SPE 145808

Three- Phase Unsteady-State Relative Permeability Measurements in

Consolidated Cores Using Three Immisicible Liquids
P. Cao* and S. Siddiqui, Texas Tech University
*Now at Penn State University

Copyright 2011, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Annual Technical Conference and Exhibition held in Denver, Colorado, USA, 30 October–2 November 2011.

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 have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract

This paper discusses results from a series of two- and three-phase coreflooding experiments on consolidated cores using three
immiscible fluids and using an unsteady-state relative permeability setup. The three-phase extension of the Buckley-Leverett
theory proposed by Grader and O’Meara (1988) and verified by Siddiqui et al. (1996) was used for calculating three-phase
relative permeabilities from the dynamic displacement data. From the results of three-phase displacement experiment, three-
phase saturation trajectories are mapped and then compared against results of DDI (decreasing of water phase and oil or heavy
phase and increasing of gas or light phase during a dynamic injection stage) runs found in the literature. The DID (decreasing
of water phase, increasing of oil or heavy phase and decreasing of gas or light phase) runs presented in the current work are
unique which map a wide range of the interior region of the ternary diagram. However, bypassing was observed during the
IDD (Increasing of water phase and decreasing of oil or heavy phase and gas or light phase) runs, possibly due to fluids
reaching the residual saturations before the dynamic water injection. In the petroleum industry, empirical models are often
used to extrapolate three-phase relative permeabilities from two sets of two-phase relative permeability data. The
experimental three-phase relative permeability data from the DDI and DID runs are compared with the model data, and it is
found out that in some cases these models cannot adequately provide satisfactory matches with the experimental data.

Introduction

The production from hydrocarbon reservoirs under primary, secondary and tertiary recovery processes usually involve the
simultaneous flow of two or more fluids. During the life of a reservoir, there may be several situations in which the
simultaneous flow of water, oil and gas is of practical importance. For instance, the displacement of oil by either gas or water,
and the enhanced oil recovery (EOR) processes such as steam injection, in-situ combustion and CO2 WAG (water alternating
gas) injection are all examples that involve three-phase flow. Also in many other engineering fields like ground water
hydrology, waste disposal sites, multiphase flow is encountered often (Honarpour, 1986).

One of the key petrophysical parameters that has been extensively used to characterize the hydrodynamics of multiphase fluids
in porous media is relative permeability. It is used to determine the energy needed to displace the fluids and the relative flow
behaviors of each phase present in the porous media. Relative permeability can provide essential information for numerical
simulation studies, reservoir recovery calculations, reservoir characterization, and it is also valuable for evaluating formation
damage.

The existence of three-phase flow in large parts of the reservoir, especially in enhanced oil recovery processes has led to
booming interest in getting reliable three-phase relative permeabilities. In the past, the use of three-phase relative permeability
data for conventional reservoir engineering calculations was not as important as it is today. In consequence, considerably less
is known about three-phase relative permeability characteristics of rocks than is known for two-phase cases. Three-phase
relative permeability is found to be quite useful in the calculation of field performance for reservoirs being produced by
simultaneous water and gas drive; and also for analyzing solution gas drive reservoirs which are partially depleted and are
2 SPE 145808

being produced by water drive. Based on this specific need, the objective of this research is to further our understanding of the
simultaneous three-phase flow in porous media, especially to quantify immiscible three-phase relative permeabilities.

It is impossible to obtain three-phase relative permeability data directly from the field or in any indirect manner and the only
way to obtain them is from experiments. Due to the complexities of the measurements and the scarcity of reliable
experimental data, mathematical models are generally used to extrapolate three-phase relative permeabilities from two-phase
relative permeability data. Most of these models are based on the assumption that each fluid establishes its own intricate
saturation path which, in turn, forms very stable flow channels (Edgar-Ricardo, 2008). However, in real reservoir cases,
relative permeabilities are either overestimated or underestimated. This paper will concentrate on coreflood experiments to
identify the relative permeabilities and to compare them with those generated from empirical models.

Theoretical Background

Buckley-Leverett Theory

Buckley and Leverett (1942) presented what is recognized as the basic equation for describing two-phase, immiscible
displacement in a linear system. The classic theory they developed is called the frontal displacement theory, which provides
the basis for describing immiscible displacement in one dimension.

A more elegant method of achieving the same result was presented by Welge in 1952. He showed, by drawing a straight line
from Swc (or from Swi if it is different from Swc) tangent to the fractional flow curve, the saturation value at the tangent point is
equivalent to that at the front Swf. The coordinate of the point of tangency represents also the value of the water cut at the
leading edge of the water front fwf. Another progress, Welge made is finding a way to calculate the relative permeability ratios
which can be expressed as:
= a e S …………...……………………………………….………………..….[1]
D
where b is the slope of the tangent line drawn in fraction flow curve.

