You are on page 1of 141

RelativePermeability of Reservoirs Petroleum

Authors

Mehdi Honarpour
of AssociateProfessor PetroleumEngineering Departmentof PetroleumEngineering Montana College of Mineral Scienceand Technology Butte, Montana

Leonard Koederitz
of Professor PetroleumEngineering Departmentof PetroleumEngineering University of Missouri Rolla. Missouri

A. Herbert Harvey
Chairman Department PetroleumEngineering of University of Missouri Rolla, Missouri

@frc')
CRC Press,Inc. Boca Raton, Florida

PREFACE
In 1856 Henry P. Darcy determinedthat the rate of flow of water through a sand filter could be describedby the equation q : K A h , - h .
L

where q representsthe rate at which water flows downward through a vertical sand pack headsat hydrostatic areaA and length L; the terms h, and h, represent with cross-sectional Darcy's experiments of the inlet and outlet, respectively, the sandfilter, and K is a constant. with were confined to the flow of water through sand packs which were 1007osaturated water. determinedthat Darcy's law could be modified to describethe flow Later investigators of fluids other than water, and that the proportionalityconstantK could be replacedby k/ p, where k is a property of the porous material (permeability)and p is a property of the fluid (viscosity).With this modification,Darcy's law may be written in a more generalform
AS

lu':*LPgos-dsl
where S v Z p
g D

dz

dPl

dP dS

Distancein direction of flow, which is taken as positive a Volume of flux across unit areaof the porousmedium in unit time along flow path S Vertical coordinate,which is taken as positivedownward Density of the fluid Gravitationalacceleration gradientalong S at the point to which v. refers Pressure

The volumetric flux v. may be further defined as q/A, where q is the volumetric flow rate to areaperpendicular the lines of flow. and A is the averagecross-sectional It can be shown that the permeabilityterm which appearsin Darcy's law has units of fluid with length squared.A porousmaterialhas a permeabilityof I D when a single-phase the a viscosityof I cP completelysaturates pore spaceof the medium and will flow through area under a pressure cross-sectional it under viscous flow at the rate of I cm3/sec/cm2 gradientof 1 atm/cm. It is important to note the requirementthat the flowing fluid must the completelysaturate porousmedium. Sincethis conditionis seldommet in a hydrocarbon reservoir,it is evident that further modificationof Darcy's law is neededif the law is to be appliedto the flow of fluids in an oil or gas reservoir. A more useful form of Darcy's law can be obtained if we assurnethat a rock which containsmore than one fluid has an effective permeabilityto each fluid phaseand that the The effective saturation. effectivepermeabilityto each fluid is a function of its percentage is saturated equal to the absolute permeabilityof a rock to a fluid with which it is 1007.o permeabilityof the rock. Effective permeabilityto each fluid phase is consideredto be to are independent the other fluid phasesand the phases considered be immiscible. of permeIf we define relativepermeabilityas the ratio of effectivepermeabilityto absolute ability, Darcy's law may be restatedfor a system which containsthree fluid phasesas tirllows:

Vo.:T(0.,*K-*)

V*.:*(o-'13-t)
Vo,:H(o-r#-k)
oil, gas' and water, respectively'Note that k,,,' o, where the subscripts g, and w represent saturations at to the threefluid phases the respective k.", and k,* arethe relativepermeabilities of the phaseswithin the rock' a hydrocarbon of Darcy's law is the basis for almost all calculations fluid flow within to is necessary determinethe relative permeabilityof reservoir. In order to use the law, it must be made throughout this determination the reservoirrock to each of the fluid phases; The problemsinvolved in measuring will be encountered. that the rangeof fluid saturations A summary and predictingrelative permeabilityhave been studiedby many investigators. in presented the following chapters' is of the major resultsof this research

THE AUTHORS
professorof petroleumengineeringat Dr. Mehdi "Matt" Honarpour is an associate the MontanaCollege of Mineral Scienceand Technology,Butte, Montana. Dr. Honarpour from the Universityof Misengineering his obtained B.S., M.S., and Ph.D. in petroleum and core in souri-Rolla.He has authoredmany publications the areaof reservoirengineering engineer,consultant, analysis.Dr. Honarpourhas worked as reservoirengineer,research organizations, and teacherfor the past 15 years. He is a member of severalprofessional of including the Societyof PetroleumEngineers AIME, the honorarysocietyof Sigma Xi, Pi Epsilon Tau and Phi Kappa Phi. Leonard F. Koederitz is a Professorof PetroleumEngineeringat the University of M fr M H Mis s ouri-Rol l a. ere ce i ve d B .S ., .S ., a n d P h .D .d egrees omt heU ni vers i t yof i ssour i Dr. Koederitzhasworked for Atlantic-Richfieldand previouslyservedas Department Rolla. severaltechnicalpublicationsand two Chairmanat Rolla. He has authoredor co-authored relatedto reservoirengineering. texts A. Herbert Harvey receivedB.S. and M.S. degreesfrom Colorado School of Mines and a Ph.D. degree from the University of Oklahoma. He has authoredor co-authored numerous on technicalpublications topicsrelatedto the productionof petroleum.Dr. Harvey is Chairman of both the Missouri Oil and Gas Council and the PetroleumEngineering Departmentat the University of Missouri-Rolla.

ACKNOWLEDGMENT
The authorswish to acknowledge the Societyof PetroleumEngineers and the American PetroleumInstitutefor grantingpermission usetheir publications.Specialthanksare due to J. Josephof Flopetrol Johnstonand A. Manjnath of ReservoirInc. for their contributions and reviews throughoutthe writing of this book.

TABLE OF CONTENTS
Chapter I Measurement of Rock Relative Permeability . Introduction.. . I. Methods.. . Steady-State il. Method Penn-State A. Dynamic Method Single-Sample B. StationaryFluid Methods C. HasslerMethod. D. Hafford Method E. DispersedFeed Method . F. StateMethods UnsteadyIII. IV. Capillary PressureMethods V. Centrifuge Methods VI. Calculation from Field Data . . R e f e r e n c e.s . . Chapter 2 Two-PhaseRelative Permeability Introduction... I. Rapoportand Leas II. III. Gates,Lietz,andFulcher... a F a t t ,D y k s t r a , n d B u r d i n e . IV. a W y llie, S pra n g l e r, n d Ga rd n e r. V. C T imme rman , o re y,a n d Jo h n so n VI. Wahl, Torcaso, and Wyllie VII. VIII. Brooks and Corey XIIX. Wyllie, Gardner,and Torcaso. . . a L a n d ,W y l l i e , R o s e ,P i r s o n , n d B o a t m a n . . . X. Knopp, Honarpouret al., and Hirasaki XI. References..... Chapter 3 Factors Affecting Two-Phase Relative Permeability Introduction... I. Curves RelativePermeability Two-Phase il. S n. Effec t sof S a tu ra ti o n ta te s Effectsof Rock Properties IV. o V. Definiti onan d C a u se s f W e tta b i l i ty. DeterminationofWettability.... VI. A. ContactAngle Method ImbibitionMethod. B. B u r e a u f M i n e sM e t h o d o C. D. C a p i l l a r i m e t rM e t h o d . . . ic FractionalSurfaceAreaMethod.. E. M Dye Adsorption ethod F. D r o p T e s tM e t h o d . . G. M e t h o d s f B o b e ke t a l . o H. Method MagneticRelaxation I. Methods ResidualSaturation J.

I I 1 I 2 4 4 5 5 6 8 9 10 t2

...... 15 .......15 .. ' 15 .....16 ...... 16 . . . . . ' . 19 . . . . . . 20

27
. . . .27 .... . .29 ...... 30 . . . . . .37 ........41

.... 45 .......45 ....45 . . . . . . 49 .... ... 50 . . . . . . . . 54 .......58 ... 58 .......60 .......63 ......63 ....64 ' ...... .64 .. ...64 ........64 ...64 .. .65

P e r m e a b i l i M e t h o d. . . . ty W a te r-P e rm e a b i l iM e th o d ty Conna te M Re lati v e rme a b i l i ty e th o d.... Pe Method Su P Re lati v e e rm e a b i l i ty mma ti o n R P Re lati v e e rm e a b i l i ty a ti oMe th o d M Waterflood ethod Method Pressure Capillary a. In Res is ti v i ty d e x M e th o d R. FactorsInfluencing Wettability Evaluation VII. VIII. Wettability Influenceon MultiphaseFlow H E f f e c t s f S a t u r a t i o ni s t o r y . . . . o IX. .. Pressure Effectsof Overburden X. P e rm e a b i l i ty ... a K ) ( I . E f f ec t sof Po ro s i ty n d Effectsof Temperature. XII. XIII. Effects of InterfacialTension and Density .;.... . o X I V . E f f e c t s f V i s c o s i t y. . Saturation XV. Effectsof Initial Wetting-Phase XVI. Effects of an Immobile Third Phase XVII. Effects of Other Factors References..... K. L. M. N. O. P. Chapter 4 Three-PhaseRelative Permeability Introduction... I. DrainageRelativePermeability... il. A. Leverettand Lewis and Wyllie Henderson, B. Corey, Rathjens, Reid. C. Snell. D. and Dean Donaldson E. Sarem F. S a r a fa n d F a t t G. WyllieandGardner... H. P I m bibit ionRe l a ti v e e rm e a b i l i ty ... m. Caudle,slobod,andBrownscombe A. N a a ra n dW y g a l . . . . . B. Land. C. D. SchneiderandOwens.... Spronsen E. ProbabilityModels IV. V. E x per im enta l C o n fi rm a ti o n U\/I . Labor at ory Ap p a ra tu s ... for PracticalConsiderations LaboratoryTests VII. VIII. ComparisonofModels References""' Appendix Symbols.

....... 65 ....... 66 .... 66 ........61 ........67 ....... 68 .... . 68 ... . ... 68 .. . 68 . . .72 ......'74 ... ' .. 78 ......79 . .. .82 . . .82 . .. ' ' 83 ... 89 . '. 90 . . .92 ..-.....97

... f 03 ......103 ..'.104 ... ' . . 104 .. 105 .. 107 .. l0g .. . . I l0 .......113 ..... I 15 .'ll5 ...117 .......117 ....I 17 .. 120 .....123 .'..123 . .123 .....126 ..127 .... ' 132 ...'133 """'134

....... 137

Chapter I MEASUREMENT OF ROCK RELATIVE PERMEABILITY

I. INTRODUCTION The relative peffneability of a rock to each fluid phasecan be measuredin a core sample "unsteady-state"methods.In the steady-state method, a fixed by either "steady-state" or ratio of fluids is forced through the test sampleuntil saturation and pressure equilibria are established.Numerous techniqueshave been successfully employed to obtain a uniform saturation.The primary concern in designingthe experimentis to eliminate or reducethe gradientwhich is causedby capillary pressure saturation effectsat the outflow boundaryof the core. Steady-state methodsare preferred unsteady-state to methods someinvestigators by wettability,' althoughsomedifficulty hasbeenreportedin applying for rocks of intermediate the Hasslersteady-state method to this type of rock.2 ln the capillary pressure method,only the nonwettingphaseis injectedinto the core during the test. This fluid displacesthe wetting phaseand the saturations both fluids change of throughout the test. Unsteady-state techniquesare now employed for most laboratory measurementsof relative permeability.3 Some of the more commonly used laboratory methods for measuringrelative perrneability are describedbelow.

METHODS II. STEADY-STATE


A. Penn-State Method This steady-state method for measuringrelative perrneability was designedby Morse et and Geffen and Yuster,6Caudleet a1.,7 al.a and later modified by Osobaet aI.,5 Henderson et al.8 The version of the apparatuswhich was describedby Geffen et al., is illustrated by Figure l. In order to reduce end effects due to capillary forces, the sample to be tested is mounted between two rock sampleswhich are similar to the test sample. This arrangement also promotes thorough mixing of the two fluid phasesbefore they enter the test sample. The laboratory procedure is begun by saturatingthe sample with one fluid phase (such as water) and adjustingthe flow rate of this phasethrough the sampleuntil a predetermined pressure gradientis obtained.Injection of a secondphase(such as a gas) is then begun at differential low rate and flow of the first phaseis reducedslightly so that the pressure a systemremainsconstant.After an equilibriumconditionis reached,the two flow acrossthe saturationof each phasewithin the test sample is rates are recordedand the percentage and weighing it. This procedure determined removing the test samplefrom the assernbly by introducesa possible sourceof experimental error, since a small amount of fluid may be that the core be One authorityrecommends lost because gas expansionand evaporation. of the problem of obtainingthe sameamountof liquid film on wgighedunder oil, eliminating the surfaceof the core for each weighing.3 by The estimationof water saturation measuringelectric resistivityis a fasterprocedure obtained by a resistivity than weighing the core. However, the accuracyof saturations measurement questionable, is sinceresistivitycan be influencedby fluid distributionas well as fluid saturations. The four-electrode assemblywhich is illustratedby Figure I was used when flow equilibriumhas been water saturation distributionand to determine to investigate of attained.Other methodswhich have been used for in situ determination fluid saturation neutron nuclearmagneticresonance, in cores include measurement electric capacitance, of scattering,X-ray absorption,gamma-rayabsorption,volumetric balance,vacuum distillation, and microwavetechniques.

RelativePermeabilin of PetroleumReservoirs
El-ectrodes Inlet

Outl-et

Differential Taps
FIGURE l.

Pressure

Inlet

Three-section core assembly.8

After fluid saturationin the core has been determined,the Penn-State apparatusis reassembled,a new equilibrium condition is established a higher flow rate for the second at phase, and fluid saturationsare determinedas previously described.This procedureis repeated sequentially at higher saturationsof the second phase until the complete relative permeability curve has been established. The Penn-Statemethod can be used to measurerelative permeability at either increasing or decreasingsaturationsof the wetting phaseand it can be applied to both liquid-liquid and gas-liquid systems.The direction of saturationchangeused in the laboratoryshould correspondto field conditions. Good capillary contactbetweenthe test sampleand the adjacent downstream core is essential for accurate measurements and temperaturemust be held constantduring the test. The time required for a test to reach an equilibrium condition may be I day or more.3 B. Single-Sample Dynamic Method This technique for steady-state measurement relative permeability was developedby of Richardsonet al.,e Josendal al.,ro and Loomis and Crowell.ttThe apparatus et and experimental procedure differ from those used with the Penn-Statetechnique primarily in the handling of end effects. Rather than using a test samplemountedbetweentwo core samples (as illustrated by Figure 1), the two fluid phases injectedsimultaneously are through a single core. End effects are minimized by using relatively high flow rates, so the region of high wetting-phase saturationat the outlet faceof the core is small. The theorywhich was presented by Richardson et al. for describing the saturationdistribution within the core may be dethrough a horizontallinear veloped as follows. From Darcy's law, the flow of two phases systemcan be describedby the equations -d P* , : Q*, F*,dL k*, A (l)

and Q. Fr" ,n - d P n : = i ^ dL

Q)

where the subscriptswt and n denotethe wetting and nonwettingphases,respectively.From the definition of capillary pressure,P", it follows that

1.0

\o \.o

> { ^ -o-i-

o
Theoretical saturation gradient

fnf low face

1>

0 5 10 15 20 25 Distance from Outflow Face, cffi


FIGURE 2. Comparison of saturationgradientsat low flow rate.e

dP.:dP.-dP*, These three equationsmay be combined to obtain

(3)

qP.
dL

_ : /Q*, Fr,*, 9"U=\ / o


\ k* , kn //

(4)

where dP"/dL is the capillary pressuregradient within the core. Since

dP. : dP. ds*, dS*, dL dL


it is evident that dS*, dL

(s)

: A |\

/Q*, Fr*, - Q"p.\

k*

L" /op.rus*

(6)

Richardson et al. concluded from experimentalevidence that the nonwetting phase saturation at the dischargeend of the core was at the equilibrium value, (i.e., the saturation at which the phase becomes mobile). With this boundary condition, Equation 6 can be integrated graphically to yield the distribution of wetting phase saturationthroughout the core. If the flow rate is sufficiently high, the calculation indicates that this saturation is virtually constant from the inlet face to a region a few centimetersfrom the outlet. Within increases the equilibriumvalue at the outlet face. to this region the wetting phasesaturation Both calculations and experimental evidence show that the region of high wetting-phase saturationat the discharge end of the core is larger at low flow rates than at high rates. Figure 2 illustrates the saturationdistribution for a low flow rate and Figure 3 shows the distribution at a higher rate.

Relative Permeability of Petroleum Reservoirs

1.0

't o I -o-o- -o--o-- :- -- : - J

Theoretical saturation gradient

I n fr o wr a c "

a>l

10

15

20

25

Distance from Outflow Face, ctrl


FIGURE 3. Comparison of saturationgradients at high flow rate.e

Although the flow rate must be high enoughto control capillary pressure effects at the dischargeend of the core, excessiveratesmust be avoided. Problemswhich can occur at very high rates include nonlaminarflow. C. Stationary Fluid Methods Leas et al.12 described technique measuring a for permeabilityto gaswith the liquid phase held stationarywithin the core by capillary forces. Very low gur flo* ratesmust be used, so the liquid is not displaced during the test. This technique was modified slightly by Osoba et al.,s who held the liquid phasestationary within the core by meansof barrierswhich were permeable gas but not to the liquid. Rapoportand Leasr3employeda similar technique to using semipermeable barrierswhich held the gas phasestationary while allowing the liquid phaseto flow. Corey et al.ra extendedthe stationaryfluid methodto a three-phar.ryri.. by using barrierswhich were permeable water but impermeable oil and gas. Osobaet to to al. observed that relative permeability to gas determinedby the stationary liquid method was in good agreementwith values measuredby other techniquesfor some of the cases which were examined. However, they found that relative permeability to gas determinedby the stationary liquid technique was generally lower than by other methodsin the region of equilibrium gas saturation. This situation resulted in an equilibrium gas saturation value which was higher than obtained by the other methods used (Penn-Siate,Single-Sample Dynamic, and Hassler). Saraf and McCaffery consider the stationaryfluid methods to be unrealistic, since all mobile fluids are not permitted to flow simultaneouslyduring the test.2 D. Hassler Method This is a steady-state method for relative permeability measurement which was described by Hasslerrsin 1944. The technique was later studied and modified by Gates and Lietz,16 Brownscombeet ?1.," Osoba et al.,s and Josendalet al.ro The laboratory apparatusis illustrated by Figure 4. Semipermeable membranesare installed at each end of the Hassler test assembly.Thesemembranes keep the two fluid phases separated the inlet and outlet at of the core, but allow both phasesto flow simultaneously through the core. The pressure

FLOWMETER

FIGURE 4.

Two-phase relative permeability apparatus.r5

barrier. By adjusting through a semipermeable in each fluid phaseis measuredseparately gradients the two phases in can be made the flow rate of the nonwettingphase,the pressure at equal, equalizingthe capillary pressures the inlet and outlet of the core. This procedure throughoutthe length of the core, even at low is designedto provide a uniform saturation flow rates, and thus eliminate the capillary end effect. The techniqueworks well under conditionswhere the porousmedium is stronglywet by one of the fluids, but somedifficulty wettability.2'r8 has been reported in using the procedureunder conditionsof intermediate The Hasslermethod is not widely used at this time, since the data can be obtainedmore rapidly with other laboratorytechniques. E. Hafford Method technique was describedby Richardsonet al.e In this method the nonThis steady-state wetting phase is injected directly into the sample and the wetting phaseis injected through a disc which is impermeableto the nonwetting phase.The central portion of the semipermeable disc is isolated from the remainder of the disc by a small metal sleeve, as illustrated by Figure 5. The central portion of the disc is used to measurethe pressurein the wetting fluid at the inlet of the sample. The nonwetting fluid is injected directly into the sample and its pressureis measuredthrough a standardpressuretap machined into the Lucite@surrounding the sample. The pressuredifference betweenthe wetting and the nonwetting fluid is a measureof the capillary pressurein the sample at the inflow end. The design of the Hafford apparatusfacilitates investigationof boundary effects at the influx end of the core. The outflow boundary effect is minimized by using a high flow rate. F. Dispersed Feed Method method for measuringrelative permeability which was designedby This is a steady-state dynamic methRichardsonet al.e The techniqueis similar to the Hafford and single-sample

RelativePermeabilin of PetroleumReservoirs
GAS

G A S P R E S S U R EG A U G E

PRESSURE

PRESSURE GAS METER

OIL BURETTE FIGURE 5. Hafford relative permeability apparatus.e

ods. In the dispersedfeed method, the wetting fluid enters the test sample by first passing through a dispersingsection, which is made of a porous material similar to the test sample. This material does not contain a device for measuringthe input pressure the wetting phase of as does the Hafford apparatus. The dispersingsectiondistributes wetting fluid so that it the entersthe test samplemore or less uniformly over the inlet face. The nonwettingphaseis introduced into radial grooves which are machined into the outlet face of the dispersing section,at thejunction betweenthe dispersingmaterialand the testsample.Pressure gradients used for the tests are high enough so the boundary effect at the outlet face of the core is not significant.

III. UNSiuoo"-STATEMETHoDS
Unsteady-state relative permeability measurements be made more rapidly than steadycan state measurements, but the mathematicalanalysisof the unsteady-state procedureis more difficult. The theory developed by Buckley and Leverettre and extended by Welge2ois generally used for the measurement relative permeabilityunder unsteady-state of conditions. The mathematicalbasis for interpretationof the test data may be summarizedas follows: Leverett2rcombined Darcy's law with a definition of capillary pressurein differential form to obtain

'*;h(*-eApsino)
f*z

r + In.&
k* Fo

(71

where f*, is the fraction water in the outlet stream;q, is the superficialvelocity of total fluid leaving the core; 0 is the angle between direction x and horizontal; and Ap is the density

7 difference between displacing and displaced fluids. For the case of horizontal flow and negligible capillary pressure,Welge2oshowed that Equation 7 implies
S*.u, S*z : f.r, Q*

( 8)

is 2 water saturation; the wherethe subscript denotes outlet end of the core, S*.ou the average in and Q* is the cumulativewater injected,measured pore volumes.SinceQ* and S*.,ucan experimentally,f", (fraction oil in the outlet stream)can be determinedfrom be measured By the slope of a plot of Q* as a function of S*,ou. definition l,z:q,,/(q,,*q*) By combining this equationwith Darcy's law, it can be shown that f,,r: t * I
I1.,/K..,

(e)

'

tlOt

tr/.,* Since p" and pw are known, the relative permeability ratio k.o/k.* can be determinedfrom Equation 10. A similar expression can be derived for the caseof gas displacingoil. The work of Welge was extendedby Johnsonet a1.22 obtain a technique (sometimes to calledthe JBN method) for calculatingindividual phaserelativepermeabilities from unsteadystate test data. The equationswhich were derived are k.. :

.(#) /,(a
f,,, t.z ttr. injectivity

(Il)

and

k.o: ltoo,,,
where I,, the ?elative injectivity, is defined as I,:

(12)

initial injectivity (q*,/Ap) (q*,/Ap) at start of injection

(l 3 )

A graphical technique for solving Equations 1l and 12 is illustrated in Reference L3.. Relationships describing relative permeabilities in a gas-oil system may be obtained by "w" with "g" in EquationslI,12, and 13. replacing the subscript In designingexperimentsto determinerelative permeabilityby the unsteady-state method, it is necessarvthat:

l. 2. 3. 4.

gradientbe large enoughto minimize capillary pressure The pressure effects. The pressure differential across the core be sufficiently small compared with total operatingpressureso that compressibility effects are insignificant. The core be homogeneous. The driving force and fluid propertiesbe held constantduring the test.2

Relative Permeabilin of Petroleum Reservoirs Laboratory equipment is available for making the unsteady-state measurements under simulated reservoirconditions.2a In addition to the JBN method, several alternative techniquesfor determining relative permeabilityfrom unsteady-state data have been proposed.Saraf and McCaffery2detest veloped a procedurefor obtainingrelative permeabilitycurves from two parameters determined by least squaresfit of oil recovery and pressuredata. The technique is believed to be superior to the JBN method for heterogeneous carbonatecores. Jones and Roszelle25 developed a graphical technique for evaluation of individual phase relative permeabilities from displacementexperimentaldata which are linearly scalable.Chavent et al. described a method for determining two-phaserelative permeability and capillary pressurefrom two sets of displacementexperiments,one set conductedat a high flow rate and the other at a rate representative reservoir conditions. The theory of Welge was extendedby Sarem to of describerelative permeabilities a systemcontainingthree fluid phases.Sarememployed in a simplifying assumptionthat the relative permeabilityto each phasedependsonly on its (particularly own saturation, and the validity of this assumption with respect the oil phase) to has beenquestioned.2 Unsteady-state relative permeability measurements frequently used to determine the are ratios k*/ko, ks/k", and kr/k*. The ratio k*/k" is usedto predict the performanceof reservoirs which are produced by waterflood or natural water drive; kr/k" is employed to estimatethe production which will be obtained from recovery processes where oil is displacedby gas, such as gas injection or solution gas drive. An important use of the ratio k*/k* is in the prediction of performanceof natural gas storagewells, where gas is injectedinto an aquifier. The ratios k*/ko, kg/ko,and kr/k* are usually measuredin a systemwhich containsonly the two fluids for which the relative permeability ratio is to be determined. It is believed that expeciallyin sandstones the connatewater in the reservoirmay have an influenceon kg/k.,, which contain hydratableclay minerals and in low permeabilityrock. For these types of reservoirsit may be advisableto measurek*/k.,in cores which contain an immobile water saturation.2a

IV. CAPILLARYPRESSURE METHODS


The techniqueswhich are usedfor calculatingrelative permeabilityfrom capillary pressure data were developedfor drainagesituations,where a nonwettingphase(gas) displacesa wetting phase(oil or water). Thereforeuse of the techniques generallylimited to gas-oil is or gas-watersystems,where the reservoiris producedby a drainageprocess.Although it is possibleto calculaterelativepermeabilities a water-oil systemfrom capillary pressure in data, accuracyof this techniqueis uncertain;the displacement oil by water in a waterof wet rock is an imbibition processrather than a drainageprocess. Although capillary pressure techniques not usuallythe preferredmethodsfor generating are relative permeability data, the methodsare useful for obtaining gas-oil or gas-waterrelative permeabilitieswhen rock samplesare too small for flow testsbut large enough for mercury injection. The techniquesare also useful in rock which has such low permeability that flow testsare impractical and for instanceswhere capillary pressuredata have been measured but a sampleof the rock is not availablefor measuringrelative permeability. Another use which has been suggestedfor the capillary pressuretechniquesis in estimating kr/k" ratios for retrogradegas condensate reservoirs, where oil saturationincreasesas pressuredecreases, with an initial oil saturationwhich may be as low as zero. The capillary pressuremethods are recommendedfor this situation becausethe conventionalunsteady-state test is not designed for very low oil saturations. Data obtainedby mercury injection are customarilyused when relative permeabilityis estimated the capillary pressure by technique.The core is evacuated and mercury(which is

9 pressures. Approxat increments increasing the nonwettingphase)is injectedin measured to testdesigned yield the complete imately 20 datapoints are obtainedin a typical laboratory relativepermeabilityby the methcurve, which is requiredfor calculating capillarypressure ods describedbelow. for have developedequations estimatingrelative permeabilityfrom Severalinvestigators presented equations the data. Purcell2e capillary pressure
fs*i

k.*, :

l,
fl

dS/pi
(l4)

t
and

dS/Pi

JSo i

I' ds/p!
fl

k.n*,:

(l 5 )

J,

dS/pi

where the subscriptswt and nwt denotethe wetting and nonwettingphases,respectively, developedsimilar equationswith n equal to and n has a value of 2.0. Fatt and Dykstra3o 3.0. developed Burdine3l by A slightly different result is obtainedby combiningthe equations The resultsare with the work of Purcell.2e

(l6)

(l7)

where S, is the total liquid saturation.

METHODS V. CENTRIFUGE
Centrifuge techniquesfor measuringrelative permeability involve monitoring liquids produced from rock sampleswhich were initially saturateduniformly with one or two phases. to tubesconnected the rock sampleholdersand production Liquids are collectedin transparent is monitored throughout the test. Mathematicaltechniquesfor deriving relative permeability are data from these measurements describedin References26, 27, and 28. the centrifugemethodshave not beenwidely used,they do offer someadvantages Although over alternativetechniques.The centrifuge methodsare substantiallyfaster than the steadystatetechniquesand they apparentlyare not subjectto the viscousfingering problems which On measurements. the other hand, the centrifuge sometimesinterfere with the unsteady-state subject to capillary end effect problems and they do not provide a means for methods are determining relative permeability to the invading phase. describean automatedcentrifuge which employs a photodiodearray O'Mera and Lease28 with a microcomputerto image and identify liquids producedduring the test. in conjunction

t0

Relative Permeabiliy of Petroleum Reservoirs

CAMER CENTRIFUGE

COMPUTER

o z
LIQUID PRODUCTION
LIJ

o U'
IJJ

o o uJ
LIJ

tr o o
J

oa)

TROBE

CONTROLLER

SPEEDDISK SPEED SET POINT

FIGURE 6.

Automated centrifuge system.28

Stroboscopiclights are located below the rotating tubes and movement of fluid interfaces is monitored by the transmitted light. Fluid collection tubes are square in cross section, since a cylindrical tube would act as a lens and concentrate light in a narrow band along the the major axis of the tube. A schematicdiagram of the apparatusis shown by Figure 6.

VI. CALCULATION FROM FIELD DATA


It is possibleto calculaterelative permeability ratios directly from field data.23Inmaking the computation it is necessaryto recognize that part of the gas which is produced at the surface was dissolved within the liquid phasein the reservoir. Thus; (produced gas) : (free gas) * (solutiongas) (18)

If we consider the flow of free gas in the reservoir, Darcy's law for a radial system may be written k h P . - Pw ln (r./r*) FrB,

9g.fr""

?.09-E-e

(l9)

ll

FIGURE 7.

Calculation of gas-oil relative permeability values from production data.

Similarly, the rate of oil flow in the same system is

(20)
where r* is the well radius and r" is the radius of the external boundary of the area drained by the well. B" and B, are the oil and gas formation volume factors, respectively.The ratio of free gas to oil is obtained by dividing Equation 19 by Equation 20. lt we expressRo, cumulative gas/oil ratio and R,, solution gasioil ratio, in terms of standardcubic foot per stock tank barrel, Equation l8 implies

* R o : s . 6 t s l ube ' * * .
Ko ltrs

(2t)

Thus, the relative permeability ratio is given by k" _ ( R o- R . ) & - ! !


ko

5.615 B. F.

(22)

to The oil saturationwhich corresponds this relative permeabilityratio may be determined we assumethere is no water influx, no water production, no from a material balance. If fluid injection, and no gas cap, the materialbalanceequationmay be written

S.: (t- too,) *,t-

s*)

(23)

where minor effects such as changein reservoirpore volume have been assumednegligible. In Equation 23 the symbol N denotesinitial stock tank barrelsof oil in place; No is number of stock tank barrels of oil produced;and B", is the ratio of the oil volume at initial reservoir conditions to oil volume at standardconditions. If total liquid saturationin the reservoir is expressedas

s,:s*+(r-s*)(\})

(*)

(24)

then the relative permeability curve may be obtainedby plotting kr/k" from Equation 22 as a function of S,- from Equation 24. Figure 7 illustrates a convenientformat for tabulating the data. The curve is preparedby plotting column 9 as a flnction of column 6 on semilog paper, with k/k" on the logarithmicscale.The techniqueis useful even if only a few highliquid-saturation data points can be plotted. These kr/k" values can be used to verify the accuracyof relative permeability predicted by empirical or laboratory techniques. Poor agreementbetween relative permeability determined from production data and from may include laboratory experiments is not uncommon. The causesof these discrepancies the following:

t2
l. 2.

Relative Permeability of Petroleum Reservoirs The core on which relative permeability is measuredmay not be representative the of reservoir in regard to such factors as fluid distributions, secondaryporosity, etc. The techniquecustomarily used to compute relative permeability from field data does not allow for the pressureand saturationgradientswhich are presentin the reservoir, nor does it allow for the fact that wells may be producing from several strata which are at various stagesof depletion. The usual techniquefor calculating relative permeability from field data assumes that Ro at any pressureis constant throughout the oil zone. This assumptioncan lead to computational errors if gravitational effects within the reservoir are significant.

3.

When relative permeability to water is computed from field data, a common source of elror is the production of water from some source other than the hydrocarbon reservoir. These possible sourcesof extraneouswater include casing leaks, fracturesthat extend from the hydrocarbon zone into an aquifer, etc.

REFERENCES
l. Gorinik, B. and Roebuck, J. F., Formation Evaluation through Extensive Use of Core Analysis, Core L a b o r a t o r i e sI,n c . , D a l l a s ,T e x . , 1 9 7 9 . 2. Saraf, D. N. and McCaffery, F. G., Two- and Three-Phase RelativePermeabilities: Review, Petroleum a Recovery InstituteReport #81-8, Calgary, Alberta, Canada, 1982. 3. Mungan, N., PetroleumConsultants Ltd., personalcommunication,1982. 4. Morse, R. A., Terwilliger, P. L., and Yuster, S. T., Relative permeabilitymeasurements small on samples, il GasJ., 46, 109, 1947. O 5. Osoba, J. S., Richardson, J. G., Kerver, J. K., Hafford, J. A., and Blair, P. M., Laboratoryrelative permeabilitymeasurements, Trans. AIME, 192, 47, 1951. 6 . H e n d e r s o n ,J . H . a n d Y u s t e r , S . T . , R e l a t i v ep e r m e a b i l i t y t u d y , W o r l dO i l , 3 , 1 3 9 , 1 9 4 8 . s 7. Caudle, B. H., Slobod, R. L., and Brownscombe, E. R. W., Further developments the laboratory in determination relative permeability,Trans. AIME, 192, 145, 1951. of 8. Geffen, T. M., Owens, W. W., Parrish, D. R., and Morse, R. A., Experimental investigation factors of affecting laboratory relative permeability Teasurements,Trans. AIME, 192, 99, 1951. 9. Richardson, J. G., Kerver, J. K., Hafford, J. A., and Osoba, J. S., Laboratory determination relative of permeability,Trans. AIME, 195, 187, 1952. 10. Josendal, V. A., Sandiford, B. B., and Wilson, J. W., Improved multiphaseflow studiesemploying radioactive tracers, Trans. AIME, 195, 65, 1952. I l. Loomis, A. G. and Crowell, D. C., RelativePermeability Studies:Gas-Oil and Water-Oil Systems,U.S. Bureau of Mines Bulletin BarHeuillr, Okla., 1962,599. 12. Leas, W. J., Jenks, L. H., and Russell, Charles D., Relativepermeabilityto gas, Trans. AIME, 189,

65,r 9s 0.
13. Rapoport, L. A. and Leas, W. J., Relative permeabilityto liquid in liquid-gassystems,Trans. AIME, 1 9 2 ,9 3 , l 9 5 l . 14. Corey, A. T., Rathjens, C. H., Henderson, J. H., and Wyllie, M. R. J., Three-phase relativepermea b i l i t y , J . P e t . T e c h n o l . ,N o v . , 6 3 , 1 9 5 6 . 1 5 . H a s s l e r , G . L . , U . S . P a t e n t2 , 3 4 5 , 9 3 5 , 1 9 4 4 . 16. Gates, J. I. and Leitz, W. T., Relative permeabilitiesof California coresby the capillary-pressure method, Drilling and Production Practices, American Petroleum Institute, Washington, D.C. 1950, 285. 17. Brownscombe, E. R., Slobod, R. L., and Caudle, B. H., Laboratory determination of relative perrnea b i l i t y ,O i l G a s J . , 4 8 , 9 8 , 1 9 5 0 . 18. Rose, W., Some problemsin applying the Hasslerrelativepermeabilitymethod,J. Pet. Technol.,8, I l6l, 1980. 19. Buckley, S. E. and Leverett, M. C., Mechanismof fluid displacement sands,Trans. AIME, 146,107, in 1942. 20. Welge'H.J.rAsimplifiedmethodforcomputingrecoverybygasorwaterdrive,Trans.A|ME95,91, 1, 1952. 21. Leverett, M. C., Capillary behaviorin poroussolids, Trans. AIME, 142, 152, 1941.

