SPE/DOESociety of

Petroleum Engineers
U.S. Department
of Energy

SPE/DOE 20202

Effects of Endpoint Saturations and Relative Permeability

Models on Predicted Steamflood Performance
M. Kumar* and T.N. Do, Chevron Oil Field Research CO.
·SPE Member

Copyright 1990, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the SPE/DOE Seventh Symposium on Enhanced Oil Recovery held in Tulsa, Oklahoma, April 22-25, 1990.

This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper,
as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society
of Petroleum Engineers. Permission to copy is restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain conspicuous acknowledgment
of where and by whom the paper is presented. Write Publications Manager, SPE, P.O. Box 833836, Richardson, TX 75083-3836. Telex, 730989 SPEDAL.

ABSTRACT Most commercial thermal simulators including

Chevron's SIS3 are three-dimensional, multi-phase,
A comprehensive numerical study was conducted to and compositional. Therefore, it is assumed here
quantify the effects of endpoint saturations and that simulator limitations do not include these.
relative permeabilities on steamflood performance Simulator parameters that do affect predictions
for a heavy oil reservoir. Results show that the include the finite difference scheme used to
region most important for performance prediction is represent the governing equations (i.e., 7-point
the high-temperature, gas (vapor)-oil relative vs. 9-point, etc.) and the use of parallel vs.
permeability near the residual oil saturation diagonal grid. Parallel grids are generally used
(S ). Furthermore, gas-oil relative permeabili- in thermal simulations 1 because they more closely
tig§g should be measured at steam temperatures predict the breakthrough times and they depict
because gas displacement occurs at steam condi- steam tongue observed in three-dimensional, scaled
tions. For the water-oil system, irreducible water laboratory models. 2 '6'7
saturation (S . ) has a greater effect on perfor-
mance than re~iaual oil saturation to water (S ) Reservoir geology and initial conditions, of
and the effect of temperature-dependent endp8i}ft course, strongly affect performance. These include
saturations is small. saturations (oil, water, and gas), permeability,
porosity, and their distributions. In addition,
The linear interpolation model (SPE 17369) for shale continuity and the presence of gas cap or
three-phase oil relative permeability is preferable aquifer also strongly influence performance. In
over the Stone's II model because it more general, geology is assumed to be known a priori
accurately predicts residual oil saturation to for simulation studies. Incorrect geological
steam. "Practical" endpoint oil saturations are representation can result in unpleasant surprises.
adequate for performance prediction because very
low values of oil relative permeability do not For a given simulator and known reservoir geology,
significantly affect recovery calculations. A 15% endpoint saturations and relative permeabilities
change in S , S . , and S resul ted in a 26, affect the performance most among the various input
13, and 7%orEhang~l.r in cumtffltive oil recovery, process modeling parameters. They affect cumula-
respectively. tive oil recovery, oil production rate, and
produced fluid ratios. Other modeling parameters
that can affect predicted performance include well-
INTRODUCTION bore fluid segregationS and grid size.
For specified operating conditions, steamflood Field measurement of residual oil saturation in the
performance predicted by numerical models is steamflooded zone are often lower than those
dependent upon (1) the numerical simulator used, measured in the laboratory. The higher residual
(2) reservoir geology and initial conditions, and oil saturation measured in the laboratory may be
(3) process modeling parameters. 1-2 Results are caused by inaccurate measurement or, more likely,
strongly affected by known process variations, such they may not be the "true" residual oil saturation.
as variation in steam quality, 3 water injection This is because residual oil saturations in the
after steam injection,4 and injection of inert gas laboratory are determined from displacement experi-
with steam. 5 ments, which are terminated when "practical"
endpoint saturations are reached, i.e., when the
References and illustrations at end of paper. oil expulsion rate (or oil relative permeability)
285
 EFFECTS OF ENDPOINT SATURATIONS AND RELATIVE
2 PERMEABILITY MODELS ON PREDICTED STEAMFLOOD PERFORMANCE
becomes small. A simulation study can assist in
quantifying the difference caused by "practical"
vs. "true" endpoint saturations.
It should be noted that, because of the high
viscosity of typical California heavy oil, accurate
measurement of even two-phase relative permeability
is difficult and is often affected by the experi-
mental procedure. This is especially true for
room- or low-temperature gas-oil relative permea-
bilities.
The use of relative perme",bility data is further
complicated for steamflood simulation because
three-phase relative permeabilities are required.
In the absence of measured data, empirical models
are used to obtain three-phase oil relative permea-
bility from two-phase data. Predicted results can
also vary with the three-phase model used. 9 Sato
and Azi z 9 showed the difference between oil
isoperms for Stone's I and II three-phase relative
permeability models. 1o - 12 They also determined the
effect of these models on steamflood recovery using
a two-dimensional reservoir simulation. Baker 13
showed that three-phase oil relative permeabilities
obtained by using a straight line or linear
interpolation between permeabilities at the two-
phase boundaries provided a better fit of the
experimental data.
Temperature dependence of endpoint saturations and
relative permeabilities is still an unresolved area
of research.l~-19 High temperature measurements
are difficult, and experimental data are limited