Later Johnson, Bossler and Naumann (1959) extended Welge’s technique. This so-called JBN method permits calculation of
individual relative permeability from displacement data obtained on normal sized reservoir core samples. The theory requires
two conditions that must be achieved before applying for use. One is that the flow velocity must be high enough to stabilize
displacement and the other requirement is that the flow velocity is constant at all cross sections of the linear porous body.
Upon meeting the requirement, relative permeabilities (water-oil system) are then derived as:
W I
= …………...………………………………………………………....….[2]
W
μ
k k …………...……………………………………….……..…….[3]
μ
where, Wi = cumulative injection of pore volumes
fo = fraction flow of oil
Ir = relative injectivity (defined as the ratio of intake capacity, u/Δp, at any time to the intake capacity at the initial
flood stage)

Extension of the Buckley-Leverett Theory to Three Phase

Grader and O’Meara (1988) proposed a method to extend the Buckley-Leverett theory to three phases and made it possible to
calculate three-phase relative permeabilities along each saturation trajectory in displacement experiments. This theory also
invokes the simplifying assumptions as Buckley-Leverett theory, one dimensional, incompressible, immiscible three phase
flow with no capillary pressure. In this method, the saturation of each phase j (water, oil or gas), can be calculated as a
function of the initial saturation of that phase, S , the pore volume injected, η, the pore volume produced of phase j, L , and
the slope of the recovery curve for that phase. The relative permeability is calculated using the equation given below:
E
L µ
k [ P] …………...…………………………………………..…….[4]
A P

From the curve, pore volume produced of phase j versus pore volume injected η, slope of the tangent yields f E , and the
intercept with y-axis will give ΔS (ΔS =S -S ).
S is the saturation of each phase and given by formula
L
S =S -L +η ……………………………………………………………..……….[5]
η is the pore volume injected of phase i and L is the recovery of phase j in pore volumes.
SPE 145808 3

P
The denominator value, ΔP η , can be obtained from the intersection of the tangent and the pressure axis by constructing
the pressure drop versus pore volume injected curve. Therefore, the relative permeability of each phase can be calculated.
Detailed procedures of getting these values can be found in Grader and O’Meara’s (1988) paper. This method is used
extensively in this research work for calculating all the three-phase relative permeabilities.

Experimental Methodology

The purpose of this research work is to further investigate the dynamics of three phases flow simultaneously through porous
media. Upon reviewing the literature in this area, it is found that lot of work has been done on the two phase cases, however,
few experimental works have been reported for the three-phase cases. A series of laboratory dynamic displacement
experiments are conducted under room condition on horizontal, one dimensional core samples using three immiscible fluids,
brine, flouronert 3M-40 (FC-40) and Soltrol-130. This research is aimed at quantifying the three-phase relative permeabilities
by employing unsteady-state methodology and comparing the experimental results with those generated from analytical
models. In this work, Stone I , Stone II, Baker and Hustad and Hansen’s models will be used for comparison.

Apparatus

A special coreflooding apparatus was set up for conducting the unsteady-state coreflood experiments. A schematic drawing
of the experimental design in use is presented in Figure 1. Fluids are kept in the accumulators and transferred at desired flow
rates by high pressure pumps (ISCO D Series Syringe Pumps 500D) into the core. The produced fluids passing through the
core are carried through the backpressure regulator (Temco®) into the fraction collector (LKB 7000 UltroRac fraction
collector) where the effluents are collected in test tubes and their volumes are measured as a function of time. During the same
period, the inlet and outlet pressures are measured using pressure transducers (Honeywell® LM model Gauge Pressure
Transducer) and the electronic output signals are transmitted to the data acquisition system (National Instrument Field Point
cFP-2020) for pressure recording.

A Berea core sample (1.5” x 6.8”) with porosity of 23.3% and average permeability of 135.5md was used as the porous
medium. Three immiscible fluids, Brine (2% KCl by weight, 1.116 cp at 21.5 ºC), Fluorinert-40 (FC-40®3M, 4.389 cp at
21.5ºC) and Soltrol-130 (1.48 cp at 21.5 ºC) are used as the three phases in this experimental work. In this liquid system, Brine
represents the water phase; FC-40 represents the heavy or DNAPL (dense non-aqueous phase liquid) phase while Soltrol
represents the light or LNAPL (light non-aqueous phase liquid) phase.