13
22. Johnson, E. F., Bossler, D. P., and Naumann, V. O., Calculationof relative permeabilityfrom displacementexperiments,Trans. AIME, 216,310, 1959. 23. Crichlow, H. B., Ed., Modern ReservoirEngineering- A SimulationApproaclr, Prentice-Hall,Englewood Cliffs, 1977, chap. 7. 24. SpecialCore Analysis, Core Laboratories, Inc., Dallas, 1976. 25. Jones, S. C. and Roszelle, W. O., Graphical techniquesfor determining relative permeability from displacement experiments, Pet. Technol., 5, 807, 1978. J. 26. Slobod, R. L., Chambers, A., and Prehn, W. L., Use of centrifugefor determiningconnate water, residualoil, and capillary pressure curvesof small core samples,Trans. AIME, 192, 127, 1952. 27 . Yan Spronsen, E., Three-phase relative permeabilitymeasurements using the CentrifugeMethod, Paper SPE/DOE 10688presented the Third Joint Symposium,Tulsa, Okla., 1982. at 28. O'Mera, D. J., Jr. and Lease, W. O., Multiphaserelativepermeability measurements using an automated centrifuge,PaperSPE 12128presented the SPE 58th Annual TechnicalConference at and Exhibition, San Francisco.1983. 29. Purcell, W. R., Capillarypressures their measurement usingmercuryand the calculation permeability of therefrom, Trans. AIME, 186, 39. 1949. 30. Fatt, I. and Dyksta, H.,,Relative permeabilitystudies,Trans. AIME, 192,41, 1951. 31. Burdine, N. T., RelativePermeability Calculations from Pore Size DistributionData, Trans. AIME, lg8, 7t,1953.

l5 Chapter 2 TWO-PHASE RELATIVE PERMEABILITY

I. INTRODUCTION to Direct experimentalmeasurement determinerelative permeabilityof porous rock has been recordedin petroleumrelatedliterature.However, empirical methodsfor deterlong mining relative permeabilityare becomingmore widely used, particularlywith the advent of digital reservoirsimulators.The generalshapeof the relative permeabilitycurves may k.* : A(S*)'; k.., : B(l - S*)"'; where A, by be approximated the following equations: B. n. and m are constants. Most relative permeability mathematicalmodels may be classifiedunder one of four categories: that a porous medium consistsof a Capillary models - Are basedon the assumption with a fluid path lengthlongerthan the sample. of capillarytubesof variousdiameters bundle natureof porous media and frequentlydo not Capillary models ignore the interconnected provide realisticresults. Statistical models - Are also basedon the modeling of porous media by a bundle of distributedrandomly. The modelsmay be described capillary tubes with various diameters to as being divided into a large number of thin slicesby planesperpendicular the axes of randomly. Again, and reassembled The slices are imagined to be rearranged the tubes. statisticalmodels have the same deficiencyof not being able to model the interconnected natureof porous media. describingexperiEmpirical models - Are basedon proposedempirical relationships haveprovi{ed the most successful and relativepermeabilities in general, mentallydetermined approximations. Netwoik models - Are frequentlybasedon the modelingof fluid flow in porousmedia using a network of electric resistorsas an analogcomputer.Network models are probably fluid flow in porousmedia'r'aa the best tools for understanding The hydrodynamiclaws generallybear little use in the solutionof problemsconcerning fluid flow through porous media, let alone multiphasefluid flow, due to the single-phase complexity of the porous system. One of the early attemptsto relate severallaboratoryThis equation equation.2 parameters rock permeabilitywas the Kozeny-Carmen to measured the expresses permeabilityof a porousmaterialas a function of the productof the effective throughwhich path lengthof the flowing fluid and the meanhydraulicradiusof the channels the fluid flows. Purcell3formulated an equation for the permeability of a porous system in terms of the the curve of that systemby simply considering desaturation porosity and capillary pressure porousmedium as a bundle of capillary tubesof varying sizes. and Purcellto the by adaptedthe relationsdeveloped Kozeny-Carmen Severalauthorsa-r6 modelson the basisof the assumption They all proposed of computation relativepermeability. that a porous medium consistsof a bundle of capillariesin order to apply Darcy's and Poiseuille'sequationsin their derivations.They used the tortuosityconceptor texture paas rametersto take into accountthe tortuouspath of the flow channels opposedto the concept of capillary tubes. They tried to determinetortuosityempirically in order to obtain a close approximation of experimentaldata.

AND LEAS II. RAPOPORT


to for presented equations relativepermeability the wettingphase. two Rapoportand Lease

16

RelativePermeabilin of PetroleumReservoirs

Theseequationswere basedon surfaceenergyrelationships and the Kozeny-Carmen equawere presented defining limits for wetting-phase tion. The equations as relativepermeability. The maximum and minimum wetting-phase relativepermeabilitypresented Rapoport by and Leas are

k.*,(max) :
fs*

(l)

t'ot Jr*,

(tj)(T#)'
.['*'
fs-

,['*'

P. dS

and P. dS
fS*,

: k,*,(min) (ti - j; )'

(2)

P . d s +|
' J

R.as

whereS- represents minimum irreduciblesaturation the wetting phasefrom a drainage the of capillary pressure curve, expressed a fraction;S*, represents saturation the wetting as the of phasefor which the wetting-phase relativepermeability evaluated, is expressed a fraction; as P. represents drainagecapillary pressure the expressed psi and S represents porosity in the expressed a fraction. as III. GATES. LIETZ. AND FULCHER Gatesand Lietzsdeveloped following expression the based Purcell'smodel for wettingon phaserelative permeability:

t. K.*r

_ -

ru I$

(3)

Fulcher et al.,ashave investigated influenceof capillary number (ratio of viscousto the capillary forces)on two-phaseoil-water relativepermeabilitycurves.

IV. FATT, DYKSTRA,AND BURDINE


Fatt and Dykstrarr developedan expression relativepermeabilityfollowing the basic for methodof Purcell for calculatingthe permeabilityof a porousmedium. They considered a lithology factor (a correction for deviation of the path length from the length of the porous medium) to be a function of saturation.They assumed that the radius of the path of the conductingpores was relatedto the lithology factor, tr, by the equation:

\ : -

a
ro

(4)

L7 Table I CALCULATION OF WETTING.PHASERELATIVE PERMEABILITY BASED ON THE FATT AND DYKSTRA EQUATION

S*, Vo

P", cm Hg

l/P"'], (cm Hg)-t

Area from 0 to S*, in.2

k.*,, Vo

100 90 80 '70 60 s0 40 30 20 ' "

4.0 4.5 5.0 5.5 6.0 6.7 7.s 8.7 13.0

0.0156 0.0110 0.0080 0.0060 0.0046 0.0033 0.0024 0.00 5 1 0.0005

n.25
7.88 5.54 3.80 2.49 t.50 0.75 0.30 0.20

100.0 70.0, 49.2b 33.8 22.1 13.3 6.1 2.7 0.4

7 . 8 8 / 1 1 . 2x 1 0 0: 7 0 . 0 . 5 x 5 . 5 4 1 1 1 . 2 5 l 0 O: 4 9 . 2 .

where r represents radius of a pore, a and b represent the material constants,and }, is a function of saturation. The equationfor the wetting-phase relativepermeability, k.*,, reported Fatt and Dykstra by is

ft*'
t

ds
-

K.*, :

Jn

P2(l

+ b)

* b) Jo P2(|

l.r

dS

(5)

Fatt and Dykstra found good agreementwith observeddata when b : Equation 5 to

r/r, reducing

TF

ft*' ds Jo P:

(6)

They statedthat their equation fit their own data as well as the data of Gatesand Lietz more accuratelythan other proposedmodels. The procedurefor the calculation of relative permeability from capillary pressuredata is illustrated by Table I and the results are shown in Figures I and 2. Burdine'3 reportedequationsfor computing relative perrneabilityfor both the wetting and nonwettingphases.His equationscan be shown to reduceto a form similar to thosedeveloped by Purcell. Burdine's contribution is principally useful in handling tortuosity. Defining the tortuosity factor for a pore as L when the porous medium is saturatedwith only one fluid and using the symbol tr*, for the wetting-phasetortuosity factor when two phasesare present, a tortuosity ratio can be defined as

T tr.*,: ;

(7)

l8

RelativePermeabilitvof PetroleumReservoirs

9 I | 7 Pol (cm Hg) 6 5 4 3 2 I

oo' lo 20
FIGURE 1.

40 50 60 70 80
Sw+

as Capillary pressure a function of water saturation.

then

/'*' {^,*,)'ds/(\)'(P.)'
kr*,

(8)

/'0r,1^;'1r.y'
If tr is a constantfor the porous medium and tr,*t dependsonly on the final saturation,then
fS*'

k.*t :

(tr.*.)' rl

ds/(P.)r
(9)

t ds/(p")l

In a similar fashion, the relative permeabilityto the nonwetting phasecan be expressed tortuosity ratio, tr,,*,, utilizing a nonwetting-phase
fl
JS*t

k.n*,:

(trrn*,)'

I dst1e.)'
(l0)

J"
Burdine has shown that
-

ds/(P.)2

S*,- S1 - S -

Arwt

(lt)

l9

r60 r50 r40 r30 t20


l l

roo
90

t
70 60 50 40 30 20

Pc3 | (CmHqi3

to

o5

lo 20 30 40 50 60 70
Sw -+

of pressure)r a function water as of Reciprocal (capillary il,;yul}: curve. the where S- represents minimum wetting-phasesaturationfrom a capillary-pressure zero at this saturation. The nonwetting The relative perrneability is assumedto approach phasetortuosity can be approximatedby Sn*t-- S' \ -r^n .w. : . . , t l-s*-s" (12)

where S. is the equilibrium saturationto the nonwetting phase. The expressionfor the wetting phase(Equation 9) fit the data presentedmuch better than the expressionfor the nonwetting phase (Equation 10).

AND GARDNER V. WYLLIE, SPRANGLER,


Wyllie and Spranglertz reported equations similar to those presentedby Burdine for computing oil and gas relative permeability. Their equationscan be expressedas follows:
fs" J os"rp;

k,,,: (iil'

(l 3 )

/' or",rl

Relative Permeabilin of Petroleum Reservoirs

A I a

o WYLLIE ond SPANGLER

GATESond LIETZ
i l | | t l

B E R E AN O . 4

o.

FIGURE 3. data.rT

Reciprocalof (capillary pressure)r a function of saturationfor normalized as

k,* -r+" )' !Y (r _ S*,/ or",r3


/'

(t4l

where S- representsthe lowest oil saturationat which the gas phase is discontinuous:S: (l - S " . ) . The above equations for oil and gas relative permeabilities may be evaluated when a reliable drainage capillary pressurecurve of the porous medium is available, so that a plot of llP"2 as a function of oil saturation can be constructed. Obviously, reliable valuesof Sand So.are also neededfor the oil and gas relative permeability evaluation. Figure 3 shows some examplesof llP.2 vs. saturation curves.rT Wyllie and GardnerrTdeveloped equationsfor oil and gas relative permeabilitiesin the presenceof an ineducible water saturation, with the water consideredas part of the rock matrix:

k,.:(H), +*
k,, (*)' f*
Jr*, Pi

ft'ds*

.s;
'6)

Jr*,Pi

f' ds*

where Sl represents total liquid saturation.Note that theseequationsmay be applied only when the water saturationis at the irreduciblelevel.

VI. TIMMERMAN,COREY,AND JOHNSON


Timmermanr8 suggests following equations the basedon the water-oil drainagecapillary pressure, the calculationof low valuesof water-oil relativepermeability. for

2l
DrainageProcess: Wetting-Phase

k.o :

S.

LTFI
f[Hl"
[l'"H 1"

fl'"H.1"

Injection Curve

(t7)

Injection Curve

InjectionCurve (l8)

k.* :

S*

LrFl
LTFj
[[H]"

lnjection Curve

Imbibition Process: Wetting-Phase Injection Curve

kro

So

(le)

Injection Curve

Curve Trap-Hysteresis

k.o :

So

Lrsl

(20)

Injection Curve

Coreyrecombined the work of Purcell3and Burdiner3into a form that has considerable for utility and is widely accepted its simplicity. It requireslimited input data (sinceresidual neededto developa set of relativepermeabilitycurves)and is saturation the only parameter it is fairly accuratefor consolidatedporous media with intergranularporosity. Corey's equationsare often used for calculationof relative permeabilityin reservoirssubjectto a was derivedfrom capillary process externalgasdrive. His methodof calculation or drainage a pressure and the fact that for certaincases,l/P"2is approximately linear function concepts i.e. rangeof saturations; , llP"2 : C [(S" over a considerable of the effective saturation greaterthan S.,,.On the S".)/(1 - S",)] where C is a constantand S" is an oil saturation the concerning natureof the tortuosityand the findingsof Burdiner3 basisof this observation were derived: function, the following expressions saturation

(2r)
(22)

\o: k,o

[S'

Lr - s * J

- S'*lo

(23)

22

Relative Permeability of Petroleum Reservoirs

where S'- is the total liquid saturation and equal to (l - Sr); S- is the lowest oil saturation (fraction) at which the gas phaseis discontinuous; and Sr* is the residualliquid saturation expressed a fraction. as Corey and Rathjens2o studiedthe effect of permeabilityvariationin porous media on the value of the S- factor in Corey's equations.They confirmed that S,,,is essentially equal to unity for uniform and isotropic porous media; however, values of S,, were found to be greaterthan unity when there was stratificationperpendicular the direction of flow and to less than unity in the presence stratificationparallel to the direction of flow. They also of concludedthat oil relative permeabilities were less sensitiveto stratificationthan the gas relativepermeabilities. The gas-oil relative permeabilityequationis often used for testing, extrapolation,and smoothingexperimental data.It is also a convenient expression that may be usedin computer simulationof reservoirperformance. Corey's gas-oil relative permeabilityratio equationcan be solved if only two points on the k,r/k,.,vs. S* curve are available.However, the algebraicsolutionof the k,g/k.., equation when two points are availableis very tediousand the graphicalsolution that Corey offers in his original paperrequires lengthygraphical construction well asnumericalcomputation. as Johnson2r offered a greatly simplified and useful methodfor determination Corey's has of constant. Johnsonconstructed three plots by assumingvaluesof Sr*, S,,, and k.s/k.., calculating by the gas saturation,(1 - S,_), using Corey's equations.The calculationwas carriedout for variousSr* and S- combinations and for k.s/k,o valuesof l0 to 0.1, 1.0 to 0.01, and 0. I to 0.001. Johnson'sgraphs may be used to plot a more completek.g/k,,, curve basedon limited experimental data. The spanof the experimental data determines which of the three figures should be selected. The suggested procedure k.g/k.,calculation,basedon Corey's equation,is as follows: for l. 2. Plot the experimentalk.r/k," vs. S, on semilog paper with k,*/k,oon the logarithmic scale. From the experimentaldata determinethe gas saturation k.r/k,oequal to 10.0 and at 0. 1, 1. 0 and 0 .0 1 , o r 0 .1 a n d 0 .0 0 1 .(T h e l i stedpai rsof val ues correspond Fi gures to 4,5, and 6 of Johnson's data, respectively, and the rangeof the experimental data dictateswhich figure is to be employed.Note that if the data do not span the entire permeabilityratio intervalof 10.0 to 1.0, Figure 4 may not be employedfirst; instead Figure 5 with the k,*/k.ointerval of 1.0 to 0.01 or Figure 6 with the k.*/k,.,interval of 0. 10 t o 0. 00 1 ma y b e u s e dfi rs t.) Enter the appropriateFigure (4,5, or 6) using the gas saturations corresponding to the pair of k.r/k.ovaluesselected step 2. in Pick a unique S.* and S- at the intersection the gas saturation of values;interpolate if necessary. Using these S.* and S- values and employing the two other figures of Johnson, determinetwo more gas saturationvalues and the k,*/k," ratio indicatedon the axes of each figure. Add thesepoints to the experimental plot for obtainingthe relativepermeabilityratio over the region of interest.

3.
4. 5.

6.

This procedure providesvaluesof gas saturation k.*/k.o at ratiosof 10.0, 1.0, 0.10, 0.01, and 0.001, which are sufficient to plot an expanded k.s/k.o curve. It should be noted that if the data cover a wide range of permeabilityratios, multiple determinations Sr* and S- can be made. If the calculatedvaluesdiffer from the experof imental data, the discrepancy indicatesthat thereis no singleCorey curve which will fit all

23

o tl I o) J

I
o) U)

20
S n , % k r g / k r o = 0 . 1O

FIGURE 4.

Corey equationconstants.2l

the points; an averageof the values for each constantshould yield a better curve fit. Figure 7 illustratesthe graphicaltechniqueof Johnson. Corey's equationsfor drainageoil and gas relativepermeabilities and the gas-oil relative permeabilityratio in the simplestform are as follows: k.o : (s".)o

(24)

k.r:(l-S".)2x(l-S3") and they are related through


I

(2s)

k.. k. : (S * X - (l - S; y

(26)

where So.representsthe lowest oil saturationat which the gas tortuosity is infinite; S". is defined as (S" - S",)/(l - S".). Corey's equationsin the presenceof irreducible water saturationtake the following form:
k,o : (s*)o

(27)

Relative Permeabilitv of Petroleum Reservoirs

q
tl -t

o)
-g

(U aQ o) U)

Sg, %, at krn /kro=0.01


FIGURE5. Coreyequation constants.2t f S 1 2 I t ;---""-^ | " )--)*iJ L

k,n : '

fl

S*)2

(28)

where S- is a constantrelated to ( I - S*") and as a first approximationS- can be assumed to be unity. This is a good approximation, sinceS*"is lessthan5Voinrocks with intergranular porosity. In theseequations,S* : S"/(l - S*,) and S" is the oil saturation represented as a fraction of the pore volume of the rock; S*, is the irreduciblewater saturation,alsoexpressed as a fraction of the pore volume. Theseequationsare linked by the relationship

+ +;-q*: | (s*), (l - s*),

(zs)

Corey et al. plotted severalhundredcapillary pressure-saturation curves for consolidated rocks and only a few of them met the linear relationship requirement. However, comparison of Corey's predicted relative permeabilitieswith experimentalvalues for a large number of samplesshowed close agreement,indicating that Corey's predictedrelative permeabilities are not very sensitiveto the shapeof the capillary pressurecurves. Equation 24 may be employed to calculatewater relative permeability if the oil saturation and the residual oil saturationare replacedby water saturationand irreducible water satu-

)< 0.9

o o o
J

o) .:< (U Ae o U)

Sg, %, at krg/kro of O.O01


FIGURE 6. Corey equationconstants.2l

ration, respectively.The exponentof Corey's water relative permeabilityequation is apon somewhat the size and arrangement rocks, but depends proximatelyfour for consolidated of the pores. The exponent has a value of three for rocks with perfectly uniform pore size distribution. Severalother authorshave proposedsimilar water relative permeabilityequations with different exponentsfor other types of porous media. Values of 3.022and 3.521 sands with a single grain structurewhich may not be were proposedfor unconsolidated absolutelyuniform in pore size but should have a nalrow rangeof pore sizes. Corey compared the calculated values of oil and gas relative permeabilities for poorly values and obtainedgood results. However, consolidatedsandswith laboratory-measured sandstone. Corey for his resultsshowedsome deviationat low gas saturations consolidated concludedthat the equationsare not valid when stratification,solutionchannels,fractures, or extensiveconsolidationis present. Application of Corey's equationpermits oil relative permeabilityto be calculatedfrom are of measurements gas relative permeability.Since k., measurements easily made while are k.o measurements made with difficulty, Corey's equationis quite useful. The procedure of involves the measurement gas relative permeability at severalvalues of gas saturationin and then performing the following steps: an oil-gas system 1. P r e p a r e a n a c c u r a t e p l o t t h e f u n c t i o n k . r : ( l - S " " ) 2x ( l - S . " ' ) b y a s s u m i n g of arbitrary values of So., the effective saturation,which is defined as

26

RelativePermeabilin of PetroleumReservoirs

o
l<

o) .:.
o n<perj-nental Xustirated Data Data of Vlelge points

--

o.lo

o.20 0.30 0.40 0.50 0.60 0.70 Sg


Example of the use of the Corey equations.rl

FIGURE 7.

2. 3. 4.

5. 6.

Preparea tabulation of k., vs. So" for values of k,, ranging from 0.001 to 0.99 in stepwisefashion. Determinevaluesof So"for eachexperimental valueof k., by usingthe above-described tabulation. Plot these values of So. againstthe values of S" coffespondingto the k., values on rectangular coordinatepaper. The plot should be a straightline between50 and 807o oil saturation. Construct a straight line through the points in this range and extrapolateto S.* : 0. The value of S" at this point corresponds S".. (SeeFigure 8.) to Employ Equation 24, k,o : (So")o and the value of S.,.obtainedin the previousstep to calculatek,o valuesfor assumed valuesof S".

Corey-typeequations drainagegas-oilrelativepermeability(gasdrive) in the presence for of connatewater saturationhave been suggested follows: as k ." : (l - S )u

(30) (31)

k.,

s3(2- s)

where S represents (Sr)/(l - S*,). Corey's equationsfor the drainagecycle in water-wet sandstones well as carbonate as formations are as follows:

, - . - : l --l l - s * 1 r K l
Ll - S*,1

(32)

27

60

50

a
o o
@

ro
Sor

ob

20

40
So,

60
o/o

80

roo

on based effective oil of FIGURE8. Determination residual saturation oil saturation. k.*: (S**)o

(33)

VII. WAHL. TORCASO. AND WYLLIE


the use of the following equationfor drainagegas-oil relative Wahl et al.2asuggested reservoirs: of permeabilityratios basedon field measurements sandstone

: +(o.o43s o.4ss6 .l,) +

(341

as ( where rf represents I - S*. - S. - Sg.)/(S,, C); Sr. is the critical gas saturation a pore space;and C is a constantequal to 0.25. fraction of total by Torcasoand Wylliett comparedgas-oilrelativepermeabilityratios calculated Corey's This Wahl et al. for variousirreduciblewater saturations. equationwith thoseobtainedfrom with values sound,sinceit agreed work was theoretically that suggested Corey's comparison by obtainedfrom field measurements Wahl et al. (seeFigure 9).t^ VIII. BROOKS AND COREY modified Corey's original drainagecapillary pressure-saturation Brooks and Corey26'27 relationshipand combined the modified equationwith Burdine's equationto develop the for relativepermeability any pore sizedistribution: that following expression predictsdrainage

28

Relative Permeabilin of Petroleum Reservoirs

roo
50 30
9 y y ;= o ' 3

to
o
.g

5 3

o,
J

to o.5 o.3 o.l o.03 o.ol o.oo5 L o.oot

o.

20

40

60

80

roo

FIGURE 9' Comparison of relative permeability calculations at three irreducible water saturations.25

: s** (l)^

for P. i Po

(3s)

where tr, and Po are constants characteristicof the media; ), is a measure of pore size distributionof the media, and Po is a measureof maximum pore size (minimum drainage capillary pressureat which a continuousnonwettingphaseexists). Using this relationship, two-phaserelative permeabilities given by are
2 + l A r
,

"rwt

- / S * l
\vw

(36)

and

k . n * ,:

(l -'t**)' [t

- (S**)

,.l J

(37)

where k.*, and k-*, are wetting and nonwettingphaserelative permeabilities respectively. The valuesof tr and Po are obtainedby plotting (S* - S*,)/(l - S*,) vs. capillary pr.rrur.

29 on a log-log scaleand establishing straightline with L as the slopeand Poas the intercept a at (S* - S* i) /(l - S * ,) : 1 . Theseequations reduceto Equations24 and 25 for \ : 2. Theoretically\ may have any value greaterthan zero, being large for media with relative uniformity and small for media with wide pore size variation. The commonly encountered rangefor L is betweentwo and Talash28 obtainedsimilar equationswith somewhatdifferent four for various sandstones.2t exponents.

IX. WYLLIE, GARDNER, AND TORCASO


Wyllie and GardnerrThave presentedthe following expressionsfor the drainage wateroil relative permeability:

k,.:(H)'H
'

(38)

ds*/P.' Jr*,

k,.:(5;)'$i11
Relative permeability to wetting phase(k,* and k,"). Nonwetting phaserelative permeability(k,r). Irreduciblewater saturation. Total liquid saturation: (l - Sr).

(3e)

/' or*,1r";'

More general expressionsfor any wetting and nonwetting relative permeability may be written where
kr*r k.n*, S*i SL

(40)

(41)

Wyllie and Gardner have also suggested following equationfor relative permeability the to water or oil when one relative permeability is available:
k.* : (S**)' k,o (S**/(1 S**))'

(42)

where S**, which is defined as (S* - S*,)/(1 - S*,), is the mobile wetting-phase saturation in a water-wetsystem. water Basedon the linear relationbetweenl/P"2and S"/(l - S*,), they obtaineda drainage relative permeability equation for water-wet rocks with intergranularporosity as follows:

Relative Permeabilitv of Petroleum Reservoirs

k,* : (s**)o

(43)

Togpaso and Wyllie2s suggestedthe following equation for calculation of gas-oil relativepermeabilityof water-wetsandstone, where l/P.2 is approximately linear function a of effective saturation.Their derivationwas basedon the relationdevelopedby Corey:

\=: k.,,

( l - s * ) ,( l - s * , )
(s*)o

(44)

where S* represents effectiveoil saturation and is equalto S.,/(l - S*,). Obviously, a reliable value of irreduciblewater saturation, S*r, needsto be known to calculate gas-oilrelative the permeabilityratio. X. LAND, WYLLIE, ROSE, PIRSON, AND BOATMAN

Land2e reportedthat an appreciable parameters adjustment experimental of was required to avoid a discrepancy betweenexperimentaland calculated two-phaserelative permeabilities. A large numberof the relativepermeabilitypredictionmethodsare basedon derivation of pore size distribution factors from the saturationand drainagecapillary pressurerelationship. Some authors3o believethat the employmentof capillary pressure relationships for the prediction of relative permeabilityis not advisable,since capillary pressureis derived from experimentsperformed under static conditions, whereasrelative permeability is a dynamic phenomenon. McCaffery3rin his thesisarguesthat the surfaceor capillary forces are ordersof magnitudelargerthan forcesarisingfrom the fluid flow and thus, predominate in controllingthe microscopicdistributionof the fluid phases many oil reservoirsituations. in Brown's32results from the measurement capillary pressureunder static and dynamic of conditionssupportMcCaffery's argument. Severalrelative permeabilityprediction methodswhich are basedon drainagecapillary pressure curves assumethat pore size distributioncan be derived from thesecurves.These proposedmodels can only be applied when a strongwetting preference known to exist. is Additionally, relativepermeabilitycalculations from capillarypressure dataare developed for a capillary drainagesituationwhere a nonwettingphase,suchas gas, displaces wetting a phase(oil in a gas-oil system,or water in a gas-water system).They are developed primarily for gas-oil or gas-condensate relative permeabilitycalculations; however, water-oil relative permeabilitycan be calculatedwith a lessercertainty. Wyllie in Frick's PetroleumProduction Handbook33 suggested simple empirical gas-oil and water-oilrelativepermeabilityequations drainage consolidated unconsolidated for in and sandsas well as oolitic limestonerocks. Theseequations presented Tables 2 and3. are in The oil-gasand water-oilrelativepermeability relations varioustypesof rockspresented for in Tables 2 and 3 may be usedto producek.g/k.o curvesat various S*, when k., measurements are unavailable. It should be noted that the k,.,/k.* values obtainedapply only if water is the wetting phase and is decreasingfrom an initial value of unity by increasingthe oil saturation.This is contrary to what happensduring natural water drive or waterflooding processes; however, Figures l0 through l4 also apply to preferentiallyoil-wet systemson the drainagecycle with respectto oil if the curves were simply relabeled. Rose6developeda useful method of calculatinga relative permeabilityrelationshipon the basisof analysisof the physical interrelationship betweenthe fluid flow phenomena in porous media and the static and residual saturationvalues. The equationsfor the wetting and nonwetting relative permeabilitesare

3l Table 2 OIL-GAS RELATIVE PERMEABILITIES (FOR DRAINAGE CYCLE RELATIVE TO OIL)33


Type of formation sand, well Unconsolidated sorted sand, poorly Unconsolidated sorted oolitic Cementedsandstone, limestone,rocks with vugular porosity" k"o (S*)' (Sxlt : (S*)' (l k.e 5x;r 5x's) 5x:1

( l - 5 x ; :( l 0 sx), (l -

Note: In theserelationsthe quantity Sx :

S,,/(l - S*,).

Application to vugular rocks is possibleonly when the size of the vugs is small by comparisonwith the size of the rock unit for which the calculation is made. The unit should be at least a thousandfold larger than a typical vug.

Table 3 WATER.OIL RELATIVE PERMEABILITIES (FOR DRAINAGE CYCLE RELATIVE TO WATER)33


Type of formation Unconsolidated sand, well sorted U n c o n s o l i d a t es a n d ,p o o r l y d sorted C e m e n t e d a n d s t o n eo o l i t i c s , limestone (l - S**)' (l (l S**)' (l S**;z (l S**'') 5"x:; k"o k.*

(s**)'
(S**)tt (S**)o

Note'. In these relationsthe quantity S** : (S* S*, is the ineducible water saturation.

S"i)/(l -

S*,), where

k*=

-s* l 6 si (s* -s* _)t(l -)


t2si(2- 3s*.) + 3S*S*-(3S*. 2) + S**(4 5S*,,)1'

(4s)
(46)

k-:

l653*,(5"*, S"-)'(I -,lr* - S.-) -2rlt*- 3S.-)+ 3S"*, + S"-(3S -2* 2,lr*) S,-(l - r!*X4-,lr* - 5S".)]' n [253*,(2

as respectively, expressed wetting and nonwettingsaturations, where S* and Sn*,represent valuesattained minimum wetting and nonwettingsaturation fractions;S*- and S.- represent of as under dynamic flow conditions,expressed fractions;they are the dynamic equivalents an S*, and S". obtainedfrom statictests.The symbol qr* represents immobile wetting-phase saturationexpressedas a fraction. It is that part of the wetting-phasesaturationwhich does not interfere with the nonwetting phase mobility and it is the maximum wetting-phase saturation at which the nonwetting relative permeability is unity. Note that Equation 46 reducesto Equation 45 for r.|l* : 0. The minimum wetting saturation,S**, dependson flow relationship S*- : (1/86.3) of conditionsand may be obtainedby the Brownell and Katz3a due to gravity, o is the interfacial o cos 0) dP/dx]-o264where g is the acceleration [V(g gradient. tension, 0 is the contactangle, k is the permeability,and dP/dx is the pressure of The principal disadvantage Rose'smethodis that the residualsaturation both phases of must be known fairlv accuratelv.

32

Relative Permeabilin of Petroleum Reservoirs

o
j

o)
.Y

20

40
Q vr

60
L

80

too

FIGURE 10. Wyllie curves for water-wetcementedsandstones, oolitic limestones, vugular systems.rl or

Pirson3s derived equationsfrom petrophysicalconsiderations the wetting and nonfor wetting phaserelative permeabilities clean, water-wet,granularrocks for both drainage in and imbibition processes. The water relativepermeabilityfor the imbibition cycle was given
AS

k.*, : later modified to k,*t : and

(S**)"'

(R.,/R,)3/2

(41)

(S**)t"

(R"/R,)3/2

(48)

k.*, :

(S**)t"

Si

(4e)

Water relative permeability for the drainagecycle was given by k .* , : (S * * )t" Si

(s0)

33

o v o)
l<

st
tt i*Ylt brine saturation expressed electricalresistivityof the test core at l00%o where R.,represents as electricalresistivityof the test core expressed ohm-meters; R, as ohm-meters; represents water saturationas a and S* represents saturation; irreduciblewetting-phase S*, represents fraction of pore space. The nonwetting phase relative permeabilityin clean, water-wetrocks for the drainage cycle was found to be
k**, : or k-*, : which was later modified to k,n*, : (l S**Xl S**t/4 Su2)tt2 (53) (1 S**) (l -S**r/4 Sr/2)2 (l S**) [1 S**r'4(R"/R,)r'4]2

water-wet unconsolidated for sorted Wylliecurves poorly

( s1 ) (s2)

The nonwetting phase relative permeability in an imbibition cycle given by

krn*, : [t

S* - S*,

l-s-,-s*J

l'

(54)

(S* -S*,)/(l - S*,) and S.*, represents irreduciblenonwetting the where S** represents phasesaturationas a fraction of pore space.Pirson also derived equationsfor the wetting

Relative Permeabilin of Petroleum Reservoirs

I,OOO Swi

roo

ro

o
J

o) jo.l

0.ol

o.ool
o.0ootoL

20

40

60

80

loo

s, L
FIGURE 12. Wyllie curvesfor well-sorted water-wet unconsolidated } cores. and nonwetting phase relative permeabilities in clean, oil-wet rocks for both drainase and imbibition processes: kr., : (S.r")"' S:

(5s1

where S.* is defined as (S" - S.,.)/( - S".) and S.. represents I irreducibleoil saturationand is the equilvalentof of ( I - S*') for a clean,water-wetrock; S" represents total oil saturation obtainedby differencesfrom (l - S*). The nonwettingphaserelativepermeabilityin clean,oil-wet rocksfor the imbibition cycle was found to be
So So,

krr*,

[' L

l-S..-S*,

t'

(s6)

and for the drainagecycle was found to be


krn-, : (

l --

s,.)u - s:,.-sl,,.l' ,

(s7)

35

3
.Y

o ra

Well--Sorted

Grarns

e -w
FIGURE 13. Wyllie curves.I

by which is determinable Albert and Butault's saturation, trapped-water whereS*, represents curve be obtainedeither suggested that a capillary-pressure These investigators method.36 with a wetting fluid or with a nonwettingfluid such as mercury to obtain irreduciblenonphasesaturation nonwetting that the irreducible They alsoestimated wetting phasesaturation. of is two thirds of net pore volume madeup of capillaries radii smallerthanthe most common the capillary size, when the nonwettingphasedisplaces wetting phase. Pirson suggesteda method to determine the in situ trapped nonwetting phase saturation by meansof microresistivitylogging devices,which respondto the flushed zone arounda well bore: Sn*r:1-(1/0) (R-,/R*.,)"t (58)

where $ representsthe porosity of the reservoir rock and R-r/R^. is the ratio of the mudfiltrate resistivityto flushed zone resistivity. water and gas relative permeabilityequationsin terms of core pesuggested Boatman3T trophysical propertiesobtained from laboratory data:
k,* : S**t'' (R"/R,)3'2

(se)

36

Relative Permeabilin of Petroleum Reservoirs

3
L

l<

o L
l<

20

40

60

80

roo

e -w
FIGURE 14. Wyllie curvesfor water-wetcemented sandstones, oolitic limestones, vugor ular systems.33

and where

k,, :

(1 -

S**t/4 Swt/2)t/2

(60)

e * \'w - S * - S * i I - S*,, Pirson et aI.38proposed equations for oil and water relative permeabilities as follows: k,* : (S**)"t (R"/R,)2 (61)

and k^, : (l - S*-)' (62)

where S*- represents - S*,-)/(l - S*,., - S.,.);S** represents - S*,*)/(l - S*,,,). (S* (S* Thornton5proposedthe following equationfor wetting-phase relative permeability: k.*, : Sl (PD/P.)2 (01)

where P"/P. represents ratio of displacement the pressure drainagecapillary pressure. to Ros e and Wyl l i e T ' 3 e ro p o se da p e tro p hysi calequati onfor w etti ng-p hase el at i ve p r permeability: k.*, : (Ir/2) (64)

where I represents resistivity index, R,/R". proposedmathematical Jonesao relationships water-oil and water-gas for relative permeabilities as function of S* and S*,, where S* may be determinedfrom well logs, S*, may be estimatedfrom an S* - $ crossplot,and d may be determined from well logs:

k.* : (s**)'

(6s)
(66)

k,-:[8H]'
XI. KNOPP, HONARPOUR ET AL., AND HIRASAKI

Knoppa' developeda correlation from 107 experimentallydeterminedgas-oil relative permeabilityratios of Venezuelan core samples.The core samples were from consolidated as well as poorly consolidated reservoirsof high porosity and permeability;the sandstone Welge gas-floodprocedurewas used for k.r/k.odetermination. A single correlationwas established the basisof the restored-state on water saturation as a correlatingparameter. The correlationis shownas a family of most probablek.s/k,., curves in Figure 15. Comparison Knopp's correlationwith experimental of valuesis more promisingwhen the geometricmeanof the suiteof k,s/k,o curvesfor a given reservoir samplegroupis compared or with the corresponding most probablecurves for the correlation.Knopp also suggested a procedurefor developing similar correlationsfor various other formations. A comparisonof Knopp's correlations with the correlationof Corey and Wahl et al. on the basisof l5%owater saturationis shown in Figure 16. Honarpour al.a2 et developed setof empiricalprediction a for equations water-oilimbibition relative permeability and gas-oil drainagerelative permeability from a large number of experimental data. Their resultsare presented Tables4 and 5. Symbolsusedin thesetwo in tables are defined as follows: : ku : ko : ko(s*i) : k., air permeability,md oil permeability,md oil permeability at irreducible water saturation,md gas relative permeability,oil and gas system,fraction k,e(so,):gas relative permeabilityat residualoil saturation, fraction k,o,* : oil relative permeability,water and oil system,fraction : water relative permeability, water and oil system, fraction k* k.o., : oil relative permeability,oil and gas system,fraction

38

Relative Permeability of Petroleum Reservoirs

roo.o
Restored State Water Saturation

o
f-

.Y

o,
f-

.Y

ooorb

24 30 36 42 48 54 60

ss
FIGURE 15. Knopp's correlationof most probablerelative permeability ratios.or

: gas saturation,fraction Ss S*. : critical gas saturation,fraction : oil saturation,fraction S. S..* : residualoil saturation gas, fraction to S..* : residual oil saturation water, fraction to : water saturation,fraction S* S*, : irreduciblewater saturation,fraction : porosity, fraction 0 The data which were usedas a basisfor the study by Honarpour et al. were derivedfr.m oil and gas fields locatedin the continental U.S., Alaska,Canada, Libya, Iran, Argentina, and the United Arab Republic. Alt of the laboratorytestswere made at room temperature and atmospheric pressure'No attemptwas madeby the authorsto group the data according to laboratorytechniques usedin measuring relativepermeability,sincethis informationwas not availablefor many of the data sets.Each set of relativepermeability data was classified

39

o l.Y
\

o.ol

tooo,o

o) l.Y

o.ool

roo.o

o.oool

ro.o

o.ooool

t2 t 8 24 30 36 42 48 54 60

r.o

s g 'o/o
FIGURE 16. Comparisonof relative permeabilitycorrelations.*'

"carbonate" or "noncarbonate", but the informationwhich was availablewas not as either sufficientfor more detailedlithologic characterization. "carbonate" or "noncarbonate", a further of In additionto the classification data setsas classificationwas made on the basis of wettability. This rough classificationwas made accordingto the following arbitrarycriteria: l. in to The rock was considered be strongly water-wetif k,,,at high oil saturations an greatly exceededk,o in a gas-oil system at the same saturations, oil-water system k,* providedk.* in a gas-oil systemgreatlyexceeded in an oil-water systemat or near water-flooding. after residualoil saturation to The rock was considered be oil-wet when k,o in the oil-water systemwas approximately equal to k,., in the gas-oil system, provided k,* in the gas-oil system was approximatelyequal to k.* in the oil-water system.