and contradictory. In addition, most of the
measurements for temperature effect have been
conducted for the water-oil system using synthetic
core and mineral oil. Even though many investi-
gators have reported that Sorw decreases and Swir
increases with temperature,14-16 others have found
that temperature does not affect endpoint satura-
tions. 17 - 19 Furthermore, temperature-dependent
endpoint saturations were believed to be an arti-
fact introduced by experimental procedures. is
In simulation studies, endpoint saturations are
often assumed to be temperature-dependent. Results
of this study can be used to estimate the effect of
temperature-dependent endpoint saturations, if they
exist, on predicted results.
In history matching of field performance, relative
permeability is often used as a matching
parameter. 20 However, no systematic study has been
conducted to identify and quantify which relative
permeability curve or endpoint saturation signifi-
cantly affects performance or how they affect
different portions of the production curve.
Results of this study will provide a more systema-
tic way for history matching.
In a recent study to evaluate field production data
and a comparison of the data with typical predicted
production curves,21 it was found that simulators
predicted a delayed production response and a sharp
production peak compared to field observations.
Although the difference could have been caused by
any of the three major parameters (simulator,
reservoir geology, and process modeling parameters)
listed earlier, this study was expected to identify
if the endpoint saturation and relative permea-
bility data used in previous studies could be one
of the causes.
286
SPE 20202
It is apparent that substantial useful results can
be obtained by a systematic study to determine the
effects of endpoint saturations and relative
permeabilities on predicted steamflood performance.
A comprehensive numerical simulation study was
conducted to quantify these effects. The specific
objectives of this study were to:
1. Quantify effects of endpoint saturations and
to determine relative permeability regions
that significantly affect predicted perfor-
mance;
2. Quantify the effect of temperature-dependent
endpoint saturations (in the water-oil system)
on performance;
3. Determine the effect of measurement accuracy
of endpoint ("practical" vs. "true") oil
saturations;
4. Determine the sensitivity of the results to
Stone's II vs. linear model for three-phase
oil relative permeability;
5. Determine if the relative permeability data
set used in previous studies was one of the
causes of the observed differences in the
shape of the field oil production curves and
earlier simulator predictions.
Results of this study can be used to conduct
laboratory experiments that focus on regions
important for predictions. They also provide a
systematic and more efficient method of history
matching when relative permeability is used as the
matching parameter.
RESERVOIR AND FLUID MODELS
Simulator
Chevron's fully implicit, compositional, three-
dimensional, steamflood simulator, SIS3, was used
in this study. It accounts for and rigorously
models important physical processes taking place
during steamflooding. It uses the seven-point
difference scheme.
Reservoir Grid
A three-dimensional model was used to represent the
symmetric element (one-eighth) of a repeated five-
spot pattern. The pattern area is 2.6 acres; the
distance between the injector and producer is
238.4 ft. The formation thickness is 100 ft. A
7x4x5 parallel grid system was employed for this
study as shown in Fig. 1. Refining the grid size
resulted in somewhat higher recovery; however, it
did not affect the relative results. Apex cells at
the three corners of the triangle were combined
with those adjoining, resulting in 22 active grid
blocks in each layer. The injector was open to the
bottom two layers (40% of the reservoir thickness).
The production well was open to the entire sand
interval.
Reservoir and Fluid Properties
The reservoir was assumed to be homogeneous for
this study· this allowed the senaraf"i nn nf nrnr'<><''''
SPE 20202 M. KUMAR AND T. N. DO
parameter effects from reserV01r geology. Table 1
lists the rock and fluid properties used in the
model. The porosity and permeability were 31% and
2,000 md, respectively. The initial reservoir
pressure and temperature were 75 psi and 90°F,
respectively.
The initial oil saturation was 52% and initial
water saturation was 48%. Reservoir (pore volume)
compressibility was 50x10- 6 pSi-I; this is within
the range of recent measurements on unconsoli
dated cores. Sand thermal conductivity was
36 Btu/D-ft-oF. The volumetric heat capacity was
35 Btu/ft3_oF.
The crude API gravity was 13° and its molecular
weight was 405. This heavy oil was represented by
a single component and was assumed to be non-
volatile. The crude oil viscosity as a function of
temperature is given in Table 2.
Relative Permeability
Two-phase relative permeability data were obtained
using the functional form given below. 22
3
calculations and its value at low temperatures does
not have any significance.
Endpoint saturations and relative permeabilities
were assumed to be independent of temperature, and
the hysteresis effect on relative permeability was
not considered because of lack of data. However,
as noted earlier, the effect of temperature-
dependent endpoint saturations in the water-oil
system on predicted performance was determined.
RESULTS AND DISCUSSION
The endpoint saturations and relative permeabili-
ties were varied systematically to determine their
effect on oil production rate, instantaneous steam
oil ratio (SOR) , cumulative recovery, and cumula-
tive SOR. In addition, relative permeability
regions that affect different segments of the
production curve were also identified. All simula-
tions were conducted for 3,400 days (9.3 years),
which corresponded to an instantaneous SOR of 10
for the base case, using the linear, three-phase
relative permeability model.