Procedures

Before each experiment, the core is carefully cleaned to ensure that the wettability is not altered and also to ensure the
repeatability to within satisfactory limits, of brine permeability measurements. The Dean-Stark extraction method was applied
to clean the core by using three solvents, toluene, methanol and FE-32®3M. Toluene is used to dissolve Soltrol while methanol
and 3M Novec fluid HFE-72DE (a hydrofluoroether solvating agent) are used to remove salt and FC-40, respectively. During
each experiment, the absolute permeability was determined by flooding the core with brine at two different flow rates and
Darcy’s law was applied for the calculations.

Three types of the unsteady-state (or dynamic displacement) experiments are conducted to map a wide range of saturation
regions. In these runs, only one phase is dynamically injected.
1. IDD: Water dynamic displacement Runs (increase the water phase, decrease the DNAPL phase and LNAPL phase)
2. DDI: Soltrol dynamic displacement Runs (increase the LNAPL phase, decrease the water phase and DNAPL phase)
3. DID: FC-40 dynamic displacement Runs (increase the DNAPL phase, decrease the water phase and LNAPL phase)

All the three-phase dynamic displacement experiments are conducted under room condition (21.5ºC, 13.2 psi) and with 500
psi backpressure. Injection fluids are 2% KCl brine, FC-40 and Soltrol-130. All DDI runs are designed to obtain water-FC-40
two phase relative permeability data and three-phase relative permeabilities as well. IDD runs are aimed at getting FC-40-
Soltrol two phase relative permeability data and three-phase relative permeabilities. DID runs though not having any
commercial significance, are performed to observe how the three-phase saturation and relative permeabilities will behave.
Another goal of all the runs is to map the three-phase saturation trajectories during the third phase dynamic displacement.

The following procedures were followed during unsteady-state relative permeability measurement in the laboratory. First of
all, the core sample was vacuum-saturated with brine and absolute permeability was measured for each run. For the Soltrol
dynamic displacement (DDI) runs, drainage and imbibition stages were carried out. FC-40 (at 10cc/min) and water (at
10cc/min) were injected respectively to obtain the water-FC-40 two-phase relative permeability data (needed for the empirical
4 SPE 145808

models), followed by simultaneously injecting water and FC-40 at fixed flow rate ratio (1:2) to establish an initial two-phase
(water-FC-40) saturation. The dynamic Soltrol injection (at 10cc/min) is conducted after that stage. The FC-40 dynamic
displacement (DID) runs were conducted in a similar manner. The only difference is that Soltrol and water were injected
during drainage and imbibition stages, as well as for establishing the initial saturation prior to the dynamic FC-40 injection.
For water dynamic displacement runs (IDD), the procedure is a bit more complicated and it involved injecting FC-40 until the
core reaches the irreducible water saturation (Swir) before performing the drainage and imbibition stages using FC-40 and
Soltrol. During all experiments, capillary pressure was not included for relative permeability measurement.

Results and Discussions

A total of eight experimental runs were conducted using the same porous medium in this work. Each run consists of a series
of displacements using the same porous medium (Berea sandstone core sample). Among all the runs, Run 1 and 2 are for
equipment familiarity and apparatus verification purpose. Distilled water was used in these runs to test the accumulator and
core holder mounting, the fraction collector and the data acquisition system. Since test data were not adequately recorded for
these runs, analysis and results are not reported here. Runs 3 and 6 are DDI runs, Runs 4 and 7 are IDD runs and Runs 5 and 8
are DID runs.

Results from the DDI Runs

Two-phase Water-FC-40 relative permeability data were obtained from the DDI runs. The relative permeability analysis
represents irreducible water saturations (Swir) of 25% and 22%, and residual FC-40 saturation (Sorw) of 47.1% and 34% for Run
3 and Run 6, respectively. After completing the drainage and imbibition cycles, steady-state saturation is performed by
flowing brine and FC-40 simultaneously at a fixed flow rate ratio. The core was saturated with both brine and FC-40, and a
final water saturation of 38.9% and FC-40 saturation of 61.1% were obtained in Run 3 from the material balance calculation.
Similarly, in Run 6, a final water saturation of 41.35% and FC-40 saturation of 58.65% were calculated.

