You are on page 1of 12

SPWLA TWELFTH ANNUAL LOGGING SYMPOSIUM, MAY 2-5,

AVERAGE WATER SATURATION FROM


CAPILLARY PRESSURE DATA

L. Paul Westbrook and W. John Lee


Mississippi State University

ABSTRACT

This paper proposes a method for estimating average water saturation


in stratified reservoirs. This method requires capillary pressure data and an
estimate of average permeability in the reservoir. In the past, there has been
some question as to how this average permeability should be determined. There-
fore, computer studies, using field capillary pressure data, were performed on
hypothetical stratified reservoirs to determine a satisfactory method of esti-
mating average permeability to determine average water saturation. Use of geo-
metric average permeability proved to yield reliable average water saturation
estimates; use of arithmetic average permeability frequently led to unacceptably
large errors.

INTRODUCTION

Average water saturation of a reservoir is frequently determined


from capillary pressure data. This is a convenient method for homogeneous
reservoirs, but in a stratified reservoir, an accurate value of average
water saturation is more difficult to determine. The difficulty arises
because water saturation depends on permeability; for a reservoir in which
permeability varies from point to point, water saturation is a complex
function of height in the reservoir. Average water saturation in a strati-
fied reservoir could be estimated rather easily from capillary pressure data
if an average permeability could be determined for the reservoir. If, in
addition to making calculations easier, use of this average permeability
results in reasonably accurate values of average water saturation, we would
have a technique of practical value to the formation evaluation engineer.
At least two methods of computing average permeability in stratified reservoirs
can be considered: arithmetic average permeability and geometric average
permeability. Both of these averaging methods have been used by engineers to
estimate the average water saturation for oil reservoirs.

We found no clear evidence in the literature supporting either of


these methods. Amyx, Bass and Whitingl suggested that average permeability
obtained with the geometric average method will yield more accurate values
of average water saturation, but they offeretino evidence based on field
capillary pressure data to support their suggestion. Therefore, the objective
of this work was to use field data to determine which of the two proposed
methods of averaging permeability will yield a value of average water satura-
tion closer to true average water saturation.

-1-
SPWLA TWELFTH ANNUAL LOGGING SYMPOSIUM, MAY 2-5, 1971

A computer program was written to calculate and compare true


average water saturation with average water saturations obtained from
arithmetic average permeability and geometric average permeability. These
computer studies verified that the geometric average permeability should
be used to find average water saturation in stratified reservoirs.

CAPILLARY PRESSURE

Water saturations calculated from capillary pressure data can


be very accurate. D. L. Luffel and R. V. Randa112 have -compared water
saturations calculated from capillary pressure data with those obtained
by coring with an oil base mud and a pressure core barrel. Figure 1 gives
depth plotted against water saturation determined by these two methods.
Each point represents a depth in an oil well with a certain water satura-
tion. Note the excellent agreement when comparing the directly determined
water saturations with those calculated from the air-kerosene capillary
pressure curve of extracted cores. Also note the complicated shape of the
resulting water distribution curve.

---
A .
---

1 1 1

-r-l
=
-._
~ -n I I
1 1 [ 1

~Me@wredDirectly by Vacuum
Distillation
of Oil-CutCores

3
-0
Calculated
from Gas-Oil
Capiilory
PressureCurves r
o 20 40
FIG. I Water SoturotionPercent

Capillary pressure data for reservoir cores are usually reported


in the form of a graph similar to Figure 2. These capillary pressure curves
were obtained from cores with different permeability but from the same res-
ervoir. These data give water saturation as a function of capillary pres-
sure. Capillary pressure is not useful as such for studying water saturation
SPWLA TWELFTH ANNUAL LOGGING SYMPOSIUM, MAY 2-5,

in an oil reservoir; therefore, capillary pressure is usually converted to


height above free water level in the following manner:*

144 Pc
H= (1)
PW-P ~

Free water level is defined as the point in the reservoir where capillary
pressure is zero. Figure 2 gives water saturation plotted against both
capillary pressure and height above free water level. For the conversion
from capillary press~re to height above free water-level in Figure 2, it
was assumed that the density of oil was 53.6 lb/ft3 and the density of
water was 68 lb/ft3.