For the water-oil system, Results for the Base Case

Relative Permeability Data

k k :. . : W:. .-_S.: .~ ~r=-._)2.

_l__ . 5 •..•.•••• (1)
Figure 4 shows temperature and gas saturation
rw rwro ( profiles at 700 days, which is approximately
orw W1r
50 days before steam breakthrough. These results
are for the base case relative permeability data

k
row
k
l-S
roiw ( l-S
orw
orw
-S
w
-So
1W
) ••••••••• (2)
using Stone's II model.

sectional profiles along

The figure shows areal
profiles at the top and bottom layers and cross-
the vertical
connecting the injector and the producer.
plane

For, the gas-oil system,

Figure 4 shows that the heated zone preferentially
grows along the top layer between the injector and
l-S. -S _S)2
k 1W org g ....•.... (3)
the producer because of steam override. Tempera-
k
rog roiw ( l-S.-S tures are highest in the upper layers near the
1W org
injector. Areas near the apex in the top layer and
most of the bottom layer see much smaller increase
S-S
g gc )1.5 in temperature. This is also evident from the gas
k
rg
k
rgro ( l-S. -S S
1W org- gc
......... (4) (vapor) saturation profiles. The gas zone prefer-
entially moves along the top layer between the
injector and the producer. The bottom layer
remains 100% liquid saturated at this time.
For the base case, irreducible water saturation,
After steam breakthrough, gas (vapor) saturation in
S . , was 0.4; residual oil saturation to water,
SW1r, was 0.2; residual oil saturation to gas, the reservoir increases and oil saturation
Sorw, was 0.15; and critical gas saturation, S , decreases as the steam-swept zone grows and moves
w~§gO.04. The values of k , k . , and k gc downward. In addition, pressure in the reservoir
declines, resulting in lower peak temperatures.
were 0.1, 0.8, and 0.4, resp~~i~vel§?1WThe relat¥~~
Because of the gravity segregation, most of the
permeability curves for the base case water-oil and
steam (gas) stays in the top half of the reservoir.
gas-liquid systems are shown in Figs. 2 and 3,
respectively. Both Stone's 1112 and linear 13 High oil saturation is limited to the unswept
regions, which are in the bottom layers, away from
models for three-phase relative permeability were
the injector, and near the apex.
used. Unless otherwise noted, results presented
are for Stone's II model. The temperature and gas saturation profiles in
Fig. 4 and also those at later times (not shown
Figure 2 includes some recent water-oil relative
here) clearly demonstrate that gas displacement
permeability data along with the base case curves
only occurs at steam conditions. That is, gas
used in this study. It is apparent that the latter
displacement occurs only in the form of vapor
represents the actual data reasonably well. The
displacement. Therefore, input gas-oil relative
gas-oil relative permeability curves in Fig. 3 are
permeability curves should be measured at steam
assumed to be at steam temperature because for
temperatures.
steamflooding in reservoirs with nearly 100%
initial liquid saturation, gas displacement only
Simulation results show that the residual oil
occurs at steam conditions. Therefore, only high-
saturation to gas (steam) is about 0.20, even
temperature S is actually used in simulator
org though the input value for two-phase S is 0.15.
org
287
 EFFECTS OF ENDPOINT SATURATIONS AND RELATIVE
4 PERMEABILITY MODELS ON PRED1CTED STEAMFLOOD PERFORMANCE SPE 2020~
The reason tor th1S d1tterence 1n::s. can be seen instantaneous SOR of 10, then the cumulative
from the three-phase oil isoperms f~gFig. 5. For recovery values will be somewhat smaller.
water saturation slightly above the irreducible,
predicted residual oil saturation by Stone's II
model (given by the zero oil isoperm line) is about Residual Oil Saturation to Water
0.20.
The effect of residual oil saturation to water,
Figure 5 also shows the saturation path for the top S ,on production rate and cumulative recovery
of the production well (cell 7,1,1). Prior to a~~ shown in Fig. 7. Similar to the irreducible
steam breakthrough the saturations are in the two- water saturation, S affects the production rate
phase, water-oil region. After steam breakthrough prior to steam break~~ough. The production rates
they move toward the residual oil saturation to gas are higher for lower S • However, after the
(steam) with nearly irreducible water saturation. breakthrough the rates a~~wessentially independent
of S
orw
The oil production rate and cumulative oil recovery
are shown in Fig. 6. The S. for the base case is The reasons for the observed S effects on the
0.40. The oil production w\-~te is for the full production rate are the same RP those for S . .
pattern. Note that results for the base case show Lower S values result in higher relative pe~€~­
a more gradual, less spiky production response than bility ~~ oil and lower for water, which results
those for lower S . . in higher oil production rate. The effect of a
W1r
change in S was smaller than that in S .
because, foror~he given initial saturations,w1~
Effects of Endpoint Saturations change in S. affects the magnitude of oil
and Relative Permeabilities relative per:Jlbility more. After steam break-
through, the production rates are governed
Irreducible Water Saturation primarily by the gas-oil relative permeability;
therefore, the effect of water-oil relative permea-
The effect of irreducible water saturation (S . ) bilities is small.
on the oil production rate and cumulative reco~~fy
is shown in Fig. 6. It should be noted that the Figure 7b shows that the cumulative recovery
S. affects both oil-water and gas-liqui4 relative decreased from 58% to 54% IOIP, when the S orw was
p:f~eability curves (for given residual oil satura- increased from 0.15 to 0.30. Note, however, that
tions) as shown in Eqs. (1) to (4) . As evident the difference between the curves remains
from Fig. 6, the S. strongly affects the oil essentially constant after steam breakthrough
production rate duri~kr the initial period prior to (approximately 750 days) because production rates
steam breakthrough. After breakthrough, its effect are the same after that time. The effect of S
on the production rate is small. It also affects on production rate and cumulative recovery ofJ
the time of steam breakthrough; a decrease in S . smaller than that of irreducible water saturation
from 0.40 to 0.25 results in an increase in bre~R~ (Fig. 6).
through time from 750 to 900 days (or 20%).

The oil production rate during the initial period Residual Oil Saturation to Gas (Vapor)
is affected by S. because, before steam break-
through, the areawJ~ar the production well stays in Residual oil saturation to gas, S ,affects oil
the two-phase, water-oil region. Higher S. production rates most strongly at o~fid after steam
causes relative permeabilities for water wf6 breakthrough as shown in Fig. 8; lower S results
0
decrease and those for oil to increase. This in higher oil production rates. This is 5@cause of
results in a higher oil production rate. In the higher oil relative permeability associated
addition, higher oil production rates result in a with lower S . The S has a very small effect
lower saturation in the oil bank ahead of the steam on the produgffon rate ~llbr to steam breakthrough.
zone. Therefore, at the time of steam
breakthrough, the peak production rate is less Figure 8b shows that S has the most dominant
spiky. Higher S. also results in higher gas effect on cumulative rec~~ry. An increase in S
relative permeabr[fty causing an earlier steam from 0.10 to 0.25 results in a decrease in cumu~~g
breakthrough. tive recovery from 62% to 46% IOIP. Therefore, an
accurate value of residual oil saturation to gas is
The effect of S. is more evident on the the most critical endpoint for performance predic-
instantaneous ste~foil ratio (SOR). For low tion.
S . IS, the SORts are unrealistically high prior to
b~€~kthrough. This also points out that oil
recovery at the end of a project is not sufficient Critical Gas Saturation
to characterize performance. The time when
instantaneous SOR reaches a value of 10 can be A change in critical gas saturation from 0.02 to
deduced from Fig. 6a. Because the injection rate 0.10 had essentially no effect on recovery. This
is held constant at 250 BID, eWE, the instantaneous is because after steam breakthrough saturations
SOR of 10 corresponds to an oil rate of 25 BID; rapidly approach and stay near the residual liquid
this value is reached somewhat sooner for higher where the effect of a change in critical gas
irreducible water saturations. saturation is small.

