PAPER 2009-066

Simulation of O/W Emulsion Flow in

Alkaline/Surfactant Flood for Heavy Oil Recovery
J. WANG, M. DONG
University of Calgary
This paper is accepted for the Proceedings of the Canadian International Petroleum Conference (CIPC) 2009, Calgary,
Alberta, Canada, 1618 June 2009. This paper will be considered for publication in Petroleum Society journals.
Publicationrightsarereserved.Thisisapreprintandsubjecttocorrection.
recovery improvement by alkaline flooding: dispersion and
entrainment, wettability reversal from oil-wet to water-wet, or
vice versa, and emulsification and entrapment. In the case of
heavy oils, the emulsification and entrapment during alkaline
flooding has been recognized as the dominant mechanism [2-4],
which can efficiently improve sweep efficiency. Jennings et
al.[2] demonstrated this mechanism through extensive
experimental studies. Visual experiments clearly showed that
the areal sweep efficiency was improved by the in-situ
generated emulsions, and the oil recovery at breakthrough was
doubled compared to that obtained in the waterflooding test.
Core flooding tests demonstrated the increase in oil recovery
and the decrease in the instantaneous water oil ratio (WOR).
The mechanism was summarized as: a drastic reduction of
oil/water interfacial tension (below about 0.01 mN/m) by the
caustic activation of potentially surface-active organic acids in
the crude oil, in-situ production of the O/W emulsions that
tends to lower the mobility of the injected water and damp
viscous fingering, and the diversion of the flow of injected
water to give improved sweep efficiency.
Dong et al.[5-6] reported comprehensive studies of the
alkaline/surfactant (A/S) flood potential for three West
Canadian heavy oils. Extensive emulsification tests, IFT
measurements, micromodel experiments and sandpack flood
tests were conducted. Their results showed that the IFT could
be reduced to be lower than 0.01 mN/m by an alkaline solution
and a very dilute concentration of surfactant, leading to easy
emulsification of heavy oil in formation brine under slight
interfacial disturbance. Tertiary oil recovery in sandpack flood
tests reached more than 20% OOIP. Liu et al.[7] studied the

Abstract
The formation and flow of emulsions during alkaline
flooding process plays an important role for improving heavy
oil recovery. In this study alkaline/surfactant (A/S) flood tests
were performed in sandpacks to demonstrate the effectiveness
of sweep efficiency improvement by the in-situ generated O/W
emulsion. High tertiary oil recoveries were obtained in all the
sandpack flood tests. Experimental results were history matched
by including the mechanisms of in-situ generation and flow of
O/W emulsion, as well as the chemical adsorption and the
reduction of interfacial tension involved in the chemical
flooding process. The decrease in local water phase
permeability caused by the entrapment of emulsion droplets was
modeled using the filtration theory. Both the pressure response
and the oil recovery improvement were fairly matched. Field
scale simulations were conducted to investigate the potential of
A/S flooding for heavy oil reservoirs. Simulations showed
promising results of chemical flooding for heavy oils. It was
indicated that certain length of waterflooding time would
benefit for the final oil recovery, and there existed an optimum
chemical slug size. These laboratory results and the simulation
technique are helpful in the simulation and design of field scale
projects of chemical flooding for enhanced heavy oil recovery.

Introduction
Both field and laboratory studies showed that caustic flood
could effectively improve oil recovery for moderately viscous
oils. Johnson[1] summarized four main mechanisms of oil