1000

000

600
500
400

300

200

z
.
I 1 11 11 II 1 1 1 \ 1 1 ~ ,~~
Im -
4
2 80
%
f:
40

30

20

OL I I I I I I 1 I I I 10
o t0203040508C 70808000 0 10 20304050307083 SOKIO

WATER SATURATION , !4 WATER SATURATION, %

Fl@URE 3 CORi+ELAT(ON OF WATER $ANRATION


F!OURE 2 CORRELATION OF WATER SAT UR4T10N WITH HEIGHT WITH PERMEABILITY FOR VARICIJS MEIQHTS
ABOVLZ FREE WATER LEVEL ANO CAPILLARY PRESSURE ABOVE FREE WATER LEVEL,
AT V&RlOUS PERME&OILiTI ES.

If the data in Figure 2 are replotted on semi-log paper with


parameters of height above free water level, straight lines result: for
example, see Figure 3. Capillary pressure data plotted as in Figure 3 are
more convenient to work with.

*symbols are defined in the table of nomenclature


SPWLA TWELFTH ANNUAL LOGGING SYMPOSIUM, MAY 2-5, 1971

220
Ooc
I I I II I I
200 tl
IoiO

20 I I I I L I I I I
WC
I I I I 1- 1 I I ,003Z10
00 10 20 30 40 50 60 70 BO SO

WATER SATURATt ON , %

FlOURE 4- CORRELATION OF WATER SATURATION WITH MEIQHT


ABOVE FREE WATER LEvEL AND SIJ8SEA CEPTM AT
TME AVERAOE RESERVOIR PERMEAC.ILITY.

h average water distribution curvefor a stratified reservoir


can be obtained easily from Figure 3, given a value of average permeability
for the reservoir. Figure 4 is an example of this water distribution curve
for an average permeability of 90.6 md as obtained from Figure 3. Figure 4
can be treated as the water distribution curve for the reservoir. This
assumes that the reservoir is homogeneous, with permeability equal to the
average permeability. Average water saturation can be obtained from Figure
4 since

I
Woc
Sw dH
GOC
Ew = (2)

WOC - GOC

This integral is normally evaluated by graphical integration; an example


problem using this method is given at the end of the paper.

COMPUTER SOLUTION

A computer program was written to compute true and approximate


water saturations (based on average reservoir permeability) for stratified

-4-
SPWLA TWELFTH ANNUAL LOGGING SYMPOSIUM, MAY 2-5, 1

reservoirs. These calculations were performed by assigning a thickness and


a permeability to each layer of the reservoir. Capillary pressure data
published by Wright and Wooddy3 were used to relate water saturation to
permeability and height above free water level.

True average water saturation for a reservoir was calculated


with Eq. 3:

~wlhl +EW2 h2 + . . .
Fwt = (3)
h1+h2+ . . .

Arithmetic average permeability and geometric average permeability were


then calculated for the reservoir. Arithmetic average permeability was
calculated with Eq. 4:

hlkl+h2k2+ . . .
Za = (4)
hl+hl+ . . .

Geometric average permeability was calculated with Eq. 5.

hllogkl+hzlogkz+. . .
log Eg = (5)
h1+h2+ . . .

Values of average water saturation were determined using each of these two
average permeabilities. Each average water saturation was compared with
true average water saturation and a percentage error was calculated.

Sixty-six test cases were studied. The number of layers and the
permeability of each layer were varied from case to case. Reservoirs had
as few as two layers and as many as five layers. The layers had perme-
ability values as low as five md and as high as 900 md. Some reservoirs
had the higher permeability on top and others had the higher permeability
on the bottom.