Figure 6b shows that cumulative recovery (at Relative Permeability Values

3,400 days) decreases from 57% to 50% initial oil-
in -place (IOIP), when the S. decreases from 0.4 Significant changes in the water relative permea-
to 0.25. If the projectsW1~ere terminated at an bility (keeping all endpoint saturations and other
288
SPE 20202 M. KUMAR AND T. N. DO
relative permeabilities the same) affects the oil
production rate (see Fig. 9a) but has little effect
on breakthrough time or on cumulative recovery at
the end of the project (Fig. 9b). As expected,
increasing water relative permeability lowers the
oil production rate prior to steam breakthrough;
however, oil production increases somewhat after
the breakthrough, resulting in effectively no
change in cumulative recovery at the end of the
project.
A significant changes in the gas relative permea-
bility affects the breakthrough time and the oil
production rate after steam breakthrough, as shown
in Fig. 10. When the gas relative permeability is
increased, the breakthrough time and the oil
production rate after breakthrough decrease.
Consequently, cumulative recovery also decreases.
Fig. lOb shows that for the gas relative permea-
bility values (at residual liquid saturation) of
0.8, 0.4, and 0.2, the oil recovery at the end of
the project was 43%, 57%, and 63% IOIP, respec-
tively.
Changing the oil relative permeability curve in-
dividually in either the water-oil system or the
gas-oil system had a very small effect on perfor-
mance. However, changing the oil relative permea-
bility curves simultaneously for both the water-oil
and gas-oil systems, by changing k. in
Equations (2) and (3) , has essentially H~~w same
effect as changing both water and gas relative
permeability curves simultaneously. For example,
the cumulative recovery was the same when k was
roiw
decreased by a factor of two or when k and
k were increased by a factor of 2. 1fcfJ'ever,
tli@r~atter resulted in earlier breakthrough.
"Practical" vs. "True" Residual Oil Saturations
Residual oil saturations determined from displace-
ment measurements on a core In the laboratory may
not be the "true" residual because experiments are
terminated when the "practical" endpoints are
reached, i.e., when the oil expulsion rate (or oil
relative permeability) becomes very small. It has
been found that "true" S obtained using centri-
fuge measurements are 19~@r than the "practical"
values measured in the displacement tests.
However, oil relative permeability values at
saturations below the "practical" endpoint are very
small (less than 0.01).
Simulation runs were made to determine the effect
of including very low oil relative permeability
values, representing the true residual oil satura-
tion with those of practical endpoint saturations
where a cutoff value of oil relative permeability
was used. Figure 11 shows the gas-liquid relative
permeability curve on a semilog plot. The true
S is 0.1 (corresponding to a residual liquid
sg£§rations of 0.5).
Two cases of practical residual saturations were
studied. In one, practical oil relative permea-
bility cutoff was assumed to be 0.001, correspond-
ing to a S of 0.15 (liquid saturation = 0.55).
That is, off~elative permeability at this satura-
tion was assumed to be zero for the gas-oil system.
In another, practical oil relative permeability
cutoff was assumed to be 0.01, corresponding to a
S of 0.20. The relative permeability curves at
org
289
liquid saturation of 0.6 and above are the same as
the base case values (Fig. 3). The water-oil
relative permeability was the same as the base case
(Fig. 2).
Figure 12a shows that for Stone's II model, the
recovery for practical S of 0.2 is the same as
true S of 0.1 during o~e steamflooding period
(3,400 °JIys). This implies that oil recovery is
affected by the absolute value of oil relative
permeability, and very small oil relative permea-
bilities do not contribute to oil production.
Therefore, practical endpoints are sufficient for
performance prediction, provided the remalnlng
portions of the relative permeability curves are
unchanged.
There was no difference in predicted recovery when
practical vs. true values of residual oil satura-
tion to water were used. This is primarily because
during the steamflooding process, there is no two-
phase (water and oil) region in the reservoir with
oil saturations near the residual.
These results demonstrate that practical endpoint
oil saturations are adequate for performance pre-
diction because generally oil relative permeabili-
ties below 0.005 (or oil to gas permeability ratio
below 0.01) do not significantly affect recovery
calculations during the steamflooding period (about
9 years). However, it should be noted that over a
much larger duration, these small relative permea-
bility values would affect residual saturations
and, therefore, oil recovery.
Stone's II vs. Linear Three-Phase Hodel
All runs reported above were also made using the
linear, three-phase oil relative permeability
model. 13 For the sets of two-phase data in this
study, the linear model generally predicts about
5% IOIP higher recovery. Figure 13 compares the
oil production rate and cumulative oil recovery for
the Stone's II and linear three-phase oil relative
permeability models. The major difference in the
oil production rates for the two models is after
steam breakthrough. The linear model shows a some-
what higher production rate (Fig. 13a) resulting in
higher cumulative recovery (Fig. 13b).
The observed differences between the results of the
linear and Stone's II model can be explained by the
oil isoperm ternary diagram and the saturation path
for the top of the production well (Cell 7,1,1),
shown in Fig. 5. It can be seen that the residual
saturation predicted by the linear model is close
to the input value and is lower than Stone's II
model. This is because, near the residual oil
saturation, the three-phase oil relative permea-
bility is higher for the linear model compared to
the Stone's II model. It should be noted that the
residual oil saturation to steam predicted by
Stone's II model can be lower than the input
value. 9 This usually occurs when the two-phase
data are ill-behaved, or their shapes are signifi-
cantly different than those used here.
In general, the conclusions reported above are the
same and are independent of the three-phase
relative permeability model used. That is, the
relative changes in recovery values are the same
for the two models. For example, when the S .
Wlr
 EFFECTS OF ENDPOINT SATURATIONS AND RELATIVE
6 PERMEABILITY MODELS ON PREDICTED STEAMFLOOD PERFORMANCE
cnanges Irom u.4U to U.L:>, tne 011 recovery uS1ngine gas ~steam vapor) re1at1ve permea0111ty curve
the linear model changes from 61% to 55% IOIP (vs. affects the time of steam breakthrough and the oil
57% and 50%, respectively, for the Stone I s II
model).
However, the linear model showed a higher depend-
ency on the true vs. practical residual oil satura- recovery.
tion to gas (see Fig. 12b) because of its higher
three-phase oil permeability at low oil satura-
tions. For an oil relative permeability cutoff of
0.01 in Fig. 11, i.e., a practical S of 0.20, Saturations for Water-Oil System
the oil recovery for linear model a~ereased by
5.5% IOIP vs. 1% IOIP for the Stone's II model.
The true S is 0.10. On the other hand, oil
g
recovery w8I unaffected (see Fig. 12) in both