mg KOH/g-oil. The salinity of the formation brine was 2.7

wt%, with 410 mg/L of Ca2+ and 256 mg/L of Mg2+.
Emulsification tests[7] showed that the oil sample could be
dispersed into the formation brine by the synergy of alkali and
surfactant. The combined application of NaOH and Na2CO3 was
chosen as the alkaline agent from the screening tests. An
anionic surfactant, alkyl ether sulfates (Stepan, Canada), was
selected as the additive because it provided the better
emulsification result for the studied heavy oil/brine system.
Oil/water interfacial tensions (IFT) were measured using
spinning-drop tensiometer (Model 510, Temco, USA). Figure 1
presents the IFT as a function of the combined alkaline
concentration with the weight ratio of Na2CO3/NaOH being 1:1,
in the presence of 100 mg/L surfactant. A minimum IFT was
observed around 0.4 wt% combined alkalis. With the
consideration of chemical loss, 0.6 wt% combined alkalis and
300 ppm surfactant were used in sandpack flood tests.
Three alkaline/surfactant flooding tests were carried out in
this study. The sandpack holder used in alkaline flood tests was
4.25 cm in diameter. Two of the three sandpacks were 14.2 cm
in length and the other one had a length of 90.4 cm to
investigate the effect of sandpack length on oil recovery. Both
ends of the sandpack holder were equipped with flow
distributors on which 200-mesh stainless steel screens were
spot-welded to prevent fine sand from flowing out and to
provide more even distribution of the injected fluid. Ottawa
sand was wet-packed in the sandpack holder. The packing
procedure was as follows: 60100 mesh sand was poured into
the sandpack holder which was vertically mounted on a vibrator
and filled with formation brine. The sandpack holder was fully
filled at a time and was vibrated for one hour. The porosity of
the sandpacks was around 36.3% and the absolute permeability
ranged from 5.5 to 7.0 m2.
The wet-packed sandpack was flooded with the heavy oil to
establish initial water saturation. The oil injection was
conducted at the velocity of 0.4 m/d and continued until water
production ceased (water cut less than 1.0%). The initial water
saturation was determined on the basis of mass balance. After
that, waterflooding was conducted by injecting formation brine
at a constant injection flow rate. Waterflooding was continued
until oil production became negligible (oil cut less than 1.0%),
and then A/S flooding was started. A 0.3-PV preflush slug (0.10
wt% Na2CO3 in brine) was used before the injection of
alkaline/surfactant slug in order to reduce surfactant loss and to
provide a Ca2+-free circumstance for emulsification of oil.
Following the injection of a chemical slug, an extended
waterflood was carried out until the oil production became
negligible.
The parameters of the sandpacks and the results of
waterflooding and chemical flooding are summarized in Table
1. Initial oil saturation ranged from 69.3 to 87.0%, and
waterflooding oil recovery varied from 29.0 to 34.1 %OOIP.
Varied injection flow rates, chemical slug sizes and sandpack
lengths were used in these three A/S flooding tests, and all tests
showed significant improvement in oil recovery after the
chemical flooding. The injection velocity was 0.4 m/d for Run1
and Run 3, and 0.8 m/d for Run 3. The chemical slug sizes were
0.5 PV for Run1 and Run 3, and 1.2 PV for Run 2. A same
sandpack length was used in Run1 and Run 2, while a longer
sandpack with more than six times of that length was used in
Run 3. Test results showed there was no evident difference in
tertiary oil recovery for these three tests, which ranged from
22.4 % to 24.4 % OOIP. This indicates that the injection rate
has insignificant effect on oil recovery in the studied flow rate
range. With approximately a same tertiary oil recovery, the 0.5

synergy of alkali and surfactant in emulsifying a heavy oil in

brine. Liu et al.[8] also presented a study of enhanced heavy oil
recovery by the in-situ produced O/W emulsion. They
conducted A/S flooding tests for five Western Canadian heavy
oils with viscosities ranging from 650 to 18,000 mPa.s, and
promising results were obtained for all the oils.
Numerical modeling the alkaline flooding for heavy oil
recovery is far from satisfactory. Unlike the case of
conventional oil, ultra-low oil/water interfacial tension is not
always the dominant mechanism during alkaline flooding for
heavy oils. Emulsion flow becomes one of the main
characteristics which must be included in the simulator for this
type of EOR process. Generally, there are three theories
describing the flow of emulsion in porous media: the
homogenous model, the droplet retardation model, and the
filtration model. The simple bulk viscosity model was
developed by Alvarado and Marsden [9, 10], in which emulsion
was viewed as a homogeneous, single-phase fluid. This model
was suitable for the description of the flow of emulsions with
smaller drop-size to pore-size ratio where the flow could reach
steady state quickly. Abou-Kassem and Farouq Ali [11] modified
the viscosity model to describe both Newtonian and nonNewtonian fluids, and their model was suitable for numerical
simulation of EOR processes. The correlation provided a
quantitative description of the effect on flow of pore size
distribution and tortuosity of porous media. Based on the
mechanism postulated by McAuliffe[12], Devereux[13] proposed
a droplet retardation model for describing the flow of stable
O/W emulsions in porous media. In this model the dispersed
drops flow slower than the continuous phase because of the
capillary retarding force is encountered when the drops are
flowing through the pore throats smaller than drops themselves.
The retardation model can arrive at larger permeability
reduction with lower flow rate and higher drop size-to-pore size
ratio, however, it cannot predict the permanent permeability
reduction observed in laboratory. Soo and Rake [14-16] proposed
a model for describing the flow of stable, dilute emulsions in
unconsolidated porous media based on deep-bed filtration
concepts. In this model dispersed drops can be captured in pores
by both straining and interception, resulting in permeability
reduction. Of all these models, the filtration model is the most
successful in representing all the experimental observations.
The objective of this study is to experimentally demonstrate
the improvement in heavy oil recovery by A/S flooding, and to
conduct history match of the experimental results by including
the main flow mechanisms in the numerical simulation. Based
on the results of history match, field scale simulations are
conducted to investigate the potential of A/S flooding for heavy
oil reservoirs.

