SPE 75502
Residual Gas Saturation to Aquifer Influx: A Calculation Method for 3-D Computer
Reservoir Model Construction
M.H. Holtz, SPE, Bureau of Economic Geology, The University of Texas, Austin
Copyright 2002, Society of Petroleum Engineers Inc.
This paper was prepared for presentation at the SPE Gas Technology Symposium held in Calgary,
Alberta, Canada, 30 April2 May 2002.
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. Electronic reproduction, distribution, or storage of any part of this paper for
commercial purposes 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
where and by whom the paper was presented. Write Librarian, SPE, P.O. Box 833836, Richardson,
TX 75083-3836, U.S.A., fax 01-972-952-9435.
Abstract
Residual gas saturation controls the volume of gas trapped
in that portion of the reservoir that has experienced water
encroachment. As water moves into a rock volume filled
with gas, the water displacement of the gas is incomplete. The
water fills pores and pore throats, causing capillary pressure and
relative permeability effects to stop the flow of gas and allow
only water to pass through the rock volume. This stoppage
results in gas being trapped behind the encroaching waterfront
as residual gas. The volume and location of the residual gas
are controlled by the distribution of the petrophysical properties.
A method based on interrelationships between petrophysical
properties is used to create a model for calculating maximum
residual gas saturation (Sgrm). The model is developed as a
function of porosity, permeability, capillary pressure, and initial
water saturation. The input to the model and its results compare
favorably with actual field data where aquifer encroachment is
verified from well production history.
Introduction
Maximum residual gas saturation (Sgrm) is what initially re-
sults from imbibition on rock at irreducible water saturation
(Swirr). Sgrm results from gas acting as the nonwetting phase
during imbibition hysteresis as pressure is depleted in a gas
reservoir and an aquifer encroaches in pore space that was once
filled with gas. Because Sgrm normally occurs from aquifer
influx into a gas reservoir, it causes a reduction in reservoir
recovery efficiency. The range of Sgrm is extensive, varying
from 0.1 to 0.7.1,2
Numerous influences may affect Sgrm: (1) how the wetting
fluid gets in (either forced or spontaneous imbibition), (2) type
of wetting fluid, (3) rate of imbibition, (4) rock type (lithology,
grain size and sorting), (5) pore type, (6) wettability and inter-
facial tensions, (7) temperature and pressure conditions, and
(8) petrophysical properties (porosity, permeability, initial gas
saturation). The effects each of these has on Sgrm have been
studied, although a unifying theory describing Sgrm has yet to be
produced.
Neither the mechanism nor the imbibing fluids seem to affect
the resulting Sgrm value. Geffen et al.3 demonstrated that Sgrm
values were similar for either forced or spontaneous imbibition.
Cromwell et al.4 demonstrated this phenomenon in Boise
sandstone. Geffen et al.3 showed that the wetting phase in
conjunction with gas did not affect the value of Sgrm obtained.
Jerauld5 and Kyte et al.6 substantiated this result.
The rate of imbibition by the wetting phase also seems
to have little effect on the value of Sgrm obtained. 4 ,7,8
This conclusion follows from the observation that neither forced
nor spontaneous imbibition significantly affects the value of
Sgrm because both occur at different rates.
Rock and pore type can have a strong affect on the value of
Sgrm, and variations in carbonate rock types can also
significantly affect Sgrm.2,9 Sgrm increases with an increase in
clay content in sandstones and decreases in sorting and grain
size.5, 10 In a dual-fractured pore network system Sgrm can be
quite high.11 These results intuitively seem correct because Sgrm
is a capillary phenomenon and these rock characteristics that
appear to increase Sgrm also increase pore network complexity.
The effect of wettability and interfacial tensions in an
oil-water system can be profound, although in a gas-water system
there is less variation and thus less effect. Crowell4 reported that a
decrease in interfacial tension between wetting and nonwetting
phases results in a slight decrease in Sgrm.
Varying temperature and pressure conditions has been found
to have no effect on the value of Sgrm.3,7,8 This result seems
reasonable insofar as the variations in conditions do not change
either the wetting condition or the pore network. It is conceivable