models when a relative permeability cutoff of
0.001, i.e., when a practical S of 0.15 was S
g
used. This is because for su2li low relative
permeability values only a small region near the
injector reached the true residual oil saturation
during the steamflooding period.
Important Relative Permeability Regions
for Performance Prediction
Figure 14 schematically shows which regions of the
relative permeability curves affect various
sections of the oil production curves. These are
based on the results discussed in the previous
sections for a heavy oil reservoir with nearly 100%
initial liquid saturation.
The water-oil relative permeability curves and the
associated endpoint saturations primarily affect
steamflood performance prior to steam breakthrough
because the area near the production well stays in
the two-phase, water-oil region before break-
through.
The irreducible water saturation affects most
strongly the shape of the production curve prior to
steam breakthrough. It also affects cumulative
recovery and the time of steam breakthrough. Both
oil production rate and cumulative recovery
increase as the irreducible water saturation is
increased. In addition, the breakthrough time
decreases. Residual oil saturation to water also
affects the shape of the production curve prior to
breakthrough and the cumulative recovery. However,
its effect is less than that of irreducible water
saturation. Lower residual oil saturation results
in a higher production rate before breakthrough and
higher cumulative recovery.
The water relative permeability curve affects the
oil production rate prior to steam breakthrough;
however, it has negligible effect on cumulative
production. Significant increase in water relative
permeability values results in lower oil production
before steam breakthrough.
The gas-oil relative permeability curves and the
associated endpoint saturations affect the oil
production rate after steam breakthrough and the
time of steam breakthrough. The residual oil
saturation to gas (steam vapor) has the largest
effect on cumulative recovery. Lower residual oil
saturation results in a higher production rate
after breakthrough and higher cumulative recovery.
The critical gas saturation has essentially no
effect on the oil production.
SPE 20202
production rate after the breakthrough. Substan-
tially higher gas relative permeability results in
earlier steam breakthrough, lower production rate
after the breakthrough, and lower cumulative
Temperature Dependence of Endpoint
The temperature-dependent residual oil saturation
to water had a rather small effect on the oil pro-
duction rate and cumulative recovery (see Fig. 15).
This figure shows results for two cases. In one,
was assumed to be temperature independent with
aO~~lue of 0.30. In the other, S was considered
to be temperature dependent; it SI: 0.30 at reser-
voir temperature and 0.20 at steam temperature.
The effect of temperature-dependent S would be
evident only in the hot-water zone ~:ad of the
steam front where no gas (vapor) exists. Because
this hot-water zone is much smaller than the steam
zone, its effect on oil production is small.
The temperature-dependent irreducible water satura-
tion affects the oil production rate and cumulative
recovery more (see Fig. 16) than the temperature-
dependent S (Fig. 15). Temperature-dependent
S. resultsO~~ increased production rate primarily
a~6ijnd the breakthrough time. The cumulative
recovery increased by about 2.5% OOIP when S. was
considered to be temperature dependent (0':25 at
reservoir temperature and 0.40 at steam tempera-
ture) .
If both endpoints, ~orw and Swir' were temperature
dependent, then the1r effects on recovery could be
obtained by adding the individual effects. In any
case, the effect of a 15% change in endpoint
saturations on recovery was relatively small.
Implications for Relative Permeability
Measurements
Based on the results of this study, the most
important relative permeability region is the
residual oil saturation to steam and the oil and
gas relative permeability values near the residual
liquid saturation. Unfortunately, gas-oil relative
permeability data are the least accurate because
these measurements are generally made at low
temperatures where the mobility contrast between
the gas and the heavy oil is very high. As a
result, relative permeability measurements may be
significantly influenced by the high mobility
ratio. 23 The relative permeability curves at steam
temperature are obtained by using normalized
relative permeability curves based on the measured
low-temperature data and temperature-dependent
residual liquid (irreducible water and residual
oil) saturations.
As noted earlier, for the steam displacement
process with nearly 100% initial liquid saturation,
actual gas displacement occurs in the reservoir
only by vapor at local steam temperatures. At
these higher temperatures, the oil viscosities are
two to three orders of magnitude lower than those
at room temperature, i.e., the mobility ratio at
steam temperature is two to three orders of magni-
290
SPE 20202 M. KUMAR AND T. N. DO
tude lower. (.;onsequently, it is very likely that 3. For the water-oil system,
the shape of the gas-oil relative permeability
curves may be different than those measured at room
temperature. Therefore, gas-oil relative perme-
ability curves should be measured at steam tempera-
tures, if possible. 4.
In the water-oil system, irreducible water satura-
tion and its temperature dependence is more
important than residual oil to water. For the data 5.
of this study, temperature dependence of Sorw had a
very small effect on recovery.
Effect of Relative Permeability Data
on the Shape of Production Curves 6.
As noted earlier, a recent study found that the
actual field production data differ from the
typical symmetric element simulation predictions
published previously.21 More specifically, simula-
tors predicted a delayed production response and a 7.
sharper peak. One of the reasons for this
difference is that many of the previous studies
(for example, Refs. 3 to 5) employed relative
permeability data of Ref. 20, which had a low Swir'
8.
and a very strong temperature-dependence. As was
demonstrated in the previous section, the former
contributes to the delayed production response; and
the latter to the sharper peak. In addition, these
studies used very high rock compressibility values,
which also contributes to a sharper peak.
7
irreducible water
saturation (S . ) has a greater effect on
performance tf{§~ residual oil saturation to
water (S ).
orw
Temperature-dependent endpoint saturations for
the water-oil system have a relatively small
effect on cumulative recovery.
The linear interpolation model 13 for three-
phase oil relative permeability is preferable
over the Stone's II model because it more
accurately predicts residual oil saturation to
steam.
"Practical" endpoint oil saturations measured
in displacement experiments are adequate for
performance prediction because very low values
of oil relative permeability do not signifi-
cantly affect recovery calculations.
A 15% change in S , S . , and S resulted
org W1r orw
in a 26, 13, and 7% change in cumulative oil
recovery, respectively.
Relative permeability data used in previous
studies are indeed one of the reasons for the
observed differences between field data and
predictions.