Alkaline/Surfactant Flooding Tests in

Sandpacks
Alkaline/surfactant flood tests were performed in sandpacks
to demonstrate the effectiveness of sweep efficiency
improvement through the in-situ formed O/W emulsions, and
more importantly, to provide experimental data for history
matching in the further numerical simulation study. The heavy
oil sample and the formation brine used in this study were from
a heavy oil reservoir in Saskatchewan, Canada. The oil sample
was centrifuged at 10,000 rpm under 35C for two hours to
remove water and solids. The viscosity and density of the oil
sample at 25C were determined to be 1,370 mPa.s and 961.8
kg/m3, respectively. The acid number of the oil sample was 1.32
2

The history matching results for these three flood tests are
shown in Figures 2 to 4 (in lines) in comparison with
experimental results. In general, the simulated oil production
and pressure drop values were very close to the test values.
During alkaline flooding, pressure drop was increased due to
the local water permeability decrease resulted from entrapment
of emulsion droplets. Oil recovery was increased mainly
because of the improvement in sweep efficiency.

PV chemical slug size is more economical compared to the 1.2

PV of chemical slug. The oil recovery efficiency was not
reduced with the increase in sandpack length, which indicated a
great potential for upscaling the laboratory tests to field
applications.
The curves of cumulative oil production and pressure drop of
these three flood tests are presented in Figures 2 to 4 (in
squares). It can be seen that the increase in oil recovery during
A/S slug injection is accompanied by an increase in pressure
drop. The mechanism involves the ultralow IFT, formation and
flow of O/W emulsions. The injected chemicals react with the
organic acids in the oil, leading to a significant reduction in
IFT, and emulsification of heavy oil in brine. Some oil is
entrained in the continuous water phase and flows out in the
form of O/W emulsion. The flowing emulsion droplets may be
trapped when arriving at smaller pore throats, which will lower
the local mobility of water phase and divert the injected water
to un-swept regions. This entrapment mechanism accounts for
the increase in pressure drop and improvement in oil recovery.

Field Scale Reservoir Simulation

Based on the history matching of the alkaline/surfactant
flooding tests, field scale numerical reservoir simulations were
conducted to investigate the potential of field application of the
EOR process. Field simulations were performed on a
rectangular area of 450 m long and 400 m wide with a pay
thickness of 4.5 m. Parallel horizontal wells spaced 200 m were
used as injectors and producers. Table 2 lists the production
parameters of the field model. A grid system of 21105 was
used to represent this reservoir model, with the grid block size
of 20 m 50 m 0.9 m. Figure 8 is a 3-D view of the
homogeneous reservoir model with color scale showing the
permeability distribution. Uniformly distributed permeabilities,
ranging from 1.0 m2 to 10.0 m2, were assigned to these 1,050
grids with the average permeability of 5.57 m2.
The homogeneous reservoir model showed in Figure 8 was
used to study various production scenarios, including varied
time lengths of waterflooding before chemical injection, and
varied time lengths of chemical injection (chemical slug sizes).
Components, chemicals and reactions were the same as those
used for the simulation of the test Run 2 in the previous section.
The effect of waterflooding time on oil recovery is shown in
Figure 9, which compares the oil recoveries for the cases with
waterflooding period of one, two and three years. There was no
primary production prior to initial waterflood for all these three
cases. Chemical flooding was continued for eight (8) months,
and then switched to extended waterflood. The production well
was shut-off when the water cut was higher than 98%. The
simulation results showed that for the case with longer time of
waterflood, the oil recovery was lower at first; however, the
final oil recovery was relatively higher. This may indicate that
certain length of time of waterflood before chemical injection
would benefit for the final oil recovery.
The effect of chemical slug size on oil recovery was studied
by simulating and comparing five cases with varied time
lengths of chemical injection ranging from 0 to 12 months.
There was no primary production, and waterflood was
continued for three years for all these cases. The production
well was shut-off when the water cut was higher than 98%. The
simulation results are shown in Figure 10. It can be seen that a
longer time of chemical injection resulted in a higher oil
recovery. However, when the chemical injection time was
longer than 8 months, the oil recovery did not show evident
increase with increasing chemical injection time. Although this
time may vary for specific reservoir conditions, the simulation
results indicate that there exists an optimum chemical slug size.
The above reservoir simulation model was used to study the
variations in oil recovery, reservoir pressure and water cut after
chemical injection. Reservoir production started at January 01,
2000, and there was no primary production. Waterflooding was
continued for 3 years before chemical flooding. After 8 months
of chemical injection (corresponding to 0.15 PV chemical slug
injection), waterflood was resumed until water cut in production
well was higher than 98%. The simulated oil recovery, average