During the three-phase relative permeability measurement, Soltrol is injected at 10cc/min rate. Take Run 3 as an example,
Figure 2 gives the normalized differential pressure curve and Figures 3 to 5 show the three recovery curves. The differential
pressure is recorded every 5 seconds. The differential pressure curves can be used to identify pump problem, pressure surge,
and flow rate changes. They can also be used for later relative permeability calculations using the Jones and Roszelle (1978)
graphical technique. Normally, the elapsed time is converted to flow rates and then to pore volumes injected for calculation
purposes. The dots in these recovery graphs are the actual data points representing the data collected from individual graduated
test tubes. In Figure 3, it is shown that the water production rate is constant during the early part of Soltrol injection and the
recovery curve is a straight line prior to breakthrough. Abrupt decline of the slope of the recovery curve, which denotes the
breakthrough, is observed around 0.33 PVI. FC-40 and Soltrol breakthrough is also observed around 0.33 PVI from recovery
curves. After breakthrough, Soltrol starts to be produced while water and FC-40 recovery starts to taper off.

Upon completion of the experiments, with the differential pressure and recovery versus pore volume injected curves plotted,
the tangent construction method is used for calculating the relative permeabilities of the three-phase system. Figures 6 and 7
give the water, FC-40 and Soltrol relative permeability as a function of its own saturation in semi-logarithmic plots for Run 3
and Run 6. The dots on the curves are the actual data points calculated at various pore volumes injected. In both runs, Soltrol
relative permeabilities are generally higher than the relative permeabilities of the other two phases. By applying the three-
phase Buckley-Leverett theory (Grader and O’Meara, 1988), the three-phase saturations during dynamic flood are calculated
and saturation trajectories are plotted for each DDI run in a ternary diagram as shown in Figures 10 and 11.

The initial (100%) water saturation is shown as point 1 in the ternary diagram (Figure 10). The Berea sandstone core is then
flooded to irreducible water saturation with FC-40 (point 2), and then brought to residual oil saturation (point 3) by flooding
water again. Point 4 corresponds to 38.9% water saturation and 61.1% FC-40 saturation at the end of steady-state saturation
by injecting water and FC-40 simultaneously and from where the dynamic Soltrol flood is started. The Buckley-Leverett
shock front is shown by the jump from point 4 to point 5, which represented by a sudden increase in Soltrol saturation and
sudden decreases in water and FC-40 saturations. The continuous saturation trajectory after breakthrough is shown by the line
from point 5 to point 6 where the Soltrol injection is stopped. The dots along the trajectory are the calculated saturations at
different pore volumes injected.

Results from the IDD Runs

Two-phase FC-40-Soltrol relative permeability data were obtained from the IDD runs. The relative permeability analysis
represents irreducible water saturation (Swir) of 39% for Run 4, and 21.2% for Run 7. This saturation of water remained
constant during the whole IDD experiment. Residual FC-40 saturation (Sorg) during the Soltrol injection is calculated to be
SPE 145808 5

19% for Run 4 and 27% for Run 7. At the end of FC-40 injection, residual Soltrol saturation (Sgr) is found to be equal to 6.3%
and 15.7% for Run 4 and Run 7, respectively.

After completing both the drainage and imbibitions cycles, steady-state saturation is performed by flowing FC-40 and Soltrol
simultaneously at a fixed ratio (1:2). A final FC-40 saturation of 24.9% and a Soltrol saturation of 36.1% were obtained in
Run 4. Similarly, in Run 7, the final saturations of FC-40 and Soltrol were 32.1% and 46.7%, respectively. During the
dynamic water displacement test (10 cc/min), neither FC-40 nor Soltrol was produced after breakthrough during the water
injection. It may be due to bypassing of water, as water has the lowest viscosity among the three phases. The saturation
trajectories for each IDD run are plotted in a ternary diagram shown in Figures 12 and 13.

Results from the DID Runs

The coreflooding experiments for DID gives irreducible water saturations (Swir) of 48% and 26.1% for Runs 5 and 8,
respectively. During the imbibition stage, Soltrol ceased to be produced after breakthrough and its saturation stayed constant
at 27.5% in Run 5, and 44.7% in Run 8. After the steady-state injection of FC-40 and Soltrol simultaneously at a fixed ratio
(1:2), a final water saturation of 60.17% and a Soltrol saturation of 39.83% were achieved in Run 5. In Run 8, the
corresponding final saturations were 52.35% and 47.65%, respectively.

During the three-phase relative permeability test, water is injected at 5 cc/min in Run 5, and at 10 cc/min in Run 8. After
calculating the relative permeabilities of the three-phase system using the graphical technique, Figures 8 and 9 show the water,
FC-40 and Soltrol reltive permeabilities as a function of their own saturations in semi-logarithmic plots for Run 5 and Run 8
respectively. The dots on the curves are the actual data points calculated at various pore volumes injected. By applying the
three-phase Buckley-Leverett theory, the three-phase saturations during the dynamic flood are calculated and the saturation
trajectories are plotted for each DID run in ternary diagrams shown in Figures 14 and 15.