2.

Relative Permeability of Petroleum Reservoirs Table 4 EQUATIONS FOR THE PREDICTION OF RELATIVE PERMEABILITY SANDSTONE AND CONGLOMERATE
(s* - s"') k... : 0.035388 - S * , - S , , , * -) o.olo874x (l 't"r- t"':' - * f . , - S " ,. l" + o.sos56(S*)rn(SS*,) (water-wer) S,,,")l Ltt k . , .: -'s*' (s* - s"'*) r . 5 8 1 4s * [ l ' " ' - 0 . 5 8 6 1 7 - s * , - S , , ,** ) (l ll-s", I (S" - S*,) - 1.24846( - S*,)(S* - S*,) (intermediately I wet)
(68)

IN

(61)

r \/l - s "'/ \ - s l , r . ' s k,,,*:0.760671 '- s,,,* [ , t-" = - t-" s,,,*J t' s*, ' : l ' " L I
+2.63180(l S , , . * ) ( S-, S " , * ) ( a n y w e n a b i l i t y ) , (69)

t,,=- t,,,,: , k n ' ,: 0 ' 9 8 3 7(l + 2 weuability) l. (any t-, ). L | - s -, - s,,r: [ ] k .*: l o ? 2(H )' u ,* ,,,,. ., + 2.i i g4*

(70)

\+

(anY k'g's,,,r' wettabilitY)

(71 )

Table 5 EQUATIONSFOR THE PREDICTIONOF RELATIVE PERMEABILITY IN LIMESTONE AND DOLOMITE


k* : 0.002oszs \f - 0.05r371 - S*,) (s* (i)"" (warer-wet)
(72)

k*: o2eer. (H)

- otztot (ffi;)'(13)

(S*-S*.)*0.4|325g(*).(intermediate|ywet)

k..* :

1.2624 (H:)

(*)'

(anywettablity)

(74)

: k*,e 0.s37s2 (jil'(ff-o'*: 'sossffik'gts,,,*t


-

,_)'

(any wettablity)

(75)

+ 8'oo53x

rS., -T-l=S , l r S . . . .-r : o'o258eo - S..)x {s.

(#)'.
- t")' ' - t" (' (t)"' _ i:: (any wettab'ity)
(76)

4r
3. wettability when it did not clearly meet to The rock was considered be of intermediate criteria. either the water-wetor the oil-wet classification

After the data sets had been classifiedaccordingto lithology and wettability, stepwise the which would approximate analysiswas employedto developequations linear regression permeability, and fluid saturations, from such factors as relative permeabilities measured porosity. of refer to displacement oil by water and the oil-gassystem All water-oilsystemequations in data were measured consolidated All refer to drainageprocesses. experimental equations rocks. The equationsthat were developedby Honarpouret al. have not yet been extensively tested.However, most of the testswhich have been made indicatedthat the equationsare which correlations of with laboratorydatathan the predictions publisfred in closeragreement were used as a basisfor comparison. et by suchas thosepresented Honarpour al., any calculated In usingempiricalrelationships equal to 1.0. If a relative relative permeability which exceeds l 0 should be assumed the permeabilityvalue is known at any water saturation, relativepermeabilitycurve may be shifted to match the known data point. Hirasakia3has suggesteda relative permeability correlation for fractured reservoirs as follows: S*:
Su So.

l-s*-So"

(77) (78)

k,a : K,o(S*)" k,, : k:" (l - S*)' where S* Sd So" So. k.a ko.o k,o k".. n

(7e)

Normalizedsaturation. Displacingphasesaturation. Immobile displacingphasesaturation. Residualoil saturation. Displacingphaserelative permeability. Displacingphaserelative permeabilityat residualoil saturation. Relativepermeabilityto oil. Relativepermeabilityto oil at immobile displacingphasesaturation. for Exponentparameter shapeof relativepermeabilitycurves, said to be equal to one in fractured reservoirs.

: : : : : : : : :

-..\

,_ trt

l) rr",/J-

'i i" ,tY

REFERENCES

tdpullien, F. A. L., Ed., Porous Media: FluidTransport and Pore Stucture, Academic Press, New York, 'r,l4lg. Akad. Wiss. Wien. Math. Naturwiss. 2. Kozeny, J., Uber Kapillare Leitung desWassersimBoden, Sitzungsber. KL., Abt. 2A, 136,2'll, 1927. using mercury and the calculationof permeability 3. Purcell, W. R., Capillary pressures their measurement therefrom.Trans. AIME, 186.39, 1949.

42

RelativePermeabilin of PetroleumReservoirs

4. Rose, W. D. and Bruce, W. A., Evaluationof capillary characterin petroleumreservoir rock, Trans. .AIME, t86, 127, t949. 5. Thornton, o. F., valuation of relative permeability,Trans. AIME, 1g6,329, lg4g. 6. Rose, W. D., Theoreticalgeneralization leadingto the evaluationof relativepermeability,Trans.AIME, 186,1il , 1949. 7. Rose, W. and Wyllie, M. R. J., Theoreticaldescriptionof wetting liquid relarivepermeability Trans. , AIME, 186,329, t949. 8. Gates,J.I.andLeitz,W.J.,RelativepermeabilitiesofCaliforniacoresbythecapillarypressuremethod, paper presented the API Meering, Los Angeles,california, May ll, 1950, 296. at 9' Rapoport, L. A. and Leas, W. J., Relative permeabilityto liquid in liquid-gassystem, Trans. AIME, 1 9 2 ,9 3 , l 9 5 l . 10. Wyllie' M. R. J., Interrelationship betweenwetting and non-wettingphaserelativepermeability,Trans. A I M E , 1 9 2 ,8 3 , 1 9 8 1 . ll. Fatt, I. and Dykstra, H., Relativepermeabilitystudies,Trans. AIME, 192,249, lg5l. 12. Wyllie' M. R. J. and Sprangler, M. B., Application of electricalresistivitymeasurements problems to of fluid flow in porous media, Bull. AApG, 36, 359, 1952. 13. Burdine, N. T., Relative permeabilitycalculations from pore size distributiondata, Trans. AIME, lgg, 7t,1953. 14. Naar, J. and Henderson, J. H., An imbibition model- its application flow behaviorand the prediction to o f o i l r e c o v e r y ,T r a n s .A I M E , 2 2 2 , 6 1 , 1 9 6 1 . 1 5 . N a a r , J . a n d W y g a l , R . J . , T h r e e - p h a s e i m b i b i t i o n r e l a t i v e p e r m e a b i l i t y , T r a n s . A I M E , 2 21 9 2 5. 4 , 2,6l 16. Land, C. S., Calculation of imbibition relative permeabilityfor two- and three-phase flow from rock properties, Soc. Pet. Eng. J., 6, 149, 1968. 17. Wyllie' M. R. J. and Gardner, G. H. F., The generalized Kozeny-Carmen equation,its applicationto problemsof multi-phaseflow in porous media, World Oit, 146, l2l, 1958. 1 8 . T i m m e r m a n , E . H . , B d . , P r a t ' t i L ' aR e s e r t , o iE n g i n e e r i n g P e n w e l lp u b r . , r 9 8 2 , l 0 l . l r , 1 9 . C o r e y , A . T . , T h e i n t e r r e l a t i o n e t w e e n a s a n d o i l r e l a t i v ep e r m e a b i l i t i e s ,r o d . M o n . , 1 9 , 3 8 , 1 9 5 4 . g b P 20. Corey, A. T. and Rathjens, C. H., Effect of stratification relativepermeability,Trons.AIME,20j, on 358,1956. 21. Johnson, C. E., Jr., Graphicaldetermination the constants the Corey equationfor gas-oil relative of in p e r m e a b i l i t y a t i o , J . P e t . T e c h n o l . ,1 0 , l l l l , 1 9 6 8 . r 2 2 . l r m a y , S . , O n t h e h y d r a u l i cc o n d u c t i v i t y f u n s a t u r a t es o i l s ,T r a n s .A G U , 3 5 ( 3 ) , 4 6 3 , 1 9 5 4 . o d 23. Averganov, S. F., About Permeabilitl, ofSubsurfuc'e Soils in Case of IncompleteSaturation, Engineenng Colfection, Vol. 7, 1950, cited by Polubarinova-Kochina, in The Theory of Ground Water Movement, P, E n g l i s ht r a n s l a t i o n y D e w i e s t ,R . J . M . , P r i n c e t o n n i v . P r e s s , r i n c e t o n N . J . . t 9 6 2 . b U P . 24. Wahl, W. L., Mullins, L. D., and Elfrink, E. 8., Estimationof ultimate recoveryfrom solution gas drive reservoirs,Trctns.AIME, 213, 132, 1958. 25. Torcaso, M. A. and Wyllie, M. R. J., A comparison calculated of k.r/k,,, ratios with field data,J. pet. Technol., 6, 57, 1958. 26. Brooks, R. H. and Corey, A. T., Hydraulic Properties of Porous Media, Hydrology papers, No. 3, Colorado State University, Ft. Collins, Colo., 1964. 27. Brooks, R. H. and Corey, A. T., Propertiesof porous media affecting fluid flow, "/. Irrig. Drain. Div.. 6.6t. 1966. 28. Talash, A. W., Experimentaland calculatedrelative permeabilitydata for systemscontaining tension additives,Paper5810, Societyof PetroleumEngineers,Dallas, Tx., 1976. 29. Land, C. S., Calculation of imbibition relative permeability for two- and three-phaseflow from rock properties,Soc. Pet. Eng. J., 6, 149, 1968. 30. Bear, J, Ed., Dynamics of Fluids in porous Media, Ersevier,Amsterdam, 1972. 31. McCafferY, F. G., The Effect of Wenability of Relative Permeabilityand Imbibition in porous Media, Ph.D. thesis,Universiry of Calgary, Alberta, Canada,1973. 3 2 . B r o w n , H . w . , c a p i l l a r y p r e s s u r en v e s t i g a t i o n s , T r a n s .I M E , 1 g 2 , 6 7 , l g 5 l . i A 33. Frick, T., Ed., PetroleumProductionHandbook,Vol. 2, Societyof Petroleum Engineers AIME, Dallas, of Tx., 1962.25. 34. Brownell, L. E. and Katz, D., Flow of fluids through porous media, Chem. Eng. prog., 43(ll), 603, 194'7. 3 5 . Pirson, S. J., Ed., Oil ReservoirEngineering,McGraw Hill, New york, 195g. 36. Albert, P. and Butault, L., Etude des Characteristiques Capillaries du Reservoir du Cap don par La M e t h o d eP u r c e l l ,P e t . A n n . C o m b u s .L i q . , 1 ( 8 ) , 2 5 0 , 1 9 5 2 . 37. Boatman, E. M., An Experimental Investigation of Some Relative Permeability-RelativeConductivity Relationships, M.S. thesis,University of Texas, Austin, 1961. 38. Pirson, S. J., Boatman, E. M., and Nettle, R. L., Predictionof relativepermeabilitycharacteristics of intergranular reservoirrocks from electricalresistivitymeasurements,Trans. AIME, Z3l,564. 1964.

43
evaluationof relatedto quantitative considerations 39. Wyllie, M. R. J. and Rose, W. D., Some theoretical physicalcharacteristics reservoirrock from electricallog data, Trans. AIME, 189, 105, 1950. of , 4 0 . J o n e s , M . A . , W a t e r f l o o dm o b i l i t y c o n t r o l :a c a s eh i s t o r y ,J . P e t . T e c ' h n o l .9 , l l 5 l , 1 9 6 6 . 41. Knopp, C. R., Gas-oil relative permeabilityratio correlationfrom laboratorydata,J. Pet. Technol.,9, llt1,1965. two-phase for 42. Honarpour, M. M., KoederitzrL. F., and Harvey, A. H., Empiricalequations estimating rock, Trans. AIME, 2'73,2905, 1982. relativepermeabilityin consolidated by by 43. Hirasaki, G. J., Estimationof ReservoirParameters History Matching Oil Displacement Water or E , G a s , P a p e r4 2 8 3 , S o c i e t yP e t r o l e u m n g i n e e r sD a l l a s ,T e x . , 1 9 7 5 . 44. Kopli.k, J. and Lasseter, T. J., Two-phaseflow in random network modelsof porous media, Sot'. Pet. Eng.J., 25, 89, 1985. on 45. Fulcher, R. A., Ertekin, T., and Stahl, C. D., Effect of cappillarynumberand its constituents twophaserelative permeabilitycurves,J. Pet. Technol., 2,249, 1985.

Chapter 3 FACTORS AFFECTING TWO-PHASE RELATIVE PERMEABILITY

I. INTRODUCTION flow of multiple fluid phases the The first publishedinformationconcerning simultaneous "relative permeability" had not yet been coined was probably by Hassleret al.r The term of and Hassleret al. studied only the flow characteristics the gas phaseas a function of conceptwas first postulated rocks. The relativepermeability in fluid saturation consolidated systems. Darcy's law to two-phase of Their work consisted extending by Muskat and Meres.2 For oil reservoirs,the relevanttwo-phasefluid combinationsare water-oil and liquid-gas (usually thought of as oil-gas). Gas-waterrelativepermeabilitycurvesare used to describe reservoirs. and gas-liquidcurvesare usedfor condensate of the performance gas reservoirs II. TWO.PHASE RELATIVE PERMEABILITY CURVES

Water-oil relative permeability is usually plotted as a function of water saturation,as (S*.), the water relativepermeability shownby Figure l. At the irreduciblewater saturation is zero and the oil relative permeabilitywith respectto water is some value less than one. by At this point only oil can flow and the capabilityof the oil to flow is reduced the presence of connatewater. The effect of connatewater in reducingoil flow rate is illustratedschematically by Figure 2. Note that data to the left of the irreducible water saturationare not useful for predicting less sincewatersaturations thanS*" arenot encountered. reservoirperformance, hydrocarbon and the oil relative the As water saturationincreases, water relative permeabilityincreases A permeability(with respectto water) decreases. maximum water saturationis reachedat zero. Obviously, aquifer becomes and the oil relativepermeability the residualoil saturation by conditionsare represented a relative permeabilityto water of unity, which occurs at a of water saturation l00%o. Unfortunately,there is an alternatedefinition of relative permeabilitycurrently in use. This terminology(illustratedby Figure 3) definesthe oil relativepermeabilityat irreducible permeabilityas the effective as water saturation having a value of one, and definesabsolute are The effectivepermeabilities identical with permeabilityat irreduciblewater saturation. of relative permeability and one set of values may be readily convertedto both definitions the other. This second definition of relative permeability (k,r) applies to both the oil and water phases. Thesealternateor normalizedvaluesof relative permeabilitymay be convertedto standard valuesby
k.srn : k,2 ku./kusrD

(l)

where k.. : k"o at S*" the definition of relativepermeability, waterrelativepermeAlso note that underthis second definition, with this alternate ability in an aquiferhas a value greaterthan unity. Essentially, relative permeability is normalized to the value at irreducible water saturation. Gas-oil relative permeabilityand gas-liquidrelative permeabilityare similar in concept to water-oil relativEpermeability. The preferredrelative permeabilityvalues are those taken with connatewater presentat the ineducible saturationvalue.

46

Relative Permeability of PetSoleum Reservoirs

I I I

\ oil
W a te r I
/ /

ftret

Swc 0

Sorw

Svrr-+ (-s o-

FIGURE l.

W a t e r - o i lr e l a t i v ep e r m e a b i l i t y u r v e s . c

,'t Rock ) oif-

;Water

L
(

-.*-t

Rock \

FIGURE 2.

of Oil flow reductiondue to the presence water.

with respect gasdecreases; to increases, oil relativepermeability the As free gas saturation however, until the critical gas saturation(Sr") is reached,the gas relative permeabilityis is zero. The critical gas saturation the point at which the gas bubblesbecomelarge enough increases, the to break through the oil and away from the rock surface.As gas saturation gas relative permeabilityincreases a and theoreticallyreaches value of unity at l00%cgas. A gas-oil relative permeabilitycurve is illustratedby Figure 4. sand procedureto determinerelativepermeabilityin an unconsolidated An experimental of was first described Wyckoff and Botset.3 by Their work consisted injectinga combination conditions.Their resultsare shown througha sampleundersteady-state of liquids and gases in Figure 5, where k.. and k,, are relative permeabilityto oil and gas, respectively.The of figure is typical of wetting- and nonwetting phase relative permeabilities,regardless whetherthe systemis oil- or water-wet. Figure 5 shows differently shapedrelative permeabilitycurvesfor the two phases.The oil relativepermeabilitycurve is concaveupward while the gas relativepermeabilitycurve at has an "S" shape.This figure also showsthat the oil relativepermeability the irreducible

47

k r e l

Sw
FIGURE 3. Normalizedwater-oil relativepermeabilitycurves

\ \

\1

Gor

oil I K rcl

I I

, ,

SwcSorg O - S L 4

Sg. |

t
FIGURE 4.

+-SG
Gas-oil relative permeabilitycurves'

at the ineducible oil (or critical) gas saturationis less than the gas relative permeability apply to water-oil same generalobservations saturation.Leverett,sworka shows that the of presence oil, the water relative permeability relative permeabilitydata. That is, in the relative permeabilitycurve or is concave curve takes on the shape of the wetting-phase upward. 5 indicatesthat, for a small The shapeof the oil relative permeabilitycurve in Figure in decrease relativepermeabilityto oil' This reductionin oil saturation,there is a sizeable pathsby the gasphase'Figure of rapid declineis due to the occupation largerporesor flow

Relative P ermeability of t etroleumReservoirs

ftrel

kts

l't:.:':)'/,.
A
SL

FIGURE 5.

Relative permeabilitycurves for an unconsolidated sancl.r

5 alsoindicates steepincrease the gasrelativepermeability the gassaturation a in as increases abovepoint "A", which is the saturation which relativepermeabilities the oil and gas at to phases becomeequal. For this unconsolidated sand,the oil relativepermeabilityat 59Vo orl saturationis equal to gas relative permeabilityat 4l%o gas saturation.The gas relative permeabilityreaches nearly l00%o a gas saturation at lessthan l007o, which 1n.un,that part of the interconnected pore spacedoes not significantlycontributeto the gas permeabilityof the porousmedium. This figure also showsthat the gasrelativepermeability remainsat iero until the gas saturation reaches critical gas saturation, the point "B". The gas phaseis not mobile at a saturationless than the critical value, but this immobile gas impedesthe flow of oil and reducesoil relative permeability.As oil saturationis increased from an initial value of zero, the oil relative permeabilityremainszero until the oil forms a continu.us phaseat the critical oil saturation, which is represented point C in Figure5. In a solutionas gas-drivereservoir, often the water saturationis small and immobile. Therefore,relative permeabilityvaluesare frequentlyplottedagainst liquid saturation the ratherthan the wetting saturation.Under such a condition, point "C" is the summationof the irreduciblewater saturation and the residualoil saturation,as previouslyindicatedin Figure 4. The sum of the relative permeabilitiesfor all phasesis almost always less than unity because interference of amongphases sharingflow channels. Thereare a numberof reasons for this interference.One of these reasonsis that part of the pore channelsavailablefor flow of a fluid may be reducedin sizeby the other fluids present the rock. Another reason in is that immobilized dropletsof one fluid may completelyplug someconstrictions a pore in channelthrough which anotherfluid would otherwiseflow. Also, somepore channels may becomeeffectively plugged by adversecapillary forces if the pressure gradientis too low to push an interface through a constriction. A fourth reason is the trapping of a group of globules that are clustered together and cannot be moved, since the grain configuraiion allows fluid to flow around the trappedglobules without developinga pressuregradient sufficient to move them. This is the phenomenon that has been referredto as the Jamin effect.

of Nowak and Kruegerstestedtwo coresin which the permeabilityto oil in the presence permeabilityto syntheticforwas considerablygreaterthan single-phase interstitialwater basedon the resultsof mation water. Yuster6and OdehTboth found the samephenomenon work. A possibleexplanationfor the high permeabilityto oil is that the distribution other causevariationsin the area of clay varies within the rock and variationsin water saturation of degrees clay swelling may betweenwater and clay minerals.Thus, increasing of contact due to the hydrationof larger amountsof clay minerals. higher water saturation occur at and the distributionof upon both the fluid saturation Relativepermeabilityis dependent of fluids in the interstices the porousnetwork.This distributionis directly related the various pheof to wettability characteristics the rock, which in turn give rise to capillary pressure curves; It is well known that hysteresisexists in capillary pressure-saturation nomena. curvescan also be expected.Thus, in therefore,hysteresis relative permeability-saturation in saturation,the relative permeabilitymeasured a rock that is given wetting-phase for a while the rock is draining. phase is not the same as that measured imbibing the wetting overburden valuesalsomay be functionsof factorssuchastemperature, Relativepermeability phaseequilibria,ro' etc. pressure, III. EFFECTS OF SATURATION STATES

Saturationis a term used to describethe relative volume of fluids in a porous medium. of At low saturations the fluid that preferentiallytends to wet the grains of a rock, the rings aroundthe grain contactpoints. Theseare called phaseforms doughnut-shaped wetting rings do not communicatewith each other and pressurecannot be pendular rings. The such a distributionmay occupy transmittedfrom one pendularring to another.Sometimes fraction of the pore space.The amount dependsupon the nature and shape an appreciable of individqal grains, distribution,as well as degreeand type of cementation. saturation,the wetting phaseis mobile through a tortuous Above the critical wetting-phase the saturationincreases, wettingpressuredifferential and as the wetting-phase path under a saturationdistribution in as permeability increases well. The wetting-phase phaserelative this region is called funicular and up to a point, the relative permeabilityto the wetting force phaseis lessthan the relativepermeabilityto the nonwettingphasedue to the adhesion wetting fluid, and the greatertortuosity of the flow path for between the solid surface and the wetting phase.The nonwettingphasemoves throughthe larger poreswithin this range of saturation, but as the saturationof the wetting phase further increases,the nonwetting phase breaks down and forms a discontinuousphaseat the critical nonwetting phase satusaturation. ration. This is called an insular stateof nonwetting-phase through that when immisciblefluids flow simultaneously Fluid flow studieshave shown its own flow path. This flow network changesfor a porous medium, each fluid follows different ranges of saturationand as the nonwetting phase saturationreduces,the network the for this phasebreaksdown and becomesdiscontinuous; remainingstationaryislandsof in at pressuregradientsencountered hydrocarbon the nonwetting phasecanrnt be displaced Similarly, as a residualnonwettingphasesaturation. This conditionis refened to reservoirs. the network through which this phaseflows breaks as the wetting phasesaturationdecreases, and discontinuous immobile. This is referredto as an ineduciblewettingdown and becomes phase saturation. sandsthe permeability that for strongly water-wetunconsolidated It has been showns-rr saturation,(i.e., a plot of k.* as a solely upon its own to the wetting phaseis dependent of function of S* has the same shaperegardless whether or not the pore spacecontains gas as well as oil). However, in the petroleumrelatedliterature,somesmall'2''3and somequite rocks. Somepublicationsr4'15 large deviationsare seenfrom thesefindings for consolidated indicate that the nonwetting phaserelative permeability dependson the wetting as well as

50

Relative Permeability of P,etroleum Reservoirs

average

o
J

o)
Y 1

minimum

.9 (u
TE

-o o

(u

E
o o. o .=

-01

(u
o (r

0.5

1.0

S L,
FIGURE 6. R e l a t i v ep e r m e a b i l i t y a t i o sf o r s a n d sa n d s a n d s t o n e s . r s r

the nonwetting phase saturationfor strongly water-wet systems.In preferentiallyoil-wet systems, oil phase the relativepermeability found to be strictlya functionof oil saturation,r6 is while in water-wet rocks, the oil phaserelative permeabilityis found to dependon both water and oil saturation. Donaldsonand Dean'7havepointedout that undertwo-phase flow, relative permeabilityto water was increased when oil, ratherthan gas was the nonaqueous phase,indicatingthat water relativepermeabilityis not solelya functionof water saturation.

IV. EFFECTS ROCK PROPERTIES OF


Relativepermeability-saturation relationsare not identicalfor all reservoirrocks, but may vary from formation to formation and from one portion to another of a heterogeneous formation. Arps and Roberts'8have presentedplots of gas-oil relative permeabilityratios for 16 consolidatedsandstones and 25 dolomites, cherts, and limestones,all with l57o connate water saturation.These plots are presented Figures 6 and 7. The maximum curve in as Figure 6 seems be typical of unconsolidated to sandstone, while the minimum curve appears to be more representative highly cementedsandstones. of The averagecurve can be consideredtypical of the averageconsolidatedsandstone. The minimum curve in Figure 7, which seemsto be the steepest and most unfavorable, from a fracturedchert core; at the is other end of the range, no well-definedmaximum case is apparent.Curve #23, adapted from Bulnes and Fitting's workrerepresenting26 samples west TexasPermiandolomite, of appears be the bestmaximum curve. The curve selected "average" on Figure7 appears to as to be typical of vugular limestones. Bulnesand Fitting as well as Stone2o have shownthat the fluid flow behaviorin uniformporositycarbonate samplesis similar to fluid flow behaviorin consolidated sandstones, but the differencebecomespronounced the rock heterogeneity as increases.

5l
10

o
-g

ma x im um

.\

a verage

nrmum

o)
. Y l

\ \

\ \

o
(!

tr .:
-o
I

(u o E
q,

oo
.: (! .01

tr

. o o1

0.5

s,
L

1.0

FIGURE 7. cherts.r*

r R e l a t i v ep e r m e a b i l i t y a t i o sf o r l i m e s t o n e sd o l o m i t e s , n d . a

Various workss''e'2r have shown that the gas-oil relative permeability of consolidated sand sandstone qualitativelysimilar to the gas-oil relativepermeabilityof unconsolidated is to of and there is a very close coffespondence the two relativepermeabilities oil at high oil sand, the wetting-phase relative permesaturation.It has been found that for consolidated ability drops sharply and the nonwetting phase relative permeabilityrises steeply as the However, Naar et aI.22 have shown that there are both wetting-phase saturationdecreases. qualitative and quantitativedifferencesbetweenrelative permeabilityof consolidated and thatpackingasmodifiedby cementation unconsolidated sands.Owensand Archerrrindicated to and consolidation affectsthe equilibrium saturation the wetting phasebut has a negligible of effect on the equilibrium saturation the nonwettingphase.Nind23statedthat an increase the in degreeof consolidationincreases nonwettingphaserelativepermeabilityin a gas-oil rangefor a mobile fluid phase have noted that the saturation system.Severalinvestigators rock than in consolidated rock. is wider in unconsolidated on Corey and Rathjens2a studiedthe effect of rock heterogeneity drainagegas-oil relative permeability.They investigated flow paralleland perpendicular obvious stratification the to in anisotropicBereasandstone coresand concludedthat the relativepermeabilityat a given value for flow persaturationfor flow parallel to bedding was greaterthan the analogous pendicularto the bedding plane, as shown in Figures 8 and 9. Huppler2s found that the when the sections water-oil relativepermeabilityof compositecore changes appreciably are arrangedin different orders. Johnsonand Sweeney'o also studiedthe effect of rock heterogeneityon the gas-oil relative permeabilityratio. changein the positionof the relativepermeabilityLeveretta found a small but systematic

52

Relative Permeabilin of Petroleum Reservoirs

o
J

c oo

FIGURE8. Relative permeability measurements from


sandstone.ra

an anlsotroprc

O O O O

- kro - porpondicul!r - kro - trg

to boddinO

- parEllol to baddlng - p.rpondlculr. to b.ddinc

- krg - p!..llol

to boddlng

o
l(

so
FIGURE 9. Relative permeability measurements from a Berea sandstone.2a

53

Time 1

Time 2

E
Time 3

otL
WATER SAND

M I

FIGURE 10. The formation of residualoil by the blocking process.

saturationrelationship due to the employment of different sizes of sand grains in his experiments.Botset2r confirmed Leverett'sfinding and concludedthat the effect of grain size distributionwas not negligible either on the relationship betweenrelativepermeabilityand saturationor on the value of the equilibrium saturation.It was found that the shape" (sphericity),roundness"(angularity),and orientation2a the grainstendedto influenceboth of the shapeof the relative permeability curve and the critical gas saturationvalue in gas-oil systems. Leverettapointed out that the relative permeability of an unconsolidatedsand to an oilwater mixture is relatedto the sandpore size distribution.Muskat et a1.27 suggested that it is necessary know the pore geometry of a reservoir rock before fluid movement through to it can be analyzed. Morgan and Gordon2sfound that pore geometry and surface area per unit volume influencedwater-oil relativepermeabilitycurves.They have shown that rocks with large pores and correspondinglysmall surfaceareashave low irreducible water saturations and therefore have a relatively large amount of pore spaceavailable for the flow of fluids. This conditionallows high relativepermeability end pointsto exist and allows a large saturation changeto occur during two-phase flow. Correspondingly, rocks with small pores have larger surface areasper unit volume and they have irreducible water saturationsthat leave little room for the flow of hydrocarbons.This condition createsa low initial oil relative permeability as well as a limited saturationrange for two-phaseflow. Gorring2e demonstrated that oil in a larger pore can be surroundedand blocked off when it is encircledby smaller pores which imbibe the displacingwater by capillary forces. He concluded that both pore size distribution and pore orientation have a direct effect on nonwetting residual equilibrium saturation, as shown by Figure l0; therefore, a perfectly uniform packing of spheres shouldgive a residualsaturation nearzero.Gorring also identified the size of channelsoccupied by the nonwetting phase as an important factor influencing relative permeability. Crowell et al.30indicated that higher initial water saturationyields a higher probability for the nonwettingphaseto be in larger channelsso-thatit can b9 recovered efficiently during wetting-phaseimbibition. Botset2' mentioned as early as 1939 that the relative perrneability-saturation relation depends the degreeand the type of interconnections the pores.Fatt,3rDodd and Kiel,32 on of and Wyllie33 also concluded that the relative permeability of porous media is a direct consequence the network structureof the media. Pathaket al.3aconcludedthat the ratio of of pore size to pore throat is a factor which controls the snapping-off of droplets of the

54

Relative Permeability of Petroleum Reservoirs

nonwettingphase,with a high ratio leadingto a high trappedoil saturation. Other workers have investigated possibilityof describingporousmedia as a network of interconnected the pore bodies and pore throats. Postdepositional alterations can form more than one type of reservoirrock from a single original rock type. Alteration may reducepore sizes,thus causinghigher irreduciblewater saturation and a natrow rangeof saturation changeduring two-phase flow. The presence of grains such as feldspar, when partially dissolved,improvesthe reservoirrock quality by forming pores larger than the pores betweengrains not containingfeldspar.This alteration causeshigher relative permeabilityvalues and a larger saturationrange during two-ph4sg flow.tt Reference35 describesalterationsin pore geometry which can occur due to the introductionof reactivefluids in the rock. Land and Baptist36 indicatedthat when a reservoirsandstone containsmontmorilloniteor mixed-layer clay mineralscontainingexpandable layers,the watersensitivity the sandstone of is not necessarily result of pore blockagedue to the increased a volume occupiedby the swollen montmorillonite.Some sandstones containingtrace amountsof clay mineralsmay exhibit sensitivityto water resultingfrom dispersionand subsequent transportation clay of mineralsto pore constrictions.Thus, permeabilityreductionmay occur in formationsthat do not contain expandableclay minerals;however, all formationscontainingexpandable clays are probably water-sensitive due to the easeof dispersionand expansionof this type of clay. Permeabilityreductionin sandscontainingsodiumclays is likely to be higher than the reductionin sandscontainingcalcium clays. Somerock properties that influencerelativepermeabilityvariations readily observable are with a binocular microscopeor even more clearly under a scanningelectronmicroscope. Therefore,microscopiccore examination can be highly usefulfor evaluating relativepermeability characteristics. Once the significantrock propertyvariationshave been identified, a reservoir can be subdivided into appropriatereservoir rock types. Within each of such reservoirrocks types, relative permeabilitycharacteristics usually similar, varying only are slightly for rather large changesin air permeabilityor mediangrain size.

V. DEFINITIONAND CAUSES WETTABILITY OF


"Wettability" is a term usedto describethe relativeattractionof one fluid for a solid in the presence other immisciblefluids. It is the main factor responsible the microscopic of for fluid distribution in porous media and it determinesto a great extent the amount of residual oil saturationand the ability of a particular phaseto flow. The relative affinity of a rock to a hydrocarbon the presence water is often described "water-wet", "intermediate", in of as or "oil-wet". Examplesof formations with strongly water-wet, strongly oil-wet, and intermediatewettability are the Spraberry formation in west Texas, the Black Bradford sand in Pennsylvania, and the Fairbanksand in south Texas, respectively. Wettability may be represented the contact angle formed among fluids and a flat solid by surfaceor the angle formed betweenthe fluids' interfaceand a glasscapillary tube, as shown by Figure I l. The angle is measured throughthe denserfluid. The wettability of a porousmedium is determined a combinationof all surfaceforces. by A sketchis shown in Figure 12, wherein two liquids, oil and water, are in contactwith a solid. The force exerted by water to spreadlaterally and displaceoil (interfacialtension betweenwater and oil) is opposedby the resultantof the solid and liquid forces (solid-oil and solid-waterinterfacialtensions). This difference opposing in forcesis calledthe adhesion tension:
A, : o.o or* : o*o cos 0*o Q)

This relationship is referred to as the Young-Dupre equation, where A, is the adhesion

55

WATER-WET

OIL-WET

INTERMEDIATE

e<goo
FIGURE I L

g >goo

g=9Oo

_w;, ='-Kfig==
,l I\
Ttrp Vian of a Drop of VJater on a Solid Surface in the Presence of Oil Ttrree

and in capillarytubes. Wettability conditionson flat surfaces

Dirrensional

Sdrernatic

View

FIGURE 12. Forcesat a water-oil-solidinterface.

tension;oso,o,*, and o*o, respectively,are solid-oil, solid-water,and water-oil interfacial tensions(usually measuredin dyne/cm); 0*" is the contact angle betweenwater and oil measured through the denserliquid phase(usually water). tensionmeansthe contactangleis lessthan 90' and the solid A positivevalue of adhesion surfaceis preferentiallywater-wet. A zerovalue of adhesiontensionindicatesthat the contact angle is equal to 90'; this is intermediatewettability. A negativevalue of adhesiontension meansthe contact angle is greaterthan 90' and that the solid surface is preferentially oil wet. There is no practical laboratory method for meaquantitieswhich can be used to suring trsoor o.*. However, o*o and cos 0 are measurable evaluate the wettability of a solid surface. A fluid is referred to as wetting or nonwetting to a surfacedependingon whether the contact angle is less than or greaterthan 90". Understandingthe causesof wettability requires a study of the chemistry of the fluids, the polarity and molecular weight of reservoir hydrocarboncompounds,and the occurrence at of surfacechemical processes the solid-fluid interfaces. experimentally found that the contact angles vary directly with Stegemeierand Jensen3T Figure l3 showsthis variation molecularweight for liquids with similar chemicalstructures. for the normal paraffin seriescompounds.