that pore and pore throat geometries could change enough under a
high-pressure compressible condition to affect Sgrm; however, to
date this has not been reported.
2 M.H. Holtz
The relationship between Sgrm and petrophysical properties,
including permeability, initial gas saturation, and porosity, has
been studied. Only a weak correlation has currently been found
between permeability and Sgrm: Sgrm shows a slight trend to
increase with decreasing permeability.12 Sgrm at face value must
be equal to or less than initial gas saturation. In general, as
initial gas saturation decreases, rock pore networks are getting
more complex, porosity is decreasing, and pore throat size is
decreasing; therefore, conditions are set for Sgrm to increase.13
Both Agarwal14 and Land15 developed relationships between
Sgrm and gas saturation.
Porosity has been shown to have the strongest relationship
to Sgrm. Nearly all studies that describe a porosity-
Sgrm relationship indicate that Sgrm, increases as porosity
decreases.1,2,5,7,8,10,1620 Some weak relationships between these
two properties have been noted, as well as an inverse rela-
tionship.10
Although the aforementioned experimental results have
resulted in a greater understanding of Sgrm, none has ap-
proached the practical engineering application of building this
petrophysical property into a reservoir model. An approach is
presented herein for applying petrophysical interrelationships
so that a model for Sgrm can be developed. This type of
practical empirical Sgrm model can then be applied so that a 3-D
geocelluar reservoir model can be populated with Sgrm values.
Residual Gas Saturation Model
Nonwetting Phase Trapping Mechanism. With water acting as
the wetting phase and gas acting as the nonwetting phase, Sgrm
results from pore scale capillary forces. Sgrm is the trapped
nonwetting phase when the wetting phase has been imbibed
into the rock from a state of irreducible water saturation to a
state of zero capillary pressure. The models that describe how
this trapping occurs are pore-geometry dependent. Three trapping
models are possible (Fig. 1). The pore doublet model is more
likely to occur in poorly sorted rock or in rock with dual-porosity
networks. The pore snap-off and dead-end models are more likely
to occur in lower porosity rocks. The trapping models thus
suggest that porosity will have an inverse relationship with Sgrm.
The relationship between capillary number (Nvc) and porosity
also illustrates an inverse relationship between porosity and Sgrm.
The direct relationship between capillary number and porosity is
shown in equation 1.22
2
cos R
Nvc j cos A
(2 )1/ 2
..................................... (1)
C A robust model of residual gas saturation was developed from
As capillary number increases, nonwetting phase residual
saturation decreases (Fig. 2). In the case of a water-gas sys-
tem, therefore, Sgrm depends on porosity. Note also how the
function is dependent on rock and pore type. As the pore
network becomes more complex from well-sorted sandstone
to typical sandstone to wide pore-size distribution, Sgrm
increases. Therefore, nonwetting phase trapping occurs in-
creasingly as porosity decreases and pore-geometry complex-
ity increases.
75502
Empirical residual gas saturation model. A method-
ology has been developed for deriving an empirical model
for Sgrm in gas reservoirs. The methodology is based on
the concept that petrophysical properties are interrelated and
that the interrelationship is a function of the rocks pore net-
work geometry. The steps in the methodology are: (1) develop
a porosity-permeability relationship, (2) develop a function
for irreducible water saturation from capillary pressure data,
(3) determine a relationship for Sgrm as a function of porosity,
and (4) combine equations to obtain Sgrm as a function of initial
gas saturation. An example of this methodologys application
from the Miocene section, offshore Louisiana, is illustrated herein
The relationship between porosity and permeability was
determined by generating a best-fit equation between the two
properties. A conventional cross plot of porosity vs. permea-
bility displays both a strong exponential and power law fit with
less than 1 order of magnitude of variation (Fig. 3). Permeability
ranged to a high end of 1 D at 0.3 porosity and a low end of 1 md
at 0.14 porosity. The power law equation was chosen, having
an R2 of 0.855 because it more realistically models permeability
approaching zero, as porosity lowers. Permeability can be
predicted from
