 . 3 to rr
4 4 C
1 9 9 3 k g /h r
4 9 5 to rr
4 1 C
2 6 6 1 0 0 k g /h r
3 8 C 4 0 m
2
T h ird sta g e
c o n d e n se r
7 4 8 k g /h r
8 9 4 . 8 to rr
5 0 C
To fire d
h e a te r
C o o lin g
wa te r
re tu rn
Figure 4 Four-stage steam ejector system with one pre-condenser (at the inlet of frst-stage steam ejector), cooled by chilled water
(Cases 3B and 3C)
reddy anhdra.indd 5 08/03/2013 17:22
130 PTQ Q2 2013 www.eptq.com
VD U o ve rh e a d
va p o u r
1 0 0 7 0 k g /h r
3 5 . 2 to rr
8 0 C
S te a m in je c to r
1 0 0 % c a p a c ity
Wa te r-o il
se p a ra tio n
ta n k
Wa te r-o il
se p a ra tio n
ta n k
4 o ff
4 8 2 m
2
1 2 5 m
2
M P ste a m
4 5 2 5 k g /h r
1 0 . 5 b a rg
2 3 0 C
C o o lin g wa te r
su p p ly
1 1 2 8 8 3 3 k g /h r
5 b a rg
3 2 C
1 4 5 9 5 k g /h r
7 5 to rr
3 8 C
S lo p o il p u m p
S o u r wa te r p u m p
F irst sta g e
c o n d e n se r
S e c o n d sta g e
c o n d e n se r
1 4 0 0 k g /h r
1 0 . 5 b a rg
2 3 0 C
S te a m in je c to rs
(x 2 )
5 0 % c a p a c ity
1 3 1 4 k g /h r
8 9 to rr
3 8 C
2 7 1 4 k g /h r
1 9 1 . 3 to rr
4 1 C
3 7 4 3 3 3 k g /h r
3 8 C
9 5 3 . 3 k g /h r
1 7 6 . 3 to rr
4 4 C
4 1 C
1 7 4 k W
3 0 0 0 0 k g /h r
3 2 C
S e a l wa te r
c o o le r
7 3 3 k g /h r
8 9 4 . 8 to rr
4 4 C
To fire d
h e a te r
C o o lin g
wa te r
re tu rn
M
Figure 5 Modifed vacuum system with steam ejectors for the frst two stages and last two stages replaced with one LRVP (Case 3D)
Figure 6 Improved vacuum system with a pre-condenser, frst two stages with steam ejector and last two stages replaced by one LRVP
(Cases 3E and 3F)
C h ille r
VD U o ve rh e a d
va p o u r
1 0 0 7 0 k g /h r
3 5 . 2 to rr
8 0 C
S te a m in je c to rs
(x 2 )
5 0 % c a p a c ity
S te a m in je c to rs
(x 2 )
5 0 % c a p a c ity
P re -c o n d e n se r
C h ille d wa te r
re tu rn
1 3 C
7 1 3 3 3 k g /h r
2 9 2 . 6 m
2
1 2 0 m
2
1 0 0 8 k g /h r
2 4 . 7 5 to rr
1 4 C
1 3 C
M P ste a m
7 9 2 k g /h r
1 0 . 5 b a rg
2 3 0 C
C o o lin g wa te r
su p p ly
C h ille d wa te r
su p p ly
6 4 2 8 3 3 k g /h r
5 b a rg
7 C
C o o lin g wa te r
re tu rn
1 8 0 0 k g /h r
7 5 to rr
3 8 C
1 4 0 0 m
2
F irst sta g e
c o n d e n se r
S e c o n d sta g e
c o n d e n se r
1 0 9 0 k g /h r
1 0 . 5 b a rg
2 3 0 C
1 2 7 2 k g /h r
6 9 to rr
3 6 C
2 3 6 2 k g /h r
1 9 1 . 3 to rr
4 1 C
2 9 0 9 0 0 k g /h r
3 8 C
Wa te r-o il
se p a ra tio n
ta n k
S lo p o il p u m p
S o u r wa te r p u m p
9 4 5 k g /h r
1 7 6 . 3 to rr
4 4 C
4 1 C
1 7 4 k W
3 0 0 0 0 k g /h r
3 2 C
S e a l wa te r
c o o le r
7 3 3 k g /h r
8 9 4 . 8 to rr
4 4 C
To fire d
h e a te r
C o o lin g
wa te r
re tu rn
M
2 6 6 1 0 0 k g /h r
3 8 C
reddy anhdra.indd 6 08/03/2013 17:22
Fertilizers
&Chemicals
Metals
&Mining Training
Sulphur
SulphurIC
ACID
26-28 J une 2013, Grand Hyatt Shanghai, China
Forging new relationships in China
China International
Sulphur & Sulphuric
Acid 2013
FErTIlIZEr | INDuSTrIAl | OIl | GAS | METAllurGy
Why choose to attend?