Verification of Saturation Trajectory

One part of this research goes to the verification of previous work. Results from the dynamic Soltrol flooding (DDI runs), i.e.,
Runs 3 and 6, are used to compare and verify Grader and O’Meara’s (1988) findings. The saturation trajectory obtained from
the two DDI runs are plotted in Figures 10 and 11. Similar trend is observed in the saturation trajectories of the seven dynamic
decane (light phase) injection runs from Grader and O’Meara’s work (see Figure 16) as well as the gas injection runs reported
by Sarem (1965) (see Figure 17).

DID runs, though not commercially feasible, are used to map the interior of the ternary diagram. These experiments being
exploratory in nature, they provide more three-phase relative permeability data and demonstrate another set of saturation
profiles, data for which have never been collected before. It is also interesting to see in Run 8, even though the water
saturation is lower than Soltrol saturation prior to dynamic FC-40 injection, a final Soltrol recovery of 0.34 PV which is higher
than the water recovery of 0.24 PV. The same phenomenon has been observed in Siddiqui et al.’s (1998) work. This may be
attributed to the strong water-wet nature of the Berea sandstone.

The objective of mapping relative permeabilities in the interior region of the three-phase ternary diagram is not fully achieved
through the IDD runs. It can be seen from the dynamic water flooding recovery curves of Run 4 and Run 7, both FC-40 and
Soltrol are displaced to their dynamic residual saturations at breakthrough during the water flood and there is a sharp increase
in water saturation. It is clearly shown in Figures 12 and 13 that the interior region of the ternary diagram is bypassed. This
bypassing phenomenon has also been reported by Thomas (1991) and Siddiqui (1994) in their three-phase dynamic water
flood experiments. The difference is that only decane (light phase) was found to be bypassed from their work while the other
two phases both ceased being produced after the breakthrough in current research. This may be due to the fact that the
viscosity difference between the displacing phase and the displaced phases is much greater in current work compared to the
work by Thomas (1991) and Siddiqui (1994).

Verification of Three-phase Relative Permeabilities

Like the IDD runs, water bypassed both FC-40 and Soltrol after breakthrough, and therefore three-phase relative
permeabilities could not be analyzed from these runs. However, FC-40-Soltrol two phase relative permeability data from these
runs were extracted for three phase relative permeability calculations using Stone’s models. It can be found out from the
three-phase relative permeability curves that the relative permeability of a particular phase increases as a function of the
saturation of that phase within the range in the experiment and the curve looks quite similar for the same type of dynamic
experiments. In general, the injecting phase always has a higher relative permeability than the other two displaced phases in
the three-phase system. It also demonstrates that the relative permeability of each phase is not constant at a given saturation,
6 SPE 145808

which indicates that the relative permeability of a phase also depends on the saturations of the other two phases during three-
phase flow.

Figure 18 shows the three-phase relative permeability experimental results by Edgar-Ricardo (2008) using Colombian
Foothills core samples. In their research work, nitrogen is injected into a water-oil two phase system and similar trend of
relative permeability change during the gas phase injection is observed. Using water, Benzyl Alcohol and decane in Berea
sandstone, Siddiqui et al. (1998) observed similar behavior during dynamic injection of decane and their result is shown in
Figure 19.

Comparison of Experimental Three-phase Relative Permeability Results with Empirical Models

With the water-FC-40 and Soltrol-FC-40 two-phase relative permeability data available from DDI and IDD runs, three-phase
relative permeability values are calculated using four different empirical models: Stone’s I (1970) and II (1973), Baker’s
(1988) and Hustad and Hansen’s (1995) and compared with those experimental results from dynamic Soltrol flood and FC-40
flood (Cao, 2010). For the DDI runs, drainage data from water-FC-40 and FC-40-Soltrol two phase system are used for
Stone’s model calculation. For the DID runs, drainage data for water-FC-40 relative permeability and imbibition data for FC-
40-Soltrol relative permeability are utilized. Detailed comparisons are shown in Figures 20 through 23 and comparison data
are available in Tables 1 through 4.

For Run 3, Stone’s model I matches better than the other models, especially at higher FC-40 saturations (from around 50% to
55%). Within the saturation range of 37% to 43%, the Hustad and Hansen model gives a better match. For Run 6, the data
matched very well with both of the Stone’s models. For Run 5, all the models under-predicted the relative permeability of FC-
40 during dynamic FC-40 flood. In Run 8, Stone’s Model I gives a better match within the saturation range of 20% to 45%
while Stone’s Model II and Baker’s model provide a better match at higher FC-40 saturations.

