See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/317660928

The Effect of Oil on Steady-state Foam Flow

Regimes in Porous Media

Conference Paper January 2017

CITATIONS READS

0 18

3 authors, including:

Sebastien Vincent-Bonnieu
Shell Global
39 PUBLICATIONS 128 CITATIONS

SEE PROFILE

All content following this page was uploaded by Sebastien Vincent-Bonnieu on 21 June 2017.

The user has requested enhancement of the downloaded file.

We P009
The Effect of Oil on Steady-state Foam Flow
Regimes in Porous Media
J. Tang* (Delft University of Technology), S. Vincent Bonnieu (Shell Global
Solutions International) & W.R. Rossen (Delft University of Technology)

SUMMARY
Foam flow in porous media without oil shows two regimes, depending on foam quality (gas fractional
flow). Complexity and limited data on foam-oil interactions in porous media greatly restrict understanding
of foam in contact with oil. Distinguishing which regimes are affected by oil is key to modelling the effect
of oil on foam. We report steady-state corefloods to investigate the effect of oil on foam through its effect
on the two flow regimes. We fit parameters of the widely used STARS foam model to data for foam-oil
concurrent flow. This research provides a practical approach and initial data for simulating foam EOR in
the presence of oil.
To ensure steady state, oil is co-injected with foam at a fixed ratio of oil (Uo) to water (Uw) superficial
velocities in a Bentheimer sandstone core. Model oils used here consist of two components: hexadecane,
which is benign to foam stability, and oleic acid, which can destroy foam. Varying the concentration of
oleic acid in the model oil allows one to examine the effect of oil composition on steady-state foam flow.
Experimental results show that oil impacts both high- and low-quality regimes, with the high-quality
regime more vulnerable to oil. In particular, oil increases the limiting water saturation (Sw*) in the high-
quality regime and also lessens gas mobility reduction in the low-quality regime. The high-quality regime
is strongly shear-thinning in the presence of oil. Pressure gradient ( p) in the low-quality regime, in some
cases, decreases with increasing Uw at fixed gas superficial velocity (Ug), either with or without oil. This
may reflect either an effect of oil, if oil is present, or easier flow of bubbles under wetter conditions.
Increasing oleic acid concentration extends the high-quality regime to lower foam qualities, indicating
more difficulty in stabilizing foam. Thus oil composition plays as significant a role as oil saturation.

A model fit assuming a fixed Sw* and including shear-thinning in the low-quality regime doesnt represent
each regime when the oil effect is strong enough. In such cases, fitting Sw* to each p contour and
excluding shear-thinning in the low-quality regime yields a better match to data. The dependency of Sw*
on oil saturation is not yet clear owing to absence of oil-saturation data in this study. Furthermore, none of
the current foam simulation models can capture the upward-tilting p contours in the low-quality regime.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
Introduction

Numerous laboratory studies and field pilots demonstrate that steady-state foam flow comprises two
regimes, depending on foam quality (gas fractional flow) (Osterloh and Jante, 1992, Alvarez et al.,
2001): the high-quality regime, which is dominated by foam stability, and the low-quality regime,
where foam strength is controlled by a mobility reduction factor. Fig. 1 illustrates this behaviour.
These two regimes are central to understanding of foam without oil and also modelling of foam with
oil, but this fundamental property is identified in the absence of oil. Owing to limited data available
and complexity of foam-oil interactions, the effect of oil on foam is not fully understood yet. This
knowledge gap greatly limits the understanding of the effect of oil on foam: in particular, its effect on
the two regimes central to our understanding of foam without oil. Furthermore, the effectiveness of
foam in presence of oil is key to successful foam for enhanced oil recovery (EOR). This issue hinders
reliable design of foam processes and effective prediction of foam performance.

Figure 1 Pressure drop across a 2-ft sandpack as a function of superficial velocities of gas (Ug) and
water (Uw). Pressure gradient in psi/ft is half the values listed. Contours of equal pressure drop are
plotted through steady-state data represented by black points; from Osterloh and Jante (1992). The
high-quality regime is to the upper left, and the low-quality regime is on the lower right.

Initial research on foam-oil interactions identifies some surface phenomenon and proposes some
coefficients to represent them (Farajzadeh et al., 2012): e.g. entering coefficient, spreading
coefficient, bridging coefficient and lamella number. These coefficients can serve for evaluating the
effect of oil on foam in bulk. However, foam behaviour identified in bulk is not necessarily consistent
with that in porous media. In some cases, foam behaviour in these two scenarios even contradicts with
each other. Besides, none of these coefficients is individually a good criterion for judging the stability
of foam interacting with oil. Moreover, these coefficients do not predict the effect of oil on foam
quantitatively.

Some studies on the effect of oil on foam stability indicate that most of oils are detrimental to foam
stability, and the lighter the oil component, the more harmful it is to foam. But again, there is no
quantitative, predictive model for the effect of oil on foam.

A modelling study on the effect of oil on foam (Tang et al., 2016) illustrates that foam-oil-interaction
parameters in current foam models shift the two regimes. Depending on the model, the presence of oil
shifts pressure gradient (p) contours in the high-quality regime to right, indicating an increase in
limiting water saturation below which foam collapses. Oil can also shift p contours in the low-
quality regime upwards by reducing resistance to bubble flow. Therefore, our goal here is to
measure the effect of oil on foam quantitatively and gather laboratory data to see whether the
effect of oil can be represented by these models. Specifically, this study examines the effect of oil
on foam through its effect on those two regimes initially identified in the absence of oil, which are

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
already verified and well understood. A series of corefloods are carried out applying co-injection of
foam and oil, with oil injected at a fixed ratio of oil (Uo) superficial velocity to water (Uw) superficial
velocity. Through this injection strategy, the effect of flow oil can be quantified. To examine the
effect of oil composition on foam, for simplicity, two oil components are selected: one component,
which is benign to foam stability, and the other one, which can destroy foam completely. It is
expected that a model oil consisting of these two components, one the one hand, can destabilize foam
to some extent, but, on the other hand, can still maintain some foam strength. Varying the proportion
of these two oil components in the model oil allows one to examine the effect of oil composition on
foam stability in porous media.

Then, applying a method similar to that of Cheng et al. (2000), for the first time, we fit the widely
used STARS foam model parameters to experimental data for local-equilibrium (LE) foam flow in
presence of oil. This research provides a practical approach for measuring the effect of oil on foam in
porous media. Initial data on foam-oil interactions obtained in this study give further insights for
simulating foam for EOR in the presence of oil. Also, these data can serve as a case study for
representing the behaviour of foam in contact with oil, and enhance the reliability of foam process
design.