- Network with buyers and sellers focused on sulphur and
sulphuric acid at this established annual conference with
a track record of bringing the largest balanced audience
of international and Chinese delegates
- Maximize your time to connect with senior executives
from across fertilizer, industrial chemicals, oil, gas,
energy, pyrites and mining sectors
- learn of the changes in demand from fertilizer
and industrial chemicals sector affecting sulphur
and sulphuric acid sales
- hear from experts on the international market
dynamics for sulphur and sulphuric acid and its
impact on the region
- Get the latest news on new capacity build up and
timing of Chinese supply of sulphur and sulphuric
acid for local consumption and export
register today at www.chinasulphuracid.com
Sponsored by: Offcial publications: Media partners:
www.energyglobal.com
EnergyGlobal
ptq
PETROLEUM TECHNOLOGY QUARTERLY
2013 > Shanghai
Fertilizers
&Chemicals
Metals
&Mining Training
4-7 November 2013, InterContinental Miami, USA
Register today at www.sulphurconference.com
International Conference & Exhibition
Oil | Gas | Fertilizers | Metallurgy | Industrial uses
Sulphur 2013
2013 > Miami
Network with 500+
professionals from the
global sulphur, sulphuric
acid and related industries
- Hear about the latest
trends and market
analysis affecting your
sulphur and sulphuric
acid businesses
- Learn how to improve
your operations from
other operator experience
and technical experts
- Meet 60+ companies
exhibiting the latest
technology, equipment
and engineering services
Sponsors: Offcial publications: Media Partners:
www.energyglobal.com
EnergyGlobal
ptq
PETROLEUM TECHNOLOGY QUARTERLY
Book now
for early bird
SAVE $295!
50%
discount for
Operators
The worlds leading sulphur and
sulphuric acid annual event
china sulphur.indd 1 07/03/2013 13:34
132 PTQ Q2 2013 www.eptq.com
replaced by one LRVP. The modi-
fed vacuum system is shown in
Figure 5.
Cases 3E and 3F
For a pre-condenser using chilled
water generated by a single-stage
absorption chiller (Case 3E) and a
mechanical chiller (Case 3F), the
frst two stages are steam ejectors
and the last two stages are replaced
with one LRVP. This improved
system is shown in Figure 6.
Equipment costs
Capital costs of steam ejectors are
derived from an available reference
chart.
4
Costs for condensers are esti-
mated using the Capcost program
(based on an Excel spreadsheet).
10
Fixed costs for absorption, mechani-
cal chillers and LRVP are based on
vendor quotations. The installed
cost for a mechanical chiller and
LRVP were taken as twice the
purchase cost, whereas the installed
cost for an absorption (single-stage-
LiBr) chiller was taken as 1.5 times
the purchase cost since the absorp-
tion chiller is a packaged unit and
involves less expensive installation.
Costs for foul water treatment are
used from reference 5. Utility costs
assumed are: MP steam = $32.14/
ton; LLP steam = $31.65/ton; cool-
ing water = $0.05/ton; and electric
power = $0.15/KW. Payback period
is defned as:
Payback period =
Capital cost for new case -
capital cost for base case
Operating cost for base case -
(10) Operating cost for base case
The results for all the cases are
presented in Table 5. For the cases
where operating costs increased
compared to the base case, payback
reduction can be achieved by using
chilled water, if more water vapour
is present in the load gas.
