Y.P.

3078

RELATIVE PERMEABILITY STUDIES

I. FATT AND H. DYKSTRA, CALIFORNIA RESEARCH CORP., LA HABRA, CALIF., MEMBERS AIME

ABSTRACT If the interval dN is made small the average flow rale

through a tube in the interval dN is given by Poiseuille's Law,
Relative wetting pha<:e permeabilities calculated from capil- 7r r'!:;,P
lary pressure-saiuration data are compared with measured qav = - - (2)
relative permeability data. The equation relating relative per- 8 JL 1
meability to capillary pressure-saturation data is derived by where !:;'P is the pressure drop across the tube of length 1
assuming that a porous medium is analogous to a bundle of and radius r, and JL is the viscosity of the fluid. Substituting
capillary tubes. The equation includes a term to correct for for qav in Equation (1),
7r r'!:;,P
the difference between fluid path length and length of core. dQ = - - dN (3)
Relative oil permeability data in the water-oil-gas system are 8 JL 1
presented, and it is shown that if the water is considered part Darcy's Law for linear flow of an incompressible fluid in a
of the rock matrix, the relative oil permeability curve is typi- porous medium is
cal of a wetting phase relative permeability curve. Apparatus KA!:;,P
for measuring two and three-phase relative permeabilities are Q=-- (4)
JLL
described. where J( is the permeability and !:;'P is the pressure drop
across a bundle of tubes which has replaced the porous me-
INTRODUCTION dium of length L and cross-sectional area A. Differentiating
Darcy's Law with A, !:;'P, JL, and L constant,
The difficulty in measuring relative permeability of cores
A!:;,P
has made it desirable to have a correlation between relative dQ = - - dK (5)
permeability and some more easily measured property of JLL
porous media. Such an easily measured property is the capil- Equating (3) and (5) gives
lary pressure-saturation relation. 7r r' L
d K = - - dN (6)
In the past there has been a tendency to separate the capil- 8Al
lary pressure concept from the very complex pore geometry If the pores are assumed to be cylinders
and to consider that the capillary pressure-saturation curve dV = 7r r'ldN (7)
gives only <:ome characteristic distribution of interfacial curva- where V is the volume of flowing fluid in the pores. Substi-
ture between two fluids.' Recently, however, it has been real- tuting for dN in Equation (6) gives
ized that if capillary pressure data are to yield information r'L
concerning pore size distribution' and fluid flow in porous d K = - dV _ (8)
media, the pore must be assumed to have a simple shape such 8Al'
as a cylinder or a sphere.
3 By definition the saturation, S, of the core is
Childs and George' showed that relative water permeability V V
s=-=- (9)
could be calculated from the capillary pressure-water satura- Vp <pAL
tion curve by assuming that the pores are cylinders in which
the fluid flowing obeys Poiseuille's Law, and that the capillary
pressure curve indicated the ,ize and number of pores. Purcell" 20
derived an equation relatin 6 t!~e permeability of a porous
medium to the capillary pressure curve by assuming the
porous medium to be analogous to a bundle of capillary tubes. \
16 /1 6
Gates and Lietz" calculated relative permeability by an equa-
tion derived from Purcell's equation.
~
o \ / If)
I
In this paper an equation is derived giving the relative
water permeability as a function of water saturation by as-
suming that the core sample can be repreEented by a bundle
W
I
a::
~
2 - ~ II I2

of capillary tubes in which t!'e fluid path length is not the

(J)
(J)
1\ /
same as the bulk length, and in which the fluid path length w
a::
0- e
\ l- f-- 8
~
varies with saturation.
>-
/~
Pc
If)
u
a:: n.
THEORY
For a bundle of N capillary tubes the flow through dN
<t
..J
..J 4 17~t- 4
'-

tubes will he
0-
<t 1/
......- V
dQ = q"v dN (1) o
where Q is the total flow rate through all the tubes and q.v is o0 o
:0 40 60 80 10o
the a'verage flow rate through the tubes in the interval dN.
PER CENT KEROSENE SATURATION
lReferences given at end of paper.
Manuscript received in the office of the Petroleum Branch, AIME. FIG. 1 - CAPILLARY PRESSURE CURVE. DATA FROM FIG. 33 OF GATES
Sept, 14. 1950, Paper presented at the Fall Meeting of the Branch in
Los Angeles Oct. 12-13. 1950. AND LIETZ."

