Simulation of bitumen upgrading
processes
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
he Canadian oil sands
industry
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 refineries.1 The production
of transportation fuels from
Canadian bitumen requires
either new, integrated upgrading and refining facilities, or the
updating of existing refineries
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 refining scheme under
different process configurations
and product scenarios, to minimise process-related energy
intensity while achieving the
best economic benefits. This can
provide guidelines to existing
upgrading and refining operations as well as to process
design for new upgraders and
refineries.
There
are
two
major
primary bitumen upgrading
technologies: coking and hydroconversion.3,4
Historically,
coking has predominated as the
www.digitalrefining.com/article/1000796
choice for primary upgrading.5
As the first step to producing a
bottomless (zero residue) 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,
ebullated-bed catalytic hydroconversion is more selective
towards a total liquid yield due
to hydrogen addition.7 In
commercial operations, the
conversion of bitumen is not
100% in the hydroconversion
unit, generating 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withdrawn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 hydroconversion,
or a combination of both.8
Both of the two primary
upgrading processes produce
liquid products with boiling
ranges similar to those of
conventional crudes. However,
they have high concentrations
of impurities, such as sulphur
and nitrogen.9 The downstream
hydrotreating units remove
these impurities to produce
sweet blending feedstocks for
SCO without changing the
boiling range of the liquid too
much. The SCO boiling range
is essentially controlled by the
primary upgrading step. In
reviewing some of the major
challenges 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 influence the selection
of the main upgrading process
Address major environmental and greenhouse gas (GHG)
emission concerns in an integrated way
Improve SCO quality with
existing facilities.
The oil sands industry is, by
its very nature, in a position to
influence technology development for its relatively unique
PTQ Q2 2013 1
Sour
gas
Naphtha hydrotreating
Diluent
recovery
Sour
gas
Diluent
H2
Gasoline
LGO hydrotreating
Vacuum
distillation
Diluted
bitumen
Sour
gas
Naphtha (IBP204C)
Diesel
LGO (204343C)
Atmospheric
topped
bitumen
HGO hydrotreating
HGO (343524C)
Coking/hydroconversion
Products
Sour
gas
Products
Cat.
Unconverted
residue
Feed
+H2
Feed
H2
Synthetic
crude oil
H2
FCC feed
Cat.
Coke
Figure 1 Generic diagram of bitumen upgrading process
needs. It is important to identify
the possible avenues for better
upgrading
technology
for
current and future projects.10
While some new primary
upgrading technologies 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 optimisation of
these integrated processes will
provide valuable and helpful
knowledge towards innovative
process design and operation.
The objective of this study is
2 PTQ Q2 2013
to evaluate commercial bitumen upgrading schemes by
performing process modelling
and simulation, which will
help the industry to improve
process
efficiency,
reduce
GHGs emissions and other
related environmental impacts
(for instance, optimised H2 and
water use) in future bitumen
upgrading
and
refining
schemes.
Bitumen upgrading process
Figure 1 presents a generic
flow diagram of the bitumen
upgrading process. The diluted
bitumen from extraction and
froth treatment plant first
passes through the diluent
recovery unit (DRU), where the
diluent is recovered and recycled to the extraction plant.
The atmospheric-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 hydroconversion 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 refineries in Canada
and the US.
www.digitalrefining.com/article/1000796
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
properties of interest (API
gravity, sulphur, nitrogen and
metals, among others) and
product yield shifts at every
conversion step were handled
by the Petroleum Refining 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 performance 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 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 (C1-C4, H2S and NH3),
naphtha (IBP-204C), LGO
(204-343C), HGO (343-524C),
vacuum residue (524C+) and
coke (for the coker unit
only). The hydroconversion
model also included hydrodesulphurisation
(HDS),
hydrodenitrogenation (HDN),
hydrodemetallisation (HDM)
and hydrogen consumption.
The rate equations accounted
for the effect of temperature
and pressure in the case of
hydroconversion, whereas they
were
only
temperature
www.digitalrefining.com/article/1000796
dependent in the case of
coking. The models were properly tuned and implemented in
a Hysys spreadsheet. This
feature enables user-defined
calculations to be performed
using any variable from the
simulation flowsheet and then
the results exported to any
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
hydrotreating 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 determined 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 hydrocracking-hydrotreating module. Rate
constants were properly calibrated in order to produce
typical operational parameters
such as reactor temperature
profiles, hydrogen consumption and product quality.20
Results and discussion
Simulation flowsheets 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
characterised 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 upgrading operations
Properties of the bitumen feedstock and its distillation products
Diluent recovery Vacuum distillationa
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.