k = 7 E 7( 9.61 ) .................................................................... (2)
An expression for initial water saturation is determined by
assuming that pore geometry controls petrophysical interre-
lationships. Because the way a data set radiates out from a low
porosity-permeability point is a function of pore geometry, the
ratio of log(k)/ should give a measure of the pore geometry as
well (Fig. 4).13
Application of this pore-geometry measure facilitates the
development of an equation for initial gas saturation. The log(k)/
corresponds to variations in capillary characteristics. Capillary
pressure curves that demonstrate low irreducible water saturations
correspond to larger log(k)/ ratios (Fig. 5). This relationship
leads to equation 3 (Fig. 6), which is used for calculating
irreducible water saturation.
1.559
log(k )
Swirr = 5.159 .......................................... (3)
field and published data. A strong relationship is documented
between increasing porosity and decreasing residual gas
saturation. The trend is linear, and the best fitting equation
is given in equation 4. This relationship predicts an Sgrm
lower bound of 0.125, corresponding to a porosity limit of
unconsolidated sandstone and an upper limit of 0.5, cor-
responding to 0.05 porosity, and an Sgrm equal to initial gas
saturation 19 (Fig. 7).
75502 Residual Gas Saturation to Aquifer Influx: A Calculation Method for 3-D Computer Reservoir Model Construction
S grm = 0.9696 + 0.5473 .............................................. (4) in Sgi. This relationship represents the end points to a set of IR
The final step in developing the Sgrm model is to integrate
equations 2, 3, and 4. They must be integrated in such a way that
Sgrm is a function of Swirr so that the initial condition of Sgrm
being less than or equal to the initial gas saturation is met.
This initial condition is met with the development of an initial-
residual nonwetting phase curve (IR curve). The gen-
eral shapes of IR curves are shown in Fig. 8 (modified from
Lake22 ). These curves represent the character of an individual
rock sample. The end point to the curve is the Sgrm value.
The shape of the initial-residual wetting phase saturation
curves displays the effect of rock type. As sandstone becomes
cleaner, better sorted, and less cemented (higher porosity),
the curves move farther away from the 1 to 1 line and increase in
slope and Sgrm decreases.
The results of Chierici et al.19 correspond to Fig. 8. They
found that unconsolidated sandstones have lower Sgrm values
than consolidated sandstones and that there is an inverse trend
between initial gas saturation and Sgrm (Fig. 9). The empirical
relationship sought after should therefore follow this trend.
To determine where the trend will intersect the 1 to 1 line,
representing the reservoir pay cutoff, equation 5 is applied.
Sgrm = Sgi = 1 Swirr ............................................................ (5)
Substituting Sgrm from equation 4 and Swirr from equation 3
and then solving in terms of porosity only by substituting from
equation 3 results in an expression that gives the porosity at
which equation 5 holds:
log( E 7( 9.6057 ))
1.56
5.32 = .467 .................... (6)
The solution to equation 6 results in a porosity of 0.18 and a
permeability of 4.9 md at an Sgrm value of 0.37. This Sgrm value
represents the highest value expected in the reservoir.
Combining equations 2, 3, and 4 with equation 7 and solving
for Sgi in terms of Sgrm results in a function to describe their
relationship:
1= Swirr + Sgi.......................................................................... (7)
1.55
Sgrm 0.547
9.606
log 7 E 7
.9696
Sgi = 1 5.159
.......... (8)
Sgrm .547
.9696
The relationship between Sgi and Sgrm is a j shape lying on
its side (Fig. 10) and generally corresponds to Fig. 9. The
interrelationship between porosity, initial gas saturation, and
Sgrm can be seen as porosity decreases slowly, along with Sgi,
and Sgrm increases rapidly. However, this rapid Sgrm increase
tapers off as small increments of porosity result in large increases
3
curves for the varying rock quality within the gas reservoir.
Generalized Initial-Residual Curve (IR curve) Model. The
shapes of the gas-water system IR curves are illustrated in Figs. 8,
9, and 10. The curves must stay below the 1 to 1 line, terminate at
a given Sgrm-Sgi position, and decrease in slope with higher
quality rock. Equation 9 modified from Land15 meets these
criteria:
1 1 Swirr
Sgr = 1 / 1 + ....................... (9)
Sgrm Sg