56

Relative Permeabilin of Petroleum Reservoirs


50 n-ct n-cl n-cl
o o o o)

40

r,u J o z o z o
< F z
F

30
n-C^ t'

n-c6 10

I 50 100
150

200

MOLECULAR WEIGHT

F I G U R E 1 3 . C o n t a c ta n g l ea s a f u n c t i o no f m o l e c u l a r e i g h t . r i w

83o -

Silica

Surface

fsooctane

fsooctane +

Isoqrinoline

Naphthenic

5.7E Isoquinoline

Calcite Sr-rrface FIGURE 14. Interfacialcontactangles.38

Benner and Bartell38 examined various multi-liquid systemsin contact with silica and calcite surfaces.Figure l4 illustratessomeof the findings of this study. It was reportedby theseinvestigators that when water and iso-octane used, the silica and calcite surfaces are are preferentially wet by water; but when water and naphthenicacid are used, water wet the silica but oil wet the calcite surface.The experimentof Benner and Bartell illustrated the effects of chemical as well as fluid compositionof phaseson wettability of a porous medium. Contact angles as low as 30o and as high as l58o were observedwhen various chemicalswere employedin the study. Salathiel3e discoveredthat the wettability of mineral surfacesmay be altered not only by adsorbedmonolayersof surface-active polar compounds,but also by much thicker layers of depositedorganic materials. Severalother workers have reportedthe formation of stable films on solid surfaces when the surfaces standin contactwith certaincrude oils. Reisberg and Doshelo described the deposition on glass or quartz surfaces of highly stable and appreciablythick films of strongly oil-wet material from Ventura crude oil. Early experimentersthought that all oil-bearing formations were strongly water-wet be-

i7 causean aqueousphasewas always the fluid initially in contactwith reservoirrock; furthermore, silica and carbonatesare normally water-wet in their clean state. Subsequent studies suggestedthat many oil reservoirsare not strongly water-wet and that the presence or agents,suchas asphaltic wax type material of crudeoils containingnaturalsurface-active by readily adsorbable solid-liquid interfaces,can render the solid surfaceoil-wet.arOther may cover a broad spectrum. studiesprovide evidencethat reservoirrock wetting preference One criticism of the idea of reservoirrock surfaces becomingmodified by the adsorption or depositionof polar organicmaterialfrom the oil phaseis that suchmaterialsshouldhave been eliminatedduring migration from the sourcerock to the reservoir.On the other hand, geochemists now finding substantial processes which affect are evidence variousalteration of crude oils subsequent their accumulationin reservoirs.[n a discussionof natural gas to deasphalting, Evanset al.a2 suggested reasonable a hypothesis that the more gas a crude has in solutionthe more of its heavyendshave come out of solution,plating out on the reservoir rock. It may be noted in this respectthat Salathiel'sstrongly oil-wet film depositionon quartzand porous rocks from a mixture of evacuated crude oil and heptanewas also probably process. the result of a deasphalting Despite uncertaintyas to the causesof reservoir wettability, much evidencehas been presented recentyearsto suggest in that many oil reservoirsare not stronglywater-wet.In particular,there are the many brine/crude contact-angle oil measurements Treiber et aI.62 of and the conclusionsof Salathielwith regard to the apparentwetting characteristics the of Woodbine reservoirin the East Texas Field. Nuttinga3 early as 1934indicatedthat some as reservoir rocks are oil-wet. Leach et al.aadescribeda reservoir believed to be oil-wet. Munganasstudied fresh carefully preservedcores from a reservoir and concluded that the formationwas oil-wet. Schmida6 shownthat stronglywater-wetcoresbecamelesswaterhas wet when equilibratedwith some crudes.Kusakov et al.a1 studiedthe thicknessof a water film left on a quartz surfaceunder crude oil drops and found that for two of the crudes, the film will rupture, bringing the crude oil into direct contactwith the quartz surface;the surface can then be describedas water-wet at some spots and oil-wet at others. Also, Craigas suggested that most formations are of intermediatewettability with no strong preferencefor to that water may not alwayscompletely eitheroil or water. There is recentevidence suggest wet reservoirrock in gas-waterflow following solvent injection. Soil scientists concerned with airlwater/soilsystemshave reportedsituations which thereis incompletewetting by in phase.ae the aqueous Authors such as Holbrook and Bernard,so and Fatt and Klikoffs' assumedthat wetting of rather than uniform. Holbrook and Bernard measured reservoir solids was heterogeneous fractional wettability by dye adsorption. Brown and Fatts2defined fractional wettability as the fraction of surface area in contact with water. This may not be a constantvalue since the water and oil saturations changeas a reservoiris produced.Schmida6 showedby means of capillary pressure-saturation data, that in preservedcores the fine pores were water-wet while the large pores were much less water-wet. This type of wetting is often referred to as "spotted", "dalmation", or "fractional". That heterogeneous wettability is a normal conby Iwankow,s3Brown and Fatt,s2 dition in oil sandshas also been suggested Salathiel,3e have suggested Gimaludinov,s4 and McGhee and Crocker.ss Severalof theseinvestigators that the wetting phasecompletely occupiesthe smaller pores of a reservoir rock in addition to the rock surface of the larger pores, while the nonwetting phaseprimarily occupies the insular regions of the larger pores. Evidence suggeststhat some oil reservoirs are partly preferentiallywater-wet and partly preferentiallyoil-wet. Such a condition could arise if someporesare lined with one type of mineral and other poresare lined with anothermineral. The existence of different minerals in porous media can create differences in surface chemistry of the grains, so all grain surfacesdo not have the sameaffinity for surfaceactive compounds. For instance, a tertiary sand reservoir in Alaska contains quartz and siderite minerals which are strongly water-wet and calcite which is strongly oil-wet. The overall

58

Relat iv e P ermeabi I i ty of Petro leum Re'rervoirs

necedi-ng l^later di spLaced by oi1 Water Static conditrlon displacing oil

|st
Advancmg

ti.

tubes. FIGURE 15. Advancing and recedingcontactanglesin capillary

siderite surfacesin rock system is water-wet, probably due to the presenceof quartz and of of presence anhydriteor gypsumin the flow channels some the main flow channels.The Thesemineralsare found to createa strongly watercarbonaterock may alter its wettability. rocks are probably oil-wet under reservoirconditions' wet system, while many carbonate to render a surface oil-wet when they are present in the Heavy metal sulfides are known flow channelsof Porousmedia' is covered by a Wagner and Leachs6stated that in some oil reservoirsthe rock surface would be preferentially bituminousor other organiccoating. Such surfaces firmly attached Boneau of of oil-wet in the presence oil and water, regardless oil and water composition. is due to reported that the oil-wet characterof the North Burbank reservoir and Clampitt.tt77o of the quartzsurface' a coating of chamositeclay which coversapproximately

OF VI. DETERMINATION WETTABILITY


qualitatively' or experimentally estimated The wettabilityof a rock can be eitherevaluated However' laboThere is no satisfactorymethod to determinein si/a reservoirwettability. in wettabilityhasbeenusedto evaluate situ wettability.Many of the widely ratory-measured methodsof wettability evaluationutilize either the reservoirrock or the used experimental shouldbe related wettabilityevaluation reservoirfluids, but not both. Therefore,a laboratory to actual reservoirconditionsusing a greatdeal of caution. A. Contact Angle Method has received the The contactangle methodis usedby a numberof laboratories; technique measurement' methodof wettability attentionin the literatureas a quantitative considerable on consistsof measuringthe contactangle 0 that a drop of pure liquid resting The method immersedin solid forms when nonporous,homogeneous a smooth, flat, incompressible, the the In most iractical situations, contactangleformedbetween solid surface anotherfluid. than a single and the water-oil interface is found to exhibit two limiting values rather into contactwith equilibrium value. The value of the contactangle when water is brought "advancingcontactangle"' oil on a solid surfacepreviouslyin contactwith oil is calledthe water on a solid surface The value of contact angle when oil is brought into contact with "receding contactangle". previouslyin contactwith water is called the in a capillary Figure l5 shows a comparisonof advancingand recedingcontactangles is referred to as tube. The fact that advancing and receding contact angles are not equal and roughness, and it is usually attributedto surfaceheterogeniety contactangle hysteresis movement. As the and rate of fluid materialsss as well as the presenceof surface-active providedthe the of roughness a rock increases, contactangle will further increase, surface however, if the contact angle measuredon the smooth surfaceof the rock is above 90o; roughness in surface is on contactanglemeasured a smoothsurface lessthan90o,the increase is found to increasein the would further decrease angle. The smoothsurfacecontactangle contact in and decrease receding,on the rough surfaceover most of the 0 to l80o advancing angle range.tn at materialsin the fluids may causeadsorptionprocesses the solid-fluid Surface-active even with a smooth, interfaceswhich give rise to appreciablecontact angle hysteresis contactangle hysline of contactincreases solid. Motion of thl three-phase homogeneous teresisas the rate of movementincreases'

59
It4ineral Flat ter plate

#1

/or,
Flat

Mineral

plate

#2

F I G U R E 1 6 . S c h e m a t r c e a s u r e m e n tf c o n t a c ta n g l e s . 5 o m o

't20
Oil-wet
o o o

. "
a

Equilibrium tContact Angle

3 8 0
!

o o,

8 E 4 0
o

Water-wet

20

40

60

Time (hours)

F I G U R E7 I

Influenceof aging on laboratory-measured contactangle.18

Advancingand recedingcontactanglescan be shownin a capillarytube for oil displacing water(receding angle)and waterdisplacing (advancing oil angle).The procedure determine to the contact angle using a contact angle cell is describedby Wagner and Leachs6 and is illustratedschematically Figure 16. Briefly, samples polishea,Rat platesof the mineral by of which is the main constituentof the reservoirrock are immersedin a sampleof formation water. A drop of reservoiroil is held betweenthe two flat samples the mineral of and the two plates are moved horizontally so that the water advances the surfaceof the plate on initially coveredby oil. The contactangleformedbetween interface the and the newly wateroccupied surfaceof the mineral is a measureof the water advancingcontact angle. The advancingcontactangle is the one that is customarilymeasured and often reportedwithout being identified as advancing. The contact angle measuredin the laboraotryis often influencedby aging. It has been shown that contactangle increases with age of the oil-solid interfaceuntil an Lquilibrium is reached.This may require severaldays and it is one of the disadvantages the of contact angle method.a8 Figure l7 shows this effect. Reliable wettability measurement requiresthat both the reservoirrock and the fluids be free from contaminants. Uncontaminated reservoirrocks can probably be obtainedif the coresare recovered with coring fluid containingno surface-active additivesor with reservoir oil that has not been exposedto oxygen. It has been reportedthat exposureof coresto air could result in alterationfrom water-wetto intermediate wettability. Uncontaminated reservoir water and oil are easierto obtain than unalteredreservoirrock. Since contactangle measurement be donewithout a sample (uncontaminated) can of reservoir rock, it hasbecome a widely used method for determiningwettability. Zismanmand other investigators studiedcontactanglesunder controlledconditionsand

60

Relative Permeabilin of Petroleum Reservoirs

expressed varying opinions concerningthe method's usefulness. Melrose and Brandner6r believed that the contact angles provides the only direct and clear specificationof the wettability property characteristic a given oil-water-rocksystem.Treiber et a|.62 of found that the water-advancing well with other wettabilityindicatorswhile contactanglescorrelate water-receding anglesdo not. Brown and Fatts2 questioned ability of the contactangle methodto provide a reliable the scale for determiningwettability and suggested that the conceptof a contact angle representation of wettability of reservoir rock be abandonedand that this method be replaced with a "fractional surfacearea" method. Morrow et al.63also observedthat severalfactors cast doubt on the utility of the contact angle method. Mungans describedsome of the limitations and pitfalls of contactangle measurement follows: as
l.

2.

3.

4.

5.

6.
7.

The mineral chosenfor the contactangle measurement the principal constituentof is the reservoirrock. For the purposeof contactangle measurement, silica or quartz is usedto represent sandstone; a calciteis usedto represent carbonate reef reservoir. a or Laboratorymeasurement contactangle or mineral surfacesmay not simulatetrue of reservoircontactangle. The contact angle at the water/oil displacement front is "advancing" while at the "receding". Thesevalues leading edge of the oil bank it is sometimesdiffer by as much as 50o. This variationcan be on the sameorder of magnitudeas the laboratorymeasured contactangle. Contactangle measurement shouldbe done when the solid surfaceand a fluid remain in contact for an adequate time before the secondfluid is introducedover the surface. This is referredto as pre-equilibriumtime and it is of different length for each crude oil-water system.Without adequate pre-equilibrium time, a stablecontactangle is not reached. somecases hasbeenreported In it that a stablecontactangleis neverobtained if the solid surfacecomes into contactwith some types of crude oils. Contact angle measurement frequentlytime consuming. is Contact angle measurement should be performedwith actual reservoirfluids, since they are in equilibrium and solubility effectsare negligible;otherwise,the fluids must be equilibratedwith one anotherso that the solubility effectsbecomenegligible. Contactanglemeasurement preferablyshouldbe donewith bottom-hole fluid samples; however,because the time and expenses of involved, flow line samples often used. are Fluid samplestaken from the storageor treating facilities are not reliable, due to the possibleaccumulation asphaltenes. of When produced water is not available,synthetic brine is commonly used. Contact angle measurements should be made under controlledconditionsso that the oxidation of crude oil can be prevented. Contact angle measurement requiresextremecare to assure cleanliness and inertness of the apparatus.

B. Imbibition Method An imbibition test is a reliabletechniqueof wettability determination providedunaltered reservoir fluids are available. The method consistsof the measurement rate of flow of a of wetting fluid spontaneously imbibed into a core and replacinga nonwettingfluid by the action of capillary forces alone. Imbibition testsmay be performedat standard conditions at reservoir or conditions.Figures 18, 19, and 20 illustrateequipment that is usedfor conducting testsat ambientconditions. the The imbibition test at standard conditionsmay be performedas follows: l. A cylindrical plug of reservoirrock I to I tlrin. in diameteris cut with water as a coolantin the cuttingprocess.

6l

Capillary Tube

M e t a l l i cS a m p l e H o l d e r
Tef lon

F I G U R E 1 8 . I m b i b i t i o nc e l l .

2. 3. 4. 5.

6. 7.

The sample is placed under water in a beaker and evacuatedto remove trapped gas. The sampleis flushed with water to reducethe oil saturation residuallevel. to The core plug is placedin an imbibition cell underoil and oil imbibition is monitored. The drained water is measured; is equal to the amount of imbibed oil. Sufficient it time should be allowed for the systemto reachequilibrium; this may take severaldays dependingon the permeability of the plug. The plug is then saturatedwith oil to reduce the remaining water to the ineducible level. The sampleis placedin an imbibition cell underwater and water imbibition is monitored by the amount of oil being drained. The fluid that imbibes into the sample (oil or water) is the wetting phase.

The imbibition test under reservoir conditions is more complex. Irreducible water saturation is establishedby flushing the core with live oil and the imbibition tests are made at reservoir pressureand temperature.

62

Relative Permeabilin of Petroleum Reservoirs

to Water Reservoir
)7

F I G U R E 1 9 . I m b i b i t i o nc e l l

Accumulated

Rubber Stopper -----+

FIGURE 20.

Imbibition cell.

Amott6s developed a quantitative techniquefor defining the degreeof water-wetnessof the by cores. He expressed degreeof water wetness a water index, which he defined as the imbibed into a core to the total volume of oil ratio of the volume of water spontaneously of displacedby a water drive (forced displacement oil by water). Similarly, an oil index was defined at the ratio of the volume of oil spontaneously imbibed to total water displaced

63 by an oil drive (forceddisplacement water by oil). Amott's test consists the following of of steps:

l 2. 3. 4. 5. 6.

Flush the reservoirsamplewith water to reducethe oil saturation its residuallevel. to Immersethe samplein water and evacuate remove gas. to Immersethe samplein kerosene(or reservoiroil) and measurethe volume of water displacedby imbibition of oil after 20 hr. Measure the volume of water displacedwhen the sampleis centrifugedunder oil. Immersethe samplein water and measure volume of oil displacedby water after the 20 hr. Measure the volume of oil displacedwhen the sampleis centrifugedunder water.

Oil index is the ratio of the volume of fluid measuredin step 3 to the volume of fluid measuredin step 4. Water index is the ratio of fluid volume from step 5 to fluid volume from step 6. The preferential wettabilityof a rock is determined the magnitude thesetwo indexes, by of i.e., strong wettability is indicatedby values approaching one and a weak preferencein indicatedby valuesapproaching zero. A water index of one indicatesa strongly water-wet surfacewhile an oil index of one indicates stronglyoil-wet surface.Valuesbetweenthese a two extremes a value nearzero for both ratioscover the rangeof intermediate or wettability. Amott's testof wettabilityof porousmediareceived high marksfrom Razaet aI.66, although Moore and Slobad,67 Bobek et aI.,68 Killens et al.,6e and RichardsonTo indicated have that the imbibition rate cannot be entirely attributedto the wettability of the core, but that it is also influencedby rock porosity, permeability,pore structure,and pore size distribution,as well as viscosityand interfacialtensionof the fluids involved in the experiment.Donaldson et al.7' tried to eliminateextraneous effectsfrom the wettabilitymeasurement comparing by the volumes of fluids imbibed into preservedreservoircores with the volumes of fluids imbibed in the same cores after extractionand resaturation. Although the use of the same core would appearto offer identicalpore size distributions, changein fluid distributions the gained. causedby the cleaningprocessmay have offset the advantage MunganT2 reportedthe use of an imbibition test to evaluatethe wettability of native-state cores. Emery et a1.73 used an imbibition test after incubationof cores for up to 1,000 hr with gas-saturated under pressure; oil water was the first phaseto contactthe rock in the test. Kyte et al.7a imbibition testsconducted reservoirtemperature pressure. described at and C. Bureau of Mines Method The U.S. Bureauof Mines methodof wettabilitydetermination a porousrock, commonly of referredto as the "Centrifuge Method", is basedon the assumption that an elementalarea of the internal surfaceof the porous medium is either wettableor nonwettableby one of the fluids involved. The problemis one of determining fractionof the internalsurfacewetted the by each fluid. A methodof measuring wettabilitybasedon the abovetheory was suggested by Gatenby and MarsdenTs and was later developedby Donaldson.Tr These investigators made use of the areasobtained from the drainageand imbibition cycles of the capillary pressure curve to producea numericalrepresentation wettability. The Bureau of Mines of method is quite rapid and it can be employedwith reservoirfluids. D. Capillarimetric Method Johansenand DunningT6 recognizedthe importanceof the liquid used in determining wettability of a rock-liquid-brinesystemand suggested use of a capillarimeterwhich the joins the two liquid phases,oil and water, through a small diameterglasscapillary tube, with a capillary pressureacrossthe interfacejoining the two phases.Adhesion tensionor

64

Relative Permeabilin of Petroleum Reservoirs

energy, was calculatedfrom the differencein height of the two liquids in the displacement in and the acceleration to gravity. due the two armsof the capillarimeter, difference densities, forceswith eitheran advancing receding or measuring interfacial is of The instrument capable Major limitations of this method are the exclusionof reservoirrock as a factor interface. influencingwettability and lack of provision to preventoil from oxidizing. E. Fractional Surface Area Method This method, developedby Brown and Fatt,s2 usesmixturesof untreated sandand sand rendered vaporsto obtainwettingconditions rangingfrom completely oil-wet by organosilane water-wetto completelyoil-wet. Wettability is represented the fraction of solid surfacemade artificially oil-wet. Alby though use of the method to evaluatefield behavioris not in evidence,the conceptof a fractionally wet surfacehas been presented the work of other writers.3e in F. Dye Adsorption Method This method,developed Holbrook and Bernard,-50basedupon the ability of reservoir is by rock to adsorba dye suchas methylene blue from aqueous solution,while rock surfaceareas covered by contaminantsfrom the oil phase remain unaffected.The test is based on a comparisonof the adsorptioncapacity of the test samplewith that of an adjacentsample similar to thoseof extractedby chloroform and methanol.This methodmakesassumptions Brown and Fatts2in their "fractional surface area" method. G. Drop Test Method This method is often used to confirm rock wettability. The procedureinvolves placing drops of oil and water on the surfaceof a fresh break in the core. The fluid that imbibes is the wetting phasewhile the fluid that forms a ball and doesnot wet the surfaceis nonwetting. misleading. The drop test is a qualitativedetermination and is sometimes H. Methods of Bobek et al. proposeda laboratorytestto ascertain preferentialwettability in a qualitative Bobek et aI.68 fashion. The techniqueconsistsof determiningwhich fluid will displacethe other from a rock sampleby imbibition. The resultsof this imbibition test are comparedwith thoseof a referenceimbibition test on the samecore sampleafter it has been heatedto 400'F for 24 hr to remove any organic materialsand to make it more water-wet. The assignmentof qualitativewettability designations basedon the relativeamountsand ratesof imbibition is in the two tests. material is In the same paper a method for estimatingthe wettability of unconsolidated slide. The oil discussed. thin layer of the unconsolidated A sandis spreadon a microscope content of the sand is increased adding a clear refined oil. Droplets of water are then by placedon the surfaceof the sandgrains and the fluid movementis observed.If the sand is water-wet, the added water will displace oil from the surfacesof the sand grains and the oil will form spherical droplets, indicating that oil is the nonwetting phase. A similar procedureis used to test for oil wettability. I. Magnetic Relaxation Method for A nuclearmagneticrelaxationtechniquewas suggesteds2 determiningthe portionsof the rock surfacearea that are preferentiallywater-wetor oil-wet. A rock sample is first exposedto a strong magneticfield, then to a much weakerfield. The magneticrelaxation rate- that is, the rate at which the initially imposedmagnetism lost - is then measured. is In sandpacks containing known mixtures of oil-wet and water-wet sand grains, a linear relationship was observedbetween the relaxation rate and the fraction of the surface area

65

0.5 Fraction of Dri-Filmed Sano

.t

FIGURE 21.

for Interstitialwater saturation sand mixtures.sr

that is oil-wet. Though the authorsreportedno studiesusing naturalcores, they proposed a testingprocedure.Their techniquerequiresspecialized equipmentnot normally found in petroleumlaboratories thereare no indications the literature in and that the methodhasfound routine use. J. Residual Saturation Methods reported correlation between residual McGheeet al.,ssLorenz et al.7e and Rezniket al.80 a oil saturationand wettability. Treiber et al.62 reportedthat the connatewater saturationin be They a native core can sometimes usedas an indicationof formationwetting preference. found that oil-wet formation have much lower connatewater saturations than the water-wet in ones. In addition, the connatewater saturation a stronglyoil-wet reservoirwas found to permeability,while in reservoirs other wettabilities of of be constantregardless the sample with increasein permeability.Iwankow53 the connatewater saturationdecreased also desandwettability in terms of a fraction of drifilmed sand. scribedthe effect of heterogenous (See Figure 21.) Drifilm is a solution commonly used in the laboratoryto make sands preferentiallyoil-wet. Coley et al.8r were not successsful using the ratio of the wetting in relationships to the nonwettingresidual saturationfrom relative permeability-saturation as preferentialwettability indicator; however, they found that the volume of mobile a rock valuesof a relativepermeability fluid shown by the spreadbetweenthe residualsaturation curve appears decrease the oil wettability increases. to as K. Permeability Method of The determination wettability of a samplefrom permeabilitydata is accomplished by with the oil permeability comparingthe ratio of water permeabilityat residualoil saturation at connatewater saturation.If this ratio is less than 0.3, the sample is consideredto be water-wet,while a value near unity indicatesthat the sampleis oil-wet.82 The relationship permeabilityand connatewater saturation has been frequentlymentioned betweenabsolute in the petroleum literature and the relationshipbetweenconnatewater saturationand rock wettability has been discussed. Rocks with low connatewater saturation considered are to water saturation normally water-wetto oil-wet, while rocks with high connate be weakly are as designated water-wet.

Relative Permeability of P etroleum Reservoirs

(Water-wet)

ft rel

Sw

(Oit-wet)

krel

Sw
FIGURE 22. Schematicwettability effecrs on relative permeabilitycurves.

L. Connate Water-Permeability Method A correlationof absolutepermeabilityas a function of water saturation corescut with in oil-base mud hasbeenusedfor qualitative identification corewettability.6s of Watersaturation is measured freshly cut coresand absolute in permeabilityis determined after extractionand drying. A plot of water saturation a function of absolute as permeabilityto air is prepared. The curve will have a gentle slope over a large saturation interval for water-wetsystems, while it will exhibit a nearlyverticalslopeover a narrowsaturation rangefor oil-wet systems. This technique applicable is primarily to thick hydrocarbon reservoirs with sufficientvariation in permeability and water saturationso the required plot can be prepared. M. Relative Permeability Method For a given water saturation, water relativepermeabilityof a water-wetrock is lower the than that of a comparable oil-wet rock. For the systemsstudiedby Owens and Archern it was found that an increase oil wetness constant (at in water saturation) producedan increase in k,* and a decrease k,.,. Treiber et aI.62 in concludedthat water-wetconsolidated porous media normally have a water relative permeabilitylessthan l5Vo at residualoil saturation, while oil-wet porous media show a 50Voor higher relative permeabilityto water at floodout. Craigas offers the following heuristicguidelines,which are illustratedby Figure 22:

67
Water-wet
S*i
. k,* -- t[r.* k,* at S.,*

Oil-wet < l 1% a,usual l yl 07o @ S*<507o > 0.5, approachi ng 1.0

>20 to 25Va @ S*>507o <0.3

block the easy flow In a water-wetrock, residualoil globulesin the large flow channels of water and causea low water relativepermeability;however,the oil in an oil-wet system occupiessmaller flow channelsand coatsthe walls of the largerones, causinga minimum disturbance water flow and a higher water relativepermeability.ttThis is why an oil-wet to rapid increase water cut, reservoirwill waterfloodpoorly, with early water breakthrough, in and high residualoil saturation. cores under steady-state The water-oil relative permeabilityrelationshipof native-state pointed Keelan82 conditionsis one of the best indicatorsof the rock wettability preference. changeaccompanied out that a sharpdrop in oil relativepermeabilityover a small saturation by a rapid rise in relative permeabilityto water, to a terminal value in excessof one third oil Careful sampleexamination the initial oil relativepermeability,often indicates wetness. or is essentialin using this technique,for heterogeneous cracked samplesyield relative permeabilitydata similar to the data obtainedfrom oil-wet cores. with the Water relative permeabilitycurves in water-oil systemsshow good agreement testsin a strongly oil relativepermeabilitycurve obtainedduring gas-oilrelativepermeability water-wet core.62'63'84 effect does not exist under any other wetting condition. In a This strongly water-wet core, the water relative permeability curve of a water-oil system also shows good agreementwith the water relative permeabilityof a gas-watersystem in the presence residualoil saturation. This agreement will occur, even though the directionof of the changein saturationmay not be the samein the two systems.In the same manner, in systemis comparable to stronglyoil-wet cores, the gas relativepermeabilityof a gas-water of systemin the presence residualoil saturation.8a the gasrelativepermeabilityof a gas-water The point of intersection of the water and oil relative permeability curves has been suggested an indication of rock wettability. Owens and Archerrr have shown that the as relative permeability intersectionpoint moves toward higher values of water saturationand lower values of relative permeability in a water-oil system as the sample wettability is changedfrom oil-wet to water-wet. As illustratedby Figure 22, a relative permeability intersectionpoint on the left of 507o water saturationindicatesthat the system is oil-wet, while an intersectionto the right of this saturationsuggests that the system is water-wet. N. Relative Permeability Summation Method The summation of relative permeabilitiesto the water and oil phaseat fixed saturations also gives some insight into the immiscible flow processes.McCafferysenoted a trend in of the minimum values of the sum of relative permeabilities samplesaccordingto their preferential wettabilities. O. Relative Permeability Ratio Method If the ratio of displacing to displacedphaserelative permeability is plotted as a function of the displacing-phase saturation,the shapeof the plot is related to preferentialwettability rock.66It has been shown that the water-oil relative permeability ratio shifts to a of the higher value as the rock becomesmore oil-wet; furthermore, a semilog plot of water-oil and gas-oil relative permeability indicatesthat the gas-oil relative permeabilityratio curve moves from under to over the water-oil relative permeability ratio curve as the rock becomes preferentially water-wet.sr The water-oil relative permeability ratio curves of rock with variousdegrees intermediatewettability are found to be practicallythe samein the presence of

68

Relative Permeabilin of Petroleum Reservoirs

of constantinitial water saturation.85 Imbibition water-oil relativepermeabilityratio curves in the absence initial water saturations of show higher valuesof residualoil saturationas the cores become more oil-wet.8sSteady-state relative permeabilitymeasurements should be usedfor determination wettability. Unsteady-state of methodsmay not allow equilibrium to occur during the flow test; therefore,they may indicatemore oil wettnessthan actually exists. P. Waterflood Method Severalattemptsto find a single correlationof wettability with waterfloodoil recovery for different porous media have failed, even though the tests were carried out under a standardset of conditions.6s However, the waterflood performance a native-state of core under carefully controlled laboratory conditions has been used as an indication of rock preferentialwettability. It is found that in a strongly water-wetsystem,a large fraction of the oil is producedprior to water breakthrough and very little additionaloil is recovered after breakthrough. For the test to be reliable,an equilibriumwetting conditionmust prevail prior to the passage the flood front through the core. of Q. Capillary Pressure Method Both displacement pressure and the ratio of drainage imbibition displacement to pressure have beenproposedas qualitativeindicatorsof preferential wettabilityof porousmedia. An increasein displacementpressureor in the ratio of drainageto imbibition displacement pressuresignifies a tendencyof the core to becomemore oil-wet. The above techniqueis applicablewhen oil-water capillary tests are made on native-state cores. However, most capillary pressure testsare either of the mercury injection or air-brinetype, which provide little information concerningwettability.8l R. Resistivity Index Method Formation resistivity obtainedfrom electric logs can be used as a qualitativetechnique for wettability identification. Resistivity index is defined as the ratio of true formation resistivityto resistivityof the formation when 1007a saturated with formationwater. A high value of resistivity index indicatesa low water saturation a discontinuous or water phase, which characterize oil-wet system.A knowledgeof the water saturation the rock may an in yield sufficient information to make a judgementabout rock wettability. There is considerable uncertaintyconcerningthe natureof the wettability characteristics of reservoir rocks in situ. Tests of wettability made on cores taken from reservoirsare not necessarily valid indicatorsof subsurface conditions, since the coring processitself may alter wettability. Cores cut in oil-base mud, for example, are often renderedentirely or partially preferentiallyoil-wet. Thereforespecialprecautions must be observed during both coring and transportingto minimize the danger of altering the true wettability of the rock. In the absence convincingevidenceto the contrary(for example,abnormallyhigh resisof tivity index) the assumption preferentialwater wettability has been frequentlyused.86 of

VII. FACTORS INFLUENCING WETTABILITYEVALUATION


It has been suggestedthat four factors may influence the results of experimental determination of rock wettability.8T One of thesefactors is core recoveryand preservation. In the processof core recovery from a reservoir, heavy hydrocarboncomponentsof crude oil become less soluble as the oil loses its associated solution gas (as a result of pressure reduction).The heavy hydrocarboncomponents can precipitate the rock grains, leading on to less water-wetor even oil-wet core behavior.8s-m Drilling fluid containingsurface-active materialsmay drasticallychangea core wettability, but it has been shown that bentonite

69 and carboxymethylcellulosehave no observable effect on rock wettability when they are used in the coring fluid.7aWeatheringand contamination coresduring preservation of and storage alsofound to influencecore wettabilities.er are Stronglywater-wetcoresmay become less water-wetas a result of air exposure,while cores with intermediate wettability show no significantchange.6s Oil-wet cores also may becomewater-wetupon exposureto air.72 It has been suggested that alterationdue to air exposurecan be minimized and native-state wettabilitycan be restored incubation the core in reservoir for two weeksat reservoir by of oil temperature.2s Crude oil is probably the best coring fluid for preservingwettability and maintaining native interstitialwater saturation;e2 however,useof the wetting phaseas a coring fluid may preserve rock properties the properly.2s NaCl brine containingCaCO, powder with no other additivesis considered good fluid for cutting cores.e3 a Care must be taken to avoid contaminationof the coring fluid with air, sediments, etc. The useof crudeoil as a coring fluid is likely to introducea fire hazardinto the coring operation,especiallyif a high API gravity oil is us ed. Native state wettability of cores is obviously the most desirablecondition, and the best techniquefor obtainingcores in this condition is by employing a pressure core barrel. The method allows coresto be cut and retrievedat reservoirpressure. the surface,the cores At are frozen, cut into sections, and sentto the laboratory.ea Although early attempts pressure at coring met with limited success, recentdevelopments indicate success a ratio of 80 to90Va. Cores that have been cleaned, dried, and restoredto some saturationand wettability condition are known as "restored state" cores.a8 This techniquehas been employed for many yearsand it is an established procedure; unfortunately, quite frequently,the coresare not restoredto their native stateand the useof thesecoresinvalidates resultsobtainedusing sophisticated measurement techniques. Put very simply, restoredstatecores are not. Factorsthat influencethe core wettability evaluationincludethe laboratorycore cleaning and preparationprocedure.Mungane2 statesthat the cleaning procedureneither changesthe pore size distributionnor the quantity of kaolinite and illite in the core. He concludesthat the changein fluid flow behavioris basicallydue to wettabilityalteration.Salathiel3e reasons that the extractionof a core with strongsolvents dissolves stronglyoil-wet surface the coating of heavy organic moleculesand therebyaltersfluid displacement behaviorof many fresh or preserved cores, as shown in Figure 23.28 Jenning'se5 resultsshow a small but measurable changein the water-oil relative permeability ratio curve after toluene extractionof a variety of core samplesfrom oil-bearing sandstones limestones. and The changes not thoughtto be caused significantchanges are by in wettability. The resultsof Richardson al.ershow a higherrateof imbibition and a lower et ineducible water saturation when East Texas Woodbinecoresare extractedby hexaneand methanol.Morgan and Gordon's28 resultsshow that the effect of cleaningprocedure core on wettability may be minimized if reservoirfluids are used as testing fluids. Richardsonet al.et believe a changein fluid flow behavioroccursas a result of repeated flooding of East Texas Woodbine cores. This change appearsas a decreasein irreducible water saturation and as an increasein residualoil saturation.Furtherwork is necessary betterunderstanding for of this problem. A third categoryof factorsthat influencecorewettabilityevaluation the testingcondition. is Stainless steel wettability can be alteredby pressure increase a methane-water in system.e6 In spite of decrease interfacialforces, the oil-water-solidsystembecamemore water-wet in with temperatureincreasesin a clean unconsolidated Houston sand and a natural unconsolidated California oil sand.eT One explanationfor the effect of temperatureon displacement behavior is that polar componentsof the crude oil may not be adsorbedas readily on the grain surfacesof a rock at elevatedtemperature,so the flow behavior becomesmore waterrA'g1.7+'9a

Relative Permeabilitv of Petroleum Reservoirs

2 air p e r m e a b i f i t y : - 2 9 m d . o Fresh A Extracted

kto

6
j

0'5

kr*

-:i -'

sw
il:|':?"il,,'t:Hiil:iJT11',:' permeabi'itv the data rrom same

A fourth category of factors that influence the core wettability evaluation is the type of surfactant compounds fluid used in the test. Carbonates very sensitiveto nitrogeneous are of containinglarge percentages silica possess containing sulfur and oxygen.arSandstones acid type surfaces.38'er Crude oil containingnormal paraffins are inert and inactive with regard to the surfacesof porous media, while naptheneand aromaticsare more active with containingoxygen, nitrogen, sulfur, and porous surfaces.Heterocyclicsand asphaltenes and Doschelo have the acid or basicsites.Reisberg metallic atomsare active with regardto probably have different proportions of these compounds indicated that different crude oils of for which are believedto be responsible the wettability characteristics surfaces. with increasing concentragas saturation and decreases that for oil increases The critical the of polar substances.ee Furthermore,increasing concentration polar compounds tions of and cumulativeoil productionto in oil causesthe cumulativewater productionto increase in laboratorytests. decrease to Oxidation of crude oil frequentlyappears modify the wettabilityof porousmedia. The of modification dependson the amount of oxidizablepolar compoundsin contact degree have and Cuiece8 Morgan and Gordon28 with air and wettability may even be reversed.e8 fluids and laboratoryhandlingon relativepermeability.Mungane2 investigated effect of the saturatedan extracted core with reservoir fluid and let it sit at reservoir temperaturefor 6 relative permeabilityvalueswere identicalto those days. He discoveredthat the measured preserved he used purified fluids in place of reservoirfluids a of freshly cores; but when more water-wetcondition in the core was developed,as indicatedin Figure 24. water alkalinityand hardness,ee in The initial fluid saturation a core,s salinityalteration,eo preferential wettability of a core. Wagner can influencethe as well as the aging processe' wet sample of have shown that the wettability of an oil- or intermediately and Leach-'6 or sandstone carbonatecan be changedto a more water-wetcondition by the addition of chemicalssuch as hydrochloricacid, sodium hydroxide,and sodium chloride. They inves-

7l

0.9

Sw
FIGURE 24.

o.7

Effect of fluid and laboratoryhandlingon relativepermeability."l

tigatedthe influenceof water pH on wettability of a quartzsampleand useda n-octylamine treatedsyntheticoil to producean oil-wet quartz surface.Their resultsindicatedthat lower pH solutionstend to producewater-wetsurfaces under controlledsalinity conditions.This effect is shown in Figure 25. Bradleyroo shown that a basic 57oNaCl solutionspontaneously has decreases contact the angleof oil-wet coresand as a resultincreases amountof imbibition. Theseeffectswere the reported be most pronounced coresof intermediate to on wettability.Morrow et al. ,63 Wagner and Leach,s6 and McCaffery and Munganror have shownthat wettabilityof typical reservoir rocks can be easily changedto any desireddegreeby adding polar compoundssuch as aminesor carboxylic acids. Bradleyrmfound that carboxylicacidssuch as stearicacid CH., (CHr)16 COOH at concentrations greaterthan 10-6 moll( alteredthe wettabilityof a waterdodecane-calcite systemtoward more oil-wetnessand stearicacid with a concentration of approximately x l0-3 mol/f causedstronglyoil-wet surfaces. found that stearicacid 5 He causedno wettability alterationwhen quarlzsamples were used. Bradley found that amines such as octadecylamine CH. (CHr),, NH, alter the wettability of both quartzand calcite toward oil wetness,especiallyat concentrations greaterthan 5 x l0-a mol/{. It should be noted that polar compoundswhich alter wettability of a given rock type may not alter the wettability of anotherrock type.