Foam models

Current foam models generally fall into two groups: population-balance models, characterizing the
dynamics of bubble creation and destruction, coupled with gas mobility as a function of bubble size
(Falls et al. 1988; Friedman et al., 1991; Kovscek and Radke, 1994; Kam et al., 2007); and implicit-
texture (IT) models, capturing effects of physical factors on steady-state foam flow, e.g. surfactant
concentration, water saturation, oil saturation, salinity, capillary number and so forth (Law et al.,
1989; Patzek and Myhill, 1989; Kular et al., 1989; Fisher et al., 1990; Islam and Farouq-Ali, 1990;
Mohammadi and Coombe, 1992; Cheng et al., 2000). This study focuses on fitting IT model
parameters to experimental data.

In the widely used STARS foam model (Computer Modeling Group, 2014), gas mobility is reduced
by modifying gas relative permeability through a mobility factor FM:

k rgf = k rg0 FM
(1)
1
FM =
1 + fmmob F1 F2 F3 F4 F5 F6 (2)

where fmmob, the reference gas mobility-reduction factor, is defined as the attainable maximum gas
mobility reduction, F1 through F6 are functions accounting for the physical effects on gas mobility of
surfactant concentration, water saturation, oil saturation, oil composition, capillary number, salinity
and etc. Here we consider three functions: F2, F3, and F5, capturing the effects of water saturation, oil
saturation and capillary number, respectively. The STARS foam simulator includes two algorithms
for the oil effect on foam: the wet-foam model and the dryout model.

In the wet-foam model, F2 represents the effect of water saturation on gas mobility
arctan( edry( S w fmdry ))
F2 = 0.5 +
(3)
where epdry controls the abruptness of foam collapses as water saturation decreases towards the
limiting water saturation, fmdry, below which foam collapses. The value of fmdry is usually assumed
to be a constant throughout high-quality regime.

In the wet-foam model, oil saturation scales fmmob by

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
 1 So floil
epoil
( fmoil So )
F3 = floil < So < fmoil
fmoil floil
0 fmoil So 1 S wc S gr
(4)
where fmoil and floil are oil-related parameters marking the boundary when oil destabilizes and
destroys foam, and epoil is the oil exponent.

F5, capturing shear-thinning behaviour in the low-quality regime, is given by

fmcap epcap
( ) N ca fmcap
F5 = N ca
1 else
(5)
kp
N ca =
wg
(6)
where fmcap and epcap are model parameters for shear-thinning, Nca is capillary number defined as a
product of absolute permeability k multiplied by pressure gradient divided by water-gas surface
tension.

In the dryout model, oil affects gas mobility not by scaling fmmob, but by controlling the limiting
water saturation sfdry. F2 in the wet-foam model is rewritten in the dryout model as F7, given by
arctan( sfbet ( S w sfdry )
F7 = 0.5 +
(7)
where sfbet and sfdry play the same roles of epdry and fmdry in wet-foam model, respectively.
Nevertheless, sfdry is not a constant in this model, but instead depends on oil saturation through
function G2 (similar to function F3 in the wet-foam model):
[(1 sfdry ) G2 + sfdry ] sfdry (8)
0 S o sloil

S sloil efoil
G2 = ( o ) sloil < S o < sfoil
sfoil sloil

1 sfoil S o 1 S wc S gr
(9)
where oil-related parameters sloil, sfoil, efoil correspond to floil, fmoil and epoil in the wet-foam
model, respectilvely.

The wet-foam model represents only the effect of oil on the low-quality regime, while the dryout
model captures only the effect of oil on the high-quality regime (Tang et al., 2016). The experimental
data shown below demonstrate clearly which regimes are affected by oil. The specific procedures for
fitting parameters of the foam models above are elaborated below.

Experimental method

Experimental scheme

To examine the effect of oil on local equilibrium (LE) foam flow in cores, first a series of
pure hydrocarbons were screened using a bulk-foam test and coreflood evaluation, respectively, so
as to select two oil components with a range of effects on foam. In particular, we wanted one
component
IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery
24-27 April 2017, Stavanger, Norway
that destabilizes foam slightly and another that substantially destroys foam. Then, modifying the
proportion of these two components provides insight on the effect of oil composition on foam
stability. Bulk-foam tests here were used to roughly characterize the defoaming potential of a variety
of pure alkanes, in terms of surfactant precipitation, initial foam height and foam decay over time.
This method can give a quick screening of the detrimental effect of those hydrocarbons examined.
Since foam behaviour in bulk is not always consistent with that in rock, several candidates potentially
meeting our expectations were examined again with corefloods for mobility-reduction factor (MRF).
MRF is defined here as a ratio of pressure drop with foam to that with only brine injection at the same
total superficial velocity:
p fm
MRF =
p nf
(10)

where pfm and pfm denote pressure drop with foam and brine injection, respectively.

Two oil components were then selected combining the evaluation results of the bulk-foam tests and
corefloods. Before carrying out foam corefloods in the presence of oil, a set of experiments with only
foam injection were conducted to identify the two regimes. These two regimes serve as a reference for
comparing the effect of oil. During corefloods, instead of monotonically increasing or decreasing gas
(Ug) or water (Uw) superficial velocity, Ug and Uw were varied in a random sequence to avoid
confusing absolute permeability loss with the effect of foam injection. Thereafter, a series of foam
corefloods were conducted with oil co-injected concurrently, to examine steady-state interaction of
foam and oil in porous media. Specifically, to measure the oil effect quantitatively and ensure steady
state, foam was co-injected with oil at a fixed ratio R of oil (Uo) to water (Uw) superficial velocity.
The variation in Uw therfore means the same change in Uo, which can qualitatively illustrate the effect
of flowing oil on foam stability. In our experiments, we didnt measure oil or water saturation, so they
were estimated through relative-permeability functions. Therefore, it is not practical here to quantify
directly the effect of oil saturation on foam. Varying the proportion of the two oil components
selected allows one to investigate the effect of oil composition on the two regimes central to
understanding of foam without oil.

Apparatus and materials

The apparatus used, as illustrated in Fig. 2, consists of five systems along the flow path:
injection system, core-holder system, data-acquisition system, back-pressure-regulating system,
and effluent-collection system. The injection system here allows for co-injection of three phases
(water, oil, and gas) through two pumps for oil and surfactant solution, respectively, and a gas
mass-flow controller. The core-holder system is installed with 5 pressure taps monitoring
pressure drop of each section along core during corefloods. The chamber between the core and the
core holder is connected with the inlet line to impose confining pressure. The whole core-holder
system is placed in a water bath maintained at 35 C. Meantime, the data monitored by pressure
transducers are recorded in the data-acquisition system through LabVIEW. The back-pressure-
regulating system allows us to carry out experiments under elevated pressure and maintains
stable outlet pressure. Produced effluent is collected through the collection system.

All corefloods conducted here were carried out in a Bentheimer core 4 cm in diameter and 40 cm in
length. The foaming agent used is AOS BIO-TERGE AS-40K, by Stepan company. NaCl is the
only salt used to provide salinity. To screen the oils for experimental study, pure hydrocarbons and
oleic acid (OA) are examined, e.g. hexadecane (C16), tetradecane (C14), dodecane (C12), decane
(C10), nonane (C9), octone (C8) and hexane (C6). The gas applied is nitrogen.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
Figure 2 A schematic overview of apparatus used for steady-state foam corefloods, allowing for co-
injection of oil, water and gas concurrently.