Using equation 8, a set of Swirr-Sgrm values can be
determined. Substituting this set of values as constants and
applying equation 9 results in a group of IR curves for varying
rock quality. This result leads to the curves shown in Fig. 11.
Notice how the ends of each line define the Sgrm line dis-
played in Fig. 10. The figure now defines the entire residual
gas saturation character of varying rock quality found within the
reservoir. Additionally, this curve, along with the capillary
pressure curves, can be used to define a capillary desaturation
curve 22 (CDC curve), such as the example shown in Fig. 2.
Wireline Application and Examples
Wireline petrophysical calculations in an aquifer-swept zone give
in situ examples of how the nonwetting gas phase saturations are
behaving during the imbibition process. Both the range of Sgrm
values and their relationship to porosity are shown. Two examples
are given, one from Miocene offshore Louisiana and the other
from the Cooper Basin, Australia.
The Star Fac T1 reservoir produced from a rollover anticline
during the late 1970s and early 1980s. Open-hole logs were
run on newly drilled wells then, concurrently with pro-
duction (Fig. 12). Well 50 D4(12) was logged in 4/1984, 4 to 5
years after the three producers had been hit by aquifer en-
croachment and shut-in. Wells 50 B2(4), 50 B1 (9), and 50 B1(2)
successively watered-out at the same structural level within the
T1 reservoir that was logged over in well 50 D4 (12). Well 50 D4
(12) is interpreted as water swept when logged because of logging
vs. aquifer-encroachment timing.
Wireline gas saturation calculations also indicate that this
well was water swept when logged. Calculated gas saturation in
well 50 D4 (12) is much lower than in preproduction wells
(Fig. 13). The calculated gas saturation, which is the Sgrm
from forced imbibition due to pressure depletion in the res-
ervoir, averages 0.195 and ranges from 0.064 to 0.376. These data
correspond very well with those that were predicted with the
Sgrm model (Fig. 10). Additionally they correspond well with
Sgrm values reported in the literature (Fig. 7).
Comparing the porosity-gas saturation character also indi-
cates that well 50 D4 (12) is aquifer swept and validates the
Sgrm model. The laboratory data indicate that Sgrm increases as
porosity decreases because greater trapping mechanisms exist at
lower porosity. This trend is recognized in wireline-log-derived
Sg for well 50 D4 (12). In contrast to initial conditions in well
50 B1(2), the Sg-porosity relationship follows an increasing
4 M.H. Holtz
trend, as is expected from the Archie equation (Fig 14). The
aquifer-swept reservoir thus verifies the Sgrm model and, further,
gives a method by which swept and nonswept zones can be
distinguished.
A second field example of in situ Sgrm calculation is from
Toolachee field, Cooper Basin, Australia. Two different structural
noses within the field experienced aquifer influx before being
penetrated by a subsequent well.
Fig. 15 summarizes a time series of how this en-
croachment occurred on the east side of the field. In 1984 ori-
ginal gas-water (OGWC) contacts defined two compartments.
The north compartment OGWC was defined by well #6 at
7,372 ft subsea, and the south compartment OGWC was
defined by wells #19 and #17 at 7,402 ft subsea. In 1987 the
north compartment demonstrated water encroachment low on
structure in well #6, whereas well #11, higher on structure,
was still producing at a prewater-encroachment WGR of
6 bbl/MMSCF. In the south compartment, well #19 had just
begun to feel the effect of water encroachment, with the
encroaching gas-water contact reaching perforations at a depth
of 7,390 ft subsea.
Water influx proceeded quickly, and by 1988 the north
compartment was experiencing the effect of encroachment
farther up structure. The WGR in well #6 had increased from
60 to 100 bbl/MMSCF, and well #11 was now experiencing water
influx with the WGR increasing from 6 to 35 bbl/MMSCF.
Water had encroached at least to the lowest perforations in well
#11 at 7,295 ft subsea. In the south compartment, water had
also encroached quickly, with the WGR in well #19 increasing
from 13 to 105 bbl/MMSCF.
By 1991 extensive encroachment had occurred in both
compartments. In the north compartment well #6 had ex-
perienced a WRG increase from 100 to 205 bbl/MMSCF, and
well #11 a WGR increase from 35 to 500 bbl/MMSCF.
Additionally, well #42 was drilled, and wireline log analysis
indicated a new GWC at 7,249 ft subsea. Thus from 1984 to
1991 wireline log data demonstrate that the aquifer encroached