72

Relative Permeabilin of Petroleum Reservoirs

o a c

O.O 25,000 50,OOO

ppm NaCl WATER PHASE

a C) O o

r30

uJ J

OIL_WET

z o
F

z
(J

90

z z o (r
uJ F
I

70

50

WATE -WET R

6 WATER_PHASE PH

FIGURE 25.

Contactangle as a function of pH.so

VIII. WETTABILITY INFLUENCE FLOW ON MULTIPHASE


The microscopicdistribution of fluids in a porous medium is greatly influencedby the degree of rock preferentialwettability. The fluid distribution in virgin reservoirsunder strongly water-wet and strongly oil-wet conditionshas been describedby Pirson.ro2 a In pores,in smallcapillaries, stronglywater-wetreservoir,most of the waterresides dead-end in and on the grain surface.In strongly oil-wet reservoirs,water is in the centerof the large poresas discontinuous droplets,while oil coatsthe surfaces the grains and occupiesthe of sm aller apillar i e s . c Under strongly water-wetconditionsthe effective permeabilityto the nonwettingphase at irreduciblewater saturationis approximatelyequal to the absolutepermeabilityof the rock. On the other hand, in strongly oil-wet systems,the effective permeabilityto oil at irreduciblewater saturation greatlyreduced the waterdropletsin the largerpores.Raza is by et aI.66 statedthat in someoil-wet reservoirs, water occupiessomeof the finer poresand is trapped as droplets in the larger ones. Raza et al. analyzedthe displacementof oil by advancingwater and the trappingof the residualoil as shown in Figure 26. In strongly water-wetreservoirs,water traps oil in the larger poresas it advances along the walls of the pore, while in strongly oil-wet reservoirs,water moves in large pores and oil is trappedclose to the walls of the pores.66 The petroleumindustry has long recognized that the wettability of reservoirrock has an important effect on the multiphaseflow of oil, water, and gas through the reservoir.API Project 27 at the University of Michigan was initiated in l92l to study this problem. The

73

Oil-Wet Sand

FIGURE 26.

The trappingprocessof oil by advancingwater'n"

---

TEST

1 waler

wet

TEST 2

weter

wel

... ... TEST 3

oil

wet

\\ \\ \
E 5 0
j

oi1

Brine

25

50 B R I N ES A T U R A T I O N

75

100

FIGURE 27.

Effect of wettability on flow behavior''r

dissymmetryof relative permeabilitycurvesis attributedlargely to the preferentialwettability As of reservoir rock.te'es'ro3 illustratedby Figure 27, Geffen et al.r2 and Donaldsonand Thomas'oahave shown the effect of fluid distributions brought about by rock preferential relationship.As the degreeof rock prefwettability on the relative permeability-saturation the erential wettability for waier decreases, oil relative permeabilityat a given saturation while the water relative permeabilityincqeases. decreases recognizedthe fact that rock type appearsto have less influence Schneiderand Owenssa However, this may not be the case than doesrock wetting preference. on flow relationships rocks or mixed wettability systems.Owens and Archerrr also confirmed for heterogeneous the importance of preferential wettability on multiphaseflow in porous media. have found that relative permeabilitybecomesprogressivelyless Some investigatorsno favorable to oil production as a rock becomesless water-wet. The residual oil saturation as increases a rock becomeslesswater-wet.Othershave shownthat weakly water-wetcores more favorable relative permeability curves and lower residual oil saturationsthan have since strongly water- or oil-wet rocks. Conceptually,this latter behaviorseemsreasonable

74

Relative Permeabilitv of Petroleum Reservoirs

S P L A C E OP H A S E S

/l

4 t/

Ni troqen Dodecane llitrogen Dioctyl HepLile

displacrng or bio.tyl displacing Ether

Heptile, eti... Water Nrtrogen

up

to l08o I3Io

49c

displacing Nitroqen

displacing

above

l38c

.8

Displaced

phase

Saturation.

pV

rt il:,yrT..

Relative permeability fluidpairs for withvarious contacr

the capillary forces in strongly water-wetcores are strong. The oil may be bypassed and trappedin larger poresby the tendencyof a water-wetcore to imbibe water into the smaller capillaries.The bypassed in the large poresis then surrounded water and is immobile oil by exceptat very high pressure gradients. The saturation intervalfor two-phase flow underthis condition is probably short. As the capillaryforcesare reducedby reductionin preferential water-wettability a rock, of the tendencytoward rapid imbibitional trappingof oil in large poresby movementof water throughsmallporesshouldalsodiminish. The zoneof two-phase flow shouldbecome broader and oil displacement a lower residualsaturation to shouldbe possible.If other factorsremain constant, higherflow ratesand lower interfacialtensions conducive higheroil recovery; are to theseare changesthat diminish the ratio of capillary forcesto viscousforces. Stegemeier Jensen3T McCafferyand Bennionr05 and and reported that wettabilityalterations over a relatively wide rangeproducea negligibleeffect on the relativepermeabilitycurve, as shown by Figure 28. However, other workers did not confirm this finding. Treiber et aI.62 found that relatively small variationsin wettabilityproduceconsiderable effectson the relative permeability curve. Figure 29 shows the effect of contact angleson relative perrneability curves for a Torpedo sandstone.

IX. EFFECTS SATURATION OF HISTORY


The relative perrneability-saturation relation is not a unique function of saturationfor a given core, but is subjectto hysteresis porous systemswith strong wetting properties. for

75 100

"Nt
\ \ 10

\ .

'/

ATER

I..

o
l<

Contact

Ancrle ^o nro

. . . . . . 9 0 o . -^o " ^^o

.1

20

40

60

80

100

sw
FIGURE 29. permeability Imbibitionrelative with various contact angles."l

depends That is, the relativepermeabilityof a porousmediumto a fluid at a given saturation it on whetherthat saturation obtainedby approaching from a higher value or a lower one. is processwhere the wetting-phase saturationis approached from a lower In a displacement value, the resulting relative permeabilitycurve is referred to as an imbibition curve (an are increase the wetting phase).Examplesof imbibition processes the injection of water in mud. On the otherhand, during waterfloodingand coring a water-wetrock with a water-base processwhere the wetting phasesaturation approached is from a higher in a displacement value, the resultingrelative permeabilitycurve is referredto as a drainagecurve. Examples of during primary depletionof of drainageprocesses the displacement oil by expansion are in of anotherexample a reservoirand the accumulation hydrocarbons oil and gasreservoirs; would be waterfloodingan oil-wet reservoir. Terwilligeret &1. andColey Josendal al.,r07 et Geffenet al. ,r2Osobaet al.,r3Levine,'oo ,'n' phenomenonand verified that both water-oil and gas-oil et al.8' describedthe hysteresis relativepermeabilityratio curvesas well as individual wetting and nonwettingphaie relative may exhibit hysteresis.rr'22'ro7'roe permeability both sandstone carbonate formations of and is to In a two-phase system,hysteresis moreprominentin relativepermeability the nonwetting ro' in phasethan in relativepermeabilityto the wetting phase. I roThe hysteresis wetting-phase difficult to distinguish relativepermeabilityis believedto be very small and thus, sometimes from normal experimental error, as indicatedin Figure 30. The drainagecurve shown in Figure 30 is a primary drainagecurve which is applicable only when drainageoccurs before imbibition. When a drainageprocessoccurs after imbidrainagecurve exists, as shown in Figure 31. bition, a secondary

Relative Permeabilin of Petroleum Reservoirs

D r a in a g e lmbibition

Sw (Water-Wet System) FIGURE 30. Primary drainage permeability relative curve

Water

e DW (water-wet system)
,t :t":rYXt Thesecurves describerelative permeabilitywhen the flow reversaloccursat one of the saturation end points. The effect of flow reversal at an intermediate saturation value is illustratedby Figure 32. As shown in Figures30 and 31, the water (wetting phase)relativepermeabilitycurve is essentially samein stronglywater-wetrock for both drainage the and imbibition processes.rr However, at a given saturation, nonwettingphaserelativepermeability a consolidated the of rock is usually less for an imbibition cycle than for a drainagecycle.t2.t3.22.to6 an For unconsolidated rock, the nonwetting phaserelative permeabilityin an imbibtion cycle is usually greaterthan the corespondingnonwettingphaserelativepermeabilityin a drainage cycle. Naar et aL.22 reportedthat relativepermeabilityrelationships poorly consolidated for formations tend to resemblethose for unconsolidated formations. Figure 33 shows the imbibition and drainagerelative permeabilities a consolidated of rock. It can be seen that the residual nonwetting phase saturationis much greater for imbibition than for drainage.That is, the nonwettingphaseloses its mobility at a higher saturation imbibition than it does in drainage.Figure 34 showsthat the imbibition cycle in k.o may lie abovek.. on the drainagecycle for some systems. This relationship probably is not typical of petroleumreservoirs. Secondary drainage curve: end-point flow

77

Secondary drainage

o
.Y

Sw (water-wet system)
FIGURE 32. reversal. Secondarydrainase curve: intermediateflow

160

.=
-o
$
L

140

E
o o. o

120

100
o

.9 =

o o oo

80

be o v

60 Water 40

20 Water 0 0 2 0 4 0 6 0 80 100

Brine saturatiofi, o/o


FIGURE 33. rock.12 Oil-water flow characteristics a consolidated of

Relative Permeabilin of Petroleum Reservoirs


1.O 1.O
--Drainage lmbibition

o . 5
l<

X*'n
\ --+---r --,^,'
.5

9.5

ta*- *rn .--Y \ t.',

/*-) r o

l(

ro
a

'

1.0

.5

1.0

so
Consolidated Sand

so
Unconsolidated Glass Spheres

FIGURE 34.

Relativepermeabilitycurves for consolidated sandsand unconsolidated glassspheres.rr

The amountof trappedoil in water-wetporousmedia is given approximately the area by betweenthe drainageand imbibition oil relative permeabilitycurves.rr2 is believedthat It the occurrence hysteresis possiblyrelatedto the pore size distributionand cementation of is of a rock. As water is progressivelyimbibed into oil-filled pores of different sizes, oil is ejected from them. The ejectionprocess continues long ascontinuous as pathsthrough escape poresstill containingoil are available.Theseescape pathsappearto be lost at oil saturations which greatly exceedthose which occur at the onset of continuity of a nonwettingphase, (e.g.,gas) on the drainagecycle. Thus, the residualoil saturation which resultsfrom waterfloodinga water-wet rock is much greaterthan the critical gas saturationthat characterizes samerock. Apparentlyoil is trapped the imbibition cycle. A similar behavior the on is observedif a preferentially water-wet rock containing free gas is waterflooded. The imbibition and drainagewetting-phase relative permeabilities a consolidated of or unconsolidated rock are retracedunder a succession imbibition and drainagecycles;in a of reversalof the saturationchangefrom drainageto imbibition, a distinct path is traced by the nonwetting phase relative permeability curve (as shown in Figure 32) to a residual nonwettingphasesaturation. This path depends the saturation on established the drainage in cycle. Also, the nonwettingphaserelativepermeabilitycurve in a drainage cycle following an imbibition cycle retracesthe imbibition curve until the previousmaximum nonwetting phasesaturationis reached.This effect is illustratedby Figure 35.22'13

X. EFFECTS OVERBURDEN OF PRESSURE


Wilsonila reportedthat a 5000 psi laboratorysimulationof overburdenpressure reservoir at temperaturereducesthe core effective permeabilitiesto oil and water by about the same extentas it reducesthe single-phase permeabilityof that core. Consequently, water and the oil relative permeabilityof a naturalcore, under 5000 psi overburden pressure,show only a moderatechange from the relative permeability measuredunder atmosphericconditions, as shown in Figure 36. Wilson alsopointedout that an overburden pressure that can produce over 5Voreduction in porosity of a core can also producea sufficiently large changein pore size distributionto affect the relative permeabilityof the core. In contrastto the work of Wilson, Fatt and Barrettrrsconcludedthat variation of rock overburdenpressuresin the range of 3000 psi does not produce any changeon gas relative permeabilityin a sandstone gas-oil system.Figure 37 shows the gas relative permeability

79

100
*\.

10

AIR

\\
b a
I

o
l.

' - air-brine system

l\
80 100

.1

.01 40 60

B r i n e s a t u r a t i o r ' ,V o
FIGURE 35. stone.r2 Air flow behaviorin two-phasesystems,Nellie Bly sand-

with and without the laboratory simulation of overburdenpressure.Similar results were reportedby Thomasand Ward"6 for a gas-oil systemin a low permeabilityrock. Geffen et al.'2 have shown that the residualgas saturation a liquid-gassystem,under atmospheric in conditions,is similar to the resisdualgas saturationmeasured under a 5000 psi laboratory simulationof overburdenpressure. Merliss et al.r17 concludedthat the effect of overburden pressure relative permeabilitywas primarily due to changes interfacialtension. on in

XI. EFFECTS POROSITY OF AND PERMEABILITY


Wyckoff and Botset3 well as Leverettand Lewis8investigated influences porosity as the of and absolute permeability on relative permeability and found them to be insignificant. Dunlaprrs usedunconsolidated sandpackshavingpermeabilities 3.0,4.5, and 8.0 D and of found no indication that the relative permeability-saturation relationshipis a function of specific permeabilityof the sand. Stewart et al.rre found that variationsin permeabilities ranging from 8.5 to 300 mD and porositiesfrom I 5 to 22Voin limestonecores with intergranularporosity, causedrelative permeabilitycurves to shift up to a maximum of 2Voof gas saturation.These investigators employeda solution gas drive, gas-oil relative permeability measurement techniquein their study. They also reportedthe relative permeability curvesto shift up to a maximum of 47o of gas saturation when fracturedlimestonecoresof variousporositiesand permeabilities were employed. Botset2r found that absolute permeabilities rangingfrom 17 to 260D had negligibleeffects

80

Relative Permeabilin of Petroleum Reservoirs

100

a
o)
L

.Y

40 s*
FIGURE 36 system.rra

60 ' o/o

100

Effect of overburdenpressureon relative permeabilityof an oil-brine

1.0

.8

OBP = 0 psig OBP = 3OOO psig

.6
o)
.Y

.4

.2

20

40

60

80

100

so
FIGURE 37. pressure gasrelative permeability.rr5 Effectof overburden on

81 1.0

o
J

40

s*
FIGURE 38. Effect of absolutepermeabilityon relativepermeability.ro

on the gas-liquidrelativepermeability-saturation relationship a consolidated of Nichols Buff sandstone. Botset'sresultswere in agreement with the findingsof Leverett,a who usedsands with permeabilities ranging from 1.04 to 6.80 D. Morgan and Gordon28 conductedtestson four sandstone samplesfrom a reservoirrock with permeabilities ranging from 109 to 213 mD. No clear effect of permeabilityon oilwater relativepermeabilitycurveswas observed. Crowell et al.30 studiedfour differentsands with absolutepermeabilities rangingfrom 3.0 to 8.0 mD and found no correlationbetween permeabilityand gas relativepermeabilityin a water-gas absolute systemas shownin Figure 38. Keelanr20 observedsatisfactorycorrelationsof sandstone permeabilitycorrectedfor air slippage and the irreducible water saturationsas well as end-point relative permeability valuesof gas-water systems. Leas et al.r2rnoteda correlation betweenabsolute permeability and gas relativepermeabilityin particularcases,but believedthis relationship to be true not in general. Felsenthalr22 tested300 sandstone cores and noted that the gas-oil relative permeability curves becameless steepas specific permeabilityincreased. This trend was also reported by McCord.'23In Felsenthal'spaper an effect of porosity on gas-oil relative permeability ratio was also noted. This effect was not generally discerniblein the study of relative permeability datafor a given reservoir became but apparent whendatafor sandstone reservoirs of similar lithology but differing averageporosity were compared.For example,a definite trend was observedin a comparisonof argillaceous and/or calcareous sandstones from I I reservoirsranging in averageporosity from l4 to 28Vo,indicatingthat for a given permeability, the gas-oilrelativepermeability ratio curvesbecame lessfavorable,(i.e., k1k., increased)

82

RelativePermeabiliryof PetroleumReservoirs

from as porosity increased.A similar trend was observedfor a group of clean sandstones and permeability, rangingin porosityfrom 15 to2lTa.For a given porosity five reservoirs ratio curves gavemorefavorablegas-oilrelativepermeability cleansandstones comparatively The leastfavorablegasor sandstones chert reservoirs. and/or calcareous than argillaceous and sandshaly sandstones, oil relative permeabilityratio curves were for conglomerates, in sandstones three catethen classified stonescontainingcarbonateinclusions.Felsenthal gories and found a correlation of gas-oil relative permeabilityratio for each class. The type, which used in the correlationwere porosity, permeability,and sandstone parameters geometrymay be characterized are all relatedto pore geometry. On the other hand, pore by the pore size distribution and Felsenthalfound a correlationbetweengas-oil relative pl.-.uUitity ratio and pore sizedistribution.He found that the morefavorablegas-oilrelative with a pore sizedistributioncurve having permeabilityratio curveswere generallyassociated a sharppeak among the large pore sizes.

OF XII. EFFECTS TEMPERATURE


with inindicatedthat ineducible water saturationincreased Severalearly studiesr2a-r28 temperature; with increasing decreased oil saturation and that residual creasingtemperature process.Difficulties in evaluating all of these studiesemployed a dynamic displacement procedure,rro wettability changesdue to the core-cleaning these results include possible and clay migration.t24't2'7't2'1 permeability, in changes absolute possible that the indicated by measurements Lo and Munganr2e relativepermeability Steady-state when using white oils, but were unafwere temperature-dependent relative permeabilities this changeswhen using tetradecane; finding agreeswith the results fected by temperature to in variations resultshavebeenattributed viscosityratio. Sufi et Other of Edmondson.'r. error due to the may have significant out that someof the previousresults pointed al.r30.r3r that and suggested at measuringrelative permeabilities elevatedtemperatures difficulty in difficulties and possibly result from a combinationof measurement t.*p..uture effects (i.e., end effectsin shortcores). phenomena, laboratory-scaling at experiments elevatedtemperaperformeddynamic displacement Miller and Ramey,32 sand packs and a Berea core. Their resultsindicatedthat changes tures on unconsolidated do in temperature not cause relative permeabilitychanges,but that changesin the flow changein pore structure,etc. are capacityat elevatedtemperatures due to clay interactions, at in was an increase oil relativepermeability irreducible they observed The only changethat is this parameter relatively unimportantfor predictingtwo-phaseflow water saturationand and Counsilr33 Chen et al-r67 relativepermeabilities, steam-water behavior.In measuring effects. of temperature also noted the absence XIII. EFFECTS OF INTERFACIAL TENSION AND DENSITY

for are The interfacialforces at fluid-fluid and fluid-solid interfaces responsible retention a described small and Leverett4 in porousmedia. Wyckoff and Botset3 of residualsaturation of interfacialtensionwithin the range of 27 to 72 dyne/cm on relative but definite effect also identified the interfacial tension of permeability. (See Figure 39.) Lefebvre du Preyro3 consolidatedsampleas a factor influencing the relative permeability and residual fluids in a values. Crowell et al.30found that a reductionin interfacialtensionof a watersaturation in air system produced an increasein gas recovery and a decrease residual gas saturation. the possibility that the interfacialtensionwithin the rangeof 27 to discounted IV1uskatr3a 72 dynelcmcan influencerelative permeability.Owens and Archer" concludedthat interfacial tension has no influence on either the water-oil relative permeability of a water-wet core or the gas-oil relative permeability of an oil-wet core. They found that water relative

83

1.0
points: o -5 \ o \ o\oo \ o o / l i n e sI o -- 24-g4 dynes/cm / o'/ / dyne/cm

/ /

o
l<

o r l \ \ .

\ o o \

./
" /

o / wArER

\ "

\1." /

;eCo

*/,')**

0
a "w
FIGURE 39. Effect of interfacialtensionon relativepermeability.a

1.0

permeabilityof the water-wet core and oil relative permeabilityof the oil-wet core were coincident. of reporteda reductionin waterfloodresidualoil saturation a waterMoore and Slobod6T relativepermestatedthat drainage wet core at lower valuesof interfacialtension.Pirsonr02 of ability is independent the interfacialtension,but imbibition relativepermeabilityis senfound that a reduction in interfacial sitive to interfacial tension. Bardon and Longeronr35 tension reducedoil relative permeabilityat constantgas saturationin an oil-gas drainage formation.(SeeFigure40.) The effectof liquid densityon relative cycle of the Fontainebleau permeability has beenfound to be insignificant.-''r2

OF XIV. EFFECTS VISCOSITY


the Leverettet al.a'8investigated effect of viscosityvariationof an oil-water mixture on with 417oporosityand 3.2to 6.8 D of sands relativepermeability artificiallycompacted of variation in relativepermeabilitywhen the absolutepermeability.He found no systematic oil viscosity was varied from 0.31 cp (hexane)to 76.5 cp (lubricatingoil) and the water phaseviscosity was varied from 0.85 to 32.2 cp. Viscosity ratios employed in the study of rangedfrom 0.051 to 90. The experiments Leverettet al. were performedunder steadystate flow at low pressuregradients.Figures41 and 42 show the effect of viscosity ratio variation on water and oil relative permeabilitycurves. of Wyckoff and Botset3found that moderatevariationsin viscosities the fluid phasesin rangingfrom 3 .2 to 6.0 D failed to produce sandpacks with permeabilities unconsolidated any changein the relative permeabilityvalues.In their experimenta mixture of water and between0.9 and 3.4 cp by carbondioxide was employedand water viscositywas adjusted addition of a susar solution to the water.

84

Relative Permeabilin of Petroleum Reservoirs


1.0

o = .0O1 mN/m

O
J

.5

ss
FIGURE 40. Effect of low interfacialtensionson gas-oil relativepermeability.r15

1.0

O MEgO. 1.80 a D 0.35 * 0.057

3
.Y

o sw

1.o

FIGURE 41. Effect of viscosity ratio (M) on water relative permeability.a

Richardsonr36 found that the water-oil relative permeability ratio is independentof fluid viscositywhere the oil viscosityvaried from I .8 to I 5 I cp (seeFigure43). Johnson al . r37 et confirmed theseresults for displaced/displacing viscosity ratios up to 37. Leviner38 found

85
1.0
o M = 9 0 o o v 1.8 .35 .057

o
-Y

r\'
%\o

s*

1.0

FIGURE 42. Effect of viscosity ratio (M) on oil relative permeability.a

that the relative permeabilityof a sandstone samplewas independent viscosity ratio in of the rangeof 1.92 to 22.6. Craigr3e reportedthat the gas-oil relative permeabilityratio of a Nellie Bly sandstone sample with 824 mD permeabilityand 28.l%oporosity showed no significantvariation with oil viscositiesin the rangeof 1.4 to 125 cp. Resultsof this study are illustrated by Figure 44. Sandberg al.'aofound that oil and water relativepermeabilities a uniformly saturated et of core are independent the oil viscosity in the range of 0.398 to 1.683 cp. Donaldsonet of al.'o' and Geffen et al.ta2 alsoconcluded that relativepermeability independent viscosity is of as long as the core wettability is preserved. Wilsonrrafound that a 5000 psi fluid pressure which caused kerosene viscosityto increase from I .7 to2.l cp and waterviscosityto increase by 17odid not produceany significanteffect on water and oil relativepermeabilityvalues. Muskat et al.27reported that the effect of viscosity on relative permeability of an unconsolidatedsand was very small and within the limits of experimental accuracy. Krutter and Day'43used methaneand air as the nonwettingphasein a two-phasesystern of oil and gas. The gas was injectedinto cores saturated with oils with viscositiesranging from 2 to 100 cP. They found that the air relative permeability values were slightly less than those for methane. Saraf and Fattroapplied Darcy's law to each of the phases a multi-phasesystemand of concludedthat relative permeability is independent viscosity. The Saraf and Fatt equation of is basedon the assumptionthat different phasesflow in different capillariesand do not come in contact with each other. Yuster,6however, concluded that relative permeability values for the systemshe studied were markedly influenced by variation in viscosity ratio, increasingwith an increaseof the ratio. This conclusionwas later supported the work of Morse et al.r44 by Odehr45 expanded Yuster's work and concludedthat the nonwetting phaserelative permeability increases with an increasein viscosity ratio. He found that the magnitudeof the effect on relative perrneability decreases with increasein single-phase perrneability. Odeh found that the deviation in nonwetting phaserelative permeability is increasedas the nonwetting phasesaturationis increased, with the deviationreachinga maximum at the nonwettingphaseresidualsaturation. He also concluded that the wetting-phaserelative permeability is not affected by variation in viscosityratios. Figure 45 showsthe effect of viscosityratio variationin the rangeof 0.5

86

Relative Permeabilin of Petroleum Reservoirs


60

o
; 4 1

3
j

q+6i.li'-qla+6

! j

ncp,eri-nent Waterflood using 151 cp. oil using

A waterflood Kerosene

30

40

50

60

70

80

90

100

sw
F I G U R E 4 3 . C o m p a r i s o n o f s t e a d y - s t a t er e s u l t s w i t h f l o o d i n g performance.r36

to 74.5 on water and oil relativepermeabilitycurves.Odeh statedthat the effect of viscosity ratio on relativepermeabilitycould be ignoredfor samples with single-phase permeabilities greater than lD. Yuster'sand Odeh'sresults have beencriticizedby other investigators.ra6 Downie and CraneraT reportedthat oil viscosity could influencethe oil effective permeability of somerocks. Later, they qualified their statement sayingthat once an increased by relativepermeabilityis obtainedby employmentof high viscosityoil, it may not be lost by replacingthis oil with one of a lower viscosity.They explained this phenomenon qualitatively in terms of the movementof colloidal particlesat oil-water interfaces. Hassler et al.r found that lower gas relative permeabilityvalues were associated with lower oil viscosity in a Bradford sand. However, they expressed doubt that the variationin relative permeabilitycould be describedby a single factor varying with oil viscosity. Pirsonro2 stated that the importanceof the effect of viscosity ratio on the imbibition nonwettingphaserelative permeabilityis of second-order magnitude.Ehrlich and Cranera8 concludedthat the imbibition and drainagerelativepermeabilities, under a steadycondition of flow, are independent viscosityratio. However,they found that the irreduciblewettingof phasesaturation following a steady-state drainage,when the interfacialeffect predominated

87 1.0

0.1

o J

0 . 01

0 . 0 01
"g
FIGURE 44.
Q

o.4

Relativepermeabilityratios for Nellie Bly sandstone.rre

over viscousand gravitationaleffects,decreases with an increase the ratio of nonwetting in to wetting-phase viscosities. McCafferyse reportedthat in stronglywettedsystems, imbibition and drainage the relative permeabilitiesare independent the viscous forces. He concludedthat even though the of relative permeabilityto a phasemight be influencedby viscosity variation of that phase, the relative permeabilityratio is independent viscosity. of Perkinsrae concludedthat flow in a porousbody is governedby relativepermeabilityand viscosity ratio when the ratio of capillary pressureto the applied pressureis negligible. Pickell et al.r-toconcluded that only a large variation in viscous forces could have any rs3 appreciable effect on residualoil saturation.Severalauthors4'67'rsr recognizedthat the wetting and the nonwetting phaserelative permeabilitymight be significantly affectedby the ratio of capillary to viscous forces, ocos0/pv, where o represents interfacialtension expressed dynes per centimeter;0 represents as contact angle; p represents viscosity expressed cp; and v represents as fluid velocity expressed centimeters second.Lefebvre as per du Prey'samade a systematicstudy of the effect of this ratio on relative permeabilityby simultaneously varying the interfacialtension,viscosity,and velocity. He found that relative permeabilitydecreases the ratio ocos0/pv increases. also concludedthat the relative as He permeabilitycurve is influencedby the viscosityratio when the wetting phaseis displaced

88

Relative Permeabilin of Petroleum Reservoirs

240

t
6
L

-Y

50

1OO

a -w
FIGURE 45. E f f e c t o f v i s c o s i t y a t i o ( M ) o n r e l a t i v ep e r m e a b i l i t y . r l s r

by the nonwettingphase.Bardon and Longelsnr'3-s found that in some gas-oil systems,the drainagerelative permeabilityand residualoil saturation stronglyaffectedby the p"vlct are ratio. An assumption that the relativepermeability valuesare independent viscosityimplies of that the system can be represented a bundle of parallel, noninterconnecting by capillary tubes,eachof which is filled with eitherthe wettingor the nonwetting phasealone.Thus, the nonwettingphaseflows throughthe largerchannels while the wettingphase flows through the smaller capillaries. However, this model probably does not completely representthe conditionsin porousmedia.An alternative model is the simultaneous flow of two immiscible fluid phasesin larger capillaries. A flow picture more compatiblewith the presentknowledgeof fluid behavioris a combinationof the two modelsdescribed above,with one dominatingover the other depending primarily on wettability. OdehTbelieved that the fluid phasesdid not flow in separate capillariesof porous media as Leverettpostulated and further statedthat the wetting phase moves microscopicallyin a sort of sliding motion impartedto it by the shearforce caused by motion of the nonwettingphase.From this modelhe concluded thata decrease interstitial in wetting-phase saturationcan be developedas a result of an increasein viscosity, thereby affectingthe relative permeabilityvalues. In view of the diverse opinions which have been expressed various investigators by concerningthe influence of viscosity on relative permeability, it seemsbest to conduct

89

Water ---102

Present 5%

at

Start

------202

\r

olL

o
J

WATER

-:='
20 40 60 80 100

sw
FIGURE 46. on Effect of original water saturation relativepermeability.''

with fluids which do not differ greatly in vislaboratoryrelative permeabilityexperiments cosity from the reservoirfluids. XV. EFFECTS OF INITIAL WETTING-PHASE SATURATION The amount of initial interstitialwater affectsthe oil-water relativepermeabilityvalues. Figure 46 shows the effect of varying the this relationship. Caudle et al.ra investigated on amountof initial water saturation water and oil relativepermeability.It can be seenthat not only the startingpoints, but also the shapeof the relativepermeabilitycurvesvary with the amountof initial interstitialwater.ro' tendedto shift water-oilrelative found that the presence initial water saturation of SaremrT2 permeabilityratio curves toward the region of lower oil saturation.The differencein the residualoil saturationcausedby this shift was reportedto be about half the differencein is Thus, a lower residualoil saturation obtainedat higher valuesof initial water saturation. initial water saturation. noted that the maximum effect of initial water saturationon the Hendersonet al.3-t'r6s 4Tc along relativepermeabilitycurve was a shift of the entire curve laterallyapproximately the oil saturationfor a given pair of the saturationaxis, in a direction which increased relativepermeabilityvalues.Craig indicatedthat up to 20Vainitial connatewater saturation However, a definiteeffect in oil-wet coreshad no effect on oil-waterrelativepermeabilities. was observedin water-wetcores. that, exceptfor specialstudies,the amountof water presentat the startof It is suggested watersaturation the sample. of shouldbe the irreducible determination a relativepermeability

90

RelativePermeabilin of PetroleumReservoirs
100

^o

*"I

3 l t

";/ :l t/

no

connate

o
j

t/
1
O l

t/

\.

I
:l

o) .:<

ilr,
;l
I

sl

tl

!t

{t
/
.01 20
40 60

80

100

e "g
FIGURE 47. Effect of connatewater on relativepermeabilityratio.'7*

XVI. EFFECTS OF AN IMMOBILE

THIRD PHASE

may The mobile phases have only two mobile fluid phases. reservoirs Many hydrocarbon be gas and oil in the upper portion of the reservoirand water and oil in the lower portion. fluid flow behaviorin are Thus, two-phaserelative permeabilities sufficientto characterize thesereservoirs. suggestthat the immobile water saturationmay be regardedas part Some investigators may be given in terms of the hydrocarbonpore of the rock, and gas and oil saturations and cleanedcores,both water-wetand testedseveralnative-state space.Owens et al.rss'r73 influence had no measurable oil-wet. and found that an immobile connatewater saturation on the gas-oil relative permeability ratio in the majority of the casesthat were studied. affect the permeability did concludedthat low water saturations not appreciably CalhounrTa to which doesnot contribute the flow capacity space the ratio, simply because wateroccupies on of the rock. Figure 47 showsthe effect of connatewater saturation gas-oil permeability porosity, the ratio. Stewartet al.'tt have also shown that in a limestonewith intergranular effect of interstitialwater on externalgas or solutiongas drive gas-oilrelativepermeability ratio is negligible. betweenthe gas-oil relative permeabilityof a Leas et al.'2' reporteda close agreement system at various values of interstitial water saturation.This agreementwas best in the

9l

\\ \ \ \ \ \
o
l< GAS

15-25eaconnate

""ut"t";' / , \ \ / / , / y/ot f /

water f

,r /

.z'\''-40 60 80 100

So-

.__ sg
FIGURE 48. Efl'ect of the presence ol' connate water on relative permeabilities.l

equilibrium gas saturationregion. They concluded that the gas relative permeability is dependent total liquid saturation.Other investigators on have suggested that even though the immobile connatewater does not appreciablyaffect the relative permeabilityratio, the amount and distribution of the interstitial water may influence the relative permeability curve. Dunlap,r18 Leverett,a Caudleet al.,'" and McCaffery''e have indicated dependency a on connatewater saturation.Figure 48 compares the permeability-saturation curves for oil and gas at l5 to 25Voconnatewater with the corresponding curveswithout connatewater. Kyte et al.t7o studieda wide rangeof corematerials and fluid properties thatcould influence residualsaturation,to determinethe mechanism oil displacement water in a partially of by gas-saturated porous system. They found that the initial gas saturationis related to the trappedgas saturation,which plays a beneficial role in reducing residual oil saturation. Mattax and ClotheirtTT concludedthat the trappedgas saturation could improve oil-water relative permeabilityvalues in consolidated water-wetsandstones. (SeeFigure 49.) Holmgrenand MorserT8 attributed oil recoveryimprovement a samplein the presence the of of residualgas to one or more of the following factors: l. 2. 3. 4. The changes physicalcharacteristics oil. in of The selectiveplugging action of the gas as indicatedby Kyte. Inclusionof mist in the free gas phase. The additionalsweepingor driving action of the free gas as indicatedby Leverett.a.s

Holmgrenand Morse concludedthat the changes physicalcharacteristics oil, within in of the pressurerange used for their experimentalwork, were not sufficient to account for the differences the residualoil saturation in which were noted.They further statedthat a change in displacement mechanism was the most importantcauseof the oil recoveryimprovement.