Case study 3
Case study 3 is the vacuum system
optimisation of a vacuum distilla-
tion column in an existing Asian
refnery. Vacuum required at the
top of the column is 35 torr, the
vacuum system suction fow rate is
10 070 kg/hr and the discharge
pressure is 895 torr. Suction gases
contain 5641 kg/hr of water
vapour, 3730 kg/hr of HC vapour
(molecular weight = 172) and 699
kg/hr of non-condensable gases
(molecular weight = 29). Details of
the existing vacuum system with
four steam ejector stages are shown
in Figure 3. Note that there are two
ejectors in parallel in each of stages
2, 3 and 4.
The following alternatives are
considered for retroftting to mini-
mise the operating costs of the
existing vacuum system in the base
case (Case 3A).
Cases 3B and 3C
Four-stage steam ejector system
with one pre-condenser (at the inlet
of the frst-stage steam ejector),
cooled by chilled water. A single-
stage absorption chiller is used in
Case 3B for generation of chilled
water using very low-pressure fash
steam at 1.5 barg and 130C. A
mechanical chiller, using R143a
refrigerant, is used for the genera-
tion of chilled water in Case 3C.
COP of the chiller is 5.68. Details of
Cases 3B and 3C are shown in
Figure 4.
Case 3D
The frst two stages are steam ejec-
tors, and the last two stages are
cooling, can further reduce the
steam requirement and the ejector
systems capital cost. This is illus-
trated by the following case study
(case study 2), solved using the
optimisation procedure in Equation
9. Different quantities of water
vapour and possible use of chilled
water are considered, and the
results are summarised in Table 2.
Load gas (Cases 2A and 2B): 45.36
kg/h of air + 97.52 kg/h of water
vapour
Load gas (Cases 2C and 2D): 45.36
kg/h of air + 453.6 kg/h of water
vapour
Suction pressure and temperature:
40 torr and 33.3C
Motive steam: 10.6 barg
Cooling water supply and return
temperatures: 32.2C and 37.8C
Chilled water supply and return
temperatures: 7C and 13C
Approach temperature at the
pre-condenser and frst-stage
condenser: 1.67C
Discharge pressure: 815 torr
Water vapour in the load gas is
increased by 365% for Cases 2C and
2D, compared to Cases 2A and 2B.
No pre-condenser (using chilled
water) is used in Cases 2A and 2C,
whereas pre-condenser, cooled with
chilled water, is used in Cases 2B
and 2D. For all the cases, inter-stage
and after-condensers use cooling
water. From Table 2, it can be seen
that the use of chilled water in the
pre-condenser reduced the steam
consumption by 28.5% in Case 2B
compared to Case 2A, and by 60.3%
in Case 2D compared to Case 2C. It
also reduced the total heat transfer
area for pre-, inter- and after-con-
densers by 30.8% in Case 2B
compared to Case 2A, and by 69.0%
in Case 2D compared to Case 2C.
More steam and system capital cost
Values from optimisation
Utilities Pre-condenser Inter-condensers Total
Number of Chilled Cooling Vent Vent Vent heat
Inter-stage Pre- MP steam, water, water, pressure, Pre-condenser, temperature pressure, Inter- and after- transfer
Case Stages condensers condenser kg/h kg/h kg/h mmHg area, m
2
C mmHg condenser area, m
2
area, m
2
2A 4 3 No 351 0 44 581 NA NA 33.9/41.7/48.9 76.5/157.6/280.9 31.4/4/2/1.4 39
28 4 3 Yes 251 9011 25 346 36 11 33.9/41.7/48.1 110.6/203.4/329.6 10.97/2.4/1.5/1.2 27
2C 4 3 No 632 0 110 202 NA NA 33.9/41.7/48.1 55.9/119.1/226 150/6.9/2.4/1.35 161
20 4 3 Yes 251 42 901 25 346 36 34.70 33.9/41.7/48.1 110.6/203.4/329.6 10.97/1.9/1.2/1 50