Results

Oil screening

In the bulk-foam test, 5 ml of surfactant solution is mixed in a 10 ml test tube with 1 ml of

different pure alkanes, e.g. C16, C14, C12, C10, C9, C8 and C6. The surfactant used is AOS at a
concentration of 0.5 wt%, with a salinity of 30,000 ppm NaCl. Initially, these test tubes were kept
in an oven at 35C to check if surfactant precipitates in contact with these alkanes. Foam then is
generated by manually shaking these test tubes. Comparing foam decay over time with and
without oil, in terms of foam height and texture, shows that hexadecane slightly destabilizes
foam; after 1010 mins foam height decays by less than half but foam texture is coarse. However,
for all the rest of the alkanes examined, foam completely collapsed after 230 mins and the fewer the
carbon atoms on the backbone, the more quickly foam decayed.

Given the concern about inconsistency between foam behaviour in bulk and in rock, C16, C10, C8 and
C6 were examined again through corefloods in a Bentheimer sandstone. The mobility-reduction factor
with and without oil in Fig. 3 shows that hexadecane C16, consistent with the bulk-foam test, is
relatively benign to foam stability. However, C10, C8 and C6 are not so effective at destroying foam in
porous media, though they are detrimental to foam stability in bulk. Oleic acid (denoted as OA
below), however, destroys foam almost completely.

Taking the evaluation results of both bulk and corefloods into account, hexadecane and oleic acid
were chosen as the two oil components: hexadecane as the one benign to foam stability, and oleic acid
as the one destroying foam completely. Figure 4 shows that foam strength decreases greatly with
increasing concentration of oleic acid in the model oil. In order to destroy foam to some extent but
still maintain some foam strength, model oils with 10 wt% and 20 wt% OA were investigated.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
 1000
937

766
750 686 703
611

491

MRF
500

250

11
0
0.9 0 0 0 0 0 0
5 fo .95 fo .95 fo .95 fo .95 fo .99 fo .95 fo
am am am am am am am
wit w w w w w w
hou ith C ith C ith C ith C ith C ith m
t oi 16 10 8 6 6 od e
l l oil
Figure 3 Gas mobility-reduction factor as a function of oil. Oil is co-injected with foam at a constant
ratio of Uo to Uw in a Bentheimer core of 1.98 darcy. Total superficial velocity of Uw and Ug is fixed
at 1.5 m/D. Foam quality in this case is defined as fg=Ug/(Ug+Uw). Model oil used consists of 10%
oleic acid and 90% C16.

600
Foam apparent viscosity, cp

500

400

300

200

100

0
0 5 10 15 20 25
Oleic acid concentration in model oil, wt%
Figure 4 The effect of oleic acid concentration on foam strength at a foam quality of 0.95 in a
Bentheimer core of 1.98 darcy. Foam quality in this case is defined as Ug/(Ug+Uw). Total superficial
velocity of Uw and Ug is fixed at 1.5 m/D. Oil is co-injected with foam at a constant ratio of Uo to
Uw.

Two foam regimes with and without oil

LE foam flow without oil

Figure 5 illustrates that steady-state foam flow in the absence of oil in a Bentheimer core again
shows two regimes as expected. The high-quality regime appears at foam qualities above
approximately 80%. The nearly vertical contours in this regime indicate independence of p from
Ug. Pressure gradient increases from about 100 to 500-600 psi/ft upon doubling Uw from 0.13
to 0.25 ft/D, indicating strongly shear-thickening rheology in the high-quality regime.

However, the p contours in the low-quality regime in Fig. 5, instead of staying horizontal as in Fig.
1, tilt upward to the top right. This means that p in the low-quality regime decreases with
increasing Uw at fixed Ug. The studies of de Vries and Wit (1990) and Kim et al. (2005) on CO2
foam in a variety of porous media suggest similar behavior in the low-quality regime. The study of

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
Hirasaki and Lawson (1985) on foam apparent viscosity in smooth capillaries and the study of Kim et
al. on LE CO2 foam suggest that this behavior may arise from easier flow of bubbles under wetter
conditions. In particular, increasing Uw in the low-quality regime probably produces an increase in
water saturation, implying a thicker water film around bubbles and therefore less flow resistance. This
consequently yields lower p at higher Uw as Ug is fixed. More efforts are still needed to examine the
exact mechanism for such upward-tilting p contours in the low-quality regime.

7 7

6 6
B B B
B
600 B
600
600 550
5 B
5 B
600

B450B B 250450 B
350
400 550
4 4
Ug , ft/D

U g , ft/D
500
B
400 500 B
350
300BB 550 B
B
BB B B
450
BB B B B
3 B 3 500 B
B B
B
550
BB
100 B B 400 B B
B B B
150 B
200 500
2 250 B B B B 2 100 450 B
B B B B
350 150300 400
B B B B B B B B
200 350
1 B B 300 1 B B
B B B B B250 B

0 0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Uw , ft/D Uw , ft/D

Figure 5 (left) Pressure gradient (psi/ft) in the absence of oil as a function of gas (Ug) and liquid (Uw)
superficial velocities (ft/D) at 35C in a Benteimer core of 1.98 darcy. Surfactant used is AOS at a
concentration of 0.5 wt%. Each filled symbol here represents a steady-state measurement.
Figure 6 (right) Expanded view of Fig. 5, with representative contour at 300 psi/ft (red dotted lines)
used in fitting model to data.

LE foam flow with C16

Figure 7 shows that steady-state concurrent foam-oil flow (C16 in this case) still produces the two
flow regimes as identified in the absence of oil: high- and low-quality regimes. The involvement of
oil greatly complicates foam-flow processes. Although hexadecane only slightly destabilizes foam in
the bulk-foam tests, apparently both high- and low-quality regimes are affected by the presence of oil.
For the same range of Ug and Uw examined, p in the presence of C16 in Fig. 7 is restricted below
about 220 psi/ft, nearly three times lower than that without oil (600 psi/ft) in Fig. 5.

The high-quality regime in Fig. 7 shifts to right for the same p compared to Fig. 5. This suggests that
oil weakens foam stability in the high-quality regime by increasing the limiting water saturation (Sw*)
below which foam collapses. As shown in Fig. 8, the uniform spacing of p contours in this regime
demonstrates nearly Newtonian rheology, reflecting an approximately constant Sw*. The low-quality
regime in Fig. 7 shifts to lower foam qualities compared with Fig. 5 (approximately below 0.75). p
in the low-quality regime is about a factor of two less than in Fig. 5, less of a reduction than in the
high-quality regime. Also, p contours in the low-quality regime of Fig. 7, similar to those in Fig. 5,
also tilt upward with increasing Uw. Since foam in this case is flowing together with oil, the upward-
tilting p contours may reflect either easier flow of bubbles under wetter conditions, or the
destabilizing effect of oil on foam strength, or a composite effect of both.