from 7 372 to 7249 ft subsea up structure, or 123 vertical
feet, within the north compartment. Further, the encroachment
at well #42, where no previous well existed, indicates that
water influx has been extensive and not localized around
producing well bores. In the south compartment, the WGR of
well #19 has again increased, changing from a WGR of 105
to 400 bbl/MMSCF. The actual height of water encroachment
was demonstrated in newly drilled well #48, which from
wireline logs indicated a GWC that had encroached to 7,258 ft
subsea. Therefore, wireline calculations indicate that water en-
croached from the original GWC of 7,402 to 7,258 ft subsea,
or 144 ft of vertical encroachment.
The resultant Sgrm values from well #48 are within the same
range as values used in the Sgrm modeling (Fig 16). This result
again indicates that laboratory measurements reflect what is
taking place in situ and lends credence to the Sgrm model inputs
and results. Higher values are coincident with the upper part of the
sandstone, which is only partly swept.
Southernmost Toolachee field also experienced extensive
VGU30 water encroachment. This encroachment is docu-
75502
mented in WGRs, wireline logs, and pressure data and is
summarized in the time series of Fig. 17. Initial development in
1984 put the GWC between the highest known water determined in
wells #7 and #20 and the lowest known gas in well #25 (fig. 17). The
WGR in well #25 was producing at 18 bbl/MMSCF, indicating
production influence from the water leg. By 1988, well #25 WGR had
increased from 18 to 106 bbl/MMSCF, and by 1991 it had increased
to 2,253 bbl/MMSCF. During 1991 wells #45 and #47 were drilled
up structure approximately 2 km from well #25. These wells were
approximately 1,500-psi pressure depleted, and wireline logs indi-
cated that the GWC had encroached to 7,216 ft subsea, or at
least 40 vertical feet and possibly 139 ft. As on the east side of
Toolachee field, southernmost area VGU30 has experienced
extensive pressure depletion and aquifer encroachment.
Again, resultant Sgrm values from well #47 are within the
same range as values used in Sgrm modeling (Fig 16), indicating
that laboratory measurements reflect what is taking place in situ,
substantiating Sgrm model inputs and results.
Application to 3-D Reservoir Model Construction
Petrophysical properties are interdependent. Porosity has been
related to residual gas saturation, permeability, capillary pres-
sure, and relative permeability such that all petrophysical prop-
erties are interrelated. This interrelationship can be applied in
constructing 3-D geocelluar models such that each petrophy-
sical property will have the proper value relative to another
within each cell in a model.
From the petrophysical models described, there is a logical
sequence for constructing 3-D geocellular models. The first step
is to populate the cells with porosity, either stochastically or
deterministically, at the discretion of the modeler. The porosity
will be the key variable on which the other petrophy-
sical properties are based. Porosity equation 2 allows the cal-
culation of permeability, which in turn populates each cell. Next,
initial water saturation can be calculated from equation 3,
resulting in the necessary petrophysical properties for cal-
culating initial gas in place. At this point the modeler can test
to see whether the total gas volume calculated is greater than
that produced. Equation 8 can then be applied to calculate an
Sgrm value for each cell. Finally, Sgrm values can be used
for material balance in order to test whether the 3-D model
results in enough mobile gas compared with the volume of
gas produced.
Conclusions
Developing an Sgrm model for use in 3-D reservoir model
construction resulted in a delineation of practical relationships
between petrophysical properties and useful wireline tech-
niques for determining aquifer swept zones. Following are the
most salient conclusions:
1. Increasing Sgrm with decreasing porosity functions is
a viable relationship for Miocene Gulf Coast gas reservoirs.
2. The log(k)/ ratio results in a useful measure of pore type
and pore geometry.
75502 Residual Gas Saturation to Aquifer Influx: A Calculation Method for 3-D Computer Reservoir Model Construction
3. Combining empirical permeability, porosity, and capil-
lary pressure equations results in a useful model for deter-
mining Sgrm.
4. Wireline log calculations in zones that have been water
swept result in Sgrm values very similar to those used in the
initial residual model.