The shape of the production curve using the base ACINOWLEDGHENTS

case data in this study is less "spiky" than that
predicted by data in Reference 20. Therefore,
relative permeability data used in the past is We wish to thank W. H. Chen and M. N. Pham for
indeed one of the reasons for the observed implementing the linear, three-phase relati ve
discrepancy between simulator predictions and field permeability model into the simulator.
performance. Furthermore, other recent studies 24
also use relative permeability data that are
significantly different from old data in Ref. 20
and have a functional form similar to the one used REFERENCES
here.
1. Coates, K. H. and Ramesh, A. B.: "Effects of
Grid Type and Difference Scheme on Pattern
SUMMARY AND CONCLUSIONS Steamflood Simulation Results," J. Pet. Tech.
(May 1986) 557-569.
A numerical simulation study was conducted to
quantify the effects and importance of 2. Prats, M.: Thermal Recovery, SPE Monograph
(1) temperature-dependent endpoint saturations, Series, Vol. 7 (1982).
(2) three-phase relative permeability models, and
(3) measurement accuracy of endpoint ("practical" 3. Gomaa, E. : "Correlations for Predicting Oil
vs. "true") oil saturations on predicted steamflood Recovery by Steamflood," J. Pet. Tech.
performance for a typical heavy oil reservoir with (Feb. 1980) 325-332.
nearly 100% initial liquid saturation. The crude
oil was assumed to be nonvolatile. Results are 4. Hong, K. C.: "Guidelines for Converting
used to focus laboratory measurements on relative Steamflood to Waterflood," SPE Res. Eng. J.
permeability regions important for predictions and (Feb. 1987) 67-76.
to explain differences between previous predictions
and field data. 5. Hong, K. C. and Ault, J. W.: "Effects of
Noncondensable Gas Injection on Oil Recovery
The conclusions of this study are summarized below: by Steamflooding," J. Pet. Tech. (Dec. 1984)
2160-2170.
1. The most important region for performance
prediction is the high-temperature, gas 6. Boberg, T. C.: Thermal Methods of Oil
Recovery, Exxon Monograph, John Wiley and
(vapor)-oil relative permeability near the
Sons, Inc. (1988).
residual oil saturation (S ).
org
7. Shen, C. W.: "Laboratory Hot Waterfloods
2. Gas (vapor)-oil relative permeabilities should Prior to and Following Steamfloods," paper
be measured at steam temperatures because gas SPE 18754 presented at the 1989 California
displacement occurs at steam conditions. Regional Meeting," Bakersfield Aoril 5-7.
291
 EFFECTS OF ENDPOINT SATURATIONS AND RELATIVE
8 PERMEABILITY MODELS ON PREDICTED STEAMFLOOD PERFORMANCR RPR ?n?n?
!S. Kumar, M. ana Hong. I!..: 1>.: '1!;ttect or 'Effect ot Temperature on Bitumen-Water
We1lbore S~~am Segregat10n on Steamflood Endpoint Relative Permeabilities and Satura-
Performance, pa~er SPE 18086 presented at the tions," J. Cdn. Pet. Tech. (Se t.-Oct. 1986)
1988 Annual Meet1ng, Houston, October 2-5. 44-50. P

9. Sato, K. and Aziz, K.: "Sensitivity of Steam

Displacement Predictions to Three-Phase 18. Miller, M. A. and Ramey, H. J.: "Effect of
Relative Permeability Models," paper SPE 16733 Temperature on Oil/Water Relative Permea-
presented at 1987 Annual Meeting, Dallas, bi1ities of Unconsolidated and Consolidated
September 27-30. Sands," SPEJ (Dec. 1985) 945-953.

10. Stone, H. 1. : "Probability Model for 19. Sufi, A. 3., Ramey, H. J., Jr., and Brigham,
Estimating Three-Phase Relative Permeability," W. E.: "Temperature Effects on Oil-Water
J. Pet. Tech. (Feb. 1970) 214-218. Relative Permeabilities for Unconsolidated
Sands," paper SPE 11071 presented at the 1982
11. Stone, H. L. : "Estimation of Three-Phase SPE Annual Technical Conference and Exhibi-
Relative Permeability and Residual Oil Data," tion, New Orleans, Sept. 26-29.
J. Cdn. Pet. Tech. (Oct.-Dec. 1973) 53-61.
20. Chu, C. and Trimble, A. E.: "Numerical
12. Aziz, K. and Settari, A.: Petroleum Reservoir Simulation of Steam Displacement Field
Simulation, Applied Science Publishers, London Performance Applications," J. Pet. Tech.
(1979). (June 1975) 765-776.

13. Baker, 1. E.: "Three-Phase Relative Permea- 21. Torabjadeh, S. J., Kumar, M., and
bility Correlations," paper SPE 17369 Hoang, V. T.: "Performance Correlations for
presented at the Sixth SPE/DOE Symposium on Steamflood Field Projects," paper SPE 20036
EOR, Tulsa, OK, April 1988. presented at the 1990 California Regional
Meeting, Ventura, CA, April 4-6.
14. Quettier, L. and Corre, B.: "Hot-Water and
Steamflood Laboratory Experiments Under 22. Aziz, K., Ramesh, A. B., and Woo, P. T.:
Reservoir Conditions," SPE Res. Eng. J. (Feb. "Fourth SPE Comparative Solution Project:
1988) 149-157. Comparison of Steam Injection Simulators,"
J. Pet. Tech. (Dec. 1987) 1576-1584.