The greater decrease in gas-mobility reduction in the high-quality regime than in the low-quality
regime reveals that foam stability in the high-quality regime is more vulnerable to oil. For instance, in
the high-quality regime of Fig. 7 with C16, (Uw, Ug) at (0.2, 2) gives a p of 160 psi/ft, 2.5 times lower
than the p obtained with the same set of flow rates in Fig. 5 without oil, 400 psi/ft. In the low-quality
regime, for the same set of flow rates, e.g. at (1, 1.5), the p in Fig. 7 yields out 190 psi/ft, nearly 1.9
times lower than 360 psi/ft obtained in Fig. 5.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
 7 7

B B
6 107 6 107
B B
201 201
5 5
120

4 B 4 B
Ug , ft/D

Ug , ft/D
104 201 104 201
B B
B
193 B
193
B 120 200 B
3 206 3 206
194 140 194
B B B B 160 B B B
200 217
204 180 204 200
2 160
180 2
100 B 200 140 100 B
204 180 204
93 140 160 93
1 B BBB220 200 B B B 1 B B B B B
181 237 198 196 158 144 181 196
237 198
0 0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
U w , ft/D Uw , ft/D

Figure 7 (left) Pressure gradient (psi/ft) in presence of oil as a function of gas (Ug) and liquid (Uw)
superficial velocities (ft/D) at 35C in a Benteimer core of 1.98 darcy. Oil used is hexadecane, co-
injected with foam at a fixed ratio R=0.25 of Uo to Uw. Surfactant used is AOS at a concentration of
0.5 wt%. Each filled symbol represents a steady-state measurement.
Figure 8 (right) Expanded view of Fig. 7, with representative contour at 160 psi/ft (red dotted lines)
used in fitting model to data.

LE foam flow with 10 wt% OA in model oil

Introduction of oleic acid, though only 10 wt% in the model oil, plays a significant role on foam
behavior and leads to some new behavior. Figure 9 illustrates that only the high-quality regime
appears for the same range of Uw and Ug as examined in Fig. 5 and Fig. 7. The generally vertical trend
of p contours in the high-quality regime of Fig. 9 suggests that Sw* is still independent of Ug.

The high-quality regime in Fig. 9, compared to that in Fig. 5 and Fig. 7, extends to much higher range
of Uw. This suggests that the introduction of OA greatly increases Sw*, weakening foam to a much
larger extent than C16. But the p contours in this regime reveal non-Newtonian behavior: p depends
on Uw, but not linearly. Based on Darcys Law, the non-linear dependence of p on Uw in the high-
quality regime indicates that the Sw*, usually represented as a constant throughout the high-quality
regime in the absence of oil, rises with increasing Uw in presence of oil. However, since oil and foam
here are co-injected at a fixed ratio of Uo to Uw. increasing Uw also means increase in oil saturation,
suggesting a greater destabilizing effect. The decrease in apparent viscosity with increasing total
superficial velocity may reflect either shear-thinning rheology or destabilizing effect of oil on foam.

Fig. 9 also illustrates that the high-quality regime in the presence of 10 wt% OA extends down to
much lower foam qualities (transition foam quality fg*< 0.44), compared with Fig. 5 and Fig. 7. This
suggests a greater destabilizing effect of oil on foam than with C16. For instance, (Uw, Ug) at (0.2, 2) in
Fig. 9 yields a p of 31 psi/ft, nearly 13 times lower than p obtained for the same set of flow rates in
Fig. 5.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
 7 B B B
7 B B B
38 32 41 B 38 32 41 B
53 53
6 6

5 5

4 B 4 B
U g , ft/D

Ug , ft/D
55 B 55 B
40 72 40 72
23 31 50 23 31 50
3 B B B B
B 60 70 B 3 B B B B B 60 70 B
30 41 48 52 73 30 41 48 52 73

2 2

1 20
B B B B B B 1 20
B B B B B B
20 36 38 46 53 67 20 36 38 46 53 67

0 0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
Uw , ft/D Uw , ft/D

Figure 9 (left) Pressure gradient (psi/ft) in presence of oil as a function of gas (Ug) and liquid (Uw)
superficial velocities (ft/D) at 35C in a Benteimer core of 1.98 darcy. Model oil used consists of 10%
oleic acid and 90% hexadecane, co-injected with foam at a fixed ratio R=0.25 of Uo to Uw. Surfactant
used is AOS at a concentration of 0.5 wt%. Each filled symbol represents a steady-state measurement.
Figure 10 (right) Contours of fixed p from Fig. 9 used in fitting model to data. Each red-dotted line
represents a specific p contour across the high-quality-regime data.

7 BBB B 7 B BB B
7 913 19 7 913 19
10 30 10
6 6 30
20 40 20 40
5 5

4 60 4 60
U g , ft/D

U g , ft/D

B B B B B B
30 50 52 77 30 50 52 77
9 9
3 BBB B B B B 3 B BB B B B B
7 13 20 32 58 79 7 13 20 32 58 79
70 70
2 2
10 30B 10 30B
1 B B B B 20
40 B
60 B 1 B B B B 20
40 B
60 71B
6 9 11 18 31 51 50 71 6 9 11 18 31 51 50
40 40
15 B B16 B 30 B 15 B B 16 B 30 B
0 22 30 0 22 30
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
Uw , ft/D Uw , ft/D
Figure 11 (left) Pressure gradient (psi/ft) in presence of oil as a function of gas (Ug) and liquid (Uw)
superficial velocities (ft/D) at 35C in a Benteimer core of 1.98 darcy. Model oil used consists of 20%
oleic acid and 80% hexadecane, co-injected with foam at a fixed ratio R=0.25 of Uo to Uw. Surfactant
used is AOS at a concentration of 0.5 wt%. Each filled symbol represents a steady-state measurement.
Figure 12 (right) Contours of fixed p from Fig. 11 used in fitting model to data. Each red-dotted
line represents a specific p contour across the data in each regime.

LE foam flow with 20 wt% OA in model oil

Fig. 11 shows the two foam-flow regimes with 20 wt% OA. The data are extended to lower Ug than
with 10 wt% OA in Fig. 9. Both high- and low-quality regimes show modestly shear-thinning
behavior. However, similar to Fig. 9, the high-quality regime in Fig. 11 shifts to the right and
extends to a higher range of Uw. The low-quality regime shrinks to much lower foam qualities (below
approximately 0.2). The increase in the concentration of the detrimental component in the oil
destabilizes foam to a greater degree.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
The maximum p achieved in Fig. 11 is about 13.3 times lower than that without oil in Fig. 5 for
same range of flow rates examined, demonstrating again that oil affects both the high-quality and low-
quality regimes. The more detrimental effect of oil on foam in Fig. 11 clearly suggests that the high-
quality regime bears greater vulnerability to oil. For instance, p at (2, 0.7), in the low-quality regime
of Fig. 11, is approximately 40 psi/ft, about 6.8 timers lower than that in the absence of oil, 271 psi/ft
in Fig. 5. But p in the high-quality regime of Fig. 11, e.g. 9 psi/ft at (0.2, 2), is nearly 40 times lower
than that in Fig. 5 (400 psi/ft) for the same pair of superficial velocities.