5. Similar Sgrm values and porosity-Sgrm trends derived
from wireline logs aid in validating the Sgrm model.
6. The linear inverse relationship between Sgrm and porosity
is in contrast to the direct relationship between Sgi and porosity.
This contrast in trends aids in the determination of aquifer-swept
zones from wireline log calculations.
7. Applying interrelated petrophysical functions allows fine-
tuning of petrophysical properties within a 3-D geocellu-
lar model.
Nomenclature
Sgrm = maximum residual gas saturation
Swirr = irreducible water saturation
Nvc = capillary number
= porosity
C = empirical capillary number constant
j = Leverett j function
A = advancing interfacial tension angle
R = receding interfacial tension angle
= tortuosity
k = permeability (md)
= porosity (fraction)
SI Metric Conversion Factors
ft 3.048* E 01 = m
md 9.869 233 E 04 = m2
*Conversion factor is exact.
Acknowledgments
Some of the ideas presented in this paper came from helpful
discussions with James W. Jennings. Lana Dieterich edited the
paper. Patricia Alfano prepared the figures under the supervision
of Joel L. Lardon, Graphics Manager. Publication authorized by
the director, Bureau of Economic Geology.
References
1. Katz, D.L., Legatski, M.W., Tek, M.R., Gorring, L., and
Neilsen, R.L.: How Water Displaces Gas from Porous Media,
Oil and Gas Journal (Oct. 1966) 55.
2. Keelan D.K., and Pugh V.J.: Trapped-Gas Saturation in Carbonate
Formations, paper SPE 4535 presented at the 1975 SPE-AIME
48th Annual Fall Meeting, April.
3. Geffen, T.M., Parish, D.R., Haynes, G.W., and Morse, R.A.:
Efficiency of Gas Displacement from Porous Media by Liquid
Flooding, paper presented at the Fall Meeting of the Petroleum
Branch, AIME (1952), v. 195, 29.
4. Crowell, D.C., Dean, G.W., and Loomis, L.: Efficiency of
Gas Displacement from a Water-Drive Reservoir, Report of
Investigations 6735 USBM (1966), 1.
5
5. Jerauld, G.R.: Gas-Oil Relative Permeability of Prudhoe Bay,
paper SPE 35718 presented at the 1996 Western Meeting.
6. Kyte, J.R., Stanclift, R.J., Stephan, S.C., and Rapoport, L.A.:
Mechanism of Water Flooding in the Presence of Free Gas,
JPT (Sept. 1956) 322.
7. McKay, B.A.: Laboratory Studies of Gas Displacement from
Sandstone Reservoirs Having Strong Water Drive, APEA Journal
(1974) 189.
8. Delclaud, J.: Laboratory Measurements of the Residual Gas Satura-
tion, Second European Core Analysis Symposium, London
(20-22 May 1991).
9. Batycky, J., Irwin, D., and Fish, R.: Trapped Gas Saturation in
Leducage Reservoirs, J. Can. Pet. Techn. (Feb. 1988) 32.
10. Hamon, G., Suzanne, K., and Billiotte, J.: Field-Wide Variations of
Trapped Gas Saturation in Heterogeneous Sandstone Reservoirs,
paper SPE 71524.
11. Pow, M., Kantzas, A., Allan V., and Mallmes, R.: Production of
Gas from Tight Naturally-Fractured Reservoirs with Active Water,
J. Can. Pet. Techn. (1999) 38.
12. Keelan D.K.: A Practical Approach to Determination of Imbibition
Gas-Water Relative Permeability, paper SPE 4988 presented at the
1976 SPE-AIME 49th Annual Fall Meeting.
13. Holtz, M.H., and Major, R.P.: Petrophysical Characterization
of Permian Shallow-Water Dolostone, paper SPE 75214 presented
at the 2002 SPE/DOE IOR Conference.
14. Agarwal, R.G.: Unsteady-State Performance of Water-Drive Gas
Reservoirs, Ph.D. Thesis, Texas A&M University (1967).
15. Land C.S.: Comparison of Calculated with Experimental
Imbibition Relative Permeability, paper SPE 3360 presented at the
Rocky Mountain Regional Meeting (Dec. 1971).
16. Bousqui P.: Texture et Porosit de Roches Calcaries, Thesis
Ecole des Mines de Paris (1979) 191.
17. Aissaoui A., Etude Thorique et Exprimentale de lHystrsis des
Pressions Capillaires et des Permabilits Relatives en Vue du
Stockage Souterrain de Gaz, Thesis Ecole des Mines de Paris
(1983) 223.
18. Fishlock, T.P., Smith, B.M., Soper, B.M., and Wood, R.W.:
Experimental Studies on the Waterflood Residual Gas Saturation
and Its Production Blowdown, paper SPE 15455 presented in 1986.
19. Chierici, G.L., Ciucci, G.M., and Long G.: Experimental Research
on Gas Saturation behind the Water Front in Gas Reservoirs
Subjected to Water Drive, Proc., World Petroleum Congress,
Frankfurt (June 1963) 17.
20. Firoozabadi, A., and Olsen, G.: Residual Gas Saturation in Water-
Drive Gas Reservoirs, paper SPE 16355, 1987.
21. Chatzis, I., and Morrow, N.R.: Measurement and Conditions
for Entrapment and Mobilization of Residual, U.S. DOE Final
Report, DOE/BETC/3251-12 (Oct. 1981).
22. Lake, Larry: Enhanced Oil Recovery, Prentice Hall (1996) 550.
6 M.H. Holtz 75502