Vol. 192, 1951 PETROLEUM TRANSACTIONS, AIME 249

T.P. 3078 RElATIVE PERMEABILITY STUDIES

20 I00 20 I00 >-

go ,: >-
~
go
X
~
..J
X
16 1 It ..J
m ~ ! m
«
~
0
I
\ II
80
«
w
0 16
\ ; 80 w
~
a:::
pc\ ~ w
\ /
,t
w a::: a::: 0 w
a:: w :::>
a..
12 60 12 60
:::> a..
~
Cf) I
Cf)
1\ w
Cf)
w
w
>
-h
Cf) / 0
w
!Jo > a:::
\ /
-
a:: a.. ~
a.. ~ «
8

--
40 « B
~
;- 40 ...J
>-
a:::
«
..J
..J
a..
OBSERVED
o
4 - - CALCULATED
GATES a LIETZ
6
""
/ /
;' 0
.1) ~
0
20
..J
W
a:::
~
Z
>-
a:::
«
..J
..J
a..
«
4
o OBSERVED
\.
l'- I- r--
L 20
W
a:::
~
Z
uJ
o
« - CALCULATED
0 W
0 0 - CALCULATED ,.,
~? V 0 YV
0
EQUATION 18 o
I
a:::
0
a::: o I I I "
I
o w
20 40 60 80 108 w o 20 40 60 80 100 a..
a..
PER CENT WATER SATU RATION PER CENT WATER SATURATION
FIG. 2 - RELATIVE PERMEABILITY AND CAPILLARY PRESSURE CURVES. FIG. 3 - RELATIVE WATER PERMEABILITY AND CAPILLARY PRESSURE
DATA FROM FIGS. 26 AND 33 OF GATES AND LIETZ." CURVES IN WATER·OIL SYSTEM. SAND: BASAL TUSCALOOSA, MISS.
PERMEABILITY: 125 MD. POROSITY: 15 PER CENT.

where V p is the total pore volume and <1> is the porosity. Differ- of the pore. Equation (14) thus gives r as a function of the
entiating Equation (9) and substituting for dV in Equation capillary pressure. Substituting for r in Equation (13) gives
(B) gives <1>(2 (f cos 0)2(1+b)
r' <1> L2 dK = dS (15)
d K = - - dS (10) Ba2 P/(1+b)
Bl2 If the constants, a and b, were known, the effective perme·
Equation (10) contains the pore radius, r, and the fluid path ability, K e , at any saturation could be calculated by integrat-
length, l. These terms are not directly measurable and there- ing Equation (15) from zero saturation to the desired satura-
fore must be related to some other property of the core. tion, S
The ratio of the fluid path length, l, to the length of the _ <1> (2 (f cos O),(1+b)
Ke-------
dS IS (16)
core, L, has been termed the tortuosity, t, or 8a~ P c:!(1+h) \I

l
t =-- (11) The relative permeability, K" however, can be calculated
L without knowing the constant, a, if the constant, b, is known
The tortuosity, t, has not yet been related directly to a meas-
urable core property. A reasonable assumption regarding t
can be obtained by considering what happens to the wetting Table I - Calculation of Wetting Phase Relative
phase in the pore spaces as the core is desaturated. During Permeability From the Equation
desaturation the wetting phase retreats into the smaller pores
and into the crevices between sand grains. Liquid in such
crevices has a small radius of curvature and by the derivation
given in this paper is considered to be in small pores. It can
be reasoned then that liquid flowing in the crevices and small
pores will travel a more tortuous path than liquid flowing
through the large pores. As a first approximation t can be
1
assumed to vary inversely as r, or S Pc Area from
a P" o to S
t = - (12) Per cent em Hg {em Hg)3 in' Per cent
rb
where a and b are constants. Substituting Equations (11) and 100 3.B IB.2 x 10-" 4.15 100
(12) in (10) gives 95 4.1 14.5 3.13 75
<1> r 2 (1+b) 90 4.4 11.7 2.31 56
d K = - - - dS (13) B5 4.B 9.0 1.6B 40
Ba'
The pressure, P" across a curved interface between two BO 5.3 6.7 1.19 29
fluids' is 75 5.B 5.1 0.B2 20
2 (f cos e 70 6.4 3.B 0.54 13
Pc=---- (14)
r 65 7.1 2.B 0.34 B.2
where r is the radius of curvature, (f the interfacial tension at
the interface, and e the liquid-solid contact angle. When Equa- 60 B.O 2.0 0.20 4.B
tion (14) is applied to porous media the pressure term is 55 9.4 1.2 0.10 2.4
known as the capillary pressure, and r is the pore radius at 50 11.6 0.6 0.04 1.0
which a non-wetting phase just displaces a wetting phase out 45 15.0 0.3 x 10-" 0.01 0.2

250 PETROlEUM TRANSACTIONS, AIME Vol. 192, 1951

 I. FATT AND H. DYKSTRA T.P. 3078

20 I00 >- 20 I 09>-

r CI
r
~ ...J ::I: ...J

~
16 80
CD
ct ~
(.) 16
"\ 80
CD
ct
W
(.) W
~
\ ! ~
a::
a:: w
W
a::
::J 12
\
I
60
w
a.. a::
::J 12
"\
! 60
w
a..

~
Cf)
I
Cf)
Cf)
r W Cf)
I
w
>
W
a::
a.. --1\ !
>
~
ct
W
a::
a.. l I ~
ct

>-
1l:
8
~ I
40 ...J
W
a::
>-
a::
8

~
\: /
40 ...J
W
a::
ct
~ I~

" f-
ct
...J ...J r
...J 4 20
r ...J 4 20 z
Z
~ I- .........
/~
a.. W a.. w
ct
(.)
o OBSERVED (.) ct
(.)
o OBSERVED
- CALCULATED - (.)