History Match of Sandpack Flood Tests

Based on the experimental studies, laboratory-scale
numerical simulations were conducted to match the production
histories of the sandpack flood tests. CMG STARS was used to
simulate the alkaline/surfactant flooding processes. The in-situ
formation of O/W emulsion was represented by a reaction of
oil, water and chemicals. Filtration Theory was used to simulate
the emulsion capture in porous media, which leading to the
decrease of local water phase permeability. A grid system of
25251 was used to represent the 14.2-cm-long sandpacks in
the tests, with the grid block size of 0.568 cm 0.567 cm 1 cm.
The longer sandpack was modeled using a grid system of
10051, with the grid block size of 0.904 cm 2.837 cm 1
cm. Uniformly distributed permeabilities were assigned to the
grids with the average values being the same as the measured
data.
Five components were used in the simulation: water,
chemical (alkali/surfactant), and O/W emulsion in aqueous
phase, dead oil in oil phase, and trapped oil in solid phase. The
trapped oil was produced when the in-situ formed O/W
emulsion drops were trapped, and then the local water phase
permeability was decreased. The measured oil/water interfacial
tension (Figure 1) was input into the simulation. Langmuir
isotherm curve was used to describe the chemical adsorption.
Tests performed in sandpacks by Liu [17] showed that the
maximum NaOH loss on sand was about 2.0 mol/g-sand. This
value was used to estimate the maximum adsorption value in
the simulation.
The oil-water two-phase relative permeability curves used
for matching these three sandpack flood tests are shown in
Figures 5 to 7. To represent the effect of IFT reduction on twophase flow, the relative permeability was interpolated as a
function of capillary number. Two sets relative permeability
curves were used, corresponding to the two cases of high IFT
(during waterflooding process) and an extreme case of low IFT.
The capillary numbers for these two cases were 2.510-7 and
3.210-3, respectively. Both oil phase and water phase relative
permeabilities increased during chemical flooding process. The
water phase permeability in the grids where emulsion drops
trapped was decreased by including a water phase resistance
factor (> 1.0). The water phase resistance factor increases with
decreasing the permeability of the porous medium and
increasing the emulsion concentration.
3

reservoir pressure and water cut in production well are plotted

as a function of time, and shown in Figure 11. The oil recovery
with only waterflood (no chemical injection) is also plotted in
this figure for comparison. It can be seen that after chemical
injection reservoir pressure was increased, and water cut in
production well was effectively decreased. Oil recovery was
greatly improved by the chemical flooding. The tertiary oil
recovery was as high as the waterfloods recovery, i.e., oil
recovery was doubled after chemical flooding. Comparing to
the final oil recovery in the case without chemical injection, the
oil recovery improvement by chemical flooding was about 20%
OOIP.
Figure 12 compares the oil saturation distribution in the
middle layer of the reservoir before and after chemical
injection. It clearly shows that oil bank was formed in the
displacement front after the chemical injection. The
corresponding pressure distribution in this layer of the reservoir
is shown in Figure 13. It shows that the pressure around the
injection well and the pressure gradient in the displacement
front are increased during the chemical flooding process.

Acknowledgements
The authors acknowledge with thanks the Petroleum
Technology Research Center (PTRC), the Natural Sciences and
Engineering Research Council of Canada (NSERC), and
Canada Foundation for Innovation (CFI) for the financial
support of this study.

REFERENCES
1. Johnson, JR., C. E., Status of Caustic and Emulsion

Methods, J. Pet. Tech., pp. 8592, January 1976.

2. Jennings, JR., H.Y., Johnson, JR., C.E. and McAuliffe,

C.D., A Caustic Waterflooding Process for Heavy Oils, J.