All the experimental conditions are exactly the same in Fig. 9 and Fig. 11 except the oil composition.
10 wt% OA in the model oil causes approximately a 2-3 times decrease in gas mobility reduction, e.g.
9 psi/ft in Fig. 11 v. 31 psi/ft in Fig. 9 at (0.2, 2), or 30 psi/ft in Fig. 11 v. 67 psi/ft in Fig. 9 at (1.25,
1). This suggests that oil composition play as significant a role as oil saturation on foam stability.
Absence of data on oil saturation throughout our experiments makes it difficult to quantitatively
compare the effect of oil saturation to that of oil composition on those two regimes. But co-injecting
oil and foam at a fixed ratio of Uo to Uw indicates that lower Uw, qualitatively, also means lower oil
saturation, suggesting less destabilization of foam.

Fit of model parameters to data

Simulating foam EOR processes in contact with oil confronts many challenges (Farajzadeh et
al., 2012): for instance, which regimes are affected by oil and how oil affects those two regimes.

Our experimental data demonstrate that oil affects both regimes but to a different magnitude.
Specifically, oil affects the high-quality regime through its effect on limiting water saturation and the
low-quality regime through its effect on gas-mobility reduction. Fitting foam model parameters to
these data provides a reference for simulating foam EOR processes with oil concurrently flowing.
Here, we present a method similar to that of Cheng et al. (2000) to fit model parameters to data of
Figs. 5, 7, 9, and 11. Our method includes shear-thinning in the low-quality regime and the effect of
oil on both regimes.

Procedure for fitting model parameters to data

The functions examined in the foam model include the following parameters: fmmob, fmdry
(sfdry), epdry (sfbet), epcap, fmcap, floil (sloil), fmoil (sfoil), and epoil (efoil). However, owing to
absence of oil-saturation data in our experiments, oil-related parameters cannot be fitted
effectively. Therefore, five parameters, in total, are fitted in this study: fmmob, fmdry (sfdry), epdry
(sfbet), epcap and fmcap. Although it is not practical to fit oil-related parameters owing to limited
data, comparing fmmob in the absence of oil to that incorating the effect of oil when foam
concurrently flows with oil gives the value of function F3. Similary, comparison of fmdry in the
absence of oil and sfdry altered by oil yields the value of function G2. The values of F3 and G2,
respectively, reflect the effects of oil on foam in the low-quality regime and foam stability in the
high-quality regime.

Reme (1999) and Cheng et al. (2000) have sketched the parameter-fitting procedures for the STARS
and UT foam models (Rossen et al., 1999), respectively, without including the effect of oil. Here we
use a method similar to that of Cheng et al. to fit model parameters to data, including the effects of
oil on fmmob and fmdry (sfdry). Before fitting the parameters, one must specificy foam-free
relative-permeability functions for each phase. Owing to absence of relative-permeability data for
the specific Bentheimer core, relative-permeability functions based on data of Persoff et al. (1991)
are applied in this study. Table 1 summarizes the fluid and transport properties used throughout
the parameter-fitting.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
Table 1 Fluid and transport properties in porous media
Quantities Parameter values
Water viscosity, w 0.710-3 Pas
Gas viscosity, g 2.0710-5 Pas
Oil vicosity, o 510-3 Pas
Water relative permeability, krw(Sw) S w 0.2 4.2
0.2( )
0.6 Sor
Gas relative permeability krg(Sg) S g 0.2
0.94( )1.3
0.6 S or
Oil relative permeability kro(So) S S or 2
0.5( o )
0.6 S or
Water-gas surface tension, wg 0.03 N/m
Temperature, T 35 C
When oil is present, Sor=0.1, otherwise, Sor=0

Specifically, one can fit model parameters to data following the preceure step by step as described
below:

Step 1. Draw p contours. For efficiently fitting parameters, one may need to smooth the p contours.
Then, pick up one representiave p contour and draw a vertical and horizontal straight line, which
can, respectively, best represent the data in high-quality and low-quality regimes for this contour, as
shown in Fig. 13.

This approach assumes the p contours in low-quality regime are horizontal. However, in some cases,
including this study, p contours in the low-quality regime tilt upward with increasing Uw (Figs. 5 and
6) (Kim et al., 2005). This means that fitted results could mismatch the low-quality regime data,
especially for higher ranges of Uw. However, none of current foam simulation models can capture this
behavior in the absence of oil. Therefore, it is necessary to improve current foam models to capture
this upward-tillting p trend for effectively fitting low-quality regime data. That is not in the scope of
this study. We apply horizontal contours to represent low-quality-regime data throughout the model-
fitting. Figs. 6, 8, 10, and 12 illustrate the contours drawn through data in both regimes in each case.

.
Figure 13 Schematic of contour treatment in each regime for fitting model parameters to data.

Step 2. Determine limiting water saturation Sw*. The vertical line in Fig. 13, representing the high-
quality regime data for the representative p contour, corresponds to a constant Uw. Applying Darcys
law for water phase (Eq. 11) at this value of Uw, along with krw(Sw) from Table 1, yields the limiting
water saturation Sw*:

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
 kk rw ( S w* )
uw = p
w (11)
Step 3. Calculate reference mobility recduction factor fmmob* (* denotes the factor without
considering shear-thinning). As shown in Fig. 13, the vertical and the horizontal lines intersect at one
point, which marks the transition between regimes. Foam quality at this point, defined as Ug/(Ug+Uw)
in the absence of oil or Ug/(Ug+Uw+Uo) in the presence of oil, is defined as the transition foam quality
fg*. Parameter fmmob* is assumed to be constant throughout the low-quality regime. Therefore this
assumption applies also at fg*. Applying Darcys law to the gas phase (Eq. 12) at fg* using krg(Sg) from
Table 1 gives fmmob*.
1
kk rg0 ( S w* )
1 + fmmob *
ug = p (12)
mw
where k rg0 ( S w* ) is the foam-free gas relative permeability at fg*, at which Sw* is already known from
step 2. The value of fmmob* obtained here neglects shear-thinning in the low-quality regime. This is
considered in step 7.

Step 4. Set epdry to an acceptable large value. Parameter epdry controls the abruptness of foam
collapse as Sw decreases towards Sw*. Consistent with experimental data, e.g. Fig. 5 or Fig. 11, an
abrupt transition is assumed as in the model-fitting of Cheng (2000).