Y ---'V
-iALCiLATiD a:: 0
a::
o o w o I I I .0
o w
o 20 40 60 80 100 a.. o 20 40 60 80 100 a..
PER CENT WATER SATURATION PER CENT WATER SATURATION
FIG. 4 - RELATIVE WATER PERMEABILITY AND CAPillARY PRESSURE FIG. 5 - RELATIVE WATER PERMEABILITY AND CAPILLARY PRESSURE
CURVES IN WATER·OIL SYSTEM. SAND: BASAL TUSCALOOSA, MISS. CURVES IN WATER-OIL SYSTEM. SAND: BASAL TUSCALOOSA, MISS.
PERMEABILITY: 100 MD. POROSITY: 26 PER CENT. PERMEABILITY: 129 MD. POROSITY: 22 PER CENT.

or asmmed, because by definition K,. = Ke/ K, thus giving The areas given in column 4 are divided by the total area
dS 1
under the - vs S curve to give the per cent relative per-
Pc'
K,.=
'00
---- (17)
meability in column 5. The calculated curve is plotted in
10dS
P 2(1+b)
Fig. 2. The relative permeability data and the curve calcu-
bted by Gates and Lietz' are also shown in Fig. 2.
Relative water permeability and capillary pressure data in
the water-oil system were obtained in. this laboratory on five
RESULTS California cores and four Mississippi cores. Relative water
Comparison of Calculated and Measured permeabilities were calculated from capillary pressure data
by means of Equation (18). The measured data and the cal-
Relative Water Permeability Data culated curve for each core, shown in Figs. 3 through 11, are
Gates and Lietz' have calculated relative permeabilities in fair agreement.
using Equation (17) with b assumed to be zero, that is, the Comparison of measured and of calculated relative perme-
tortuosity independent of the pore radius and thus independent ability data indicates that the constant, b, may not be the
of the saturation. This paper compares measured data with same for all types of sands. For the Basal Tuscaloosa sand b
calculated data for b = Y2, that is the tortuosity is a function
of pore radius and therefore dependent on the saturation.
20 I00 >-
For b = % Equation (17) becomes r
I ...J
[£
Pc3
-- ~--

CD

to
o
ct

~·±f
16 .. - I- ~- 80 W

~l
K, = dS (18) i ~
o Pc3 w I-~---L- -_.- a::
w
a:: J a..
1

-+\1 .\-- f
::J 12 60
To calculate relative permeability from Equation (18) - Cf)
, P," w
is plotted against saturation, S and the area under the -
P,"
1
vs
Cf)
W
a::
a..
---1- -_.- _. -.-

0
>
r
ct
8 40 ...J
--.l f-- W
S curve is obtained by graphical integration or by planimeter-
ing. As an example, capillary pressure-saturation data from
>-
a:: '- "'1i
I a::
Gates and Lietz" (their Figs. 26 and 33) are used to calcu-
late relative permeability of the wetting phase, which for the
ct
....J
::! 4 --- +
~
~
/ 20
r
z
w
a.. o OBSERVED / ~ (.)
example given is kerosene in the kerosene-air system. The ct -

1
(.) - CALCULATED / a::
capillary pressure, P" and -3 vs the saturations, S are 0 I I I ni: /" o w
a..
Pc o 20 40 60 80 100
1
shown in Fig. 1 and Table 1. The area under the - vs S PER CENT WATER SATURATION
P,.'
FIG. 6 - RELATIVE WATER PERMEABILITY AND CAPILLARY PRESSURE
curve from zero saturation to saturation, S, was obtained bv CURVES IN WATER-OIL SYSTEM. SAND: BASAL TUSCALOOSA, MISS.
graphical integration and is shown in column 4 of Table I. PERMEABILITY: 58 MD. POROSITY: 18 PER CENT.

Vol. 192, 1951 PETROLEUM TRANSACTIONS, AIME 251

T.P. 3078 RELATIVE PERMEABILITY STUDIES

I 00 >- vs 5., where Ko is the effective oil permeability at gas satura-

20
~ tion, 5 g • The above ratio is obtained by dividing the effective
.J oil permeability at each water saturation by the effective oil

,
II)
permeability at the same water saturation and zero gas satu-
16 8o <
W ration. The data fall approximately on the same curve. The
, ~
0:::
curve is typical of a wetting phase relative permeability curve.
W
This seems to indicate that in th~ water-oil-gas system the

12
0 I 60
0.. water can be considered part of the rock matrix with oil as
the wetting phase and gas as the nonwetting phase.
\ 1 Considering interstitial water as part of the rock matrix

8
\ I 40 suggests the possibility of calculating
Ko
Ko at 5"=0
in the water-
>-
0::: \ II oil-gas system by using Equation (18) from the oil-gas capil·
<
.J
.J 4
~ )-- 20
~
Z
lary pressure curve with interstitial water present. The method
Ko
0..
<
o
o OBSERVED
- CALCULATED
"--- i i--
W
o
of calculating
Ko at 5<=0
is illustrated by using the data