Effect of pre-condenser with chilled water cooling on steam consumption in the multi-stage steam ejector system (case study 2)
Table 2
reddy anhdra.indd 7 08/03/2013 17:22
Researched and produced by:
TO BOOK
Please visit www.refiningsummit.com or contact Richard Joneson
+44 (0)207 202 7690 or email richard.jones@wtgevents.com
Leading experts include:
Lars Rosenlv,
Vice President & Refning Manager, Statoil
Ramon Nieto,
Director of Engineering, Repsol Spain
Radolaw Gospos,
Head of Technology (Crude Distillation Unit), PKN Orlen
Andrs Pntek,
Vice President, Strategy Development, MOL
Dmitry Spivakovsky,
Director of Strategy, Investment Governance &
Planning Department of Downstream, TNK-BP
Chris Beddoes,
Director General, EUROPIA
In association with: Global industry developments
Globalrefningindustryroadmap:Europe,MiddleEast,
LatinAmerica,Russiainfocus
InnovativestrategicsolutionsfortheEUrefningindustry
Majorrefningprojectsandinvestments
Operational excellence
Costleadershipstrategiesintherefningbusiness
Increasingmarketcompetitivenessthroughoptimisation
ofprocesses
Masteringplantmaintenance
Reinforcingsafeenvironmentthroughcreationofapersonal
commitmenttosafetyamongemployees
Developinganeffcientturnaroundstrategyandfawlessexecution
Production optimisation
Integrationofrefningandpetrochemicals
Excellinginproductionfromheavyoils
Maximisingfeedstockfexibilityatarefnery
Increasingcapacityofexistingequipment
Energy effciency
Examiningadvancedenergysavingstrategies
Learningfrompioneeringcasestudiesoncarbonmanagement,
carboncaptureandstorage
AchievingcommercialisationofCCS
Clean fuels production
Fosteringlowandzero-sulphurfuelproduction
Biofuelsroadmapininternationalmarkets
Biofuelsproduction,technologyandsupplychainoptimisation
Fuelsblending
Presents
C
o
n
ta
c
t u
s
to
d
a
y
a
n
d
a
tte
n
d
fo
r
ju
s
t
9
9
5
Event topics include:
wtg.indd 1 26/02/2013 16:47
134 PTQ Q2 2013 www.eptq.com
Accurate estimation of a vacuum
systems suction load for various
plant operating scenarios is diffcult.
Hence, considerable safety margin
with respect to suction load is often
allowed in the system design. In this
case study, the frst-stage steam ejec-
tor is a single unit, hence it is very
diffcult to reduce the steam
consumption if the operating suction
gas load is lower than the design
rate. However, for the case of a
steam ejector system with a chilled
water pre-condenser, electrical
power required for the chiller can be
reduced by capacity control or by
operating a few chillers arranged in
Depending on the suction pressure
of the frst stage, chilled water may
be required. For the refnery case
study, the mechanical chiller option
(Case 3C) has a payback period of
5.28 years, which will reduce to
2.25 years if the power cost is lower
at 75% compared to the base case
(see Table 3). If waste low-pressure
steam (~1.5 barg) is available at
zero cost, a single-stage absorption
chiller (Case 3B) will have a
payback period of 2.45 years. Thus,
the benefts of pre-condensers using
chilled water depends on availabil-
ity of waste low-pressure steam or
low electricity prices.
is shown as not economical in
this table.
Analysis of the results
Annual operating costs of a
vacuum system are very signifcant
compared to the installed costs.
Hence, optimum vacuum system
confguration is essential to mini-
mise operating costs. Even for the
optimum design, the majority of
the operating cost arises from the
frst stage, which handles the maxi-
mum fow rate of gases/vapours.
The only way to reduce this cost is
to condense the vapours before
they reach the frst-stage ejector.