PRESENTATION OF RESULTS

Table 1 gives the results obtained in this study. Each hypothet-


ical reservoir is denoted by a case number. The layers in each reservoir
are numbered from bottom to top; i.e. , layer one is closest to the water-
oil contact.
SPWLA TWELFTH ANNUAL LOGGING SYMPOSIUM, MAY 2-5, 1971

TABLE L
RESULTS OF COMPUTER STUDIES ON STRATIFIED RESERVOIRS

Average Average
~ermeabilities of layers, mcl Permeability Water Saturation Error

Case K K K K K R i s s s E E
1 2 3 4 5 a g Wt wa w a g

1 15 200 108 55 47.1 40.4 47.8 14.1 1.4


2 200 15 108 55 45.6 40.4 27.8 11.4 4.7
3 15 200 900 370 140 37.5 27.0 37.6 27.9 0.2
h 200 900 15 370 140 35.6 27.0 37.6 24.2 5.5
5 900 15 200 370 140 35.7 27.0 37.6 24.3 5.3
6 10 5 20 100 34 18 59.0 53.0 60.0 10.1 1.7
7 10 200 600 900 428 181 34.8 25.4 34.8 26.9 0.1
8 10 200 900 600 383 145 37.1 26.6 37.2 28.2 0.3
9 15 75 300 200 148 91 41.7 37.0 b2.3 11.3 1.4
10 15 300 74 200 148 91 41.7 37.0 h2.3 11.2 1.5
11 15 300 200 75 148 91 41.5 37.0 42.3 10.8 2.0
12 75 15 300 200 148 91 41.2 37.0 62.3 10.2 2.5
13 75 200 15 300 148 91 41.2 37.0 42.3 10.2 2.6
14 75 300 200 15 148 91 40.6 37.0 42.3 8.9 4.1
15 200 15 75 300 148 91 41.0 37.0 42.3 9.7 3.2
16 200 15 300 75 148 91 40.7 37.0 42.3 9.2 3.3
17 200 75 15 300 148 91 40.9 37.0 42.3 9.6 3.4
18 300 75 200 15 148 91 40.2 37.0 h2.3 8.0 5.1
19 300 200 15 75 148 91 40,5 37.0 42.3 8.6 4.5
20 300 200 75 15 148 91 40.2 37.0 62.3 7.9 5.2
21 10 21.5 100 225 900 251 85 42.8 31.2 43.0 27.0 0.7
22 10 21.5 100 900 225 251 85 42.6 31.2 43.0 26.8 1.0
23 10 21.5 225 100 900 251 85 42.7 31.2 43.0 26.9 0.8
24 10 900 21.5 100 225 251 85 42.2 31.2 43.0 26.1 1.9
25 10 900 225 21.5 100 251 85 40.7 31.2 43.0 23.3 5.7
26 21.5 100 10 900 225 251 85 42.3 31.2 43.0 26.2 1.7
27 21.5 100 225 10 900 251 85 42.2 31.2 43.0 26.1 1.9
28 21.5 100 225 900 10 251 85 41.9 31.2 43.0 25.4 2.8
29 21.5 100 900 225 10 251 85 41.8 31.2 43.0 25,3 3.0
30 21.5 225 10 900 100 251 85 42.1 31.2 43.0 25.9 2.1
31 21.5 225 100 900 10 251 85 41.8 31.2 43.0 25.3 3.0
32 100 10 900 225 21.5 251 85 41.7 31.2 43.0 25.1 3.3
33 100 21.5 10 225 900 251 85 42.2 31.2 43.0 25.9 2.1
34 100 21.5 225 10 900 251 85 42.0 31.2 43.0 25.6 2.5
35 100 225 10 900 21.5 251 85 41.5 31.2 43.0 24.8 3.7
36 100 225 900 10 21.5 251 85 41.2 31.2 43.0 24.3 4.4
37 100 900 10 21.5 225 251 85 41.5 31.2 43.0 24.8 3.6
38 100 900 225 10 21.5 251 85 41.1 31.2 43.0 24.1 4.6
39 225 10 21.5 100 900 251 85 42.o 31.2 43.0 25.6 2.6
40 225 10 900 21.5 100 251 85 41.5 31.2 43.0 24.8 3.6
41 225 10 900 Loo 21.5 25L 85 41.4 31.2 43.0 24.6 3.9
42 225 21.5 100 10 900 251 85 41.8 31.2 43.0 25.2 3.0