0 I I I .nQ. ~ o
0:::
W
0..
from Figs. 14 and 15 of Gates and Lietz." The calculation'
are carried out as illustrated in the previous example, except
o 20 40 60 80 100
1 Ko
PER CENT WATER SATURATION that p", ~p 3 ' and are plotted against gas satura-
F:G. 7 - RELATIVE WATER PERMEABILITY AND CAPillARY PRESSURE
" K" at 5"=0
CURVES IN WATER-OIL SYSTEM. SAND: GATCHELL, NORTHERN SAN tion. The calculated and experimental results'x- are given in
JOAQUIN VALLEY, CALIF. PERMEABILITY: 180 MD. POROSITY: 18 PER Figs. 15 and 16. The agreement between the calculated curves
CENT. and the observed data is good.
Calculation of relative permeabilities from capillary pres-
less than Y2 will give a better check, whereas for the Gatchell sure-saturation data greatly simplifies the problem of obtaining
sand b = Y2 appears to give the best check. For most of the relative permeabilities. With some assurance of fair to good
data presented by Gates and Lietz" b greater than % will give agreement between calculated and observed relative perme-
a better check. As more relative permeability data are ob- abilities, much time and expense can be saved by obtaining
tained, an average b term may be chosen for each type of capillary pressure-saturation data and calculating relative
sand to give the best fit between calculated curve3 and experi- permeabilities by Equation (18).
mental data.

Relative Oil Permeability Data in *The relative permeability 'Yith interstitial water present as p'otted
by Gates and Lietz" in their Figs. 12 through 15, is the same as
Water-Oil-Gas System Ko
- - - - - used in this paper. To obtain the relative oil p2,:meabiHty
Relative oil permeability data in the water-oil-gas system Ko atS. == 0
are shown in Figs. 12 and 13 for two Mississippi cores. The (defined in the API Code 27) as a function of gas saturation it is on1y
Ko
Ko necessary to multiply by the effective oil permeability with
data of Figs. 12 and 13 are replotted in Fig. 14 as - - - - - Ko atS< == 0
/("at5,=0 interstitial water present and to divide by the permeability of the core

20 I00 >- 20 I 00 >-

~ I-
co CO
:z: .J :z: .J
II) II)
~
0 16 80
«
w
~
0 16 1\ I 80 «
w

W
\ I ~
0::: W
~ / ~
0:::
0:::
::> 12 '\ ! 60 W
0.. 0:::
::> 12 \\ I 60
W
0..

I
(/)

c-\\
(/)
(/)
W
0:::
0..
8
--r-~---
-~
I

L 40
w
>
~
«
.J
(/)
W
0:::
0..
8
•
"\'\ I
40
w
>
~
<
>-
0:::
I 1\ Ir-- W
0:::
>-
0::: ~ /
.J
ILl
0:::
< .J
~ 4 I
.....
1\ J. 20
~
«
.J
~ 4
-~ ~ II-I--- 20
I-
Z Z
~ ~
0..
<
0
o OBSERVED
- CALCULATED
J- W·
0
0..
«
0
o OBSERVED
- CALCULATED
/' ~
ILl
o
o
o
I
20
I I
40 60
/
80
"o 100
0:::
W
0..
o
o
I I
20
I
40
o~ ~
60 80
o
100
0:::
ILl
0..
PER CENT WATER SATURATION PER CENT WATER SATURATION
FIG. 8 - RelATIVE WATER PERMEABILITY AND CAPILLARY PRESSURE FIG. 9 - RELATIVE WATER PERMEABILITY AND CAPillARY PRESSURE
CURVES IN WATER·OIL SYSTEM. SAND: GATCHElL, NORTHERN SAN CURVES IN WATER-OIL SYSTEM. SAND: GATCHEll, NORTHERN SAN
JOAQUIN VALLEY, CALIF. PERMEABILITY: 140 MD. POROSITY: 14 PER JOAQUIN VAllEY, CALIF. PERMEABILITY: 80 MD. POROSITY: 11 PER
CENT. CENT.

252 PETROLEUM TRANSACTIONS, AIME Vol. 192, 1951

 I. FATT AND H. DYKSTRA T.P. 3078

20 I00 >- 20 I00 >-

I- 0 I-
...J Z ...J
m
2
(.) 16 \ I 80 ct
I.LI
2
(.) 16
I
I 8o
m
ct
I.LI
\ / ::E
I ::E

,,
I.LI
a:: I.LI a::
a::
;:)
en 12
\ I
il
60
I.LI
a.. a::
;:)
en 12 ! 6 o
I.LI
a..
en
I.LI ....
I.LI
>
en
I.LI I I.LI
>
a::
a.. /r-- ~
ct
a::
a..
8 " /
~
ct
8 40 ...J o
>-
a:: "- I r-...
/ ........
I.LI
a::
>-
a::
c(
'\ ...J
I.LI
a::
c(
...J I-
...J
::! 4
-\ f f-
2o
I-
::! 4 20 Z
'-- ~ >; Z
a..
C(
(.)
o OBSERVED
- CALCULATED
/ I.LI
(.)
a..
ct
(.)
o OBSERVED
- CALCULATED
.0
I.LI
(.)