92

Relative Permeability of Petroleum Reservoirs


100 GAS SATURATION, PV %
MOBILE TRAPPED

o
5

10

9 't2

3
-Y \ 1

o
L

.o1
1.O

so
lfrU.}?"a:.,, water-oil permeability improvement to relative ratio due

Schneider and Owenssa investigated effect of trappedgas saturation sandstone the in and carbonate rocks and concludedthat the trappedgasaffectedwater relativepermeabilitymore than oil relative permeabilityin oil-wet rocks. These effects are illustratedin Figures 50 and 5l . They also concludedthat the trappedgas saturation loweredthe maximum value of oil relative permeability. Water relative permeabilitywas also lowered as a result of an increasein trappedgas saturation.Theseeffectsare illustratedby Figure 52. XVII. EFFECTS OF OTHER FACTORS The effects of displacement pressure,pressuregradient,and flow rate on the shapeof relative permeability curves have long been a controversialsubject in petroleum-related literature.Someauthorsbelievethat the effect of displacment pressure and pressure gradient may be due to the changes imposedon viscosity,interfacialtension,and other fluid or rock properties.Others believe that the changesin relativepermeability,which appearto result from changesin displacmentpressureand pressure gradient,are actually due primarily to an "end effect" developedduring laboratorytests. End effect or boundary effect refers to a discontinuityin the capillary propertiesof a systemat the time of relative permeabilitymeasurement. a stratumof permeablerock, In the capillaryforcesact uniformly in all directions,and thusnegate eachother. In a laboratory sample,however,thereis a saturation discontinuity the end of a sample.When the flowing at phases discharged are into an open region under atmospheric pressure, net capillary force a persists the sample;this force tendsto preventthe wetting phasefrom leavingthe sample. in The accumulation the wetting phaseat the outflow face of the samplecreates saturation of a gradientalong the sample which disturbsthe relative permeabilitymeasurements. exFor ample, a large difference in saturationat the displacement front causesa large capillary pressure gradient, which in turn causesthe water to advanceaheadof the flood front and to reducethe capillary pressure gradientin the measured region.The advancing watercannot be producedwhen it first reachesthe outflow face of a core, because the pressurein the

93

100
\ *r* vs. S,

10
be o
lz
kro vs. So

Trapped Gas Sat. o 0 % .a 1 1.8%

.1

1'o s*.-so

FIGURE 50. carbonate)Ea .

Effect of trapped gas saturation (oil wet Grayburg

waterjust insidethe core is lower than the pressure the oil-filled spacearoundthe outflow in face. This differencein pressure equal to the capillary pressure the existingsaturation is for at the outflow face. Therefore,water accumulates the outflow end of the core, causing at a reduction in the capillary pressure.The water will not be produceduntil the capillary pressureis overcome and the residual oil saturation(at the outflow face of the core) is reached.The calculation of relative permeability basedon the averagesaturationof the sampleproduces erroneous resultsin this case,sincethe relativepermeability variesthroughout the core due to the saturation gradientcreatedby the wetting phaseaccumulation the at outflow face of the core. Owens et al.,r-5s Sandberg al.,to"Kyte and Rappoport,rs6 Perkinsrae et and believethat the most convenientway of minimizing the boundaryeffect is the adjustment capillary of forcesto insignificantvalues, as comparedto the viscousforces. This is usually done by a flow rate adjustment.However, the adjustedrate must be low enoughso the inertial forces do not disturb the laboratory measurement. is suggested It that the higher flow rate also increases fluid dispersion the inflow end of the sample,so that fluid mixing is enhanced. the at An equation has been developedr-57 predict the extent that a core can be disturbedby to boundaryeffect, at a given rate. Another convenient way of minimizing the boundaryeffect at the outflow end of a core is to use a more viscousoil in a longercore.rs6 Leverett et al.a'8reported,then refuted, the influenceof flow rate upon relative permeability. They eventuallyattributedthe observeddeviationsin their resultsto an end effect.

Relative Permeability of Petroleum Reservoirs

100

be o
Y kro vs. So

T r a p p e dG a s S a t . o 0 % a 10.3% o 19.2%

s,rt_ So

(oil Tensleep sandstone).Ea gas of RGURE 51. Effect trapped saturation wet found that a 50-fold variation by suchas that previouslydescribed Hassler.rCrowell et al.3o of injectionrate, within the limits of viscousflow of waterand gas,had no effect on residual also concludedthat, at reasonable Geffen et al.ra2 gas saturationof an Arizona sandstone. flow rates, the effect of waterflooding rate on the efficiency of gas displacementwas found that relativepermeabilitywas and Morse et al.r-58 negligible. Henderson and Yuster3s also found that were studied.Wyckoff and Botset3 in rate-dependent all gas-liquidsystems were when the two phases were rate-dependent that the gasand liquid relativepermeabilities gradient. allowed to flow through the core under the samepressure in with increase flow rate when Caudleet al.'a found that relativepermeabilitydecreased the was a gas. Labastieet ol.,'-tnhowever, investigated effect of one of the flowing phases coresand concludedthat relative and oil-wet carbonate flow rate in a water-wetsandstone of were independent flow rate exceptnear residualoil saturation.Sandberg permeabilities rel O Ric e et al. , r ao ha rd s o n t a l .,r-t7 s o b ae t a l .,r3and Leas et al .r2rfoundthat drai nage aof tive permeability is independent the flow rate as long as a saturationgradient is not concludedthat relativepermeability introducedin the core by the inertial forces. Pirsonr02 Ehrlich and Crane'otexaminedthe effect of flow in is not rate-sensitive drainageprocesses. rate variation on steady-staterelative permeability and concluded that both imbibition and of drainagerelative permeabilitywere independent flow rate. found that the imbibition relativepermeabilityvalueswere dependent Handy and Dattar62 that is, the relativepermeabilityvaluesunder free imbibition on the imbibition procedures;

95

100

kro vs

so

o\s o
-Y Kr* us. S*

s*---_
-so
Effectof trapped saturation gas (warer wet Tensleep il*"r,::.,i,]. were largerthan thoseundera controlledprocess. The difference between free and controlled imbibition was found to be smaller for more permeable samples.Perkinsrae found that the residualoil saturationafter flooding was independent the flooding rate and concluded of that capillary forces controlled the microscopicfluid distribution in the core. Moore and Slobod6T reportedthat waterflood recoveryfrom a water-wetcore was practically independent of flooding rate. However,they observed that a significant recoveryincrease may be obtained at extremely high rates. Hupplerr6s stated that the waterflood recovery from cores with significantheterogeneity was sensitiveto flooding rate. Lefebvredu Preyrsa concludedthat the relativepermeabilitywas a function of velocity (v), throughthe ratio (ocosO/pv),when the viscousforces predominate. Wyckoff and Botset,3Leverett,a and Henderson al.3-s'r6s et studiedthe possibleeffectsof displacement pressureand pressure gradienton water-oil relative permeability.They concluded that the water and oil relative permeabilityvalueswere slightly influencedby these factors. Muskatr3a and Krutter and Day,'oohowever, reportedthat the gas and oil relative permeabilityvaluesof a consolidated sandstone were not affectedby changes differential in pressure.McCafferyrTe indicated that the drainagerelative permeabilityvalues were not influencedby the flow rateswhich result from apressuregradientin the rangeof 1.0 to 5.0 psi acrossa 12 in. core. Delclaudr60 concluded also that relativepermeabilityis independent of displacement pressure. Pirson,ro2 however, suggested that the relativepermeabilityin an imbibition cycle is sensitiveto pressure gradient.

96

Relative Permeabilin of Petroleum Reservoirs

Krutter andDayr66 found that ultimaterecoveryincreases with increasing pressure gradient, althoughthe ratio of increased recoveryto increased pressure gradientdecreases the region in gradients. of high pressure Brownell and Katzr68 reported that an increase pressure in gradient decreased residualsaturation the toward zero in the systems that were investigated. Geffen et al.ra2also confirmed that residual gas saturationwas a function of pressuregradient. Stegemeier and Jensen3T believedthat the residualwetting phasein a drainageprocesswas held in pendularrings interconnected with only thin wetting-phase layers. They concluded that this residualwetting phasewas trappedby capillary forces and that a higher pressure gradient might overcome capillarypressure reduce residual the and the wetting-phase saturation. Stewart al.rreobserved et that the rateof pressure declinein a nonuniform limestone might influencethe gas-oilrelativepermeability ratio when the solutiongasdisplacement technique for relative permeabilitymeasurement was employed.Wall and Khuranar6e found the gas saturation developedin a sand pack, at a given rate of pressure decline, was a function of the meanparticlesize and probablya functionof permeability.They found that a finer grain sandpack gave rise to higher gas saturations the solutiongas displacement in technique. Crowell et al.30 studiedthe effect of coredimensions laboratory on measurement relative of permeability.They found that the residualgas saturation water-gassystemswas almost in independent length of the core, within limits of the laboratory-scale of models used. They also examinedcylindrical and rectangular samples,and observedthat a 100-foldchangein the ratio of core length to core cross-sectional area of Berea and Boise sandstones not did alter the residualgas saturationof the samples.Moore and SlobodoT also found that fluid recovery from water-wet cores was not affected by the sample length. Perkinstae and McCafferyrTe recommended use of longer cores, to reduceinfluenceof the end effect. the RoserTo studied the effect of gas expansion,createdby the pressure gradientalong the sample, on gas-liquid flow characteristics. concludedthat a necessary He condition for correct steady-state measurements liquid-gasrelativepermeabilitywas the establishment of of a uniform fluid saturation distributionin the core. Osobaet al. '3 found that gasexpansion affected gas and oil relative permeability values in tests conductedat near-atmospheric pressure.Richardsonet al.,r-'7however, found that the effect of gas expansionon gas and oil relativepermeabilityvalueswas insignificantat the low pressures which were employed in their study. In the laboratorygas displacement methodof relativepermeabilitymeasurement, "staa ' bilized zone' tendsto form when the wetting liquid saturation sufficientlyhigh to permit is its readjustment faster than the imposed displacement the externaldrive. The relative by permeabilityvaluesobtainedprior to passage the stabilizedzone arenot valid. Therefore, of it is advantageous reduce the range of saturationinfluencedby the stabilizedzone, to to obtain valid measurements over as wide a saturation rangeas possible. It can be shown from the Buckley-Leverettequation that the saturationat which the stabilizedzone passes out of a systemis inverselyrelatedto the viscosityof the displaced liquid. This relationshipis based on an assumptionthat a true stabilizedzone forms in laboratorygas drives on short cores. It can also be shown that the length of the stabilized zone is inverselyrelatedto the injection rate or differentialpressure. has been suggested It that the stabilizedzone will be sufficiently small if the pressuredifferential is of such a magnitudethat a volume of gas approximatelyequal to one half the pore volume of the samplewould be producedin lessthan 60 sec. This flow rate insuresthat the portion of the core in which the capillary effects predominate will be a negligibly small fraction of the total pore space. Loomis and CrowellrT' showedexperimentallythat the influence of the stabilizedzone fluid flow is much less marked with relatively viscousoil as the displaced phase. Botset2' investigatedthe effect of saturationpressureon gas-oil permeability values and pressure negligibleeffecton laboratory concluded that the saturation had relativepermeability

97 measurement. Stewart studiedthe effect of gas supersaturation laboratorysolution gas on displacement relative permeabilitymeasurements. indicatedthat even though very liitle He supersaturation existsundermost field conditions,the effect may be significantfor laboratory testsconductedat high flow rates.He found that the gas-oil relativepermeabilityratio was generallyindependent the degreeof supersaturation rock with intergranular of in porosity. The influenceof dispersionon relative permeabilitywas studiedby Chilingarianet al.ee They concludedthat an increase degreeof dispersionincreased relativs permeability in the of the porousmediumto both the continuous discontinuous and phases. They alsoconcluded that the degreeof dispersionincreasedwith decreasing interfacialtension and increasine time of coalescence dispersed-phase of droplets.

REFERENCES
l' Hassler, G. L., Rice, R. R., and Leeman, E. H., Investigations recovery of on the oil from sandstones b y g a s - d r i v e , r a n s .A I M E , I 1 8 , l 1 6 , 1 9 3 6 . T 2. Muskat, M. and Meres, M. W., phv-sics, 346. 1936. j, 3 ' W y c k o f f , R . D . a n d B o t s e t , H . G . , F l o w o f g a s l i q u i d m i x t u r e st h r o u g hs a n d s , p h y s i c s j,, 3 2 5 , 1936. 4 ' L e v e r e t t , M . C . , F l o w o f o i l - w a t e r m i x t u r e s t h r o u g h u n c o n s o l i d a t e d s a n da n s . A I M E , Tr s, l32, l49,lg3g. 5' Nowak, T. J. and Krueger, R. P., The effect of mud filtratesand mud particles upon the permeability of cores, Proceedings the Spring Apr Meeting, Los Angeres, 1955. of 6' Yuster, S. T., TheoreticalConsideration MultiphaseFlow in ldealizedCapillary of System, proceedings of the Third world Petroleumcongress, Hague, Netherlands,lgsl, (z\ 43i. 7. Odeh, A. S., Relative PermeabilityStudies,Mastersthesis,Universityof California, Los Angeles, 1953. 8' Leverett, M. C. and Lewis, W. 8., Steadyflow of gas-oil-water mixturesthroughunconsolidated sands, Trans. AIME, 142. 107. t94t. 9. Sarem, A. M., Three-phase relativepermeabilitymeasurements unsteady-state pet. Eng. by methods, Soc. J.,9. t99. t966. l0' Saraf, D. N. and Fatt, I., Three-phase relativepermeabilitymeasurement using a N.M.R. techniquefor estimatingfluid sarurarion, Soc. pet. Eng. J., 9,235, lg6j. ll' Owens, W. W. and Archer, D. L., The effect of rock wettability on oil-water relative permeability relationships Trans. AIME, 251, 8j3, lgjl. , l2' Geffen, T. M., owens, W. W., Parrish, D. R., and Morse, R. A., Experimental investigation factors of affectinglaboratoryrelative permeabilitymeasurements, Trans. AIME, lg2, gg, lg5l. 13. Osoba, J. S., Richardson, J. G., Kerver, J. K., Hafford, J. A., and Blair, p. M., Laboratorymeasurementsof relative permeability,Trans. AIME, 192, 47, lg5l. 14' Caudle, B. H., Slobod, R. L., and Brownscombe, E. R., Further developments in the laboratory determination relative permeability,Trans. AIME, 192, 145, 1951. of 15. Snell, R. W., Measurement gas-phase of saturation porousmedia, J. Inst.pet.45, (4Zg), lg5g. in l6' Emmett, W. R., Beaver, K. W., and McCaleb, J. A., Little Buffalo basin Tensleep heterogeneity and i t s i n f l u e n c e n d r i l l i n g a n d s e c o n d a r ye c o v e r y J . p e r . T e c h n o l . , 2 , 1 6 l . l g 7 l . o r , l7' Donaldson, E. C. and Dean, G. W., Two- and Three-Phase Relative PermeabilityStudies, report# 6g26, u.s. Departmentof the Interior, Bureauof Mines, Bartlesville,okla., 1966. 18. Arps, J. J. and Roberts, T. G., The effect of the relative permeabilityratio, the oil gravity, and the solutiongas-oil ratio on the primary recoveryfrom depletiontype reservoir,Trans.AIME,204,120, lg5l. 19' Bulnes, A. C. and Fittings, R. U., An introductory discussion of reservoir performance of limestone formations, Trans. AIME, 160, 179, 1945. 20' Stone, H.L., Probabilitymodel forestimating three-phase relativepermeability,Trans.AIME,24g,Zl4, t970. 2 1 . B o t s e t , H . G . , F l o w o f g a s l i q u i d m i x t u r e st h r o u g hc o n s o l i d a t e d a n d , T r a n s . s AIME,136,91 ,1940. 22' Naar, J., Wygal, R. J., and Henderson, J. H., Imbibition relativepermeability in unconsolidated porous media, Trans. AIME, 225. t3. t962. 23. Nind, T. E. w., Ed., Principles of oil production, McGraw Hill, New york. 1964. 24' Corey, A. T. and Rathjens, C.H., Effect of stratification relativepermeability,Trans. on AIME,207, ( 3 s 8 ) ,6 9 , 1 9 5 6 . 25' Huppler, J. D., Numerical investigation of the effects of core heterogeneities on waterflood relative permeability,,Soc.Pet. Eng. J., 10, 381, 1970. ,

98

Relative Permeabilin of Petroleum Reservoirs

26. Johnson, C. E., Jr. and Sweeney,S. A., Quantitativemeasurement flow heterogeneity laboratory of in core samplesand its effect on fluid flow characteristics, papersPE 3610 presented the SpE 46th Annual at M e e t i n g ,N e w O r l e a n s ,O c t o b e r3 , l g l l . 27. Muskat, M., Wyckoff, R. D., Botset, H. G., and Meres, M. W., Flow of gas-liquidmixturesthrough sands,Trans. AIME, 123, 69, 193i. 28. Morgan, T. J. and Gordon, D. T., Influenceof pore geometryon water-oilrelativepermeability J . pet. , T e c h n o l . ,l 1 9 9 , 4 0 7 . 1 9 7 0 . 29. Gorring, R. L., Multiphase Flow of Immiscible Fluids in porous M e d i a , P h . D . t h e s i s ,U n i v e r s i t yo f Michigan, Ann Arbor, 1962. 30. Crowell, D. C., Dean, G. W., and Loomis, A. G., Efficiency of gas <lisplacement from a water drive r e s e r v o i r , . S . B u r e a uM i n e s , 6 i 3 5 . 3 0 . 1 9 6 6 . U 31. Fatt, I., Network model of porous media, dynamic properties networks with tube radius distribution. of Trans. AIME, 20'7, 164, 1956. 32' Dodd, C. G. and Kiel, O. G., Evaluationof Monte Carlo method in studying fluid-fluid displacement a n d w e t t a b i l i t yi n p o r o u sr o c k s ,J . p h v s .C h e m . , 6 3 , 1 6 4 6 , 1 9 5 9 . 33. Wyllie' M. R. J., Interrelationship betweenwetting and non-wettingphaserelative permeability Trans. , A|ME, 192,38t , l95l. 34. Pathak, P., Davis, H. T., and Scriven,L,E,, Dependence residualnonwetting of liquid on pore topology, p a p e rS P E l l 0 l 6 , p r e s e n t e d t t h e S P E 5 7 t h A n n u a l F a l l M e e t i n g ,N e w O r l e a n s ,1 9 8 2 . a 3 5 . H e n d e r s o n ,J . H . a n d Y u s t e r , S . T . , R e l a t i v ep e r m e a b i l r t y t u d y , W o r l t l O i \ , 3 , 1 3 9 , 1 9 4 g . S 36. Land, C. S. and Baptist, O. C., Effect of hydration of montmorilloniteon the permeabilityto gas of water-sensitive reservoirrocks, _/.pet. Technol., 10, 1213, 1965. 37. Stegemeier, G. L. and Jensen, F. W., The Relationshipof Relative Permeabilityto Conract Angles, Theory of Fluid Flow in PorousMedia Conference, University of Oklahoma, 1959. 38. Benner, F. C. and Bartell, F. E., The effect of polar impuritesupon capillary and surfacephenomena in petrofeumproduction,Drill. Prod. Pract., 341, 209, 1941. 39. Salathiel, R. A., Oil recoveryby surfacefilm drainage mixed wettabilityrocks, paperSpE 4104 presented in at SPE 47th Annual Meeting, San Antonio, Calif., October g, 19'72. 40. Reisberg, J. and Doscher, T. M., Interfacialphenomena crude oil-water systems,prod. Mon., 10, in 43. t956. 41. Denekas, M. O., Mattax, C. C., and Davis, G. T., Effect of crude oil components rock wettability, on T r a n s .A I M E , 2 1 6 , 3 3 0 . 1 9 5 9 . 4 2 . E v a n s ,C . R . , R o g e r s , . A . , a n d B a i l y , N . J . L . , C h e m .G e o l . , g , l 4 l , l g l l . M 43. Nutting, P. G., Some physicaland chemicalproperties reservoirrocks bearingon the accumulation of and dischargeof oil, Probl. Pet. Geol. AAPG, lZ, 127, 1934. 44. Leach, R. O., Wagner, O. R., Wood, H. W., and Harpke, C. F., A laboratoryand field study of wettability adjustmentin waterflooding,"/. pet. Te<.hnot., 44,206, 1962. 45. Mungan, N., Interfacialeffects in immiscible liquid-tiquiddisplacement porousmedia, Sot'.pet. Eng. in J.,9,247, 1966. 46. Schmid, C., The wettability of petroleumrocks and resultsof experiments study effects of variations to i n w e t t a b i l i t yo f c o r e s a m p l e s E r d o e l K o h l e , l 7 ( 8 ) , 6 0 5 , 1 9 6 4 . , 47. Kusakov, M. M. et al., Researchin SurfoceForce.s,Deryagin, B. U., Ed., Consultants Bureau, New York, 1963. 48. Craig' F. F., Jr., The Reservoir EngineeringAspectsof WaterfloodingMonograph, yol.3, SpE of AIME, H e n r y L . D o h e r t yS e r i e s ,D a l l a s ,T e x . , 1 9 7 1 . 49. DeBano, L. F. and Letey, J. L., Symposiumon Water RepellentSoils, University of Calif., Berkeley, t969. 50. Holbrook, O. C. and Bernard, G. G., Determination wettability by dye adsorption,Trans. AIME, of 2t3.261.t958. 5 I . Fatt' I. and Klikoff, W. A., Jr., Effect of fractionalwettabilityon multiphase flow throughporousmedia, J . P e t . T e c h n o l . ,1 0 , I l , 1 9 5 9 . 5 2 . B r o w n , R . J . S . a n d F a t t , I . , M e a s u r e m e n t s o f f r a c t i o n ae t t a b i l i t y o f o i l f i e l d r o c k s b y n u c l e a r m a g n e t i c w l refaxation method, J. Pet. Technol., ll , 262, 1956. 53. Iwankow, E. N., A correlationof interstitialwater saturation and heterogeneous wettability, prod. Mon., 24, t8, t960. 54. Gimaludinov, Sh. K., The natureof mineral surfaces oil bearingrocks, Neft. Gazov.2., 12,37,1963. in 55. McGhee, J. W., Crocker, M. E., and Donaldson, E. C., RelativeWetting Properties Crude Oils in of BereaSandstone, BartlesvilleEnergyTechnology Center,Department Energy,Bartlesville, of Okla., BETC/ R I - 7 8 1 9 ,J a n u a r y ,1 9 7 9 . 56. Wagner, O. R. and Leach, R. O., Improvingoildisplacement efficiencyby wettabilityadjustment,Trans. AIME,216.65. 1959.

of Systemfor the Oil-Wet Sandstone the North Burbank 57 . Boneau, D. F. and Clampitt, R. L., A Surfactant T o U n i t , S y m p o s i u m n I m p r o v e dO i l R e c o v e r y , u l s a , A r i z o n a , M a r c h , 1 9 7 6 . recovery, reference petroleum to on roughness contactanglewith special 58. Morrow, N., The effectsof surface J . C a n . P e t . T e c h n o l . .1 0 . 4 2 , 1 9 7 5 . 59. McCaffery, F. G., The Effect of Wettability on Relative Permeabilityand Imbibition in Porous Media, P h . D . t h e s i s ,U n i v e r s i t yo f C a l g a r y ,A l b e r t a ,C a n a d a ,1 9 7 3 . i , 6 0 . Z i s m a n , W . A . , C o n t a c tA n g l e W e t t a b i l i t ya n d A d h e s i o nA d v a n c e s n C h e m i s t r y A m . C h e m . S o c ' . ,4 3 , t. 1964. displacement of 61. Melrose, J. C. and Brandner, C. F., Role of capillaryforcesin determination microscopic fbr efTiciency oil recoveryby water flooding, J. Can. Pet. Technol., 10, 54, 1914. 62. Treiber, L. E., Archer, D. L., and Owens, W. W., A laboratoryevaluationof the wettability of fifty , o i l p r o d u c i n gr e s e r v o i r sS o t ' . P e t . E n g . J . , 1 2 ( 6 ) ,5 3 1 , 1 9 7 2 s i 6 3 . M o r r o w , N . R . , C r a m , P . J . , a n d M c C a f f e r y , F . G . , D i s p l a c e m e n tt u d i e s n d o l o m i t ew i t h w e t t a b i l i t y c o n t r o lb y o c t a n o i ca c i d , S o r ' .P e t . E n g . / . , l 3 ( 4 ) , 2 2 1 , 1 9 1 3 . o 6 4 . M u n g a n , N . , E n h a n c e d i l r e c o v e r yu s i n g w a t e r a s a d r i v i n g f l u i d . W o r l d O i l , 3 , 1 7 . 1 9 8 1 . relating to the wettability of porousrock. frrut.r. AIME,216, 156. I959. 65. Amott, E., Observations 66. Raza, S. H., Treiber, L. E., and Archer, D. L., Wettability of reservoirrocks and its evaluation,Prod. Mon.. 32. 156.1968. oi 67. Moore, T. F. and Slobod, R. L., The effect of viscosity and capillarity on the displacement oil by w a t e r ,P r o d . M o n . , 8 , 2 0 , 1 9 5 6 . 68. Bobek, J. E., Mattax, C. C., and Denekas, M. O., Reservoirrock wettability - its significanceand e v a l u a t i o nT r a n s .A I M E , 2 1 3 , 1 5 5 , 1 9 5 8 . , and Imbibition in Porous 69. Killens, C. R., Nielsen, R. F., and Calhoun, J. C., Capillary Desaturation U U B S , R o c k M i n e r a l I n d u s t r i e sE x p e r i m e n t a l t a t i o n u l l e t i n# 6 2 , P e n n S t a t e n i v e r s i t y , n i v e r s i t yP a r k , l 9 - 5 3 , 55. /6, Sectiort McGraw-Hill. 70. Richardson, S. G., Flow Through PorousMedia, Hundbookof Ftuid Dt'namic's New York. 1961. and 7l . Donaldsol, E. C., Thomas, R. D., and Lorenz, P. B., Wettabilitydetermination its effecton recovery e f f i c i e n c y ,S o c . P e t . E n g . J . , 3 , 1 3 , 1 9 6 9 . i , . 7 2 . M u n g a n , N . , R o l e o f w e t t a b i l i t ya n d i n t e r f a c i atle n s i o n n w a t e r f l o o d i n gS o c ' P e t . E n g . J . , 6 , | 1 5 , 1 9 6 4 . 73. Emery, L. W., Mungan, N., and Nicholson, R. W., Causticslug injectionin the Singletonfield, J. Pet. T e c h n o l . ,1 2 , 1 5 6 9 , 1 9 7 0 . on 74. Kyte, J. R., Nuamann, V. O., and Mattax, C. C., Effect of reservoirenvirclnment water-oil dis1 p l a c e m e n t ,. P e t . T e c h n o l . , 6 , 5 ' 7 9 , 9 6 1 . J of 75. Gatenby, W. A. and Marsden, S. S., Some wettability characteristics syntheticporous media, Prod. Mon.. 22. 5. 1957. of 76. Johansen, R. T. and Dunning, H. N., Relative Wetting Tendencies Crude Oils by Capillarimetric M e t h o d . U . S . B u r e a uo f M i n e s . 1 9 6 1 . 5 ' 7 5 2 . 3rd and Chemistryof Surfat'e.s, ed., Oxford Univ. Press,London, 1959, 192. 77. Adams, N. K., The Physics 78. Slobod, R. L. and Blum, H. A., Method for determiningwettability of reservoirrocks, Truns. AIME, t95. t. t952. of 79. Lorenz, P. 8., Donaldson, E. C., and Thomas, R. D., Use of CentrifugeMeasurements Wettability t o P r e d i c tO i l R e c o v e r y ,U . S . B u r e a uo f M i n e s , 1 9 1 4 , ' 7 8 7 3 . imbibition model with fractional80. Reznik, A. A., Fulton, P. F., and Colbeck, S. C., Jr., A mathematical w e t t a b i l i t yc h a r a c t e r i s t i c P,r o d . M o n . , 3 l ( 9 ) , 2 2 , 1 9 6 7 . s 8l . Coley, F. H., Marsden, S. S., and Calhoun, J. C., Jr., Study of the eff'ectof wettabilityon the behavior of fluids in syntheticporous media, Prod. Mon., 20(8), 29, 1956. t J 8 2 . K e e l a n , D . K . , A c r i t i c a lr e v i e w o f c o r e a n a l y s i s e c h n i q u e s ,. C o n . P e t . T e c h n o l . , 6 , 4 2 , 1 9 1 2 . 83. Poettmann, F. H., Caudle, B. H., Craig, F. F., Jr., Crawford, P. 8., Bond, D. C., Farouq Ali, S. M., Holott, C. R., Johansen, R. T., Mungan, N., and Dowd, W. T., Secondaryand Tertiary Oil lnterstate Oil CompactCommission,OklahomaCity, Okla., September1974. RecoveryProcesses, relativepermeability and two- and three-phase 84. Schneider; F. N. and Owens, W. W., Sandstone carbonate, c h a r a c t e r i s t i c S,o c . P e t . E n g . J . , 3 , 1 5 , 1 9 7 0 . s 85. Scrom,H.M.,Significanceof Water-OilRelativePermeabilityDataCalculatedfromDisplacementTests, Theory of Fluid Flow in PorousMedia Conference,University of Oklahoma, 1959, 189. McGraw-Hill, New 86. Amyx, J. W., Bass, D. M., and Whiting, R. L., PetroleumReservoirEng,ineering, York. 1960. of 87. Colpits, G. P. and Hunter, D. E., Laboratorydisplacement oil by water under simulatedreservoir conditions,J. Can. Pet. Technol., 3(2), 64, 1964. 88. Haddenhorst, H. G. and Koch, R., Effect of temperatureand pressureon the separationof solids from petroleum,Erdoel Kohle, 2, 12, 1959.

100

Relative Permeability of Petroleum Reservoirs

89. Luks, K. D. and Kohn, J. P., The Effect of MethaneUnder Pressure the Liquid Solubility on of Heavy HydrocarbonComponents,Liquid-Vapor and Solid-Liquid-VaporBehavior, progressReport II, Apl Res e a r c h r o j e c t1 3 5 , N o t r e D a m e , I n d i a n a ,J u l y , t 9 7 1 . P 90' Rathmell, J. J., Braun, P. H., and Perkins, T. K., Reservoirwaterfloodresidualoil saturation from l a b o r a t o r y e s t s ,J . P e t . T e c h n o t . , 2 2 5 , l i 5 . l g j 3 . t 91. Richardson, J. G., Perkins, F. M., Jr., and osoba, J. S., Differencein behaviorof fresh and agedeast Texas Woodbine cores, Truns. AIME,204, 86. 1955. 9 2 . M u n g a n , N . , R e l a t i v ep e r m e a b i l i t y e a s u r e m e n rs i n g r e s e r v o i r l u i d s , S o c .p e t . E n g . J . , l 2 ( 5 ) , 3 9 g , m u f 1972. 93. Ehrlich, R., Hasiba, H. H., and Raimondi, P., Alkaline waterfloodingfor wettability alteratione v a f u a t i o n f a p o r e n t i a f i e l d a p p l i c a r i o n , , /p e t . T e c h n o t . , 2 6 , 1 3 3 5 ,l g j 4 . o t . 94. DeterminQtion of Residual Oil Saturatior?,Interstate Oil Compact Commission, Oklahoma City, Okla., t978. 95. Jennings, H. H., Surfaceproperties naturaland syntheticporousmedia, prod. Mon., 2l(5). of 20. 1957. 96' Hough, E. W., Rzasa,M. J., and Wood, B. 8., Interfacial tensions reservoir at pressures temperatures, and apparatus and the water-methane system,Trans. AIME, 192, 5i, lg5l. 97. Poston, s. w., Ysrael, s., Hossain, A. K., Montgomery, E. F., and Ramey, H. J., Jr., The Effect of Temperature Relative Permeabilityof Unconsolidated on Sands.paper SPE 1897 presented the SpE at 4 2 n d A n n u a l F a l l M e e t i n g ,H o u s t o n ,T e x a s, 1 9 6 7 . . 98. Cuiec, L. E., Restoration the Natural Stateof Core Samples, of paperSPE 5634 presented the SpE 50th at A n n u a l M e e t i n g ,D a l l a s ,T e x . , 1 9 7 5 . 99. Chifingarian, G. V., Mannon, R. W., and Rieke, H. H., Eds., Oil and Gas productionFrom Cqrbonctre R o c k s ,E l s e v i e r ,A m s t e r d a m ,1 9 7 2 . 100. BradleY, D. J., The Applicability of WettabilityAlterationto NaturallyFracrured Reservoirs anrllmbibition W a t e r f l o o d i n gM a s t e r st h e s i s ,U n i v e r s i t yo f T u l s a ; ' O k l a h o m a1 9 g 3 . , , l0l' McCafferY, F. G. and Mungan, N.' Contactangle and interfacialtensionstudiesof some hydrocarbon w a t e r s o l i d s y s t e m sJ . C u t t . p e t . T e c h n o t . ,j , l g 5 , 1 9 7 0 . , 1 0 2 . o i l R e s e r v ' o iE n g i n e e r i n g P i r s o n ,s . J . , E d . . M c G r a w - H i l l , N e w y o r k . 1 9 5 g . 6 g . r , 103' Lefebvre du Prey, E., Deplacements non-miscibles dans les millieux poreux influence des parameters interfaciauxsur les permeabilites relatives,c.R. IV Cotoq. ARTFp puu, 196g. 104. Donaldson, E. C. and Thomas, R. D., MicroscopicObservations Oil Displacement Water-Wet of in and O i l - W e t F o r m a t i o n s S P E 3 5 5 5 p r e s e n t e d t t h e 4 6 t h S P E A n n u a l F a l l M e e t i n g ,N e w O r l e a n s , c t . 3 - 6 . , a O 197t. 105' McCafferY, F. G. and Bennion, D. W., The effect of wettability on two-phaserelative permeabilities. J. Can. Pet. Techno1.10.42. 1974. , 1 0 6 . L e v i n e , J . S . , D i s p l a c e m e ne x p e r i m e n t sn a c o n s o l i d a t e p o r o u ss y s t e m ,T r a n s .A t M E , 2 0 l , t i d 57, t9-54. 107. Josendal, V. A., Sandford, B. 8., and Wilson, J. W., Improved multiphaseflow studies employing r a d i o a c t i v er a c e r s , r o n s .A I M E , I 9 5 , 6 5 . 1 9 5 2 . t T 108. Terwilliger, P. L., wilsey, L. E., Hall, H. N., Bridges, p. M., and Morse, R. A., Experimental and theoreticalinvestigation gravity drainageperformance Trans. AIME, l92, 285, 1951. of . - 109. Johnson, E. F., Bossler, D. P., and Naumann, V. O., Calculationof relative permeability from displacementexperiments,Trans. AIME, 216. 370. lg5g. l l 0 . L a n d , C . S . , C o m p a r i s o n f c a l c u l a t e d i t h e x p e r i r n e n t am b i b i t i o nr e l a t i v ep e r m e a b i l i t y , T r u n s . o w il AIME, 2 5 t . 4 1 9 . t 9 7| . I I l ' Gardner, G. H. F., Messmer, J. H., and Woodside, W., EffectivePorosityandGas Relativepermeability on Liquid Imbibition Cycle. Theory of Fluid Flow in PorousMedia Conference, University of Oklahoma. Norman. 1959. 173. ll2. Shelton, J. L. and Schneider, F. M., The effect of water injectionon miscible flooding methods using hydrocarbons and CO,, paper SPE 4580 presented the SPE 48th Annual Meeting, Las Vegas, 1973. at 5 I l3' Land, C., Calculation imbibitionrelativepermeability two- and of for three-phase flow fiom rock properties, Soc'. et. Eng. J., 6, t49. 1968. P I 1 4 . W i l s o n , J . W . , D e t e r m i n a t i o n f R e l a t i v eP e r m e a b i l i t y n d e r S i m u l a t e d e s e r v o i r o U R Conditions. AIChEJ, 2(t), 4. 1956. - f l5' Fatt, I. and Barrett, R. E., Effect of overburdenpressure on relative permeability Truns. AtME, lgE, , 325. t953. I l6' Thomas, R. D. and Ward, D. C., Effect of overburden pressure and water saturation gas permeability on of tight sandstone cores,-/. Pet. Te<'hnot., 120, lg'/2. 2, ll7. Merliss, F. E., Doane, J. D., and Rzasa, M. J., Influenceof rock and fluid properties and immiscible fluid-flow behaviorin porous media, paper 510-G presented the AIME Annual Meeting. New orleans. at I955. l l 8 . D u n l a p , E . N . , I n f l u e n c e o c o n n a t e w a t e r o n p e r m e a b i l i t y o fn d s o o r l , T r a n s . A I M E , l 2 j . 2 1 5 . l g 3 g . f sa t