Details of the vacuum system Case 3A Case 3B Case 3C Case 3D Case 3E Case3F
Steam ejectors MP steam, ton/hr 7.814 3.635 3.635 5.9 1.882 1.882
Condenser area Total area, m
2
2145 1893 1893 2020 1813 1813
LRVP Power consumption, KW 0 0 0 174 174 174
Cooling water, ton/hr 0 0 0 30 30 30
Single-stage Cooling water, ton/hr 0 1126 0 0 1126 0
absorption chiller LLP steam, ton/hr 0 10.58 0 0 10.58 0
Power consumption (includes 0 160 0 0 160 0
power for chilled water pumps), KW
Mechanical chiller Power consumption, KW 0 0 1010 0 0 1010
Cooling water, ton/hr 0 0 618 0 0 618
Total utility requirements MP steam, ton/hr 7.814 3.635 3.635 5.9 1.882 1.882
LLP steam, ton/hr 0 10.58 0 0 10.58 0
Chilled water, ton/hr 0 642.8 642.8 0 642.8 642.8
Cooling water, ton/hr 1099 1197 689 1129 1227 719
Power requirement, KW 0 160 1010 174 334 1184
Savings at foul water stripper LP steam saved at foul water stripper, ton/hr 0 0.72 0.72 0.33 1.02 1.02
Power saved at foul water stripper, KW/hr 0 11.41 11.41 5.23 16.19 16.19
Cooling water saved at foul water stripper, ton/hr 0 2.56 2.56 1.17 3.64 3.64
Operating cost, million $/year Base case 2.68 4.48 2.44 2.29 4.13 2.10
Considering LLP steam for 2.68 1.54 2.44 2.29 1.20 2.10
absorption chiller is zero cost
Considering electric power cost is 2.68 4.43 2.11 2.23 4.03 1.71
75% of the cost considered
If MP steam cost is 75% of that previously due 2.13 4.22 2.18 1.87 4.00 1.96
to credit from power generation at steam turbines
If MP steam cost is 75% of that previously due 2.13 1.29 2.18 1.87 1.07 1.96
to credit from power generation at steam turbines
and LLP steam to absorption chiller is at zero cost
Capital cost, million $ Steam ejectors 0.371 0.159 0.159 0.330 0.118 0.118
Surface condensers 1.691 1.691 1.691 1.497 1.497 1.497
LRVP 0.000 0.000 0.000 0.692 0.692 0.692
Single-stage absorption chiller 0.000 3.000 0.000 0.000 3.000 0.000
Mechanical chiller 0.000 0.000 1.500 0.000 0.000 1.500
Total installation cost 2.062 4.850 3.350 2.519 5.307 3.807
Payback period Base case - Not 5.28 1.16 Not 2.98
economical economical
If LLP steam for absorption chiller is - 2.45 5.28 1.16 2.19 2.98
available at zero cost
If electricity cost is 75% of that - Not 2.25 1.01 Not 1.80
considered for the above cases economical economical
If MP steam cost is 75% of that - Not Not 1.75 Not 10.37
considered in the base case (due to economical economical economical
credit from power generation at steam turbines)
If MP steam cost is 75% of that - 3.30 Not 1.75 3.05 10.37
considered in base case and LLP economical
steam to absorption chiller is available at zero cost
Analysis of alternatives for the vacuum system of a vacuum distillation column in a petroleum refnery
Table 3
reddy anhdra.indd 8 08/03/2013 17:22
www.eptq.com PTQ Q2 2013 135
T Temperatureofsuctiongas/vapour,C
TC
a
Temperatureentrainmentratioforair
TC
w
Temperature entrainment ratio for
watervapour
U Overallheattransfercoeffcientforthe
condenser,W/m
2
K
V
s
SuctionvolumetricfowforLRVP,m
3
/h
W
OG
Flow rate of gases/vapours other than
watervapour,kg/hr
Ww Watervapourfowrate,kg/h
e EffciencyofLRVP
References
1 Aliasso J, Choose the right vacuum pump,
Chemical Engineering, March 1999, www.
graham-mfg.com/usr/pdf/TechLibVacuum/
222.PDF,accessedinJan2012.
2 ASHRAE Handbook, Refrigeration (I-
P) edition, American Society of Heating,
Refrigerating and Air-Conditioning Engineers,
2010.
3 Bannwarth H, Liquid Ring Vacuum Pumps,
Compressors and Systems, Wiley-VCH Verlag
GmbH&CoKGaA,Weinheim,2005.
4 PowerRB,Steam Jet Ejectors for the Process
Industries,2nded,McGraw-Hill,2005.
5 Prakash S, Refning Processes Handbook,
GulfProfessionalPublishing,2003.