Two separate sets of experiments have been done for both DDI and DID type of runs. The only difference between each DDI
run or DID run is the different initial saturations before the dynamic injection of the third phase. Relative permeability of each
phase and residual saturation vary from run to run which is mainly due to the fact that the initial saturations are different.
Changes of the initial steady state condition of the porous media can play a significant role in the whole three-phase fluid flow
process.

The discrepancy between experimental results and model predicted ones may result from the hysteresis effects caused by
oscillating saturations. In the current research, before each run of the third phase dynamic injection, the porous medium has
experienced both drainage and imbibition processes. In that case, these models selected for comparison may not be capable of
treating the complicated hysteresis effects. Therefore, inconsistency is exposed between the three-phase relative permeability
data obtained upon finishing back and forth saturation and those estimated directly from models.

Conclusions and Recommendations

1. A novel coreflood apparatus has been constructed and the laboratory data successfully generated three-phase relative
permeabilities by applying the extension of Buckley-Leverett three-phase theory.
2. Same behaviors as literature data are seen in saturation paths from the DDI experiments.
3. The DID experimental data, usually rare in three-phase relative permeability tests, contributed to mapping more of the
interior region of the ternary diagram.
4. Bypassing was observed in the case of the IDD runs. After breakthrough, neither FC-40 nor Soltrol was observed. This
may be due to the specific immiscible liquid system used in current the research in which water has the lowest viscosity.
The reason behind this bypassing phenomenon needs further investigation.
5. Discrepancy between model derived results and physical experiments may be due to the complex saturation processes
conducted in this research work. The back-and-forth saturation of the core before the third phase injection procedure is
another unique exploration of relative permeability and saturation trajectory behaviors during three-phase flow. Empirical
models that are generally used for three-phase flow are not always reliable for predicting three-phase relative
permeabilities, especially when complicated hysteresis effects exist.
6. More research needs to be conducted in the area of the hysteresis experiences and relative permeability models which are
more appropriate for treating the hysteresis effect still needs to be studied.
7. With the capability of current experiment apparatus, later experiments can be conducted under reservoir conditions with
higher temperature and pressure. Effects from viscosity changes due to higher temperature on three-phase relative
permeabilities are to be explored.
SPE 145808 7

Acknowledgement

The authors wish to acknowledge the Petroleum Engineering Department of Texas Tech University for supporting this
research project financially. The authors also acknowledge Mr. J. McInerney for helping with the core preparation and
equipment setup.

References
1. Baker, L. E.: “Three-Phase Relative Permeability Correlations,” paper presented at the SPE Enhanced Oil Recovery Symposium, 1988.
2. Buckley, S. E. and Leverett, M. C.: “Mechanism of Fluid Displacement in Sands,” SPE paper 942107, 1942.
3. Cao, P.: “Three-Phase Unsteady-State Relative Permeability Measurements in Consolidated Cores Using Three Immiscible Liquids,” MS Thesis, Texas Tech University, May 2010.
4. Edgar-Ricardo.: “A New Method for the Experimental Determination of Three-Phase Permeability,” CT&F Cienacia, Tecnologia' y Futuro, 2008.
5. Grader, A. S. and O’Meara Jr., D. J.: “Dynamic Displacement Measurements of Three-Phase Relative Permeabilities Using Three Immiscible Liquids,” SPE paper 18293 presented at the
63rd Annual Technical Conference and Exhibition of the SPE in Houston, Texas, October 2-5, 1988.
6. Honarpour, M, Koderitz, L.F and Harvey, A. 1986. Relative Permeability of Petroleum Reservoirs. CRC press, Florida.
7. Hustad, O.St. and H. A. G.: “A Consistent Correlation for Three-phase Relative Permeabilities and Phase Pressures Based on Three Sets of Two Phase Data,” paper presented at 8th
European IOR Symposium in Vienna, Austria,1995.
8. Johnson, E. F., Bossler, D. P. and Naumann, V. O.: “Calculation of Relative Permeability from Displacement Experiments,” SPE paper 1023, 1959.
9. Jones, S.C. and Roszelle, W.O.: “Graphical Techniques for Determining Relative Permeability from Displacement ExpAeriments,” JPT, May, 1978, pp. 807-817.
10. Sarem, A. M.: “Three-Phase Relative Permeability Measurements by Unsteady-State Method,” SPE paper 1225 presented at SPE Annual Fall Meeting held at Denver, Colorado, Oct 3-6.
1965.
11. Siddiqui, S.: “Three-phase Dynamic Displacements in Porous Media”, PhD Dissertation, The Pennsylvania State University, University Park, Pennsylvania,1994.
12. Siddiqui, S., Hicks Jr., P.J. and Grader, A.S.: “Verification of Buckley Leverett Three-phase Theroy Using Computerized Tomography,” J.Pet.Sci.Eng, 1996, 15(1), pp.1-21.
13. Siddiqui, S., Yang, J., and Ahmed, M.: “Dynamic Displacement Measurement of Three-Phase Relative Permeabilities in Berea Sandstone Cores,” SPE paper 49306 presented at the SPE
Annual Technical Conference and Exhibition, New Orleans, Louisiana, September 27-30, 1998.
14. Stone, H. L.: “Probability Model for Estimating Three-Phase Relative Permeability,” SPE paper 2116, Journal of Petroleum Technology, Vol.22 (2), pp.214-218, 1970.
15. Stone, H.L.: “Estimation of Three-Phase Relative Permeability and Residual Oil Data,” The Journal of Canadian Petroleum Technology, pp.53-61, 1973.
16. Thomas, L. W.: “Three-Phase Dynamic Displacement of Relative Permeability in Porous Media Using Three Immiscible Liquids”, MS Thesis, The Pennsylvania State University, University
Park, Pennsylvania, 1991.
17. Welge, H. J.: “A Simplified Method for Computing Oil Recovery by Gas or Water Drive,” SPE paper 124 presented at the Petroleum Branch Fall Meeting in Oklahoma City, Oklahoma,
Oct.3-5, 1952.