t0l
characof ll9. Stewart, C. R., Craig, F. F., Jr., and Morse, R. A., Determination limestoneperformance teristicsby model flow tests,Trans AIME, 198, 93, 1953. relativepermeability J. Pet. of , to 120. Keelan, D. K., A practicalapproach determination imbibition gas-water T e c h n o l . ,4 , 1 9 9 , 1 9 7 6 . t l 2 l . L e a s , W . J . , J e n k s , L . H . , a n d R u s s e l l ,C . D . , R e l a t i v ep e r m e a b i l i t yo g a s , T r a n s . A I M E , 1 8 9 , 6 5 ,

r950.
Trans.A1ME,216,258,1959. core characteristics, 122. Felsenthal,M., Correlationof k*/k,,datawith sandstone maintenance gravity drainageand gas pressure predictionsincorporating 123. McCord, D. R., Performance LL-370 Area, Bolivar coastal field,Trans. AIME' 198, 231, 1953. on 124. Edmondson, T. A., Effect of temperature waterflooding,Can. J. Pet. Tec'hnol.,10, 236, 1965. 125. Poston. S. W.. Ysrael, S., Hossain, A. K. M. S., MontgomerY, E. F., IV, and Ramey, H. J., Jr.' The effect of temperatureon irreducible water saturationand relative permeability of unconsolidatedsands, S o c .P e t . E n g . J . , 6 , l 7 l , 1 9 7 0 . on 126. Davidson, L. B., The effect of temperature the permeabilityratio of different fluid pairs in two-phase s y s t e m sJ . P e t . T e c h n o l . ,8 , 1 0 3 7 , 1 9 6 9 . , 1 2 7 .S i n n o k r o t , A . A . , R a m e y , H . J . , J r . , a n d M a r s d e n , S . S . , J r . , E f f e c t o tf e m p e r a t u r e l e v e l u p o n c a p i l l a r y p r e s s u r e u r v e s ,S o c . P e t . E t t g . J . , 3 . 1 3 . 1 9 7l . c on 128. Weinbrandt, R. M., Ramey, H. J., Jr., and Cass6, F. J., The effect of temperature relative and p S o a b s o l u t e e r m e a b i l i t y f s a n d s t o n e s ,o r ' .P e t . E n g . J . , 1 0 . 3 7 6 , 1 9 1 5 . in on 129. Lo, H. Y. and Mungan, N., Effect of Temperature Water-Oil RelativePermeabilities Oil-Wet and r S W a t e r - W e tS y s t e m s , P E # 4 5 0 5 , L a s V e g a s ,N e v . , S e p t e m b e3 0 , 1 9 7 3 . Effects on Relative Permeabilities 130. Sufi, A. S., Ramey, H. J., Jr., and Brigham, W. E., Temperature r S o f O i l - W a t e rS y s t e m s , P E # l 1 7 0 1 , N e w O r l e a n s ,L a . , S e p t e m b e2 6 , 1 9 8 2 . o E 13l. Sufi, A. S., Ramey, H. J., Jr., and Brigham, W. E., Temperature ff'ects n Oil-Water Relative Sands, U.S. Departmentof Energy, Technical Report, 12056-35.Defor Permeabilities Unconsolidated c e m b e r .1 9 8 2 . of on t 3 2 .Miller, M. A., and Ramey, H. J., Jr., Effect of Temperature Oil/Water Relative Permeabilities , S U n c o n s o l i d a t ea n d C o n s o l i d a t e d a n d s ,S P E # l 2 l 1 6 , S a n F r a n c i s c oC a l i f . , O c t o b e r5 , 1 9 8 3 . d RelativePermeability,Ph.D. thesis,StanfordUniv., Stanford,Calif., 1979. 1 3 3 Counsil, J. R., Steam-Water . t 3 4 .Muskat, M.,, PhvsicalPrinciples of oil Production, McGraw-Hill New York. 1949. on tensions relativepermeability,paper 1 3 5 Bardon, C. and Longeron, D., Influenceof very low interfacral . a S P E 7 6 0 9 p r e s e n t e d t t h e S P E 5 3 r d A n n u a l M e e t i n g ,H o u s t o n ,T e x . , 1 9 7 8 . relativepermeabilitydata,Trans. 136. Richardson, J. G., Calculationof waterfloodrecoveryfrom steady-state AIME. 210.373. 1951. 137. Johnson, E. F., Bossler, D. P., and Nauman, V. O., Calculationof relativepermeabilityfrom displacement experiments,Trans. AIME, 216. 370, 1959. d s t 1 3 8 . L e v i n e , J . S . , D i s p l a c e m e ne x p e r i m e n t i n a c o n s o l i d a t e p o r o u ss y s t e m ,T r s n s .A I M E , 2 0 1 , 5 7 , 1 9 5 4 . from laboratory data,J. Pet. Techrutl., of 139. Craig, F. F., Jr., Errors in calculation gas injectionperformance 8.23, 1952. 140. Sandberg, C. R., Gourney, L. S., Suppel, R. F., Effect of fluid flow rate and viscosityon laboratory T d e t e r m i n a t i o o f o i l - w a t e rr e l a t i v ep e r m e a b i l i t i e s ,r s n s .A I M E , 2 1 3 , 3 6 . 1 9 5 8 . n 14l. Donaldson, E. C., Lorenz, P. 8., and Thomas, R. D., The effect of viscosity and wettability on oilat water relativepermeability,paper SPE 1562 presented the SPE 4lst Annual Meeting, Dallas, Oct. 2-5, t966. from 142. Geffen, T. M., Parrish, D. R., Haynes, G. W., and Morse, R. A., Efficiency of gas displacement porous media by liquid flooding, Trans. AIME, 195,29. 1952. consolidated cores../.Pet. Technol., on 143. Krutter, H. and Day, R. J., Air-drive experiments long horizontal t2, t, t943. on 144. Morse, R. A., Terwilliger, P. K., and Yuster, S. T., Relativepermeabilitymeasurements small core samplesOil Gas J., 46. 109, 1947. , 145. Odeh, A. S., Effect of viscosityratio on relativepermeability,Trans. AIME, 216,346, 1959. , o 1 4 6 . B a k e r , P . E . , D i s c u s s i o n f e f f e c t o f v i s c o s i t yr a t i o o n r e l a t i v ep e r m e a b i l i t yJ . P e t . T e c h n o l . , 2 1 9 , 6 5 , I 960. 1 59' 1 4 7 .D o w n i e , J . a n d C r a n e , F . E . , E f f e c t o fv i s c o s i t y o n r e l a t i v e p e r m e a b i l i t y , s o t ' . P e t . E n g . J . ' 6 '9 6 1 . materials Trans. Al M E , 246, flow in consolidated E. , A model for two-phase , 148. Ehrlich, R. and Crane, F. 22t, t969. 149. Perkins, F. M., Jr., An investigationof the role of capillary forces in laboratorywaterfloods.J. Pet. T e c h n o l . ,l l , 4 9 , 1 9 5 7 . 150. Pickell, J. J., Swanson, B. F., Hickman, W. B., Applicationof air-mercuryand oil-air capillarypressure S a d a t a i n t h e s t u d y o f p o r e s t r u c t u r e n d f l u i d d i s t r i b u t i o n , o c .P e t . E n g . J . , 4 , 5 5 . 1 9 6 6 . l5l. Warren, J. E. and Calhoun, J. C., A study of waterfloodefficiency in oil-wet systems,Truns. AIME. 204.22. t955.

102

Relative Permeabilin of Petroleum Reservoirs

i o a 1 5 2 . C a r o , R . A . , C a l h o u n , J . C . , J r . , a n d N i e l s e n ,R . F . , S u r f a c e c t i v ea g e n t s n c r e a s e i l r e c o v e r y .O i 1 GusJ., 12. 6. 1952. 153. Ojeda, E., Preston, F., and Calhoun, J. C., Jr., Correlationof residualsfollowing surfactantfloods, Prod.Mon., 12,20, 1953. porous of 154. Lefebvre du Prey, E. J., Factorsafl'ectingliquid-liquid relative permeabilities a consolidated m e d i u m .S o c . P e t . E n e . J . , 2 , 3 9 . 1 9 ' 1 3 . o 1 5 5 . O w e n s , W . W . , P a r r i s h , D . R . , a n d L a m o r e a u x , W . E . , A n e v a l u a t i o n f a g a s d r i v e m e t h o df o r Truns. AIME, 201,275, 1956. determiningrelative permeabilityrelationships, in 156. Kyte, J. R. and Rapoport, L. A., Linear waterfloodbehaviorand end ef'fects water-wetporousmedia, Trans. AIME, 213. 423. 1958. of 157. Richardson, J. G., Kerver, J. K., Hafford, J. A., and Osoba, J. S., Laboratorydeterminations , r e l a t i v ep e r m e a b i l i t y T r a n s .A I M E , 1 9 5 , 1 8 7 , 1 9 , 5 2 . on Measurements Small Core 158. Morse, R. A., Terwilliger, P. K., and Yuster, S. T., RelativePermeability SamplesOil GasJ., 46. 109, 1941. , 159. Labastie, A., Guy, M., Delclaud, J. P., and lffly, R., Effect of flow rate and wettability on water-oil paperSPE 9236 presented the SPE Annual Meeting, Dallas. at and relativepermeabilities capillarypressure, T e x . , S e p .2 l - 2 4 , 1 9 8 0 . 159a. McCaffery, F. G., The Effect of Wettability on Relative Permeabilityand Imbibition in PorousMedia. P h . D . t h e s i s ,U n i v e r s i t yo f C a l g a r y ,A l b e r t a ,C a n a d a ,1 9 7 3 . of I 6 0 . Delclaud, J. P., New resultson the displacement a fluid by anotherin a porous medium, paper SPE 4 1 0 3 p r e s e n t e d t t h e S P E 4 7 t h A n n u a l M e e t i n g .S a n A n t o n i o , T e x . , 1 9 1 2 . a at l 6 l . Fetkovitch, M. J., The isochronaltestingof oil wells, paper SPE 4529 presented the 48th Annual Fall M e e t i n go f t h e S P E , L a s V e g a s ,N e v a d a . 1 9 7 3 . in during immiscibledisplacements porous media. Sot'. t 6 2 . Handy, L. L. and Datta, P., Fluid distributions Pet.Eng./.,, 10.261, 1966. on t 6 3 . Huppler, J. D., Numerical investigationof the effects of core heterogeneities waterflood relative p e r m e a b i l i t,yS o c ' .P e t . E n g . J . , 1 2 , 3 8 | , 1 9 7 0 . 164. Stewart, C. R. and Owens, W. W., A laboratorystudy of laminar and turbulentllow in heterogeneous p o r o s i t y i m e s t o n eT r u n s .A I M E , 2 l 3 , 1 2 l , 1 9 5 8 . l , P s r H 1 6 , 5 . e n d e r s o n , J . H . a n d M o l d r u m , H . , P r o g r e s s e p o r to n m u l t i p h a s e - f l o w t u d i e s , r o d . M o n . , 4 . 1 2 , t949. ., cores,-/. Pet. TeL'hnol on 166. Krutter, A. and Day, R. J., Air-drive experiments long horizontalconsolidated l l. l. r943. P . G R 1 6 7 . C h e n , H . K . , C o u n s i l ,J . R . , a n d R a m e y , H . J . , J r . , S t e a m - W a t e r e l a t i v e e r m e a b i l i t y1 9 7 8 e o t h e r m a l R e s o u r c e C o u n c i l A n n u a l M e e t i n g ,H i l o , H a w a i i , J u l y 2 5 - 2 7 , 1 9 1 8 . s fluids, 168. Brownell, L. E. and Katz, D. L., Flow of fluids through porous media - single homogeneous C h e m . E n s . P r o s , . ,4 3 ( 1 0 ) , 5 3 7 . 1 9 4 ' 7 . p . r 1 6 9 . W a l l , C . G . a n d K h u r a n a , A . K . , S a t u r a t i o n e r m e a b i l i t ye l a t i o n s h i p alto w g a ss a t u r a t i o n . J . l n s t P e t . , 5 1 .2 6 1 .1 9 7 1 . f g s 1 7 0 . R o s e , W . D . , F l u i d d i s t r i b u t i o n c h a r a c t e r i z i ng a s - l i q u i d l o w , T r a n s . A I M E , 1 9 2 , 3 1 2 , 1 9 5 1 . - l 7 l . L o o m i s , A . G . a n d C r o w e l l , D . C . , R e l a t i v e e r m e a b i l i t y t u d i e s I I . W a t e r o i l s y s t e m sP r o d . M o n . , 8 . . p s . r8. 1959. tests, 172. Sarem, A. M., Significanceof water-oil relative permeabilitydata calculatedfrom displacement o , P r o < ' . ,T h e o r y o f F l u i d F l o w i n P o r o u sM e d i a C o n f e r e n c eU n i v e r s i t y f O k l a h o m a .N o r m a n , 1 9 5 9 , 1 8 9 . and of 173. Owens, W. W., Parrish, D. R., and Lamoreaux, W. E., A comparison field k*/k,,characteristics at laboratoryku/k,,test results measuredby a new simplified method. paper 518-G presented the AIME 3 0 t h A n n u l M e e t i n g , N e w O r l e a n s ,1 9 5 5 . 174. Calhoun, J. C., Jr., Fundamentalsof ReservoirEngineering,University of Oklahoma Press,Norman, t94'7. performance characteristics of | 75. Stewart, C. R., Craig, F. F., and Morse, R. A., Determination limestone by model flow tests. Truns. AIME, 198, 93, 1953. 176. Kyte, J. R., Stanclift, J. R., Stephan, S. C., Jr., and Rapoport, L. A., Mechanismof waterflooding of in the presence free gas,Trans. AIME, 101, 215, 1956. p d 1 7 7 . M a t t a x , C . C . a n d C l o t h e i r , A . T . , C o r e A n a l y s i so f U n c o n s o l i d a t ea n d F r i a b l eS a n d s , a p e rS P E 4 9 8 6 p r e s e n t e d t t h e S P E 4 9 t h A n n u a l M e e t i n g ,H o u s t o n ,T e x . . 1 9 7 4 . a on 178. Holmgren, C. R. and Morse, R. A., Effect of free gas saturation oil recoveryby waterflooding.Trans. A I M E , 1 9 2 , 1 3 5 ,1 9 5 1 . 179. McCaffery, F. G., The Effect of Wettability on Relative Permeabilityand Imbibition in PorousMedia. Ph.D. thesis, University of Calgary, Alberta, Canada, 1973. 180. Gornik, B. and Roebuck, J. F., Formation Evuluetion through Extensive Use of Core Analysi,s,Core L a b o r a t o r i e sI,n c . , D a l l a s ,T x . , 1 9 7 9 .

103
Chapter4 THREE-PHASE RELATIVE PERMEABILITY

I. INTRODUCTION Recent innovationsin the field of oil recovery have led to a renewedinterestin threephaserelative permeability. Three-phase flow occurs when the water saturationis higher than the irreducible level, and oil and gas are also presentas mobile phases.Detailed engineering of calculationsof the performance reservoirsunder recoverymethodssuch as carbon dioxide injection, in situ combustion,steamdrive, micellar fluid injection, and relativepermeabilitydata. nitrogen injection frequentlyrequirethree-phase All factors which influenceflow in systemscontainingtwo mobile phasesalso apply to potentialthree-phase systems, constitute systems.Virtually all oil reservoirs three-phase sincereservoirrocks invariably contain interstitialwater, and naturallyoccurringoils completely devoid of gas are rare. In fact, a two-phasesystemof oil and gas may be regarded as a three-phase systemin which the water phaseis immobile. The numberof reservoirs in which oil, gas, and water are simultaneously mobile during primary productionis probably small. Nevertheless,three-phase mobility is always possible when a producing interval includespart of the oil-water transitionalzone in a reservoir.It is probable,however, that in most caseswhere oil and free gas are producedwith an appreciable water cut, the water is being producedfrom layers of the reservoir in which relative permeabilityto water is high and not by true three-phase flow. relative permeabilitydata for conventionalreservoir In the past, the use of three-phase In engineeringcalculationshas seldom been necessary. consequence, considerablyless is known about three-phaserelative permeability characteristics rocks than is known for of comparabletwo-phasecases.The realizationthat detailedengineeringcalculationsof the performance reservoirs producedby in sitrzcombustion processes of requirethree-phase data is quite new. Three-phase relativepermeabilityis useful in the calculationof field performwaterandgasdrive, and alsoin analyzing ancefor reservoirs beingproduced simultaneous by solution gas drive reservoirs which are partially depletedand are being produced by water drive. An increasinginterestin three-phase flow phenomena anticipated. is perrneabilitydata: ( I ) that pertaining There are two distinct classesof three-phase relative to drainage;and (2) that pertainingto imbibition. Drainagerefersto the direction of saturation changein which the wetting-phase Imbibition refers to an increasing saturation decreases. wetting-phase saturation.For the relative permeabilitydata to yield correct reservoirpredictions, the directionof saturation changein the reservoirmust correspond the direction to of saturationchange for which the data were derived. Drainagerelative permeabilitydata should be used in the following situations: l. Enhancedrecovery processes involving the injection of dry gas, flue gas, carbon dioxide, and other gasesinto watered-outreservoirs. in Miscible flood processes which liquified petroleum gas (LPG) is injected into watered-out reservoirs. Productionfrom reservoirs which the water saturation greater in is than the ineducible saturation. Imbibition relative permeabilitydata should be used under the following conditions: l. Reservoirsproduced by natural water drive.

2.
3.

104

Relative Permeabilin of Petroleum Reservoirs

I OOI

w!tor

oil

FIGURE l.

Three-phase relative permeability.r

2. 3.

Reservoirs developed water flood, as well as by processes by wherethe injectedwater containssurfactants, polymers, or other additives. Reservoirsdevelopedby recovery processes where water is used to push a slug of chemicals, LPG, etc.

II. DRAINAGERELATIVEPERMEABILITY
A. Leverett and Lewis Much of the credit for the classicalwork in three-phase relativepermeabilityis accorded to Leverettand Lewis' who were the first to measure three-phase relativepermeabilityof a water-oil-gassystem in an unconsolidated sand. These investigators used a steady-state single-core dynamic method and ignored end effects and hysteresis.Errors from ignoring capillary end effects were probably significant,since low flow rateswere used. Ring electrodes were spacedalong the length of a sand pack to measureresistivity of the sample and brine saturationwas assumed be directly relatedto resistivity. Gas saturation to was determined from pressure and volume measurements. sdturation Oil was obtainedby a material balancetechnique.Leverett and Lewis obtainedthree separate triangulargraphs showing lines of constantrelative permeability("isoperms") to the three phases; thesewere plotted againstthe saturations the three fluids, as shown by Figure l. They also obtaineda plot of showing the region of three-phase flow; Figure 2 shows the region where each component comprisesat least5Voof the flow stream.As shown by the figure, three-phase flow occurs in a rather confined region. Relativepermeabilityto water, k,*, was found to be dependent only on water saturation, S*, and was not affected by the introduction of an additional nonaqueous phase. Relative permeability to gas, k,r, was found to be slightly less than would be expectedfor the same gas saturation, S* in two-phase flow. The k., isoperms are convex towards the 1007oS, apex of the triangulardiagram. As gas becomes one of the two flowing nonwettingphases, whenboth oil and waterarepresent, relativepermeability gasdecreases oil saturation the to as

105
lOO%gas

100% water

lOO%oil

FIGURE 2.

Region of three-phase flow.l

approaches water saturation the value, becominga minimum when roughly equal saturations of oil and water are associated with the gas. The relative permeability to oil is seento vary in a more complex manner. Starting with a gas saturation zero, oil relative permeabilityat constantoil saturation of increases gas as (except at low oil saturations saturationincreases where k,, remainsconstant)then decreases to its original value as more gas is introduced,finally falling well below this value when gas saturationis further increased.In a water-wet system,the presence gas leavesthe of mode of water flow unchanged,but since the gas tends to occupy the central portions of (where the oil is also driven by capillary forces)interference the intergrainspaces between oil and gas flow is likely. Visual examinationunder the microscope showsthe presence of an oil film (in some casescontaininga very small amountof finely divided water) through which oil flows aroundeachgasbubble.It is not clearwhetherall gasbubblesare connected. However, the gas bubblesare observed move jerkily, as opposed the generallysmooth to to flow of water (and of oil when gasbubblesare absent are stationary). or This unevenmotion of the gas implies a similar motion of at leastpart of the oil, which would be expectedto move faster than in the absenceof gas at the same oil saturation.We see a decreasein k,. at constantS" as S, is increased, especiallyat low S*. Also, there is an increasein k.o at constantS" as S* is increased low values of S*. at This effect is evidently due to ihe shifting of oil into parts of the intergrain spacewhere it may flow more freely. The water introduced tends to occupy the sharply curved parts of the pores,forcing oil into the centralspacevacated gas. Sincefluid in the sharplycurved by parts of the poresmoves only with difficulty and that in the centermoves more readily, the result is an increasein k,o. Leverettand Lewis pointed out that they found no effect of oil viscosityon the isoperms for various saturationsof the three phases. B. Corey, Rathjens, Henderson, and Wyllie The resultsof the work of Corey et al.2are shown by Figure 3. Theseinvestigators used a calcium chloride brine. Capillary end effects were minimized by using a core with semi-

106

Relative Permeabilitv of Petroleum Reservoirs

gas FIGURE 3. relative permeability.r Three-phase

gravimetricallyand saturations permeable membranes mountedat eachend. They measured rather than resacores for each measurement effects by using separate avoided hysteresis of turating the samecore. In an initial conclusion,they reportedthat when the saturations the wetting phaseswere equal, the nonwettingphase relative permeability,k,n, was unregardless whetherthe nonwettingphasewas oil or gas.They usedthe equivalent of changed liquid permeabilityas the basevalue. The oil isopermsof Corey et al. are similar to those obtainedby Leverettand Lewis, exceptthat Corey's oil isopermshave a greatercurvature. that on but Relativepermeabilityto water was not measured, was calculated the assumption aloneand that waterpermeabilityin a water-wetsystem it was a function of water saturation was the sameas the oil permeabilityin an oil-wet system.It shouldbe noted that the data that the behavior of Cltrey et al. are for oil drainagein an oil-gassystem.They alsoobserved of the nonwetting phaseswas more sensitiveto changesin pore geometry than was the in in behaviorof the wetting phase.The increase k." (at low S*) with the increase S* (and samplesthan sandstone in decrease S*) is higher in Corey's consolidated a corresponding of of samples.This is because the dependence k..,on the ratio in the unconsolidated ft' I dsl/P.r

JS*

l,

dsL/P:

which is usually higher in consolidated rocks than in unconsolidated rocks. Corey et al. extendedtheir two-phaserelative permeabilityrelationshipto three-phase flow on the basisof the following approximation:

107

N
2
FIGURE 4. Three-phase relative permeability.r

I
Sr_,

l/Pl : g

IS,_-(S*,..+S,,.)]

for S, t '

St' (soi,, + s.,,)

t''=, :ofbrS.s'^ (S*,,, * S",)

(l)

The drainage phaserelativepermeabilityin a water-wetsystemcontaininggas is given oil bv

(2)

where S., is residualliquid saturation. As in Leverett's data, the oil isopermstend to be parallel to the oil isosaturation lines, especiallyat high S*. At increasingS* and constantS.,, the gas which was previously in the systemis no longerpresent.Thus, the rate of increase k..,with increasing decreases of S* at higher valuesof S*. Corey et al. proposeda methodto obtain k.,,and k,*, basedon k.* alone. Incidentally, k.* was found to be a function of S* and independent the relative of wetting propertiesof the fluids within the rock. C. Reid Using the samemethodemployedby Leverettand Lewis (single-core dynamictechnique), Reid3obtaine{ the isopermsshown in Figure 4; He eliminatedend effects, but hysteresis was ignored.'frine saturation was measured resistivity, by and'oil and gas saturations were obtainedby gamma ray absorption.His saturation measuremenis possiblywere affectedby differentialabsorption gamma rays by oil and water. While Leverettand Lewis obtained of

108

Relative Permeability of Petroleum Reservoirs

straightlines for the water phasebehavior(showingk,* to be independent the distribution of of the nonwettingphases) and oil isoperms concavetoward the l}OVaSoapex, Reid's results indicatedconcavewater isoperms,convex oil isoperms,and slightly concavegas isoperms. These results were interpreted as indicating that the relative permeability to each phase is dependent both upon its own saturation and the saturations the other phases.His results of showeda greateroil permeabilitywhen threephases were presentthan with two phases,at a given oil saturation. Reid made no attempt to correlate the three-phaseresults with those from two-phase experiments.He placed emphasison his conclusionsfor the oil isopermsand noted that Leverett'soil phasedata showeda substantial amountof scatter.For this reason,he believed that his oil isopermswere more valid than Leverett's.The work of Rose seemsto confirm Reid's findings. D. Snell Three-phase behavior in a water-wet unconsolidated sand was investigated Snell,a-6 by who used radio frequency detection for the determination of S* and a neutron counting methodfor measurement Sr. Oil saturation of was obtainedby materialbalancecalculation. His experiments had a repeatability within Ilr%o relativepermeability for values,with a better repeatability the saturation for values.He found that when the wetting phasesaturation was uniform over a length of the test sample,the saturations the other two phases of were also uniform over the samelength. Although Caudle et al. " did mention hysteresis their work, the first significantstudy in on the effect of saturationhistory on three-phase relative permeabilitywas done by Snell. In describingSnell's work, it is convenient definefour typesof liquid saturation to histories: l. 2. 3. 4. Imbibition of water with oil saturation increasing(II). Imbibition of water with oil saturation (ID). decreasing Drainageof water with oil saturation increasing(DI). (DD). Drainageof water with oil saturation decreasing

As seen from his results in Figure 5, k.o values were lower for DD than for the other saturation histories.Since, in two-phaseflow, drainagecausedthe wetting phapeto lose its mobility at highersaturations, hasbeensuggested thereis a partialchange wettability it that in from water-wetto oil-wet during DD. When the systemwas oil-wet, a largerS,,was required for the samek,., because someof the oil was trappedin the smallerpores.This oil increases S.,,but it is immobile. He further suggested that this changein wettability may be caused by polar compoundsin the oil. Snell's resultsdo not show good agreement with those of Leverettand Lewis except in the caseof k.*. Oil and water isoperms reported by Snell are similar to those determinedby Reid, but Snell's k.. valuesare higher than Reid's, especiallyat low water saturation. In a later work, Snell reinterpreted resultsof four earlier studiesdoneon unconsolidated the sands.In these investigations,no hysteresis was found for water isoperms.Oil isoperms showed hysteresisonly when keroseneor a kerosene/lubricating mixture was used as the oil oil phase.Nonpolaroil gaveno hysteresis. Reinterpretation the earlierresultswas possible of because Leverettand Lewis indicated possible enors in their saturation measurements. Reid's saturationdata might also have been inaccuratebecause differential absorptionof gamma of rays by oil and water. Relativepermeabilityto oil was found to be dependent only on the historiesof the liquid phasesaturations, althoughSnell did not rule out dependence gas on phasesaturationhistory. Snell reinterpreted Leverett'sdata to obtain oil isopermsconvex toward the l00%oS,,apex. Oil isopermsthen followed the samepatternin all four investigations.Theseresultsareshownin Figure6. The curvature the isoperms both nonwetting of of

109

I OO*

watet

10Oi

oil

water

10OS water

M
'l0O% gas

R)
oil I OOi oil

roOS

gas

FIGURE 5.

Three-phase relative permeability.5

100% gas

lmbibition _ D ra ina ge - - - -

Rsults of Snell

1OO%water

/I,a'

il!to,Dt oD

_ ----

100% oil

100% water

10 0 %

1O0% oil

non-polar

1OO%water

1O0% oil

'l

OO% water

100% oil

FIGURE 6.

Reinterpretation resultsby Snell.6 of

ll0

Relative Permeabilin of Petroleum Reservoirs

---

_ 4 o :;,--=-- _ ) .
\ Gas iniectsd, pore volumes

i---- to-..\..
(L E

zj+

o o o
O_ o .z

-'--\:\:1..,, i /,
[--'"-\..

.1

F---,:-:N*,
*-'
^-.5
< - L

-l
I

-A

L-.25

---A

J . 1

0
O tNtlAL wATERsATURATtoN

100
0

100
FIGURE 7.

{-

rNrrrALrL sATURATToN o

Fluid flow experimental aata for Berea sandstone.T

(oil phases and gas)are convextoward the corresponding phase-apex, whereas wetting phase isopermsare straightlines or are concavetoward the l00%o apex of the wetting phase. E. Donaldson and Dean An extensionof Welge's two-phase unsteady-state technique was usedby Donaldsonand DeanTto determinethree-phase relative permeabilities Berea sandstone of and Arbuckle limestone.Oil and water in the core were displacedby gas and the flow ratesof all three phaseswere measuredsimultaneously. Their resultsfor the displacement testson the two cores startingwith various S*, and S.,,are shown in Figures7 and 8. They minimized end effects by using a high pressure differential and high flow rates, and they did not account for hysteresis effects. The volumes of oil and water displacedwere less in the limestone than in the sandstone the same S,,(or S*) and the samepore volumes of gas injected. for This effect is presumably caused the largerflow channels the limestone. by in The efficiency of a gas displacement processis greaterfor a matrix with smallerpores.There is a narrower rangeof saturations three-phase for flow in the limestonebecause large vugs may allow the gas to flow without transferingenergy to oil or water. The isoperms presented functionsof terminalratherthanaverage are as saturations, because the former govern the flow of fluids throughthe core. The resultsof Donaldsonand Dean, shown in Figures9 through 14, indicatethat, at low and constant k., for Bereasandstone S' initially decreased with increasing until S" reached valueof about50Vo. S" a Furtherincreases in Socausedan increase k,r. At S* greaterthan I 3Vo,k," increased the isoperms in so became concavetoward the gas apex. No explanationof this phenomenon was suggested the by authors.At a given S* the k., was lower in the presence water than in the presence of of oil, probably because water adheredmore stronglyto the rock surfacethan did the oil. The flow path of gas is more restrictedin the presence water, sincegas can displaceoil more of easily than it can displacewater. For the limestone,k,, was alwaysconcavetoward the gas apex.

lll

DataPoints:

.4
A)

oE q) O (g O A -

.3
t

e "oo,
.r, /-oas Inlecteo. pore votumes

o U, o

A .Z

i5 o

o o -o_ !

A'

100
FIGURE 8.

INITIAL OIL SATURATION

Fluid flow experimentaldata for Arbuckle limestone.T

uw
FIGURE 9.

so
Gas relative permeabilityfor Berea sandstone.T

The water isoperms are concave toward the water apex. Relative permeability to water was generallyhigher in the presence oil than in the presence gas, but k.* was higher of of in the presence gas than in its absence a constanthigh S*. Both k." and k.* increased of at

tt2

Relative Permeability of Petroleum Reservoirs

o^ ao/

sw
FIGURE IO

so
Gas relative permeabilityfor Arbuckle limestone.i

A6

ieo

1o

sw
FIGURE I l.

so
Oil relative permeabilityfor Berea sandstone.T

at constantS,,and S*, respectively,when S, was increased from 0 to 8Vapossiblybecause gas was trappedin poreswhich would otherwisebe occupiedby immobile wetting phases. Also, k.* increased the presence oil because in of there may be partial oil wetting, so that water was displacedinto larger pores;this was not the casewhen gas was present.

113

sw

so

I F I G U R E 2 . Oil relative permeabilityfor Arbuckle limestone.T

ss

-w

oo

f F I G U R E 1 3 . W a t e r r e l a t i v ep e r m e a b i l i t yo r B e r e as a n d s t o n e ' -

F. Sarem data for a Bereacore. He obtainedthree-phase method, Sarems Using an unsteady-state history, but his methoddid accountfor wettability. did not considerend effectsor saturation technique,is relatively fast. Sarem'smethod, which is an extensionof Welge's two-phase

tt{

Relative Permeabilin of Petroleum Reservoirs

sw

so

FIGURE 14. Water relative permeabilityfor Arbuckle limestone.T

first with one liquid and then flooded with an immiscible unreactive The core is saturated by Then, both liquids are displaced gas. In the derivation breakthrough. liquid, at leastuntil each relative permeabilityto be a function of the colreSarem assumed of his equations, lines. The spondingsaturationalone. Isopermswere thereforeparallel to the isosaturation on permeability to gas was assumedto be dependent total liquid saturationand relative of independent the relative wetting properties. The saturationequationsare
Soz : S*z : S..o,, * f", Q

(3)
(4)

S*.ou* *

f*, Q

S r r : 1 - S * r - S . , 2

(5)

where Q : cumulativevolume of injectedfluid (per pore volume)

e : 9 LAO rt
and f : fractionalflow. and q, : total volumetric flow rate (cclsec),t : time (seconds), The relative o, Subscripts w, g, and 2 standfor oil, water, gas, and outlet, respectively. permeabilitiesare computed from the following relationships: d (l/Q) k.* : I*z -l-4pp4 1 tl l I \L p " q' Q/ (6)

115

k.* Or,,H
d ( l/Q) : K"' t"'(-4r!4-r*,, o,J

(7)

(8)

Saremalso concludedthat initial saturation conditionsaffect k..,and k.*, but havelittle effect on k,r. He found that k."/k.* was influencedby initial saturations three-phase in studiesin the samemanneras in two-phasestudies.Sarem'sresultsdiffered from thoseof Donaldson and Dean even though both used the sametype of sandstone. G. Saraf and Fatt (NMR) techniques A dynamic method using nuclearmagneticresonance was used by Sarafand Fatteto determineliquid saturations Boise sandstone. volumetricmethodwas in A used to obtain gas saturations. The experimental techniquewas designedto minimize end effects.To maintaina constantpressure differential,the gas flow rate was increased the as justification for Sarem's oil flow rate was decreased. Saraf and Fatt found no theoretical assumption that three-phase relative permeabilityto each phasewas a function only of the saturation that phase.In the water-wetBoise sandstone, of however, they did find that k,* was a functionof S* alone.Using waterpermeability the base,they found thatk,* depended as only on the total liquid saturationand was independent the relative wetting properties. of Oil isopermsdeterminedby these investigators were convex toward the oil apex. Their resultsare shown in Figure 15. The explanationgiven by the authorsfor this unexpected shapeof the isopermsseemsless than convincing.They did state,however,that in studies where k,* was a function of both S* and S", the systemwas not 1007awater-wet.In such a case,it seemslikely that S* did not remain constant when Soor S* was increased and that the assumption constantS* could be a sourceof experimental of error. H. Wyllie and Gardner Three-phase relative permeabilityequationsfor preferentiallywater-wet systemswhere water and oil saturations were determinedby the drainagecycle rather than by imbibition have been given by Wyllie and Gardnerro and are presented Chapter 2, Table 3. The in following factors should be taken in consideration when using the equations presented into this table. l. 2. The k,* valuesare normalizedto absolutepermeability. The values of k.o and k., calculatedfrom theserelationships both normalizedto are the effective hydrocarbon permeability at irreducible water saturation. Inasmuch as they are normalized to the same base, k.s/k.. values may be calculateddirectly by using these equations.This is not true, of course, for water-hydrocarbon relative permeabilityratios. The gas and oil relative permeabilityequations not include provision for residual do oil saturation.When S* equals S*i.,, k,o is equal to [S"/(l - S*,.,)]ofor cemented sandstone, oolitic limestones,and vugular rocks. To handle residualoil saturation, this relationshipshould be alteredto [(S" - S.,,)/(l - S*,'..)].0

3.

The correlations developed by Wyllie and Gardner can be used to construct a ternary diagramshowing the relative permeabilities oil, gas, and water. In general,the valuesof to relative permeability(10, 20,30Vo, etc.) are chosenfirst and then the valuesof saturation are obtained from the correlations. As can be seen from Chapter 2, Table 3, some of the

r16

Relative Permeabilin of Petroleum Reservoirs 10 0 % g a s

orL

100% water

100% oil

gas

WATER

water gas

GAS

water

oil

FIGURE 15. Three-phase relative permeability.'

equations nonlinear.Hence, numericalmethods(suchas Newton-Raphson) required are are to solve theseequations.Manual interpolationis also possiblefor plotting relative permeability isoperms.

tl7

WATER

wut6l olL GAs

FIGURE 16. Three-phase relative permeabilitydata of Caudleet al.rr

Empirical relationships provide reasonable resultsin some casesand very disappointing ones in other situations;consequently, they must be used carefully. Note that most of the previousrelationships were developed media with intergranular for porosity.This points out the huge problem of determining relative permeability curves for naturally fractured reservoirs. The difficulty arisesprimarily from the difficulty (or impossibility)of making this type of measurement a fracturedcore sample. on For totally oil-wet three-phase systemsin which oil is the wetting phase, water the nonwettingphase,and gas nonwettingwith respectto both, the substitution S" for S* in of the Wyllie and Gardner equationscan be made for estimationof the relative permeability to each phase.