6 RyansJ,BaysJ,Runcleanwithdryvacuum
pumps, Chemical Engineering Progress, 32-41,
Oct2001.
7 ReddyCCS,RangaiahGP,NaiduSV,Waste
heat recovery methods and technologies,
Chemical Engineering,28-38,Jan2013.
8 Standards for Steam Jet Vacuum Systems,
HeatExchangeInstitute,6thed,2007.
9 Trambouze B, Petroleum Refning, Vol. 4,
Materials and Equipment, Editions Technip,
Paris,1999.
10 TurtonR,BailieRC,WhitingWB,Shaeiwitz
JA,Analysis, Synthesis, and Design of Chemical
Processes, 3rd ed, New Jersey, Prentice Hall,
2009.
CChandraSekharaReddyistheLeadProcess
Design Engineer with Singapore Refning
Company and a PhD scholar at Andhra
University, Visakhapatnam, India. He holds
bachelors and masters degrees in chemical
engineering from Andhra University and IIT
Kanpur.
SV Naidu is a Professor in the Department
of Chemical Engineering and Dean, Planning
and Resource Mobilisation with Andhra
Universitys College of Engineering. He holds
bachelor and doctoral degrees in chemical
engineering from Andhra University, and a
mastersfromR.E.C.,Warangal.
GPRangaiahisProfessorandDeputyHeadin
theDepartmentofChemical&Biomolecular
Engineering with the National University of
Singapore. He holds bachelors, masters and
doctoral degrees in chemical engineering,
from Andhra University, IIT Kanpur and
MonashUniversity,respectively.
power and extent of waste heat
recovery vary greatly. A detailed
economic study similar to the one
shown in the present study can be
conducted to decide the best strat-
egy for a particular plant by
considering site-specifc factors such
as energy cost, plot size, capital cost,
acceptable payback period, opera-
tional reliability, maintenance and
safety issues.
Nomenclature
A
i
Areafortheithstagecondenser,m
2
Dp
i
Process gas-side pressure drop across
thecondenserattheoutletoftheithejector,
torr
L
i
Latent heat of vapour in the ith stage
condenser,kJ/kg
LMTD
i
Logarithmicmeantemperatureforthe
ithstagecondenser,C
MW
CWV
Molecularweightentrainmentratiofor
watervapour
n Numberofejectorstages
NC
i
Mole percentage of non-condensable
gasesintheithstagecondenserinlet
nw
i
Mole fraction of water vapour in the
inletoftheithstagecondenser
Pa
i
,Pd
i
Suctionanddischargegaspressuresfor
theithstageejector,torr
Ppw
i
Partialpressureofwatervapourinthe
ithstagecondenserinlet,torr
P
v
Motivesteampressure,barg
Pv
i
Saturation pressure of water
corresponding to vent temperature of the ith
stagecondenser,torr
TAP
i
Approachtemperaturefortheithstage
condenservent,C
Tv
i
Saturation temperature of water
vapourattheithstagecondenserinlet,C
Tvo
i
Gas temperature at the ith stage
condenservent,C
T
wc
,T
wco
Cooling water supply and return
temperatures,C
U
i
Overallheattransfercoeffcientforthe
ithstagecondenser,W/m
2
K
Wa Airfowrateinloadgas,kg/h
Wcw
i
Coolingwaterfowratetotheithstage
condenser,kg/h
Wv
i
Steam fow rate for the ith stage
ejector,kg/h
Ww
i
Water vapour fow rate at the inlet of
theithstageejector,kg/h
General
M Molecularweight
MW
c
Molecularweightentrainmentratio
MW
COG
Molecularweightentrainmentratiofor
thegasesotherthanwatervapour
NC Mole percentage of non-condensable
vapour/gas
P PowerforLRVP,KW
Pa,Pd Suctionanddischargepressures,torr
R
a
RatioofmotivesteamfowratetoDAE
fowrateofsteamejectorsuctiongas/vapour
parallel operation. Thus, operating
cost benefts can be further
increased. Availability of plot space
and maintenance costs are the other
critical issues for installing a chiller.
Installation of a mechanical chiller
and LRVP may require modifcation
costs at a power intake substation.