Pet. Tech., pp. 13441352, December 1974.
3. FAROUQ ALI, S.M., FIGUEROA, J.M., AZUAJE, E.A.
and FARQUHARSON, R.G., Recovery of Lloydminster
and Morichal Crudes by Caustic, Acid and Emulsion
Floods, JCPT, pp. 53-59, JanuaryMarch 1979.
4. BRYAN, J. and KANTZAS, A., Enhanced Heavy-oil
Recovery by Alkali-Surfactant Flooding, SPE 110728,
paper presented at the 2007 SPE Technical Conference
and Exhibition held in Anaheim, California, U.S.A., 11
14 November 2007.
5. Dong, M., Displacement of Heavy Oil through Interfacial
InstabilityA Study of an EOR Method for Murphy East
Bodo Heavy Oil Reservoir, Petroleum Technology
Research Centre (PTRC) Technical Report, 2004.
6. Dong, M., Liu, Q., Ma, S., and Zhou, W., Displacement
of Heavy Oil Through Interfacial Instability: A Study of
an EOR Method for CNRL Brintnell Heavy Oil
Reservoir, Petroleum Technology Research Centre
(PTRC) Technical Report, 2005.
7. Liu, Q., Dong, M., Yue, X. and Hou, J., Synergy of
Alkali and Surfactant in Emulsification of Heavy Oil in
Brine, Colloids and Surfaces A: Physicochem. Eng.
Aspects, 273, pp.219-228, 2006.
8. Liu, Q., Dong, M., and Ma, S., Alkaline/Surfactant Flood
Potential in Western Canadian Heavy Oil Reservoirs,
paper SPE 99791 presented at the 2006 SPE/DOE
Symposium on Improved Oil Recovery held in Tulsa,
Oklahoma, U.S.A., 2226 April 2006.
9. ALVARADO, D.A., The Flow of Macroemulsion
through Porous Media; Ph.D. Thesis, Department of
Petroleum Engineering, Stanford University, Stanford,
California, 1975.
10. ALVARADO, D.A. and MARSDEN, JR., S.S., Flow of
Oil-in-Water Emulsions through Tubes and Porous
Media, Soc. Pet. Eng. J., pp. 369377, December 1979.
11. ABOU-KASSEM, J.H. and FAROUQ ALI., S.M.,
Modelling of Emulsion Flow in Porous Media, JCPT,
Vol. 34, No.6, pp. 3038, 1995.
12. MCAULIFFE, C.D., Oil-in-Water Emulsions and Their
Flow Properties in Porous Media, J. Pet. Tech., pp. 727
733, June 1973.
13. DEVEREUX, O.F., Emulsion Flow in Porous Solids I. A
Flow Model, Chem. Eng. J., Vol. 7, pp. 121128, 1974.
14. SOO, H. and RADKE, C.J., The Flow Mechanism of
Dilute, Stable Emulsion in Porous Media, Ind. Eng.
Chem. Fundam., Vol. 23, No. 3, pp. 342347, 1984.
15. SOO, H. and RADKE, C.J., A Filtration Model for the
Flow Of Dilute, Stable Emulsions in Porous Media I.
Theory, Chem. Eng. Sci., Vol. 41, No. 2, pp. 263272,
1986.

Conclusions
Three alkaline/surfactant flooding tests were conducted in
sandpacks using a heavy oil to demonstrate the effectiveness of
sweep efficiency improvement by the in-situ generated O/W
emulsions, as well as to provide experimental data for
numerical simulation study. Numerical simulation was
performed to match production history by including the in situ
generation and flow of the O/W emulsions. Using the
parameters obtained in history matching, field scale simulations
were conducted. On the basis of experimental and simulation
results, the following conclusions were drawn:
1. A properly designed alkaline flooding for heavy oil
recovery could effectively improve sweep efficiency
through in-situ generated O/W emulsion. The formation
of O/W emulsion lowered the mobility of water phase,
diverted the injected water to un-swept regions and
improved oil recovery. The injection flow rate had
insignificant effect on oil recovery in the studied range
corresponding to the flow velocity of 0.4 to 0.8 m/d.
The oil recovery efficiency was not reduced with the
increase in sandpack length, which showed a great
potential for upscaling the laboratory tests to field
applications.
2. Laboratory scale numerical simulations, including the
adsorption of chemicals, interfacial tension reduction,
and in-situ generation and flow of emulsions fairly
simulated the pressure response and oil recovery
improvement in chemical flooding tests in sandpacks.
The entrapment of oil drops was represented by
decreased local water phase permeability.
3. Field scale simulations showed promising results of
chemical flooding for heavy oils. It was indicated that
certain length of time for water injection would benefit
for the final oil recovery, and there existed an optimum
chemical slug size. For the studied case in this study,
0.15 PV of chemical slug would achieve a more
economic result, and nearly a doubled oil recovery was
obtained by the application of alkaline/surfactant
flooding.