I I I / a::
I.LI o 1 I I ~V--r----'----. a::
0 n
o a.. o 0 I.LI
o 20 40 60 80 100 20 40 60 80 100 a..
PER CENT WATER SATURATION PER CENT WATER SATURATION
FIG. 10 -- RELATIVE WATER PERMEABILITY AND CAPILLARY PRESSURE FIG. 11 -- RELATIVE WATER PERMEABILITY AND CAPILLARY PRESSURE
CURVES IN WATER-OIL SYSTEM. SAND: McADAMS, NORTHERN SAN CURVES IN WATER-OIL SYSTEM. SAND: VAQUEROS, NORTHERN SAN
JOAQUIN VALLEY, CALIF. PERMEABILITY: 13 MD. POROSITY: 10 PER JOAQU:N VALLEY, CALIF. PERMEABILITY: 47 MD. POROSITY: 12 PER
CENT. CENT.

across the core and porous end plates. The slope of the line
APPARATUS
Relative Water Permeability Apparatus obtained is [ l::,. (l::,P) .] The effective water permeability
total
l::,. ( ql'-)
The relative water permeability apparatus is shown in Fig.
is calculated from the equation
17 for the water-oil system. The apparatus is all glass except
the nickel flow lines between the core and the pipets. Flow 1
L
[
rates as low as 10-' cc per sec are measured in 0.1 ml gradu-
ated pipets, the flow rate into the core being measured in one K w =- l::,. (l::,.P) l::,. (l::,.P)
A
pipet and the flow rate out of the core being measured in the l::,. ( ql'- ) l::,. ( ql'- )
other pipet at the same level. When the flow rates are identi-
cal, the core is at saturation equilibrium. Bypass stopcocks
are used to return the water in the measuring pipets to the Use of the [ l::,. (l::,.P) ] plot for calculating permeability by
l::,. ( qiL )
zero mark. A fine sintered glass filter is used to filter the
water before it passes through the core. The end plugs are
Monel metal to which porous plates are cemented with De 60
Khotinsky cement. The porous plate* has a permeability of >-
about 50 md and entry pressure of about three psi in the I-
...J
water-oil system. The porous plate covers the whole face of
the core. m 50
C(
Before effective water permeabilities can be measured in I.LI
~
the apparatus shown in Fig. 17, the pressure drop across the
porous end plates must be known as a function of flow rate.
a::
ILl 40 /f
a.. I
To obtain this calibration, flow rates at different pressure
...J
/ I
drops are measured with the porous plates in capillary con-
~
~fl
(5
tact. The product of the flow rate, q, and the viscosity, 1'-, is 30 -(,;
plotted against the pressure drop, l::,.P, and the slope of a I.LI
~ Q:-
•
straight line thus obtained give, [l::,. (l::,.P) ]
. The plate ~ 1 1--./ ~I
_Q-/ '<!:/ "-
l::,. ( ql'-) plates
calibration is independent of capillary pressure.
...J
ILl
a::
20
~-q/
~'?"
&r-!
Q:- I
To measure effective water permeability, a water saturated I- .$:- +--+- <t' 0
core is placed between the porous end plates of the core holder Z
ILl 10
(Y/ ~/ ~
and the assembly is immersed in oil. The desired water satu-
ration is obtained by imposing a capillary pressure on the
(.)

a:: ~~/
q,45'V V V
vr~v.~V
core. The water removed from the core is measured in the I.LI
0..
volume changer shown in Fig. 17. The saturation at equilib- o
rium is calculated from the volume of water removed and 10 20 30 40 50 60 70
the pore volume. Thus a capillary pressure-water satura- PER CENT OIL SATURATION
curve is also obtained in the apparatus. At each water satura·
tion, flow rate is measured as a function of the pressure drop FIG. 12 -- RELATIVE OIL PERMEABILITY IN WATER-OIL-GAS SYSTEM.
SAND: BASAL TUSCALOOSA, MISS. PERMEABILITY: 124 MO. POROSITY:
'Allen filter disk made by Allen Filter Co., Toledo, Ohio. 26 PER CENT.

Vol. 192, 1951 PETROLEUM TRANSACTIONS, AIME 253

T.P. 3078 RELATIVE PERMEABILITY STUDIES

60 50

...
>- / \ • 0
FROM FIGURE 12
FROM FIGURE 13 -
-I
Q3 50 ~/
<:i
f 0q i\
40
I
I~
<t
W ~ CI
~
0::
W 40 1/ til~/-~;
(.)
0
~
C/)

o~
~
Q..