15. Maini, B. and Okazawa, T.: "Effect of 23. Peters, E. J. and Khataniar, S.: "The Effect
Temperature on Heavy Oil-Water Relative of Instability on Relative Permeability Curves
Permeability of Sand," J. Cdn. Pet. Tech. Obtained by the Dynamic-Displacement Method,"
(May-June 1987) 33-41. SPE Formation Evaluation J. (Dec. 1987)
469-474.
16. Honarpour, M., DeGroth, C., and Manjnath, A.:
"How Temperature Affects Relative Permeability 24. Johnson, R. S., Chu, C., Mims, D. S., and
Measurement," World Oil (May 1986) 116-126. Haney, K. L.: "History Matching of High and
Low Quality Steamfloods in Kern River Field,
CA," paper SPE 18768 presented at the 1989
17. Polikar, M. , Ferracuti, F., Decastro, V., California Regional Meeting, Bakersfield,
Puttagunta, V. R., and Farouq Ali, S. M.: April 5-7.

292
 SPE 20202

TABLE 1
RESERVOIR DATA AND INPUT PARAMETERS

Model Grid 7x4x5

(for 1/8 of as-spot)

Pattern Area, acres 2.61

Sand Thickness, ft 100

Crude API Gravity, °API 13

Molecular Weight of Crude Oil 405

Porosity 0.31

Permeability, md 2,000

Initial Reservoir Temperature, OF 90

Initial Reservoir Pressure, psi 75
Initial Oil Saturation 0.52
Initial Water Saturation 0.48
Initial Gas Saturation 0.00

Oil Compressibility, 1/psi 5 x 10-6

Rock Compressibility, 1/psi 50 x 10-6

Reservoir Thermal Conductivity, BTu/O-ft_oF 36

Sand Volumetric Heat Capacity, BTU/ft3_oF 35

Injection Pressure, psi 200

Injected Steam Quality 0.5
Injection Rate, BID CWE 250 (for full pattern)

TABLE 2
CRUDE OIL VISCOSITY

Temperature, OF Viscosity, cp

75 4,200
100 1,100
150 130
200 33
250 12.5
300 6.4
350 3.8
400 1.6

293
 SPE 20202

vLx 1.0
,
[
.o.x=.o.V= 29.4H

0.8
-- Base Case I-
.0 Core 1
..2
I;
i . :\ , 80 Core 2
Core 3 1-
:\
u

~
.:J: 0.6
Injector Producer
\ [
[

,
,
iI

'\
I
i.;. [

0.4 I------
\.
zr
--,--
x
l .o.Z=20H t 1-- .-

-\:
.. ,
•

!
0.2 1---
, . .~ .
t IIIIII t ~.- i
......, .. i

I
I
~
i· ...:
o ! . -.tIti
o 0.2 0.4 0.8 0.8 1.0
Wate, Saturation

Figure 2
Figure 1
One-Eighth of a Five-Spot Model Water-Oil Relative Permeability (Base Case) and
Used for Simulation Recent Measurements

1.0

0.8
~
...I;
~
~
0.6 /
i f7
1
:. 0.4
/
k,og

!.!! ''\ krg

II
/
,!
0.2
"'"
)<"
o
o 0.2 0.4
V
0.8
"'-,---
0.8 1.0
Liquid Saturation

Figure 3
Gas-Oil Relative Permeability (Base Case)

294
 SPE 20202

Temperature G•• Saturation

.) Are.l Temperature Profil.s at Top Layer .) Ar•• l Ca. Saturation Profile. at Top Layer

.:1'-:'--:::::=:::::=:::::::=---:/
~y-0.l0
~.
0
0:..

b) Cron-Sectional Temperature Profiles b) Cross-Sectional Ga. Saturation Profil••

Detween Injector and Producer Detween Injector and Producer

....

c) Areal Temperature Profiles at Bottom Layer c) Areal Gas Saturation Proftles at BoUom Layer

Figure 4
Temperature and Gas Saturation Profiles at 700 Days

Saturation Path Sg Isoperm

---- linear ---------- Linear Model

- - Stone's II - - Stone's II Model

Sw 0.8 0.4

Figure 5
Three-Phase Oil Isoperms and Saturation Path at the Top of
Producer (Cell 7,1,1) Using Linear and Stone's II Models

295
 SPE 20202

'N.---~----,----,----~---'----~---'
'''r----,----,_---,----,_---,----,----...,

'28
__--_+----t------+-I
,31-----.----+----,--1
~
~Ji "I----L--A6~+===~~
1
i"~--~~.L-~~~~==~=L~-~
~

t~~.____L-------i
!
i .. I-----,P-¥--I+-+---c-:~~~__+__-----l
3 8

500 '000 ,...

nme,Dap
.) 011 Production Rate
2000 .... - .... TIme, oay.
.) 011 Production .....
3500

'.7 .----,----,-----,-----,-----,----,----, ur---------~--_,----,_--_,----,_--~

I••.• ~---lr------;!_c_----;;c;i;_.---t---_+----t----' IcO.•

g ~
1o 0•5
i'"
o
i •.• f--+--+--+--T~I-L-+--+---I
i"i •.3f--+--+--:A-.r'7Lf--+--+--i l
J
1. 3

5 0.2 1__--_+------,j<--.,,<,4-----t------+----t------I 8 O~f---+--~~~+----j_--+----j_----i

!€ 0.1 f_--+,L----rbL--+----j_--+----j_----i
~~ •., 1-----:J?;;;.<:.--r----+---+-----j----+-----1
a u

500 'OlIO '500 2000

Time, DIY'
2500
_ 3500 .~~-L

•
___ J_____'____
'000 '500
~

2000
TIme, 011,..
....
__~_____'____~
3000 >NO

b) Cumulative Recovery b) Cumulative Recovery

Figure 6 Figure 7
Effect of Irreducible Water Saturation (Swirl on Effect of Residual Oil Saturation to Water (Sorwl on Oil
Oil Production Rate and Cumulative Recovery Production Rate and Cumulative Recovery

..
,
I I
,.. r---------~--------~--------r_---,

I ---Sorg .. 0.25
,3r----,----+----,-1
12. II----
---Sorg 0.15
I =
If---
I
1
1\ :, --Sorg .. 0.10
~.. "t--
I --'~~~=+=:'__._+___~
J\~~I ! ~