Step 5. Specify reference capillary number fmcap. As shown in Eq. 5, no matter what value of fmcap
is used in models, it has to keep F5 less than 1 for shear-thinning to come into effect. Parameter fmcap
in this study is calculated using Eq. 6, based on a value of p smaller than all measured data in each
case.

2.85
Pressure gradient, psi/ft (log)

2.80 Uw fixed at 0.75 ft/D

2.75
2.70
2.65
2.60 y=0.3535x+2.5174
2.55
2.50
2.45
2.40
-0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Ug, ft/D (log)
Figure 14 log-log plot for p v. Ug based on data of Fig. 5 at fixed Uw=0.75 ft/D.

Step 6. Capture shear-thinning in the low-quality regime through epcap. The model proposed by
Rossen et al. (1999) characterizes shear-thinning in the low-quality regime by a parameter less than
1. The slope of a plot of log(p ) v. log(U g ) is . Parameter epcap here can be obtained through Eq.
13:
1
=
1 + epcap (13)

Plotting log(p) v. log(Ug) based on data of Fig. 5 at Uw=0.75 ft/D, as illustrated in Fig. 14, gives a
slope of 0.3535. Then, epcap is 1.83 from Eq. 13.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
Step 7. Adjust the reference mobility-reduction factor fmmob accounting for shear thinning. When
considering shear-thinning in the low-quality regime, the value of fmmob* obtained above is the
mobility reduction that incorporates the effect of shear thinning. Model parameter fmmob can be
obtained from fmmob* through Eq. 14, using fmcap and epcap obtained in steps 5 and 6, based on p
of the representative p contour:
fmcap epcap
fmmob * = fmmob ( )
N ca (14)
where Nca is calculated from Eq. 6 using the value of p from the characteristic p contour.

Step 8. Fit the effect of oil through functions F3 and G2. As noticed in Figs. 7, 9 and 11, oil impacts
both high- and low-quality regimes, respectively, through its effect on Sw* and reference-mobility-
reduction factor fmmob in each regime. Experimetal data here suggest that oil has different
magnitudes of effects on those two regimes, with the high-quality regime more susceptible to its
influence. Therefore, oil-related parameters should be fitted seperately based on data in each regime.

Specifically, when oil is present, steps 2 and 3 give Sw* and fmmob*, each of which now incorporates
the effect of oil. Based on the fitted values of Sw* and fmmob* in Fig. 5 obtained in the absence of oil,
applying Eq. 8 yields the value of G2 in the dryout model, accounting for the final effect of oil on the
limiting water saturation. Parameter fmmob* obtained in the presence of oil divided by the value of
fmmob* obtained in the absence of oil, gives the value of F3 in wet-foam model, accouting for the oil
effect on gas mobility reduction in the low-quality regime. However, the non-linearly distributed
contours in the high-quality regime of Fig. 9 and Fig. 11 suggest that Sw* is not fixed throughout the
high-quality regime, but varies with oil saturation. Therefore, in order to fit the high-quality-regime
data in presence of oil, one may have to fit Sw* corresponding to each p contour in the high-quality
regime, instead of just fitting one representative p contour. Nonlinear spacing of contours in the low-
quality regime is fit by adjusting epcap for oil, but observed behavior may reflect both shear-thinning
and the effect of oil. The fit to epcap in the absence of oil may not apply when oil flows with foam.

Step 9. Plot fitted results using model parameter values. The experimental data demonstrate clearly
that oil shifts both regimes. However, the modeling study of Tang et al. (2016) suggests that the oil in
the wet-foam model affects only the low-quality regime, while in the dryout model, it shifts only the
high-quality regime. Therefore, the wet-foam model and dryout model need to work together to
represent low- and high-quality-regime data.

Results of parameter-fitting

Table 2 summarizes the fitted parameter values, with and without accounting for shear thinning, to
the data of Figs. 5, 7, 9 and 11, applying the fitting procedure described above. Figs. 6 and 8 illustrate
the representative p contour for 300 psi/ft with fg* at (0.17 ft/D, 0.75 ft/D) from Fig. 5 and for 160
psi/ft with fg* at (0.2 ft/D, 0.9 ft/D) from Fig. 7, respectively. Figs. 10 and 12 illustrate the contours
plotted from Figs. 9 and 11, particularly for fitting Sw* to each p contour. The mobility-reduction
factor for the case with 20% OA in Fig. 12 is fitted based on the 30 psi/ft contour with fg* at (1.17,
0.45). The superscript of fmmob* indicates that it excludes shear-thinning, while fmmob is the
reference mobility-reduction factor adjusted for shear-thinning. When oil is introduced, both
fmmob* and fmmob incorporate the effects of oil. Model parameter epdry/sfbet is specified to give
an abrupt transition between regimes. fmcap in Figs. 5 and 7 is based on the contour for 50 psi/ft, and
in Figs. 9 and 11 on the contour for 3 psi/ft. Fitted results below are obtained using the same value of
epcap value based on the data of Fig. 5, accounting for shear-thinning in the low-quality regime. In
other words, in our initial fit we assume that oil doesnt alter the shear-thinning nature of the low-
quality regime.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
Table 2 Overview of fitted parameter values using STARS foam model for all the investigated studies
Cases Fitted parameters
* * epdry/
fmmob F3 fmmob F3 fmdry /sfdry epcap fmcap
sfbet
Without
1.94105 5.10106 0.274 2.00104 1.83 7.5010-5
oil
C16 1.09105 0.563 9.08105 0.178 0.275 2.00104 1.83 7.5010-5
0.309 at 20 psi/ft
0.309 at 30 psi/ft
Low-quality regime doesnt appear for the 0.313 at 40 psi/ft
10% OA 2.00104 1.83 4.4810-6
range of Uw and Ug examined 0.319 at 50 psi/ft
0.330 at 60 psi/ft
0.339 at 70 psi/ft
0.334 at 7 psi/ft
0.354 at 10 psi/ft
0.360 at 20 psi/ft
0.370 at 30 psi/ft
20% OA 2.24104 0.115 1.51106 0.296 2.00104 1.83 4.4810-6
0.375 at 40 psi/ft
0.378 at 50 psi/ft
0.380 at 60 psi/ft
0.381 at 70 spi/ft
Model fit assumes large value of epdry/sfbet; here, we use a value of 2104.

Figures 15 and 16 show the model fit in the absence of oil. In the presence of oil, Figs. 17 through 22
are plotted using both the dryout model and the wet-foam model, capturing respectively the effects of
oil on the high- and low-quality regimes. The model fit without oil or with oil slightly destabilizing
foam, as C16 does here, assuming a constant Sw* throughout the high-quality regime and including
shear-thinning in the low-quality regime, gives good agreement with data. However, after oil that is
substantially detrimental to foam is introduced, this fit doesnt show a good match to data. Instead, a
better fit to data is obtained, with Sw* fitted respectively to each contour in the high-quality regime and
excluding shear-thinning in the low-quality regime. In other words, the presence of these oils (10 wt%
and 20 wt% OA) changes the extent of shear-thinning in the low-quality regime.