-I
(5
§!
~I
-14Jf-t.u
Q. Q.
Q: 'Q:I
0
30

~ 0

_ R'-/- "?/ ~

W
~
30
it I I t-
Z
20 \
~
-I ~
~/ / / W
0 ~
W
0::
20
-
"'J / i/V a:: ~ 0

... f
w 10 •
Z
W 10 / / Q..
\~
V /
-
0 /
0:: / ~
k ~ v.---- /
W r' o
Q.. o 10 20 30 '" 40 50
20 30 40 50 60 70 PER CENT GAS SATURATION
PER CENT OIL SATURATION
Effective Oil Permeability Ko
FIG. 13 - RELATIVE OIL PERMEABILITY IN WATER·OIL·GAS SYSTEM. F I G . 1 4 - - - - - - - - - - - - - - - - - - - = - - - - - - - IN WATER-
SAND: BASAL TUSCALOOSA, MISS. PERMEABILITY: 100 MD. POROSITY: Effective Oil Permeability at Sg=O Ko at Sg=O
26 PER CENT. OIL-GAS SYSTEM.

Darcy's law gives a more accurate permeability from the data in series, the pressure drops are directly proportional to the
than a permeability obtained by averaging permeabilities cal-
I::. (I::.P) ] terms. Table II shows that the ratio of the pres-
culated from each point. On the [ I::. (I::.P) ] plot an obvi- [ I::. (qiJ-)
I::. (qiJ-) sure drop across the plates to the total pressure drop ranges
ously erroneous measurement can be easily seen and omitted from 0.23 to 0.02, so that the probable error in the calculated
from consideration in drawing the best straight line. pressure drop across the core is only slightly greater than the
The above method of measuring effective water permeabil- probable error in the total pressure drop.
ity has certain advantages over the method using a static man-
For cores of high permeability the probable error in the
ometer with a pressure tap in the center of the porous end
pressure drop across the core may be appreciably greater than
plates. The advantages are: (1) No time is lost while waiting
that for cores of low permeability. However, as the water
for a static manometer to come to equilibrium with the core;
saturation decreases the effective water permeability decreases
(2) no correction is needed due to constriction of flow caused
by a center tap; (3) the pressure drop across the core is
obtained by subtracting one direct measurement from another, I00
whereas the static manometer does not measure the true pres- pc\
sure drop across a core.'

The magnitude of the

I::. (I::.P) ]
[ I::. ( OiJ- )
. .
terms Involved m cal-
CI
:I:
~
0
I
16
-\ 80
0
W \ c
culating effective water permeability are shown in Table II 0:: 0
for a 47 md core. Inasmuch as the core and porous plates are ::J
(/)
(/) 12 \ <i 60 x

Table II - Calculation of K From the I::.P vs qiJ- Plot

W
0::
0..
1\ ~ 0
"CI
pw

Core of Fig. 11, K = 47 md, L = 3.08 cm. A = 5.05 cm'

>-
0::
8
\ .1 40 0
(/)

...<t
I::. (I::.P)] [I::. (I::.P)] [I::. (1::.!:2] K
«
-I
-I
ll\\ :\I::

Sw [ " ( OiJ-) total 1\ ( OiJ-) plate~ 1\ ( qiJ-) ror~w 0..

o OBSERVED
GATES a LIETZ 6
;i 0
:\I::

v:
4 20
<t r----
0 - CALCULATED
Per cent cm-md cm-md cm-md md EQUATION 18 I
/ocP
90 2.90 X 10' 0.67 x 10-' 2.23 X 10-' 27.2 o I I I I o
88 3.81 0.67 3.14 19.3 100 80 60 40 20 o
85 5.0 0.67 4.3 14.1 PER CENT GAS SATURATION
80 7.8 0.67 7.1 8.5
12.3 0.67 11.6 S.2 Effective Oil Permeability Ko
76 ., ..., FIG. 15 - - - - - - - - - - - - - - - - - - - = - - - - - - - IN WATER-
70 19.0 0.67 ]8.3 ~ ") Effective Oil Permeability at Sg=O Ko at 5.=0
62 34.4 x 10-' 0.67 x 10' 33.7 x 10 2 1.8 OIL-GAS SYSTEM. OBSERVED DATA FROM FIG. 14 OF GATES AND
LIETZ." INTERSTITIAL WATER 46 PER CENT. =
254 PETROLEUM TRANSACTIONS, AIME Vol. 192, 1951
 I. FATT AND H. DYKSTRA T.P. 3078

more rapidly, so that the probable error in the calculated

pressure drop across the core again approaches the probable
TO GRADUATED
error in the total pressure drop. VACUUM PIPETTES

Relative Oil Permeability Apparatus

The apparatus for measuring relative oil permeability in the
water·oil·gas system is the same as the relative water perme·
ability apparatus shown in Fig. 17 with the exception of the
core holder. Detail of the core holder is shown in Fig. 18.
The porous diaphragms are made oil·wet by treating ordinary
TO
VACUUM

o AIR
RESERVOIRS

water·wet porous diaphragms with "Dri-Film_'"