".j)
f-:
V'''--(~~
~"~"-.
I"~~~----I
!
/ i , ,~...
~
~" ~
• 500
i
!
'OlIO
!
'500 2000
Tim., Day.
....I ---
3000 500 'OlIO ,... .... 3000 """

.) 011 Production Rete

.) 011 Production Ret.
'.7 r--------------r----.---------C---, '.7 r---------.,-----------,----,----,----,
I
~ 0.'
is
1 0.5

~
g
1 04
•
j •.. f----c---+-----;--:~~-_j_-_t_----1
i
i5
'.3 1__---------C--#7L---+-----t-----'------i ,u
0.2 f_----'----;oP~--------;_---+----f_--__1 5 ~f_--~--_.r'~--_+----+_--_+----+_--_i

!
~ 0.1 I-------'c~------~----'----------I__--_i
a
ia . ,f------;t~~.:---------T----+----+---_i
.~~~--~----~--~--~-----'----~
o 500 1000 1500
Tlm.,Dap
2000 2SOO 3000 3500 ... 'OlIO ,...nme,c.,..
.... 3000 3500

b) Cumulallve Recovery b) Cumulellve Recovery

Figure 8 Figure 9
Effect of Residual Oil Saturation to Gas (Sorgl on Oil Production Effect of Water Relative Permeability on Oil Production Rate and
Rate and Cumulative Recovery Cumulative Recovery

296
 SPE 20202

,~~------~--------~--------~--~

'211----------+-----1

~.2S"I-I ___~~~~~--~========~+---~
j M~-~~---+----I
is

.L-__ ~ __ ~ ____ ~ __-L__ ~ ____ ~ __ ~

o IlOO 1000 1100

TImo,O.,.
2000 2500 3000 3100
•• '\1' Practical Endpoint Saturations X
.) 011 Production Rat.
krg==

I krog ,/
r"
i··'
I ~.
~

1
f. 3

!ul--------~~~~---_r--~----~---~
I I
I~,~-'-----'-----+----1_______i I
,
.~~~

o !iCIO
__ ~

1000
____ ~

1500
__-L__
2000
~ ____
2500
~

3000
__ ~

3500
i
TIme.""
b, Cumulattve Reconry
0.1 0.2 0.3 0.4 0.5
liquid Saturation
0.6 0.7 0.8 0.9 1.0

Figure 10 Figure 11
Effect of Gas Relative Permeability on Oil Production Rate and Gas·OIl Relative Permeability ("Practical" VS. "True" Endpoint
Cumulative Recovery Saturations)

,
,.,--------------------------------,

../" 4-- ".

•
•
V
V ,.
~

j
-

r---~_,~~------------~
• i Mr-----f=--~-----"'-~~~__i
/ I·-s.,. - 0.15
O.'O'nd~
I"·'·"'· $0'0 - !
, / 0.20

• ../ ... 1000 1500 2000 260D 3DOO .... • ..... ,... ....
TI..... O..,. 0 '000 '500nm., D.,. 3000

• ) Ston... II Moda. .) 011 Production Rat•

..,,----------------------------------, ..,
I ~. I~.
Ci -EJ·-~----
-slOne'.n
~ ... _______. .. __.- U,..., ____
'

i··'
~
~

i •.• ------

t~·~----.-c-.-~~------~~ l
f··3
f·"
11 .~
a: 0.2

!
5

J. , ; 0.1

..
u

...
~~----~------------------~--~
1000 1500 2000 2500 3000 3500
•• '000 ...
, ..... ,...
Time, Days
3000 ....
TIme, Da,. b) Cumulative Reco.,.,.,
b) Un...-M...I
Figure 13
Figure 12
Comparison of Oil Production Rates and Cumulative Recovery
Effect of "Practical" vs. "True" Endpoint Saturation on Oil Recovery for Linear and Stone's " Models (Base Case)

297
 SPE 20202

1Nr-------------------------------____--,
--.-----::-----:--:--:-:l
I Breakthrough Tim. Affactad by
128
---~---
- - Temp. Independent
SON =0.30
YSWlr and Krgro -_.- Temp. DeperHMnt
Affected Sorw (10- F) " 0.30
Primarily
byWater-On
R.latlve
I ~

t ..
Sorw (_oF) "0.20
-.~- .-~~~-.-~--------

Permeability .!
Curve,

IMI------,:'F-----=~-_
3
32 1---fJ'---------

Time
nme.D.ys
a) on Production Rate a) 011 Production R,tl

'.7 r---------------------,------,------,-----,
I-o~ 0.' - Temp. Independent
SOIW " 0.30
I 0.5 - - Temp. Dependent
.!i Sorw (ltOO F) " 0.30
'0 Sorw (38O"F)" 0.20

Shape
j '.4 f-----''======'--..;,:z..-----'---.j
Determined by
Swlr and
Water-Oil
Relative
Pannaabllltl••
Cumulatlva Recovary
Affected by
Sorgo Krgrot Swift and Sorw
I"
S 0.2

I 3
0.1 1---------:<i'-----------------------+----1

·.~~~~~---,~...
=---~,NO=---~~~----,--2NO--,----~OOOO~--~3NO
Tim.
nme.Da"
b) Cumulatlva Recovary b) Cumulativi Recovlry

Figure 14 Figure 15
Schematic of Production and Recovery Curves Affected by Various Effect of Temperature-Dependent Residual Oil Saturation (Sorw)
Regions of Relative Permeability on Oil Production Rate and Cumulative Recovery

1~r-------------------------------------,

- - T-.rIp. Indrlpendltnt
121 Swtr: G.2I
- - Temp. o.p.ncs.m
..... CWF)· ....
....«....
F)= •. 40

32

••c===~------,------,~-----",,-----2IOO----~~----~~
nrne,O.ys
.) 011 Production Ratl

'.7 r--------------------.,.-----,-----~----,

, •.3

is ...

I
U
'.1 1-----------#'------------------1----1

·.~--~~~~----,--,NO~--~....
~---2IOO----- L-----'~
. ...
nme,Oaye
b) Cumulative Recovery

Figure 16
Effect of Temperature-Dependent Irreducible Water Saturation
(Swir) on Oil Production Rate and Cumulative Recovery

298