Figure 15 (left) Model fit to data of Fig. 5 in the absence of oil, excluding shear-thinning in the low-
quality regime: fmmob*=1.94105, fmdry=0.274, epdry=2.00104.
Figure 16 (right) Model fit to data of Fig. 5 in the absence of oil, including shear-thinning in the low-
quality regime: fmmob=5.10106, fmdry=0.274, epdry=2.00104, epcap=1.83; fmcap=7.5010-5.

Specifically, Figs. 15 through 18 illustrate that the high-quality regime fitted assuming a constant Sw*
throughout this regime shows good agreement with the data of Figs. 5 and 7, respectively, although
not representing the shear-thickening behaviour in Fig. 6. Figure 15 (excluding shear thinning) and
Fig. 16 (including shear thinning) clearly demonstrate that considering shear-thinning in the low-

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
quality regime gives much better match to the data of Fig. 5 in the absence of oil. The same
conclusion can also be drawn for model fit to data of Fig. 7 in presence of C16, which is slightly
detrimental to foam. However, in both cases, either with or without oil, fitted results fail to capture the
upward-tilting p contours in the low-quality regime, greatly overestimating the p in this regime.
Therefore, this issue requires improvement of the current models to represent these contours in the
low-quality regime.

Figure 17 (left) Model fit to data of Fig. 7 in the presence of C16, excluding shear-thinning in the low-
quality regime: fmmob*=1.09105, sfdry=0.275, sfbet=2.00104.
Figure 18 (right) Model fit to data of Fig. 7 in the presence of C16, including shear-thinning in the
low-quality regime: fmmob=9.08105, sfdry=0.275, sfbet=2.00104, epcap=1.83; fmcap=7.5010-5.

Figure 19 shows that when 10 wt% OA is introduced, the high-quality regime fitted using a fixed Sw*
completely fails to represent the data of Fig. 9. It suggests that the assumption of a constant Sw*
throughout the high-quality regime is not applicable there, in particular when oil effect is strong
enough. Therefore, Sw* was fitted respectively to each p contour in Fig. 10. The model (Fig. 20)
shows much better agreement with the data than Fig. 19, suggesting that Sw* here is now a function of
oil saturation or p. Owing to absence of low-quality-regime data in Fig. 9, fitted results in this
regime of Figs. 19 and 20 do not reliably represent the low-quality regime.

Figure 19 (left) Model fit to data of Fig. 9 with 10 wt% OA in model oil based on a representative p
contour of 30 psi/ft, excluding shear-thinning in the low-quality regime: fmmob*=3.00104,
sfdry=0.309, sfbet=2.00104.
Figure 20 (right) Model fit to data of Fig. 9 with 10 wt% OA in model oil, excluding shear-thinning in
the low-quality regime: fmmob*=3.00104, sfbet=2.00104. Table 2 gives limiting water saturation
sfdry corresponding to each p contour in the high-quality regime.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
The model fit in Fig. 21 also assumes a fixed Sw*, showing again a great mismatch to the high-quality-
regime in Fig. 11. Nevertheless, Fig. 22, fitting Sw* based on each specific p contour in Fig. 12,
agrees well with the data, suggesting again that Sw* is a function either of p or of some other
property, such as oil saturation, that depends on p. Surprisingly, the model fit in the low-quality
regime including shear-thinning fit to data without oil as illustrated in Fig. 21, does not match the data
of Fig. 11. Excluding shear-thinning in this regime as shown in Fig. 22, gives much better agreement
to the data of Fig. 11 than does Fig. 21.

Figure 21 (left) Model fit to data of Fig. 11 with 20 wt% OA in model oil based on a representative
p contour of 30 psi/ft, assuming shear-thinning in the low-quality regime fit to data without oil:
fmmob=1.51106, sfdry=0.370, sfbet=2.00104, epcap=1.83; fmcap=4.4810-6.
Figure 22 (right) Model fit to data of Fig. 11 with 20 wt% OA in model oil, excluding shear-thinning
in the low-quality regime: fmmob*=2.24104, sfbet=2.00104, epcap=1.83, fmcap=4.4810-6. Table 2
gives the values of limiting water saturation sfdry corresponding to each p contour in the high-
quality regime.

The model fit, as suggested in Figs. 19 through 22, shows varying Sw* (sfdry) in the high-quality
regime in the presence of OA, suggesting that Sw* may be a function of oil saturation (So) or of p.
Thus, next we consider the dependency of Sw* on oil saturation, which is represented in the STARS
model by Eqs. 8 and 9. Owing to absence of data, oil saturation here is calculated using the oil-
relative-permeability function in Table 1. For a case where Uo/Uw is fixed, illustrated by the solid
black and red curves in Fig. 23, all the results must lie on a curve defined by Uo/Uw=1/4, which is
determined by the relative-permeability functions for water and oil. Although the changes in Sw* are
modest, the corresponding changes in krw(Sw*) are large. The dryout model (Eqs. 8 and 9), with
constant parameters, gives a second curve (illustrated by the orange curve in Fig. 23) defining Sw* as a
function of So. The intersection of the two curves would define the high-quality regime. The range of
points in Fig. 23 suggests that the dryout function with fixed parameters does not explain our results.

Furthermore, a comparison of Figs. 9 and 11 suggests that 20 wt% OA in the model oil is more
detrimental to foam than 10 wt%. However, the data for 20 wt% OA (red curve in Fig. 23)
substantially overlap the data for 10 wt% OA (black curve). This could indicate either a failure of the
foam model or that the Corey relative-permeability function for oil phase doesnt reflect the true oil
saturation. If a more complex kro function is assumed, depending not only on oil saturation but also on
the other two saturations, there might no longer be a single trend like black or red curves shown in
Fig. 23.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
Figure 23 The dependency of limiting water saturation (Sw*) on oil saturation based on data of Figs.
9 and 11. Oil saturation on black and red curves is calculated using relative-permeability functions in
Table 1. Parameters used for illustration in the dryout model: sloil=0.1, sfoil=0.16, efoil=3.

Conclusions

This study provides an initial approach to simulating foam EOR using steady-state data for flow in the
presence of oil.

Experimental data show that the high- and low-quality regimes, central to understanding of foam
without oil, also appear for LE concurrent foam-oil flow. The presence of oil affects foam flow in
both regimes, with the high-quality regime more vulnerable to oil. Oil composition plays as
significant a role as oil saturation on LE foam flow.

Specifically, oil, in the high-quality regime, affects foam stability by increasing the limiting water
saturation (Sw*) below which foam collapses. In the low-quality regime, oil lessens gas mobility
reduction.