FIG. 17 - RELATIVE WATER PERMEABILITY APPARATUS.
In the water-oil-gas system, water is present as an immobile
phase to correspond to connate water in the re3ervoir. A rela- 5. Purcell, W. R.: Trans. AIME (1949) 186, 39.
tive oil permeability curve may be obtained for several differ-
6. Gates, J. I., and Lietz, W. Tempelaar: "Relative Perme-
ent constant water saturations.
abilities of California Cores by the Capillary Pressure
Before effective oil permeabilities can be measured, the Method." Paper presented API Meeting, Los Angeles,
pressure drop across the oil-wet porous end plates must be Calif., May 12, 1950.
known as a function of flow rate. The porous end plate cali- 7. Dykstra, H., and Parsons, R. L.: Trans. AIME (1951)
bration is determined exactly as is done for the relative water 192,227, (1.0/ Pet. Tech., Aug., 1951).
permeability. 8. Norton, F. J.: "Organo-Silicon Films," General Electric
The desired water and oil saturations are first obtained in Review (August, 1944).
a capillary pressure apparatus. The core is then placed in the
core holder of the relative oil permeability apparatus. A cap- DISCUSSION
illary pressure is imposed on the oil to give a desired gas
saturation. Air enters the core through the hole in the center
By lohn I. Gates, Shell Oil Co., Los Angeles, Calif., Member
AIME
of the porous end plates. Effective oil permeabilities and oil
saturations are measured and calculated as is done for rela- This paper is another valuable addition to our growing
tive water permeability. knowledge concerning relative permeability, and we are m·
debted to the authors for its presentation.
To me the paper makes three important contributions.
REFERENCES First, it presents a simplified apparatus for measuring rela-
tive wetting phase permeabilities by the capillary pressure
1. Leverett, M. c.: Trans. AIME (1941) 142, 152. method. Those of us who have used this method know that
2. Ritter, H. L., and Drake, L. c.: Ind. Eng. Chem. Anal. Ed. the accurate measurement of the pressure drop across the core
(1945) 17, 782. is the most difficult part of the operation. The authors have
3. Ritter, H. L., and Erich, L. c.: Anal. Chem. (1948) eliminated this measurement by determining the pressure
20, 665. drops across the diaphragms at various flow rates. This pres·
4. Childs, E. C., and George, N. c.: Discussions 0/ the F ara- sure drop is then subtracted from the total pressure drop
day Society (1948) 3, 78. across the system (which is easily measured) to obtain the
pressure drop across the core. The advantages of this tech-
20 I00 nique are enumerated in the paper.
00 There are, however, some difficulties which should be pointed
::I: out. The end diaphragms should have permeabilities of the
~ 0 same order of magnitude as the core so that the pressure drop
0 16 80 across the diaphragms will be a fraction of the total pressure
o drop. The entrance pressure of the diaphragms must also be
W o
-- ~
a:: high enough to allow desaturation of the core to the desired
~
CJ) 12 60 x degree without penetration of fluid. Diaphragms meeting these
CJ) \ requirements are not easy to find. Care must be taken to see
W o that there is nothing in the flowing liquid which will tend to
a:: •
a..
~ 8 \ j 40 0
plug the diaphragms; otherwise, the permeabilities of the dia·
phragms will change during the test and therefore the pres-
a::
ct ~~ ~I­
<t sure drops for given flow rates will increase.
..J
..J
a.. 4
o OBSERVED
GATES Il LtETZ6
)f\ 20
o
~
Finally, some consideration must be given to the pressure
drop across the capillary contact material. In our laboratory
ct
0 -CALCULATED Ii \ we have used finely ground diatomaceous earth for contact
material and have found that the combined permeability of
EQUATION 18
~o the diaphragm and contact material decreased as higher pres-
I I I .- sures were applied to the gas phase. Possibly the authors have
80 60 40 20
some technique for eliminating this difficulty which they would
PER CENT GAS SATURATION be willing to explain.
Effective Oil Permeabilily Ko The second important contribution in the paper is the mod-
FIG 16--------------------=------- IN WATER-
- Effective Oil Permeability at S.=O Ko at 5.=0 ification of the formula for calculating relative wetting phase
OIL-GAS SYSTEM. OBSERVED DATA FROM FIG. 15 OF GATES AND permeabilities from capillary pressure curves. In our formula.
LIETZ: INTERSTITIAL WATER =
43 PER CENT. the relative permeability curve of a system of capillary tubes