Waste LLP steam can be recov-
ered economically from the steam
condensate system. Petroleum refn-
eries often discard a lot of waste
heat through furnace stacks, fn-fan
coolers and product coolers, which
can also be used to generate
low-pressure steam. If MP steam
cost is lower (~75% compared to the
base case, due to credit from power
generation at steam turbines), a
mechanical chiller will not be
economical. For this case, a single-
stage absorption chiller (Case 3B)
will be feasible (payback ~3.3 years)
only if the low-pressure steam is
available at zero cost (see Table 3).
The option involving a pre-
condenser (cooled by chilled water),
frst two stages with a steam ejector
and the last two stages replaced by
one LRVP (Cases 3F and 3E), is a
very attractive investment with a
payback period of 1.8 to three years
(with an incremental capital cost of
$1.75 million to $3.25 million). Thus,
replacement of the last two stages of
a multi-stage steam ejector system
with one LRVP is highly benefcial.
Conclusions
This article analysed the techniques
available to reduce the energy
consumption of vacuum systems
used in petroleum refneries. Key
requirements, benefts and
constraints for implementation of
these techniques are highlighted,
and strategies for selection and
implementation of a suitable
method are outlined. It can be
concluded from the analysis that use
of chilled water at the pre-condenser
reduces the energy costs of vacuum
systems. As refneries operate many
steam ejector vacuum systems,
considerable potential exists for
reducing energy consumption for
vacuum generation, thus improving
energy effciency and also reducing
carbon emissions. The economics of
such an optimisation vary from one
site to another as the costs of steam,
reddy anhdra.indd 9 08/03/2013 17:23
7th Annual Global Refning Summit 133
ABB Consulting 85
Aerzener 63
AFPM Reliability & Maintenance 115
Air Liquide Global E&C Solutions IFC
Anguil Environmental Systems 79
Axens OBC
BASF Corporation, Catalysts Division 67
BETE Fog 108
Bryan Research & Engineering IBC
Burckhardt Compression 11
Cameron Process & Compression System 64
CB&I 7
China International Sulphur & Sulphuric Acid 131
ContiTech Schlauch 73
Criterion Catalyst & Technologies 4
CS Combustion Solutions 43
DeltaValve 35
DigitalRefning.com 104
DuPont Sustainable Solutions 36
Enersul LP 48
EuroPetroleum 112
ExxonMobil Research and Engineering Company 28
Foster Wheeler 94
GE Oil & Gas, Surface Pumping Systems 77
Grabner Instruments 6
Hermetic-Pumpen 10
Hoerbiger Kompressortechnik Holding 55
ITW Technologies 56 & 101
Johnson Matthey Catalysts 38
Johnson Screens 80
KBC Advanced Technologies 2
KBR 19
Koch-Glitsch 86
Kurita Europe 103
Linde 51
Merichem Company 31
Metso Automation 90
OHL Gutermuth Industrial Valves 41
Onis International 98
Optimized Gas Treating 21
Petrochemical Asia 2013 124
Process Consulting Services 12 & 15
Prosim 82
Rentech Boiler Systems 32
Sabin Metal Corporation 59
Saint-Gobain NorPro 74
Samson 26
Sofraser 8
Spectro Analytical Instruments 53
Sulphur 2103 131
Sulzer Chemtech 22
UOP 9
UOP Callidus Technologies 68
Weir Minerals Lewis Pumps 44
Yokogawa Europe 24
Zwick Armaturen 61
Alphabetical list of advertisers
136 PTQ Q2 2013 www.eptq.com
For more information on these advertisers, go to www.ptqenquiry.com
ad index.indd 1 11/03/2013 16:33
^K'Z/
h^
ProMax is a powerful and versatile process simulation
package that is used worldwide to design and optimize
gas processing, refning and chemical facilities. Totally
integrated with Microsoft Visio
, Excel
and Word
,
ProMax is the best simulation tool for designing gas
plants and predicting performance under varying
process conditions.
Discover why engineers around the world depend on ProMax!
M NGL LNG LG
k
D CC
2
S
U L
C MCnn
2
CnC
S
M
Process Focus: Gas Processing
ProMax
with TSWEET
& PROSIM