(a) Pore doublet model

1000
y = 7E7()9.606
Wetting
R2 = 0.855

Permeability (md)
100
Nonwetting

10
(b) Pore snap-off model
Low-aspect ratio Nonwetting

1
Wetting 0 .10 .20 .30 .40

Porosity (fraction) QAd502c

High-aspect ratio
Fig. 3A power-law equation gives a give a strong empirical fit to
Miocene StarFac porositypermeability data.

Nonwetting trapped Collar of water

by snap-off

(c) Dead-end model 10,000

1000 Fractures and touching-

Nonwetting
vug pores

Permeability (md)
phase
trapped 100

Interparticle pores
10
Wetting
phase
QAd501c
1
Fig. 1Three possible models indicate how a wetting phase can
trap a nonwetting phase as a result of variations in pore
21 0.1
geometry. ( Modified from Chatzis et al. )
Intercrystalline pores
0.01
0 2 4 6 8 10 12 14
Porosity (percent) QAb436c

Fig. 4The log(k)/f relationship is a function of pore type and pore

13
geometry. ( From Holtz and Major )
Nonwetting residual saturation (percent)

Sgrm Wide pore size distribution 160 11.7 9.5 8.3 7.7 7.3 4.6 = Log(k)
(e.g., carbonates)
Capillary pressure (psi)

Sgrm
94

30 Typical sandstone
34 m

120
md
78 m

Sgrm
134 m

9m
706 md

Well-sorted sand
d

d
d
d

20 80

10 40

0 0
107 106 105 104 103 102 0 20 40 60 80 100
NVC Brine saturation (percent) QAc9830c
QAd509c

Fig. 2The relationship between capillary number (Nvc) and Fig. 5The character of capillary pressure curves displays a strong
nonwetting phase residual saturation shows how Sgrm will ratios because they indicate pore type and
correlation with log(k)/
22
decrease as Nvc increases (from Lake ). geometry.
75502 Residual Gas Saturation to Aquifer Influx: A Calculation Method for 3-D Computer Reservoir Model Construction 7

Irreducible water saturation (fraction)

1.0 1.0

Residual gas saturation (fraction)

0.9 0.9
0.8 0.8 y = 1.0199x +1.0255
y = 5.1593(x)1.9594
0.7 R2 = 0.7409
0.7 R2 = 0.9933
0.6 ine
0.6
o 1l
0.5
0.5 1t Consolidated ss
0.4
0.4
0.3
0.3
0.2 Unconsolidated ss
0.2
0.1
0.1
0
0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0 2 4 6 8 10 12 14 Initial gas saturation (fraction) QAd507c
Log(k)/porosity (md) QAd508c Fig. 9Unconsolidated sandstones display lower Sgrm values than
consolidated sandstones, and initial gas saturation is inversely
Fig. 6An empirical power-law fit (R =0.99) between log(k)/ and 2
related to Sgrm.
Swirr results in a strong predictive tool because it takes into account
pore geometry variation.
1

Residual gas saturation (fraction)

0.9
0.8
0.6
Katz et al. (1966) 0.7
Chierici et al. (1963) e
0.5 0.6 lin
saturation (fraction)

o1
Fishlock et al. (1986)
Firoozabadi Pay cutoff t
Residual gas

0.5 1
0.4 and Olsen (1987) = 0.18
BEG data 0.4 = .19 .21 .23
k = 4.9 md .25
0.3 Linear regression
0.3
.33
0.2 0.2
0.1
.43
0.1 y = -0.9696x + 0.5473
R2 = 0.6604 0
0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0 0.1 0.2 0.3 0.4 0.5 0.6 Irriducable gas saturation (fraction)
Porosity (fraction) QAc9651c QAd510c

Fig. 7Residual gas saturation displays an inverse linear relation- Fig. 10The Sgrm model results in values ranging from 0.37 to 0.12,
ship with porosity. with Sgrm rapidly increasing as porosity and Sgi decrease and then
tapering off.

1
1.0 Sgrm = 0.1
0.9 Sgrm = .2
Residual gas saturation (fraction)

0.9
Sgrm = .3
0.8 Sgrm = .34
0.8
Incre
and

Sgrm = .36
Dec

0.7
Residual gas saturation (fraction)
Dec

asin ize
grain
reas

0.6 0.7
reas

g so

ine
ing c shale

l
s

ne
o1
ing

0.5 sto
rting

1t and
eme

ds
V

ne 0.6
nte ndsto
0.4 m e ea sa
ntat

Ce rage Ber
e
lin

Ave
ion

0.3 ndstone
1

nsolidated sa 0.5
Uniform unco
to

0.2
1

0.1 0.4
Sgrm
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.3
Initial gas saturation (fraction) QAd506c
0.2
Fig. 8The shape of the initial residual wetting phase saturation
curves displays the effect of rock type. As sandstone becomes
cleaner, better sorted, and less cemented (higher porosity), the 0.1
curves move farther away from the 1 to 1 line and Sgrm decreases.
22
(Modified from Lake ). 0
0 0.2 0.4 0.6 0.8 1
Gas saturation (fraction) QAd523c

Fig. 11From the empirically derived equations initial residual

nonwetting phase curves (IR curves) can be generated to display
petrophysical rock character at varying rock quality.
8 M.H. Holtz 75502