Table 1: Relative Permeability Comparison with Empirical Models for Run 3

krFC krFC krFC krFC krFC

Sw SFC SSol
(Stone I) (Stone II) (Baker) (Hustad&Hansen) (Experiment)
0.2905 0.5260 0.1835 0.1503 0.4011 0.4067 0.0598 0.1549
0.2804 0.4650 0.2546 0.0633 0.2223 0.2481 0.0413 0.0932
0.2729 0.4350 0.2921 0.0307 0.1428 0.1658 0.0313 0.0625
0.2669 0.4160 0.3171 0.0205 0.1296 0.1444 0.0269 0.0412
0.2622 0.4120 0.3258 0.0181 0.1245 0.1332 0.02608 0.0412
0.2583 0.3890 0.3527 0.0060 0.1186 0.1211 0.0209 0.0227
0.255 0.3850 0.3600 0.0037 0.1163 0.1145 0.0200 0.0193

Table 2: Relative Permeability Comparison with Empirical Models for Run 6

krFC krFC krFC krFC krFC

Sw SFC SSol
(Stone I) (Stone II) (Baker) (Hustad&Hansen) (Experiment)
0.331 0.4065 0.2625 0.1641 0.1930 0.4308 0.0878 0.1532
0.3145 0.3715 0.3140 0.1427 0.1688 0.3665 0.0702 0.1310
0.3055 0.3535 0.3410 0.1300 0.1543 0.3311 0.0603 0.1307
0.2975 0.3365 0.3660 0.1227 0.1446 0.3032 0.0505 0.1234
0.2935 0.3065 0.4000 0.0900 0.1038 0.2523 0.0304 0.0871
0.2835 0.2935 0.4230 0.0695 0.0784 0.2075 0.0231 0.0761
0.2765 0.2865 0.4370 0.0547 0.0593 0.1756 0.0192 0.0697
0.2655 0.2561 0.4784 0.0171 0.0109 0.1049 0.0030 0.0385
0.2575 0.2335 0.5090 0.0006 0.0109 0.0671 0.0020 0.0161
0.2525 0.2265 0.5210 0.0005 0.0100 0.0574 0.0013 0.0114
8 SPE 145808

Table 3: Relative Permeability Comparison with Empirical Models for Run 5

krFC krFC krFC krFC krFC

Sw SFC SSol
(Stone I) (Stone II) (Baker) (Hustad&Hansen) (Experiment)
0.4152 0.3215 0.2633 0.1066 0.3116 0.3732 0.0017 0.4232