Either with or without oil, in some cases pressure gradient (p) decreases with increasing water
superficial velocity (Uw) at fixed gas superficial velocity (Ug). This may reflect either easier flow of
bubbles under wetter conditions in the absence of oil or an increasing destabilizing effect of oil on
foam with increasing Uw and Uo.

Without oil, foam flow in this study shows strongly shear-thinning behaviour in the low-quality
regime. This behaviour changes greatly when oil is present. However, foam rheology in the high-
quality regime in some cases shows Newtonian behavior, but in some others shear-thickening or
shear-thinning behaviour.

We present an approach of fitting LE foam model parameters to data for p as a function of Ug and
Uw, including shear-thinning in the low-quality regime and the effect of oil on both regimes.
Assuming a fixed Sw* throughout the high-quality regime and including shear-thinning in the low-
quality regime with or without oil doesnt give a good match to data, particularly when the oil effect is
strong enough. In such cases, one has to fit Sw* seperately to each p contour in the high-quality
regime and exclude shear-thinning in the low-quality regime to represent data in both regimes.

The variable Sw* shown in our model fits suggests that Sw* is a function of oil saturation or pressure
gradient. But it is not clear yet whether the dependency of Sw* can be represented by the dryout model
function owing to absence of oil-saturation data in our study. None of currently foam simulation
models can capture the upward-tilting p trend in the low-quality regime.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
Acknowledgements

This project is part of Joint Industry Project (JIP) on Foam for EOR. Authors would like to
acknowledge the sponsorship of sponsor companies, and particularly the technical support from
Michiel Slob for the laboratory experiments. Sebastien Vincent-Bonnieu also thanks Shell Global
Solutions International B. V. for permission to publish this article.

References

Alvarez, J. M., Rivas, H., and Rossen, W. R., "A Unified Model for Steady-State Foam Behavior at
High and Low Foam Qualities," SPE Journal 6 (2001), 325-333.
Computer Modeling Group, STARS User's Guide, Version 2015, Calgary, Alberta, Canada.
Cheng, L., Reme, A. B., Shan, D., Coombe, D. A., Rossen, W. R., "Simulating Foam processes at
High and Low Foam Qualities", SPE 59287 presented at the SPE/DOE Symposium on Improved
Oil Recovery (2000), Tulsa, Oklahoma, 3-5 April.
De Vries, A. S. and Wit, K., "Rheology of Gas/Water Foam in the Quality. Range Relevant to
Steam Foam. " Soc. Pet. Eng. Res. Eng. 5(2), 185-192 (1990).
Farajzadeh, R., Andrianov, A., Krastev, R., Hirasaki, G. J., and Rossen, W. R., "Foam-Oil Interaction
in Porous Media: Implications for Foam Assisted Enhanced Oil Recovery," Advances Colloid
Interface Sci., 183-184 (2012), 1-13.
Falls, A. H., Hirasaki, G. J., Patzek, T. W., Gauglitz, D. A., Miller, D. D., and Ratulowski, J,
"Development of a Mechanistic Foam Simulator: The Population Balance and Generation by
Snap-Off," SPE Reser. Eng. 3, 884-892 (1988).
Friedmann, F., Chen, W. H., and Gauglitz, P. A., "Experimental and Simulation Study of High-
Temperature Foam Displacement in Porous Media," SPE Reser. Eng. 6, 37-75 (1991).
Fisher, A. W., Foulser, R. W. S., and Goodyear, S. G., "Mathematical Modeling of Foam Flooding,"
SPE/DOE 20195 presented at the 1990 SPE/DOE Symposium on Enhanced Oil Recovery, Tulsa,
OK, Apr. 22-25.
Hirasaki, G. J. and Lawson, J. B., "Mechanisms of Foam Flow in Porous Media: Apparent Viscosity
in Smooth Capillaries," SPE Journal 25 (1985), 176190, SPE-12129-PA.
Islam, M. R. and Farouq-Ali, S. M., "Numerical Simulation of Foam Flow in Porous Media," J.
Canadian. Pet. Tech. (July-Aug. 1990) 47-51.
Kovscek, A. R. and Radke, C. J., Fundamentals of Foam Transport in Porous Media, in Foams:
Fundamentals and Applications in the Petroleum Industry, L.L. Schramm (ed.) ACS Advances in
Chemistry Series, Am. Chem. Soc., Washington, D.C. (1994) 3, No. 242.
Kam, S. I., Nguyen, Q. P., Li, Q., and Rossen, W. R., "Dynamic Simulations With an Improved
Model for Foam Generation," SPE Journal 12, 35-48 (2007).
Kular, G. S., Lowe, K., and Coombe, D., "Foam Application in an Oil Sands Steam Flood Process,"
paper SPE 19690 presented at the 1989 SPE Annual Tech. Conf. and Exhibition, San Antonio, TX,
Oct. 8-11.
Kim, J. S., Dong., Y., and Rossen, W. R., "Steady-State Flow Behavior of CO2 Foam," SPE Journal
10, 405-415 (2005).
Law, D. H., Yang, Z.-M., and Stone, T., "Effect of Presence of Oil on Foam Performance: A Field
Simulation Study," paper SPE 18721 presented at the 1989 SPE Symposium on Reservoir
Simulation, Houston, TX, Feb. 6-8.
Mohammadi, S., and Coombe, D. A., "Characteristics of Steam-Foam Drive Process in Massive
Multi-Zone and Thin Single-Zone Reservoirs," paper 14030, presented at the 1992 SPE Western
Regional Meeting, Bakersfield, CA, March 30-April 1.
Osterloh, W. T., and Jante, M. J., "Effects of Gas and Liquid Velocity on Steady-State Foam Flow at
High Temperature", SPE 24179 presented at the SPE/DOE EOR symposium (1992), Tulsa, OK,
April 22-24.
Patzek, T. W. and Myhill, N. A., "Simulation of the Bishop Steam Foam Pilot," paper SPE 18786
presented at the 1989 SPE California Regional Meeting, Bakersfield, Apr. 5-7.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway
Persoff, P., Radke, C. J., Pruess, K., Benson, S. M., and Witherspoon, P. A.: "A Laboratory
Investigation of Foam Flow in Sandstone at Elevated Pressure," SPERE (Aug. 1991) 185-192.
Reme, A. B.: "Parameter Fitting and Calibration Study with a Commercial Foam Simulator," Diploma
Thesis, Faculty of Applied Earth Sciences, Norwegian University of Science and Technology
(1999).
Rossen, W. R., Zeilinger, S. C., Shi,, J.-X., and Lim, M. T., "Mechanistic Simulation of Foam
Processes in Porous Media," SPEJ (Sept. 1999), 279-287.
Tang, J., Ansari, M. N., Rossen, W. R., "Modeling the Effect of Oil on Foam for EOR," ECMOR
XV, 2016.

IOR NORWAY 2017 19th European Symposium on Improved Oil Recovery

24-27 April 2017, Stavanger, Norway

View publication stats