10,725 10,775

Logged 2/78 10,800 50_D3

50_D2 (7) Logged 5/83
50_D4 (12)
10,756
10,756
Logged 4/84 50_E1 (9)
50_C3 50_D1 (6) Logged 5/78
10,789 10,746 Logged 11/77
Logged 4/77 H2O hit 7/80
50_C2 (3)
Logged 7/83 50_C1 (1) 10,731 50_F1
10,765
Logged 9/77 Logged 3/81 10
,75
775
10,800

0
750
10,

50_B2(4)
10,

50_B1(2) 10,703
10,752 Logged 7/77
10,728 Logged 4/77 H2O hit early 79
High H2O cut after 5 years of
12/81 production
50_C4
N 10,800
Logged12/83

10,825

50_5

10,850

0 0.2 mi
0 0.3 km
Contour interval 25 ft QAd505c

Fig. 12The 50 D4(12) well was drilled and logged after wells 50 B2(4), 50 B1(9), and 50 B1(2) had watered-out at the same structural level.

20
50_D4(12) 50_B1(2) 0.7
50_B1(2), Sgi
50_B1(2) Mean = 0.20 Mean = 0.60
0.6 50_D4(12), Sgr
16 50_D4(12) Min = 0.06 Min = 0
Gas saturation (fraction)

Max = 0.38 Max = 0.64 Initial conditions

0.5
Frequency

12
0.4
8
0.3
Water swept
4 0.2

0.1
0
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
0.55
0.6
0.65
More

0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Gas saturation (fraction) Porosity (fraction) QAd504c
QAd503c

Fig. 13Wells logged before (50 B1(2)) and after (50 D4(12)) aquifer Fig. 14Water-swept zones follow the porositySgrm trend predicted
encroachment display significantly different gas saturation values. by the Sgrm model and differ from that of initial conditions.
75502 Residual Gas Saturation to Aquifer Influx: A Calculation Method for 3-D Computer Reservoir Model Construction 9

(a) (b)
1984 1987
GWC (7372) of ? GWC (7350)

24 24 6, but below 11

35 11
11 35

WGR 6
38 38
6
6

42 WGR 60

42
0 4 km WGR
13
13 13

GWC (7402) of ? GWC (7390)

19 in 30-30 at the
39 19
perfs in 19

48 48

17 17
? GWC (7390)
GWC (7402) of of 30-20 WGR
30 bbl/MMcf
30 QAa8014(a)c

(c) (d)
1988 1991

? GWC above 7295 WGR in 1991

(lowest perf in 11) 500
24
24 of 30-30 in 11

35
WGR 35 35 11
11

WGR in 1989
6 WGR 100 205
38
38 6

42 GWC (7249) 42
of 30-30
in 42
WGR in 1990
13 13 400
? GWC (7360)
N
above perfs 19
in 13, 19 GWC (7258)
of 30-20 19
WGR 105 in 48

48 48

? GWC (7360) 17 0 4 km 17
of 30-20

30 30 QAa8014(b)c

Fig. 15A time series of water/gas production ratios shows aquifer encroachment into the gas reservoir and indicates that
when well T48 was drilled, water had swept the targeted sand.
10 M.H. Holtz 75502

18
16 Toolachee No. 47
14 Toolachee No. 48
12

Frequency
10
8
6
4
2
0

10

20

30

40

50

60

70

80

90

100
Gas saturation (percent) QAa7834c
Fig. 16Gas saturations calculated in aquifer-swept zones are similar to those used in the Sgrm model
development, thus indicating that in situ field Sgrm values are similar to those measured in the laboratory.

(a) (b)
November 1984 August 1986
? GWC
LKG in 25 7300
7256
25 25

HKW in
7 & 20 N 0 1 km
7355
7 7

N
7 300

7
40
7

45

40
45

7
0
7

0
30
7
20

WGR in 25

0
20
WGR in 25
0

0
18 bbl/MMcf 18 bbl/MMcf 47
47

20 20
0 1 km

(c) (d)
September 1988 August 1991

25 GWC
25 ? GWC 7216
7250

7 0 1 km
7 0 1 km

N
N
7
7 00

40
40

45
7

30

45
7

0
0
3

0
7

2 00
20
0

WGR in 25 47
106 bbl/MMcf 47
WGR in 25
20
20 253 bbl/MMcf

45 drilled Aug 91, depleted 1500 psi

47 drilled Feb 92, depleted 1400 psi
extensively swept QAa7636c

Fig. 17A time series of well water/gas production ratios shows aquifer encroachment into the gas reservoir
and indicates that when well T47 was drilled, water had swept the targeted sand.