Vol. 192, 1951 PETROLEUM TRANSACnONS, AIME 255

T.P. 3078 RELATIVE PERMEABILITY STUDIES

ance to those obtained in practice. They would possess the

feature, however, that at every saturation, the sum of the oil
and water relative permeabilities would equal 1. In actual
C====~4IO-A: R
experience, this is far from true. It is possible to account for
C===::::;;)--OiL the deficiency by a "tortuosity factor," and admittedly it is
logical to expect each phase to interfere with the freedom of
flow of the other in an interconnected system. To use only
the pore radius as the measure of this factor does not seem
quite adequate, however, especially if one considers the prob-
lem of applying the concept to the calculation of the non-
wetting phase permeability. Here we would need a "tortu-
osity factor" of infinity when the oil is in the largest pores
END RING (up to the equilibrium oil saturation), with a decreasing fac-
PLUG tor as the oil is introduced to smaller and smaller pores.
With respect to the experimental technique, I would like to
ask if it is not likely that the effective permeability to the wet-
FIG. 18 - CORE HOLDER. ting phase of the end plates decreases at increasing capillary
pressure. Such variation would introduce an error into the
having the same capillary pressure curve as the core was cal-
method of calculating pressure drop used by the authors. The
culated. When these calculated curves were compared to the
fact that a plate 1/16-in. thick will have a lower breakthrough
measured curves, it was found that they deviated more and
pressure than one Vwin. thick indicates that there is probably
more a~ the core was desaturated. The calculated relative per-
penetration of a plate by the non-wetting phase over much of
meability values were made to approach the measured values
its surface at capillary preswres approaching the break-
by lIlultiplying by an average tortuosity coefficient which de-
through pressure, and, of course, there would be virtually
creased with decreasing saturation. The authors have assumed
complete displacement of the non-wetting phase from any
that the tortuosity is related to pore radius and have developed
regions between the plate and core not in good contact. In
a revised formula for calculating relative permeabilities from
our laboratory we use side taps to measure pressure drop,
capillary pressure curves. It is interesting to note that their and have often observed appreciable decreases in effective
formula is similar to ours and differs only in the power to
permeability in the end plates at the lower wetting phase
which the capillary pressure is raised. The assumption that
saturations.
the tortuosity is related to the pore size has received some
verification in our laboratory. We have noted that the change
of the coefficient of tortuosity with saturation is related to AUTHORS' REPLY TO MESSRS. GATES AND WILSON
the change of the pore size with saturation. The use of the Wil50n's comments imply that, if the bundle of capillary
revised formula would therefore appear justified. Numerou~ tubes analogy were an accurate picture of q porous medium,
tests must be made, however, before it will be known whether our method of calculating relative wetting phase permeability
the constant b is determined by a particular field or zone or should also be applicable to the calculation of the nonwetting
whether it is related to some measurable property of the core. phafe relative permeability. If one carries out the calculation
The third contribution of the paper is the confirmation that for, say, the relative oil permeability in the water-oil system,
the oil acts as the wetting phase in the core even though inter- the calculated curve does not have the form of the measured
stitial water is present. The work in our laboratory has also relative oil permeability curve. Upon closer examination, how-
confirmed this conclusion. This is an important point since. ever, one notes that, if the calculated curve is shifted to the
as the authors point out, we can then calculate relative oil left, it takes on the appearance of part of the nonwetting
permeabilities from oil capillary pressure curves when inter· phase curve.
stitial water is present. It also has other valuable implications. The simple bundle of tubes analogy predicts symmetry be-
The interstitial water can then be considered part of the rock tween the wetting and nonwetting phase relative permeability.
matrix for some problems. The manner in which the oil is Although few data are available, laboratory results indicate
displaced by other flowing fluids is influenced by the way in a tendency toward symmetry for the nonwetting and wetting
which the oil wets the core. Probably many other implications phase relative permeability curves. Cores for which the rela-
will be become apparent as we study the problem. tive water permeability curve is far to the right on the usual
These contributions are definite advances in the study of plot also yield relative oil or gas permeability curves which
relative permeability relations. are to the right of the center of the plot.
We realize that the bundle of capillary tubes concept is too
DISCUSSION simple. The results of its w'e to date, however, have given us
By J. W. W'ilson, Union Oil Co. of Calif· encouragement.
The authors have made an interesting modification in the To answer the question about the change in combined per-
equation used by Purcell and by Gates and Lietz in an effort meability of the porous diaphragms and contact material we
to eliminate the correction factor called "lithology factor" by can say that we have found no change in combined perme-
Purcell for homogeneous flow. and "coefficient of tortuosity" abality up to one psi capillary pressure. The maximum capil-
by Gates and Lietz for multi phase flow. Since the new "con- lary pressure that we have med is three psi at which pressure
stant" seems to vary less than the old "constants," progress the decrease in combined permeability may be as much as
has been made. 10 per cent. However, by referring to Fig. II and Table II
There is one disturbing feature of the "bundle of capillaries of the paper. one can see that at capillary pressures greater
concept." It is assumed that all capillaries above a certain than one psi the pressure drop across the diaphragms and
size contain only the non.wetting phase (oil, let us say). while contact material is less than two per cent of the total pressure
all smaller capillaries ("ontain only water. For this idealized drop. Consequently a decrease of 10 per cent in the combined
system we could calculate relative permeability curves from permeability of diaphragms and contact material causes a
the capillary pressure relation somewhat similar in appear- negligible change in overall pressure drop. * * *
256 PETROLEUM TRANSACTIONS, AIME Vol. 192, 1951