Table 1
PTQ Q2 2013 3
Coker and hydroconverter performancea
Product yields, wt%
H2S
NH3
Gas (C1-C4)
Naphtha
LGO
HGO
Vacuum residue
Coke
Total
H2 consumption, scf/bbl
Conversion ,wt%
Sulphur
Nitrogen
Metals (Ni+V)
Vacuum residue
Coker
Hydroconverter
1.5
0.1
6.1
15.5
20.8
23.5
1.5
31.0
100.0
-
4.3
0.5
4.0
12.6
34.0
27.2
19.2
101.7
1218
39.1
68.4
98.5
98.5
67.3
55.2
76.6
78.0
a) Product yield values are based on total VTB feed.
Table 2
Properties of coker and hydroconverter liquid productsa
Property
Naphtha
Yield, wt%
26.8
SG 60/60F
0.7363
API gravity
60.7
Sulphur, wt%
1.79
Nitrogen, wppm
316
Aromatics, wt%
27.1
Nickel, wppm
-
Vanadium, wppm
-
CCR, wt%
-
Coker Hydroconverter
LGO
HGO Naphtha LGO HGO Vacuum residue
34.1
39.1
14.0
37.9
27.8
20.4
0.8715
0.9736 0.7365 0.8710 0.9715
1.1240
30.9
13.8
60.6
31.0
14.2
-5.6
3.67
4.43
0.25
0.71
1.37
6.90
1694
3973
393
1582 3063
8905
58.0
66.0
20.5
42.5
53.0
99.8
-
8.2
-
-
3.0
262.9
-
27.4
-
-
6.2
458.3
-
1.5
-
-
0.2
14.3
a) Yield values are based on total liquid product.
Table 3
very difficult, costly and environmentally
detrimental.
Practically all metals and
carbonaceous
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
hydroconversion reactor performance are shown in Table 2. It
is clear from the product distri-
4 PTQ Q2 2013
bution that hydroconversion 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
H2 addition upgrading is about
78%, and there is a substantial
catalytic removal of sulphur,
nitrogen
and
metals.
Hydroconversion also generates
fewer C1-C4 compounds, but
more H2S and NH3 due
to
sulphur
and
nitrogen
conversion. Conversely, coking
offers almost complete residue
conversion but with significant
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
fractionation 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 the ebullated-bed reactor or fed to a
coker unit. For this study, it
was considered that this stream
is recycled back to the hydroconversion reactor, meaning
that there is complete overall
residue conversion 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 distillates, require further
hydrotreating
to
reduce
sulphur, nitrogen and aromatic
contents for producing suitable
blending components for SCO.
The properties of the hydrotreated products and final 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
www.digitalrefining.com/article/1000796
the details of the HDT units. It
is observed that the hydrotreating of coker products requires
higher 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 distribution 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 hydrogenation is partial saturation of
polyaromatics to form di- and
mono-aromatics.
Finally, an overall comparison of the two process schemes
is presented in Table 6.
The
hydroconversion-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 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 H2/bbl
bitumen vs 426 scf H2/bbl bitumen), which in return makes
this scheme more energy
intensive.
Summary
This article presents a detailed
www.digitalrefining.com/article/1000796
Properties of hydrotreated products and SCO
Coker-based scheme
Hydroconverter-based scheme
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
Table 4
Hydrotreater performancea
Coker-based scheme Hydroconverter-based scheme
Naphtha
LGO HGO Naphtha
LGO
HGO
Operating conditions
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
H2/oil, scf/bbl
1680
1680 3370
1680
1680
3370
Product yields, wt%
H2S
1.8
2.7
3.6
0.3
1.0
2.5
NH3
0.03
0.1
0.2
0.04
0.1
0.2
Gas (C1-C4)
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
H2 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.
Table 5
Overall performance of the two schemesa
Products, wt%
Gases
Naphtha
LGO
HGO
Coke
Total
SCO, wt%
SCO, vol%
H2 consumption, scf/bbl
Coker-based scheme
Hydroconverter-based scheme
7.0
10.3
30.3
35.5
17.6
100.6
76.1
87.1
427
8.0
11.0
42.5
40.4
101.9
93.9
107.3
1269
a ) Product yield and H2 consumption values are based on total ATB feed.
Table 6
simulation of two typical bitumen upgrading schemes using
Aspen Hysys. Process models
for every major unit were
constructed and calibrated
based on representative data.
PTQ Q2 2013 5
The simulations predicted typical upgrader performance and
allowed tracking of the transformation 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 cokingbased scheme (76 wt%).
However,
hydrogen
consumption in hydroconversion-based upgrading is much
higher than that in the cokerbased scheme (1269 scf H2/bbl
bitumen vs 426 scf H2/bbl bitumen), which makes the former
option more energy intensive.
In general, the simulation flowsheet 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, operation 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 &
Very Heavy Crude Upgrading Long Term
R&D Opportunities, 1994.
2 Yui S, Chung K H, Syncrude upgrader
6 PTQ Q2 2013
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, 4th 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.
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/
Al2O3 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 diolefins 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 refinery 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.
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.
LINKS
More articles from the following
categories:
Coking
Hydroprocessing
Process Modelling & Simulation
www.digitalrefining.com/article/1000796