0.4054 0.3353 0.2593 0.1117 0.3188 0.3770 0.0052 0.4458

0.3772 0.3875 0.2353 0.1654 0.4408 0.4435 0.0209 0.5561

0.3646 0.4181 0.2173 0.1748 0.4524 0.4487 0.0292 0.6274

0.3567 0.4440 0.1993 0.1819 0.4619 0.4514 0.0355 0.6904

0.3517 0.4580 0.1903 0.1858 0.4675 0.4537 0.0389 0.7219

0.3437 0.4800 0.1763 0.1917 0.4764 0.4567 0.0440 0.7546

0.3367 0.4900 0.1733 0.1944 0.4802 0.4596 0.0470 0.7868

0.3317 0.5000 0.1683 0.1980 0.4865 0.4623 0.0498 0.8113

0.3267 0.5120 0.1613 0.2008 0.4905 0.4638 0.0525 0.8340

0.308 0.5517 0.1403 0.2063 0.4965 0.4673 0.0610 0.8574

0.2977 0.5800 0.1223 0.2107 0.5032 0.4688 0.0654 0.8776

0.2869 0.5998 0.1133 0.2133 0.5068 0.4723 0.0692 0.8965

0.2757 0.6140 0.1103 0.2153 0.5095 0.4795 0.0734 0.9142

0.2642 0.6505 0.0853 0.2171 0.5114 0.4773 0.0747 0.9310

0.2525 0.6692 0.0783 0.2189 0.5143 0.4981 0.0778 0.9473

Table 4: Relative Permeability Comparison with Empirical Models for Run 8

krFC krFC krFC krFC krFC

Sw SFC SSol
(Stone I) (Stone II) (Baker) (Hustad&Hansen) (Experiment)
0.4025 0.2050 0.3925 0.3625 0.6836 0.7738 0.0070 0.2559

0.3585 0.2630 0.3785 0.4681 0.7177 0.7900 0.0080 0.3604

0.3475 0.2810 0.3715 0.5016 0.7357 0.7983 0.0125 0.3952

0.3355 0.3020 0.3625 0.5290 0.7428 0.8017 0.0376 0.4662

0.3255 0.3182 0.3563 0.5511 0.7506 0.8053 0.0576 0.5127

0.3185 0.3305 0.3510 0.5713 0.7627 0.8108 0.0731 0.5334

0.3085 0.3440 0.3475 0.5879 0.7673 0.8118 0.0911 0.5561

0.2745 0.3810 0.3445 0.6299 0.7739 0.8102 0.1466 0.5918

0.2535 0.4155 0.3310 0.6618 0.7853 0.8127 0.1958 0.6970

0.2475 0.4470 0.3055 0.6882 0.8066 0.8296 0.2322 0.7729

0.2345 0.4690 0.2965 0.6752 0.8243 0.8421 0.2668 0.8400

0.2285 0.4850 0.2865 0.7321 0.8400 0.8529 0.2891 0.8843

SPE 145808 9

Figure 1: Schematic of the coreflooding system.

Figure 2: Normalized differential pressure Figure 3: Pore volume of water produced versus
during Soltrol injection for Run 3. pore volume of Soltrol injected during DDI for Run 3.

Figure 4: Pore volume of FC-40 produced versus Figure 5: Pore volume of Soltrol produced versus
pore volume of Soltrol injected during DDI for Run 3. pore volume of Soltrol injected during DDI for Run 3.
10 SPE 145808

Figure 6: Three-phase relative permeability of Figure 7: Three-phase relative permeability of

each phase during Soltrol injection for Run 3. each phase during Soltrol injection for Run 6.

Figure 8: Three-phase relative permeability of Figure 9: Three-phase relative permeability of

each phase during FC-40 injection for Run 5. each phase during FC-40 injection for Run 8.

Figure 10: Saturation trajectory for DDI Run 3. Figure 11: Saturation trajectory for DDI Run 6.

Figure 12: Saturation trajectory for DID Run 4. Figure 13: Saturation trajectory for DID Run 7.
SPE 145808 11

Figure 14: Saturation trajectory for DID Run 5. Figure 15: Saturation trajectory for DID Run 8.

Figure 16: Saturation trajectories for Run 5,6,7,8,9,14

Figure 17: Saturation trajectories for Run 1-7 by Sarem.
and 15 by Grader and O’Meara.

Figure 18: Three-phase unsteady-state relative permeabilities obtained by Edgar-Ricardo (2008).

12 SPE 145808

Figure 19: Three-phase relative permeabilities vs. saturation in Run 1 obtained by Siddiqui et al. (1998).

Figure 20: Comparison of FC-40 relative Figure 21: Comparison of FC-40 relative
permeability for Run 3. permeability for Run 6.

Figure 22: Comparison of FC-40 relative Figure 23: Comparison of FC-40 relative
permeability for Run 5. permeability for Run 8.