16. SOO, H. and RADKE, C.J., A Filtration Model for the

Flow of Dilute, Stable Emulsions in Porous Media II.
Parameter Evaluation and Estimation, Chem. Eng. Sci.,
Vol. 41, No. 2, pp. 273281, 1986.
17. Liu, Q., Interfacial Phenomena in Enhanced Heavy Oil
Recovery by Alkaline Flood, Ph.D. Thesis, University of
Regina, Canada, 2006.

Table 1 Summary of alkaline/surfactant flooding tests

Run number

14.2

14.2

90.4

36.6

36.8

35.4

5.5
69.6

6.5
69.3

7.0
87.0

Waterflooding recovery, %OOIP

29.0

34.1

34.0

Injection velocity, m/d

0.4

0.8

0.4

Na2CO3, wt%

0.30

0.30

0.30

NaOH, wt%

0.30

0.30

0.30

Surfactant, ppm

300

300

300

Chemical slug size, PV

0.5

1.2

0.5

Tertiary recovery, %OOIP

22.4

24.4

23.4

51.4

58.5

57.4

Length, cm
Porosity, %
Sandpack

Permeability, m
Initial oil saturation, %
2

Chemical
formula

Chemical
flooding

Final oil recovery, %OOIP

Table 2 Parameters of field scale model

Pattern

Line drive

Kh, m2

5.57

Injectors

Horizontal

Kv/Kh

0.5

Producers

Horizontal

Initial Oil, Soi

0.755

Length, m

450

Oil Viscisity, mPa.s

1000

Width, m

400

Initial Pressure, kPa

2500

Thickness, m

4.5

Maximum Injection Pressure, kPa

10,000

Depth, m

400

Injection rate, m3/d

150

Porosity

0.3

Production rate, m /d

200

Interfacial Tension, mN/m

10

0.1

0.01

0.001
0

0.2

0.4

0.6

0.8

Combined Alkaline Concentration, w t%

Figure 1 Interfacial tensions of heavy oil/brine as a function of combined alkaline concentration. Weight ratio of
Na2CO3/NaOH being 1:1, and surfactant concentration in brine being 100 mg/L.

Figures 2 Experimental and simulated cumulative oil production and pressure drop for test Run 1 of A/S flooding in
sandpack.

Figures 3 Experimental and simulated cumulative oil production and pressure drop for test Run 2 of A/S flooding in
sandpack.

Figure 4 Experimental and simulated cumulative oil production and pressure drop for test Run 3 of A/S flooding in sandpack.

Waterflooding
Ca = 3.2E-3
0.8
Kro

Kr

0.6

0.4
Krw

0.2

0
0

0.2

0.4

0.6

0.8

Sw

Figure 5 Simulation used oil-water two-phase relative permeability curves for test Run 1 of A/S flooding.
8

1
Waterflooding
Ca = 3.2E-3
0.8
Kro

Kr

0.6

0.4
Krw
0.2

0
0

0.2

0.4

0.6

0.8

Sw

Figure 6 Simulation used oil-water two-phase relative permeability curves for test Run 2 of A/S flooding.

Waterflooding
Ca = 3.2E-3

0.8
Kro

Kr

0.6

0.4

Krw
0.2

0
0

0.2

0.4

0.6

0.8

Sw

Figure 7 Simulation used oil-water two-phase relative permeability curves for test Run 3 of A/S flooding.

Figure 8 3-D view of reservoir model with uniformly distributed permeabilities.

Figure 9 Effect of waterflooding time on oil recovery.

10

Figure 10 Effect of chemical slug size on oil recovery.

Figure 11 Simulation results for A/S flooding in homogeneous reservoir.

11

(a)

(b)

Figure 12 Oil saturation distributions in the middle layer of the reservoir model.
(a) Before chemical injection; (b) after chemical injection.

12

(a)

(b)

Figure 13 Reservoir pressure distributions in the middle layer of the reservoir model.
(a) Before chemical injection; (b) after chemical injection.

13