III. IMBIBITION RELATIVEPERMEABILITY


A. Caudle, Slobod, and Brownscombe Using a dynamic displacementmethod on a consolidated core sample, Caudle et al.rl obtainedisopermsfor k.o, k.*, and k,*, as shown in Figure 16. They useddistillationto find the water and oil saturations eachdata point, and usedmaterialbalancefor determination at of gas saturation.Caudle et al. employed a pressuredifferential of 5 to 50 in. of water across core and usedwater permeabilityas the basevalue. Relativepermeabilityto water the k.* was found to be dependent S*, Sr, and S". These workers recognizedthe presence on of some form of hysteresisin the three-phase studies,but they ignored the capillary end effect. They found all relative permeabilities be approximately minimum valueswhen to at So was maintainedat the value of S*.. B. Naar and Wygal Naar and Wygal12developeda set of equationsthat was discussed Chapter2. Based in

l18

RelativePermeabilin of PetroleumReservoirs

water-wet

S' w
m F I G U R E 1 7 . T h r e e - p h a sie b i b i t i o n . r r

S;

at reducedsaturations the apexes,as on theseequationsthey plotted isopermswith 1007a that at the beginning indicated mechanism shown in Figures17 and 18. The displacement of increased the expense S, at at of the imbibition process,S** (reducedwater saturation) at constantS", until no more gas was trapped.Thereafter,S* increased the expenseof S., at constantS*. This path is tracedin Figure 17. The locus of all such paths is also shown. Unlike the findings of other workers, Naar and Wygal concludedthat k,.,/k,* is not a flow. On the other hand, the function of S*, for equal valuesof oil recoveryin three-phase is ratio was found to be a function of Sr, and wettability. This dependence shown in Figure the 19. The higher the initial gas saturation, lessthe influenceof wettabilityon k.,,/k.*.Also, and initial at the water saturation a given recoverywas a function of initial water saturation gas saturation. The ratio of S*, for a water-wetsystemto S*, for an oil-wet systemincreased was an incresingfunction of S*,. For a given S* ratio and with Sri, and the rate of increase Sri. with increasing With higherS*,,thereis lesspore space a given recovery,S*, decreased of availableand the oil is alreadypushedout into larger channelsbecause the higher S*i; therefore,less water is requiredfor the samerecovery. developedby Naar and Wygal, The imbibition water-oil relative permeabilityequations that l/P.2 equalsCS*, are basedon the assumption
""intt' So,i,,,.imb /s*

S * * . , n-b s * d s * , P:

kr*.irrb

f' t - s* ds* J,, P:

(e)

119

o2

water

FIGURE 18. Three-phase drainage.rl

and
kr,r.irrb

: s;:i(s..,, 3 Si") +
+ S"o)t(S,, + 2S.,n 3S*, - 3S*') ( I - S * ' )"

(s,,

(l0)

Naar and Wygal suggestedthe following approximationfor imbibition gas relative permeability: k,* where
S**,.i-t,
: S**.druin -

0.5

S**i.in,r,

0.5

( I - S*i'",n)

(ll)

l/2 S::, drain;

S* :

S _ S* '. l -S * ,

s.l,:S"-S"t' l - S * ,

Sl*-S*lS*'
l -

S.t,

S*i

and S"ois the trappedoil saturation.

t20

Relative Permeability of Petroleum Reservoirs


1 A

S*1r,= o'3

't2
o 3
I (!

3 ro = 3 ; ;
L 8
-l

s w i r r - o . ' l5

^t^ (Dl
{ o =
o

I o

u) l { l= lo:

{ lq - l t lr
l o

.1

.2

.3

.4

.5

.6

INITIAL ASSATURATION G FIGURE 19. Influenceof wettability at 40o/c recovery.'l

Thesemodelswere derivedby assuming randominterconnection straightcapillaries, the of with a provision for blocking of the nonwettingphaseby the invading wetting fluid. The imbibition water, and drainage and gas, relativepermeabilittequations oil developed by Naar and wygal were also presented the following form: in
kr*,i,r,b : (S;)o

(t2)
(l 2S*x;,,2t

k,..i*u :

(l -

2S*x;t't {2 -

(l 3 )

and

k,* s_ir: - 2srr)


where

(r4)

SF -' , : s F - s * '
l - S * ,

"t" In theseequations subscript the stands "trapped" and,,f" for.,free,, for C. Land In Land's13work, equationsfor imbibition two- and three-phase relative permeabilities were obtainedfrom rock properties. Land considered residualgas saturation after imbibition to be directly related to the initial gas saturation.The gas and water imbibition relative permeabilities were reportedto be the samein three-phase systems in two-phase as systems,

12l
Gas

orL

Water
FIGURE 20. Imbibition k.,,for a mobile gas saturation.rl

since the totally nonwetting and wetting phasesoccupiedthe samepores regardless the of natureof the other phasespresent. His plots fork.o in the II and ID casesare shown in Figures20 and21. The ID plots are similar to the plots obtainedby Naar and Wygal,12 their systembeing an II case. Land's final equationsare ^*r It ds*

5 i I ts'Jr*sgr p:

k,*

It
_ -

(ls)

t. Kr*

II''E

It

(l6)

k.. :

r,'fl.,": Y f'ds*
J,, P:

(t7)

For S* increasingand S* constant: k.o : S:i [2(S** * S",*) * S"r*] (18)

122

Relative Permeabilitv of Petroleum Reservoirs


Gas

GAS

Water

oil

F IGU R2 I. E

Imbibition k,,,for a trappedgas saturation.rl

This equationis similar to the one obtainedby Corey et al.2 for the drainagecondition. When all the gas is trapped: k.. : where S* : (l _ S*r*) S. - S.l-S*, -S,,,.

s:i(2s** +

S"r*) - S,,r*[S*i.+ 2/C(S;, + llC{lnSr,/Sr,})J

(l e )

S .* :

S", Sr,* : l -S *,
Sot

S.u* :

l-s*,
S. - Sru l-S*, I (sr,* ),.u*

S.r* :

(--

123

s**:
s
v()m

S* - S*. l-S*.

: minimum residual saturation oil


trappedgas saturation trappedoil saturation

Sr, : S.r, :

Land's correlations not considerhysteresis did sincehis derivationwas basedon the work of Corey et al., which did not include hysteresis effects. D. Schneider and Owens Schneiderand Owensraperformed steady-state and unsteady-state tests on a variety of carbonateand sandstonesamples, and found the relative permeability to oil during an imbibition processin a water-wet systemto be insensitiveto the flowing gas phasewhen gas saturation was increasing. Oil relativepermeabilitywas found to be primarily dependent on oil saturation.It was reported that residual oil significantly reducedthe gas relative permeabilityin a water-wetsystem.The gas relativepermeabilityin an oil-wet systemwas found to be insensitiveto the presence a residualoil saturation. of The nonwettingrelative permeability-saturation relationshipin three-phase flow was reportedto dependon the saturation history of both nonwetting phasesand on the ratio of the saturations the two of wetting phases.In some casesthe nonwettingrelative permeabilitywas found to be lower then the two-phasevalue due eitherto trappingof a nonwettingphaseor to flow interference between the nonwetting phaseswhen both were mobile. For some tests the nonwetting relative permeabilityvalue for three-phase flow was found to be higher than the two-phase value. The authorsdiscussed reasons the why their resultsdid not fullv asreewith thoseof Corey et al. B. Spronsen The centrifugemethod, alreadyproven for two-phaseflow, was extendedby Spronsenrs to drainagethree-phase flow in a water-wetsystem.Oil isopermsdetermined Spronsen by are concavetoward the l00Vo oll apex. He discussed adverseinfluenceof immiscible the CO, injection on the shapeof three-phase isoperms.The resultsof his investigationare oil shown in Figure 22.

IV. PROBABILITY MODELS


Sincethe experimental problemsassociated with three-phase flow aredifficult to surmount, a mathematical model appears be an alternate to approach. The correlations discussed earlier required some type of experimental three-phase flow data. On the other hand, probability models as formulated by Stoner6'r7 and modified by Dietrich and Bondor'8 and later by Nolen as cited by Molina,'e assume that two-phase flow behaviorcan usedas a limiting be condition for three-phase flow. Water-oil-gas flow can be boundedby water-oil flow at one extremeand oil-gas flow at the other. While someof thesemodelscan considerhysteresis, the earlier correlations,such as Land's,r3cannotdo so. Most probability models assumethat gas relative permeabilityis dependent only on the gas saturation:
kry k., (Sr)

(20)

124

Relative Permeabilin of Petroleum Reservoirs gas

water OIL ISOPERMS

gas

.0o002 - o o o1 -ooo5 - o o1

water WATER ISOPERMS


FIGURE 22. Data of Spronsen for Berea sandstone.r5

oil

Similarly, it is assumedthat the relative permeabilityto water is dependent only on the water saturation: k,* : k.*(S*)

(2t)

Oil relative permeability,however, varies in a more complex manner.Theseassumptions have been confirmed in laboratoryinvestigations a water-wetsystem. for In a water-wet system, gas behavesas a completelynonwettingphase, but oil has an intermediate ability to wet the rock. The relativepermeability oil in a water-oil-gas to system

12s
will thereforebe boundedby relative permeabilityto oil in a water-oil systemat low gas saturations and by relative permeabilityto oil in a gas-oil systemat low water saturations. Stoneattempted combinethesetwo terminalrelativepermeabilities obtaina three-phase to to result by using the channel flow theory in porous media and simple probability models. Water and gas three-phase relativepermeabilities, accordingto Stone,are the sameas their corresponding two-phaserelative permeabilities.In his first model, Stone developedthe expression:

k.. : s;P",F*
where S.,* : S. - S.,, l-S*,-S.. k_. P*: r=; (2-Phase)

(22)

S*.*

s * - S*i
l-S*, -S.,

9, : +T
and

(2-phase)

S,:

s .[-s*, -s..-Sr.

Fayersand Matthews26 suggested that S.. : where : lr S ' I -S*.-{ c{.S,,,* + (l - o,) S,,.o

Stone'searliermodel did not agreewell with datainvolving the dependence waterflood of residualoil saturationon trappedgas saturations. Stone's secondmodel gave three-phase oil relative permeabilityas k,,, : (k..,* + k,*)(k",* + k,s) - (k.* + k,s) (23)

where k,o* and k.* represent and water relativepermeabilities oil from two-phase, oil-water relativepermeabilitydata;k,o,and k,* represent and gasrelativepermeabilities oil from twophase,oil-gas relative permeabilitydata. Equation 23 may yield unrealisticresultsat low k.o values. Although it seemsreasonable that one should be able to combine the two two-phase relativepermeabilities arrive at three-phase to dataat leastfor water-wetsystems, manner the in which they have been combined in thesemodels may not accountfor the total physics of the process.Theseprobability modelsstronglydependon the assumption that there is at most one mobile fluid in any channel.That is, Stone's assumption implies that water-oil

126

Relative Permeabilin of Petroleum Reservoirs

capillary pressure and water relative permeabilityare functionsof water saturation alone in the three-phase system,regardless the relativesaturations oil and gas. Moreover, they of of are the samefunction in the three-phase systemas in the two-phase gas-oil system.Stone's second model generallypredictsthe correctoil relativepermeability the three-phase in system if the relative permeability at the end points is equal to one. Stone points out that when his secondmodel yields a negativek,,,, this implies a completeblockageof oil and as a result k." equals zero. The Stone models accountfor hysteresis when water and gas saturations are changingin the samedirection. Dietrich and Bondorr8applied Stone's models to publishedthree-phase data and found them to be only partially successful. They found that it was necessary modify Stone's to secondmodel for the casewhere gas/oil relative permeabilityis measured the presence in of connatewater. They pointed out that, at irreduciblewater saturation and zero gas saturation, Equation 23 reducedto k,. : (k..,*)(k.,,*)

This expressioncan be valid only if both k..,* and k.,,, equal unity. Since k,,, at S*. is frequentlyless than one, Stone's secondmodel has some limitations. Dietrich and Bondor adjustedStone'smodel by normalizingit with k..,.* to obtain:

k..:

fr

(k.", * t,0,"* k,*) * k.r)l (k.*+ k.g)

(24)

where k.o.* is the oil relative permeabilityat connatewater saturation. irreduciblewater At saturationand zero gas saturation this equationreducesto:
k-. : (k"'o )(k"'t ) k..r.*

This model tends to predict incorrectoil relative permeabilityvalues (magnitudelarger than unity) for valuesof k,,,.* < 0.3. Nolen, as referenced Molina'e has taken into accountthis problem and suggested by the following model which remainsboundedas k.,,.* approaches zero: k.,, : k..,.*)*
k_ t.

+ k,*:: '*
k..r.*

* k.s - (k.*, + k.s)

(25)

Na,,.o

V. EXPERIMENTAL CONFIRMATION
Three-phase relativepermeabilitystudiesare still in an early stageof development. Little has been done on the experimental confirmationof imbibition correlations and most of the correlationsavailableare for imbibition. Early work was done primarily on unconsolidated sandsand the effectsof wettability and hysteresiswere not recognizeduntil recently. Donaldsonand Kayser2o have categorized the reasonsfor divergenceof experimental three-phase relative permeabilitydata as follows: l. 2. 3. 4. 5. Errors introducedin saturation measurements variousexperimental in methods. Errors introducedby neglectof capillary end effectsand saturation hysteresis phenomena. Variations caused by use of different oils, brines, and cores which could exhibit different wettability characteristics. Assumptionsmade to facilitate experimental procedures calculations. or Inadequacy mathematical of formulationsto represent three-phase flow conditions.

The empirical methods,though seeminglysimpler, suffer from simplifying assumptions that have limited the rangeof saturation historiesthat can be simulated.

r27
determinedthree-phase relative Table I is a chronologicallisting of the experimentally permeabilities In that have been reported.2r all of the studiesincludedin the tabulationthe authorsusedrefined oils in order to minimize oil-wetting;they assumed totally water-wet a history of the system.In caseswhere a single core was used,the influenceof the saturation rock samplewas frequently ignored. The gasesused in the studieslisted in Table 1 were air, carbon dioxide, and nitrogen.

VI. LABORATORY APPARATUS


relativepermeabilitystudieshavebeenconducted using refinednonpolaroil, Three-phase hydrocarbon fractions,brine, nitrogen,air, and carbondioxide. dieseloil, Soltrol, kerosene, as Berea, Boise, Torpedo, Tensleep,and Weeks Island sandstones, well as Arbuckle limehave beenusedfor the flow media. Bereasandstone stoneand unconsolidated sandsamples is often preferredbecause its uniformity and generalacceptability an industrystandard. of as in communications with researchers this field have indicatedthat the most pracPersonal measurement gravimetric.Othermodernmethods,suchas gamma is tical meansof saturation ray absorption,X-ray absorption,NMR, etc., are unnecessarily expensiveand elaborate. measurements sufficientlyaccurate relativelyinexpensive. are and The gravimetricsaturation Problems may be encountered with gravimetricsaturation measurements, however,especially of when gas is used in the presence volatile oil. Therefore,core holderswhich permit rapid removal of cores (without the removal of rubber sleeves)should be used when relative permeabilityis determined steady-state by methods.Wettability of the core shouldbe monitored either by the centrifugaltechnique23 an alternativemethod. or Brine saturationmay be determinedwith satisfactoryaccuracyby',electricalresistivity when nonpolaroil is employed.Oil saturation may be obtainedgravimetrically measurement and the gas volume may be computedas the differencebetweentotal pore volume and total liquid volume. The oil and water flow ratesmay be obtainedby a simpleburettearrangement or by flowmeters.The gas flow rate may be obtainedby use of a gas flowmeter. The effect of wettability on the relative permeabilities an importantfactor that should is be studied.The changeof wettability in a core from oil-wet to a water-wethas beenknown to influencerelative permeabilities, but no definite conclusions are found in the literature relativepermeabilities. concerningthe influenceof wettability on three-phase Boundary effects should be eliminatedby using core plugs at either end of the test core high flow rates.A semipermeable membrane and performing the experimentat reasonably A may precedethe core plug at the inlet end for properdistributionof the phases. modified may be used, since most investiPenn Statemethod of relative permeabilitymeasurement gatorsbelieve that the Penn Statemethodgives betterresultsthan any of the other steadystatemethods. for In addition to saturationmeasurements each phase,one needsto measurethe flow rate of each fluid and the pressure relativepermeability drop while making the steady-state A regulator measurements. gas dome may inject fluids into the core and a back-pressure may be used to maintain a constantpressureat the outlet end. Also, the gaseousphase shouldbe bubbledthrough the oil supply tanks. This procedure ensures that the oil is saturated with the gas before it enters the core. As a result, there should be no masstransfer between the gaseousphase and the oil inside the core. The gas flow rate may be regulated with a needle valve, with a large pressuredifferential acrossthe valve. The rate of gas flow may of be measured either by collecting gas by displacement water for a known length of time, differentialmeasurements, it or with a soapbubble meter or a wet-testmeter. For pressure devicebe as small as possible of fluid into the measuring that the displacement is desirable is so as to minimize error. Hence, use of a regularmanometer not possible,but differential pressuretransducersmay be used. The connectionsfor measuringpressuredifferential can

r28

Relative Permeabilin of Petroleum Reservoirs


a
t J U ) (h .

U)

U)

9 = E . =
:'': 9

(n 5 v i (t

U) (n

i a

.jl

u)

V)

F
rrt

F (n
r-l
E

E ; E = q E z -

o -'. r E , l

; ^, )
J

c..l

z
H

3
-l

" c 3

J
r-l -

tr.9 9 q F Y U gF L
F b

.= ;
t 5 E

E U J

<
r 9

n 2 X F

. ! E
^ : ? ;
L ;

tv. r-l

O C)

Q , : = ! . .
, : ' i q g

t r tL s E ; '

3
11

F E 3 EEtE F 3.E
L Fi E

? X
bO

q) L

= !EEi ' "Eri jI a S = r .


9 E J i l H : : i; h.d i..5 = :E--

s &
! r-r
rF
H H I

: a

n
FF H

i t g

i E iE *
E E

E ! E : t H i H EE E t 3 : ; E Z ; ; : f i ; ! r
* E E d ' = O o # = o - : + = 3 - i E j 5 : g ;
i Es

E EE e = i = Es i ! r e ! t g s r H s ; : F

EIEEE;IgEgEili;
2E
EE
( ) L

Fr

rh

z
3 a
FI

.E: 8ir

F- FeE
I g

g Ei F ;

iie=

tfrE
#ytE
s
(
t A

Fl j Q

-h r
J

. E E - r r_ 9 { q , c i * o . - . .

E - o 3 F : s- > a I c p t l
A ^ v

j
v l l

.- c 2
u v f

>
)
.

z
(.)
il fF

t = ! o - :
O t r A

t$e; rr
^ l l " L ! ! f q J )

-Pidr\
;

F i l a - Y , . .

-aE
- i '

) C

( ! c a ^
rr

ss : g ; :

= c ' t(O o
=

Eu ''
f,u
v g ; H

;,,
-

gl

,c!

f V

a+:
l

i rr

\n O,

\n o\

0 E =
!

;
.

o
:

9 . = > q O -.1

E z ' =E P .i Fco \./

e # F
> e ? = g.qr.>, .1ia3 \-,/
E a )

129

(n

U)

2
(n a
q;

l
3
-v .o a > . =

U)

(h

R
,,. J
F - =
; t

O :

-i,

E : 7 E P

E E y , E
> . - Q

'j ,n u

j ; E?; i 5 e.e 9Eyep .ef; . ; , v t E . , . iE > : .


E OrJ9 0J 0J = E o!> =

E
E
bO

" 3 3 .= 3 : ' E - :a ;o 9 _ E E i
' - ' =L F ^ *
! >

: . Er

g'rtil,
E

r -i

s.)

E . = = = = t E . = - t z
. "
q) L

. = ! 2 e. ?- Z ? *

= 13E i i E E jz = UI e
P B g

t > 9 ! = g

-a'*
: 1

i-e-+E=32
oo

E E E ; E - . E E
I-lJ
I

9 . = F i i o o l = E - 4 r i ; . l b P

Eg U E E : E " f = E H E I E = 3

Z a i t ' Z i . ' .

. . : = - i

.=
O -

9 r y >. g)

> =
cJ
L J

E 9

c!i; sElEeEr=:;EEEg ' 5 E : gi ; s : : * E ! ; s ; : * E s s + ! $ P E* ; tFE e3:; ::!BE ; *' s ;E EsEr E !3 i ; *$ F 3 E E E= > i


i

Ere

*=E*

i=E f

u o
F ' v) .

t r
=

l F
qJ
r v

i
s t

=
-

u
4 a

x t z
=
-

V a

",

biH
F . - 4 _
C)

'E*,H:
F . - E 4 l

e.,

brJ=

gEe? lri=-=
l

i sE_ a i "! F i

Z&y?EE i=E ' P . r -I ,; ' i E ! +=.E Zi-==.7=;='=

F . g *

t t Z

: : i ; E T q ?* U !
c.l o\
|r)

- l t -

1 a

t * [ i t n l i . , ; = s= 3i-;; =l : $ t[ g g; tI ; - n j=
l x e l : g i $ $ i g l E
n

HE f i - . i E H *
t--

(.)

q..)

I (h

trc)

U)

o -

(n

130

Relative Permeability of Petroleum Reservoirs


(n (t) >. U) ._i A L (n

:itr . o E . !

(na

FI

z
3
rh

j , ;
o .r, u Y v t

3 a r-'! -l

T ; E
J

= g E ) z -

) ? -

z
3
Fl

& -

O F F J

tr .c)

()
O

>';

9 -

; J E t r ( D

Fr

rrt

r -

a 0 )

^)

E F
O F

= <
= n x Y r-r -t U) o ( .o* Enr F 'Jr -

E#*
:': >, v

FL t8' , E :

; ; g 9 J - t r H > ' 13 o) 92i-= = d - t E t r i F


LIJ

u - - - 2 F i 9'6 '
>.-'

X
9

c : i
S U .

=
' d

d
A

i E "
U)

E 3

i.= t

-5? :; $ ! s 5 1

3
q) r Q ) o =

8.s

.r, I .l
. Y Z

. = e

E E E :

>: i/ i
. o ; :

il L

X c t r

& 3 FF H

F U 3 g ,;,() E F,9

ri

rfr

E-r (t) tl J
lr( rh Fl

z -

.E;
x

E X

9 q r S olJ

ii ' U)

o.r
O

(.)
2

rco
q)

o Z x 9 r "
@ V ^ N

t 't , ,=gF E Egl; lE: i , ( E E ' g* Y i


E : x . E , . ,I , -: ; E_y;X
2 = " d s I
E

H J?o

a E =

'r t

t = r " 3F

z
tl rF FI

L q )

g " A
lv

9 i x : Y! i l ' , , , a ; s I :;
a.l co

i fi: _

5s=3;g5geTU;e ,,

I
q,,

: F A o ,

t^

q)

E 4

E 9 E :
v (h v

U)

131

CORE HOLOER

FIGURE 23.

relative permeabilityapparatus. Schematicdiagram of three-phase

ports. The capillary tubesconnecting transmembrane the be made through semipermeable in ducermay be insertedand cemented placeabout I in. from eachend of the testapparatus. relative permeabilitymeasurements have the admethodsof three-phase Unsteady-state gas by vantage beingrapid. Oil and watermay be displaced gasto duplicate drive processes of recoverymethods.However, the calculationof isopermsfrom laboratory used in enhanced data requiresanalytical solutionsof the partial differentialequationsdescribingthe threephase fluid flow. Some early studies have made erroneoussimplifying assumptionsin process.Reliablevaluesof relative describingthe dynamic condition of the unsteady-state permeabilityas a function of saturations may be obtainedby mathematical simulation of laboratory data using finite difference calculations.20 Capillary pressuredata should be for obtainedfor gas-oil, water-oil, and water-gassystemsto provide basic data necessary relative permeabilitycalculations. Solubility of the gas in the liquids employed three-phase are in the study should be determinedbefore thesecalculations performed. relative permeabilitymeasA schematicdiagram of the apparatus used for three-phase urementis shown in Figure 23. The core holder, which has ports for differentialpressure is allows rapid retrievalof the core. Temperature controlledwith a Propormeasurements, tional Controller connectedto a heatingtape wrappedaroundthe core holder. In order to with pumps, fluids are injectedby applying eliminatepulsationof flow normally associated gas pressure top of the fluid in a tank equippedwith appropriate relief valves. Solenoid on valves and level controllersmaintain a constanthead of fluid in the supply tanks. Filters are provided in the supply lines of each phasebeing injectedinto the core holder. Check valves preventbackflow of each of the three phases.A cross sectionof the core holder is shown in Figure 24. balance,electricalresistivitymeasurement sysAuxiliary equipmentincludesan accurate transducer, cylinders,compressed tem, level controller,chart recorder,differentialpressure air and regulators,and a humidity oven.

r32

Relative Permeabilin of Petroleum Reservoirs

D I F F E R E N T I AP R E S S U R E O R T S L P
ANNULAR

<-._

2"

P R ES S U R E PORT

rO N cv)

-19"
FIGURE 24.

-1.125>

+1JJ\

Diagram of a core holder.

VII. PRACTICALCONSIDERATIONS FOR LABORATORYTESTS


The literature cited contains a large amount of information on factors affecting the.laboratory investigationof relative permeability.The following listing, however, cites iome practicalconsiderations that have not been widely discussed the literature: in l. If a pump is used to inject fluids into the core, the packingmaterialshouldpreferably be Teflon@.Most other packing materialscontain silicon and carbon which may dissolvein injected fluids and affect the wettability of the core. When brine is used as one of the fluids, all metal parts of the systemshould be of stainless steel. One-eighth-in.tubing offers excellenthandlingcharacteristics. Tygon tubing is recommended the pressure not too high. if is Most electronic differential pressuretransducershave good linearity and hysteresis characteristics; however,if possible,the transducer shouldbe recalibrated leastonce at per month. While changingpressures the liquid storage on tanks, it is importantnot to exceedthe backpressure rating of the solenoidvalves. Every effort should be made to ensure l00%osaturationof the wetting phase before startinginjection of the nonwettingphase. In a steady-state experiment, input flow rate should equal the output flow rate for each phase.In many cases,this condition is tediousto achieve. Some extraneous material may be noticed in the output lines. It must be determined whether the particles are fines from the test sampleor bacterial matter. A bactericide may be used with caution not to alter either the wettability or the resistivity of the core. Often the resistivity meter utilizes chamoisleathercontactsat either end of the core holder. The contacts should be kept immersed in brine to prevent changesin the readings. It has been noticed that the position of the outlet tubes going into the measuring cylinders affects the pressuredifferential readings.It is recommended that the tubing outlet be kept at the same level as the core holder to eliminate gravitational effects.

2.

3.

4. 5.
6.

7.

8.

9.

r33

FIGURE 25.

Comparisonof three-phase relative permeabilitydeterminations. oil

10.

Gas in the transducer lines seriouslyaffectspressure differentialreadings.The transducer should be bled of gas at frequentintervals. I l. Every effort should be made to eliminateend effectsas described Batycky et a1.22 by 12. If possible, the wetting characteristics the core should be frequently monitored of during the relative permeability experiments.The centrifugemethod23 may be employed for monitoring wettability.

VIII. COMPARISON OF MODELS The following sectionpresents comparisonof some of the modelsdiscussed a earlier. The equationof Corey et al.2 for three-phase valuesis valid for a systemin which oil k,,, is displacedby a gas. Donaldsonand DeanTobtainedthree-phase valuesfollowing the k.., same displacementmechanism.Thus, we have an opportunity to observehow well the equation of Corey et al.2 fits data provided by other workers. Three-phase relative oil permeability valuescalculated the equation Coreyet al.2werecompared by of with Donaldson and Dean's data. The isopermsobtainedare shown in Figure 25 along with Donaldsonand Dean's data as a basis for comparison.The Corey et al.2equationgives higher k,. values than those obtainedby Donaldsonand Dean. Isopermsby Corey et al3 are less concave towards l00%o saturation.Both methodsare in agreement predictingthat the isoperms oil in becomeconcavetoward l00Vo S" and decreasing The Donaldsonet a1.23'2a show S*. data k..,increasingup to an optimum S, value and then decreasing. This is evidentfor valuesof S" between30 and 6OVo this Berea core. The Corey et al. correlationsgive isoperms on which show k,o to increaseas S.,increases the expense S*. The discrepancy at of between the two methodsis larger at low S.,values. In the second comparison, data of Schneiderand Owens2s have been used to obtain

134

Relative Permeabilin of Petroleum Reservoirs

Nolen,s Model -o--<rMeihod of Naar & Wygal

FIGURE 26.

Comparisonof three-phase relativepermeabilitydeterminations. oil

isoperms Nolen's modelreand by Naar and Wygal's correlation.r2 by Few dataare available in the literature that show how the latter methodcompareswith experimentalvaluesor other correlations.Figure 26, however, provides such a comparison.Schneiderand Owens obtainedgas-oil drainagedata in the absence connatewater; their oil-water imbibition data of is for a water-wet system. Theoretically, the Dietrich and Bondorr8or the Nolen model should give the same results as Stone's second model, since gas-oil data used in this comparisonhave beenobtainedin the absence connatewater, i.e., k.o"*equalsunity. As of in the earlier comparison,the discrepancybetweenthe two methodsis evident at low S" values.Another point to note is the evidencethat k," depends only on Sovalues,especially at low S" in Naar and Wygal's correlations.There is a slight indication in both methods that k," isopermsbecomeconvex towards the l}OVo So apex at high S".

REFERENCES
l. Leverett, M. S. and Lewis, W.8., Steadyflow of gas-oil-watermixtures throughunconsolidated sands, T r a n s .A I M E , 1 4 2 . 1 0 7 . 1 9 4 1 . 2. Corey, A. T., Rathjens, C. H., Henderson, J. H., and Wyllie, M. R. J., Three-phase relativeperrneability, Trans. AIME, 201,349. 1956. 3 . R e i d , S . , T h e F l o w o f T h r e eI m m i s c i b l eF l u i d si n P o r o u s e d i a , P h . D . t h e s i s ,U n i v e r s i t v f B i r m i n s h a m . M o E n g l a n d1 9 5 6 . 4. Snell, R. W., Measurements gas-phase of saturationin a porous medium, "/. Inst. Pet., 45(428), 259, l 959. 5. Snell, R. W., Three-phase relative permeabilityin an unconsolidated sand, "/. Inst. Pet., 48(459), 80, t962.

135
6. Snell, R. W., The saturation history dependence three-phase relativepermeability,J. Inst.pet.,59, of oil 4 ' 7 1 .1 9 6 3 . 7. Donaldson, E. C. and Dean, G. W., Two- and Three-Phase RelativePermeabilityStudies,U.S. Bureou of Mines, Washingron,D.C. , #6826, 1966. 8. Sarem, A. M., Three-phase pet. Eng. relativepermeabilitymeasurements unsteady-state by methods,Soc.. J . . 9 . 1 9 9 .1 9 6 6 . 9. Saraf, D. N. and Fatt, I., Three-phase relative permeability measurement using a nuclear magnetic r e s o n a n c ee c h n i q u e o r e s t i m a t i n g l u i d s a t u r a t i o ns o r ' . P e t . E n g . J . , 9 , 2 3 5 . 1 9 6 7 . t f f , 10. Wyllie' M. R. J. and Gardner, G. H. F., The generalized Kozeny-Carman equation,its applicationto p r o b l e m s f m u l t i - p h a s el o w i n p o r o u sm e d i a , W o r l d O i l , 1 4 6 , l 2 l . 1 9 5 8 . o f ll. Caudle, B. H., Slobod, R. L., and Brownscombe, E. R., Further developmentsrn the laboratory determination relative permeability Trans. AIME, 192. 145, l95l . of , 1 2 . N a a r , J . a n d W y g a l , R . J . , T h r e e - p h a s i m b i b i t i o nr e l a t i v ep e r m e a b i l i t y ,S o < 'P e t . E n g . J . , 1 2 , 2 5 4 . e .

r96r.

13. Land, C. S., Calculation of imbibition relative permeability for two- and three-phase flow fiom rock properties,Soc. Pet. Eng. J., 6, 149, 1968. 14. Schneider, F. N. and Owens, W. W., Sandstone carbonate and two- and three-phase relativepermeabrlity c h a r a c t e r i s t i c S,o c . P e t . E n g . J . , 3 , 7 5 , 1 9 1 0 . s 15. Van Spronsen, E., Three-Phase RelativePermeability Measurements Using the CentrifugeMethod. Society o f P e t r o l e u m n g i n e e r s / D e p a r t m e n tE n e r g y ,T u l s a ,O k l a . , # 1 0 6 8 8 , 1 9 8 2 . E of 1 6 . S t o n e , H . L . , E s t i m a t i o n f t h r e e - p h a sre l a t i v ep e r m e a b i l i t yJ . P e t . T e c h . , 2 , 2 1 4 , 1 9 7 0 . o , e 1 7 . S t o n e ,H . L . , E s t i m a t i o n o f h r e e - p h a s e r e l a t i v e p e r m e a b i l i t y a n d r e s i d u a l o i l d a t a ,C a n . P e t . T e c h n o l . , t J. of t2, 53, t913. 18. Dietrich' J. K. and Bondor, P. 8., Three-phase oilrelative permeability models,paperSPE 6044 presenred at the 5lst Annual Fall TechnrcalConference and Exhibition of the SPE, New Orleans. 1976. 19. Molina, N. N., A systematic approach the relativepermeabilityproblemsin reservoirsimulation,paper to SPE 9234 presented the 55th Annual Fall TechnicalConference at and Exhibition of the SPE, Dallas, 1980. 20. Donaldson, E. C. and Kayser, M.8., Three-Phase Fluid Flow in PorousMedia. DOE/BETCilC-8)t4. r e p o r tp u b l i s h e d y t h e U . S . D e p a r t m e no f E n e r g y .B a r t l e s v i l l eO k l a . , A p r i l . 1 9 8 1 . b t , 21. Manjnath, A. and Honarpour, M. M., Investigationof three-phase relative permeability, SPE 12915 presented the Rocky Mountain RegionalMeeting of the SPE, Casper,May 20-23, 1984. at 22. Batycky, J. P., McCaffery, F. G., Hodgous, P. K., and Fisher, D. 8., Interpreting capillary pressures a n d r o c k w e t t i n g c h a r a c t e r i s t i c s f r o m u n s t e a d y - s t a t e d i s p l a c e m e n t m e a s u r e P e tn E n g oJ . . 6 , 2 9 6 , me . ts,s . r', t 9 8 l. 23. Donaldson, E. C., Thomas, R. D., and Lorenz, P. 8., Wettabilitydetermination its ef-fect recovery and on e f f i c i e n c y ,S o < 'P e t . E n g . J . , 3 , 1 3 , 1 9 6 9 . . 24. Donaldson, E. C. and Dean, G. W., Two- and Three-Phase RelativePermeability Studies,U.S. Bureau of Mines, WashingtonD.C., #6826. 1966. , 25. Schneider, F. N. and Owens, W. W., Sandstone carbonate and two- and three-phase relativepermeability c h a r a c t e r i s t i c S,o r ' .P e t . E n s . J . , 3 , 7 5 , 1 9 1 0 . s 26. Fayers, F. J. and Matthews, J. D., Evaluationof normalizedStone'smethodsfor estimatingthree-phase r e l a t i v ep e r m e a b i l i t i e s . o r ' .P e t . E n g . J . , 4 . 2 2 4 , 1 9 8 4 . S

r37
APPENDIX SYMBOLS
: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :

A A, a B b C F g h I k L m N n P

a
q R r S

s*
SL

T Z
ct

p
0 \ f.r
o

0 .1,

area constant adhesiontension materialconstant formation volume factor constant materialconstant constant fraction gravitationalacceleration thickness injectivity resistivityindex permeability length exponent number of barrelsof oil exponent pressure volume volumetric rate radius resistivity radius saturation distance directionof in flow reducedsaturation total liquid saturation time velocity vertical coordinate constant constant angle lithology factor viscosity surfaceor interfacial tension porosity immobile saturation

Subscripts : absolute a av : average - critical c : capillary cw : connatewater : displacement D : displacingphase d de : immobile displacingphase - equilibrium e : external(radius) : effective : free f :gas g - initial i : index number : irreducible imb : imbibition irr : irreducible : liquid L LR : residualliquid : minimum m mf : mud filtrate : nonwetting n :oil o : measured 1007oS* at (resistivity) ob : trappedoil : produced p - relative r : residual - solution s SL : total liquid STD : standard condition : total T - trapped t : water w : well wt : wetting xo : flushed zone

You might also like