43 225 100 900 10 21.5 251 85 41.1 31.2 43.0 24.o 4.7
44 225 900 10 21.5 100 251 85 41.2 31.2 43.0 24.2 4.4
45 900 10 21.5 100 225 251 85 41.4 31.2 43.0 24.6 4.0
46 900 10 21.5 225 100 251 85 41.3 31.2 43.0 24.4 4.1
47 900 21.5 100 225 10 251 85 40.9 31.2 43.0 23.7 5.2
48 900 100 21.5 10 225 251 85 41.1 31.2 43.0 24.0 4.7
49 900 LOO 21.5 225 10 251 85 40.8 31.2 43.0 23.5 5.4
50 900 225 10 21.5 100 251 85 41.0 31.2 43.0 23.8 5.1
51 900 225 21.5 100 10 251 85 40.7 31,2 43.0 23.3 5.7
52 5 900 800 500 10 433 112 39.0 25.1 40.0 35.8 2.4
53 10 5 900 800 500 433 112 39.8 25.1 40.0 37.1 0.3
54 500 10 5 900 800 433 112 38.9 25.1 40.0 35.5 2.8
55 800 500 10 5 900 433 112 38.2 25.1 40.0 34.3 4.7
56 900 800 500 10 5 433 112 37.4 25.1 40.0 32.9 6.9
57 20 700 600 500 30 370 166 34.7 27.0 35.7 22.2 2.9
58 40 20 700 600 500 370 166 35.3 27.0 35.7 23.5 1.2
59 500 30 20 700 600 370 166 34.6 27.0 35.7 21.9 3.2
60 600 500 30 20 700 370 166 34.2 27.0 35.7 20.9 4.6
61 700 600 500 30 20 370 166 33.6 27.0 35.7 19.6 6.3
62 40 500 400 300 50 258 164 34.7 39.0 35.8 10.9 3.2
63 50 40 500 400 300 258 164 35.1 39.0 35.8 12.0 2.0
64 300 50 40 500 400 258 164 34.7 39,0 35.8 10.9 3.2
65 400 300 50 40 500 258 164 34.4 39.0 35.8 10.1 4.2
66 500 400 300 50 4f.1 258 164 33.9 39.0 35.8 8.9 5.5

-6-
SPWLA TWELFTH ANNUAL LOGGING SYMPOSIUM, MAY 2-5, 1

INTERPRETATION OF RESULTS

Table 1 shows that the average water saturation obtained using


arithmetic average permeability varied from 7.9%, in case 20, to 37.1%,
in case 53. Error in average water saturation obtained using geometric
average permeability varied from 0.1%, in case 7, to 6.9%, in case 56.

Table 1 also shows that reservoirs which have high permeability


layers near the water-oil contact, overlain by layers of low permeability,
yield larger errors in average water saturation calculated using geometric
average permeability. Cases 20, 51, and 56 are examples of this.

Cases 8, 22, and 53 in Table 1 show the large errors caused in


the average water saturation calculated using arithmetic average permeability
when very low permeability layers are near the water-oil contact.

However, in every case, geometric average permeability yielded


average water saturation closer to true average water saturation than did
arithmetic average permeability.

CONCLUSIONS

Geometric averaging yields permeabilities which can be used to


obtain more accurate estimates of average water saturation than arithmetic
averaging. If a relatively small error can be tolerated in average water
saturation, then the geometric average permeability can be used with con-
fidence to determine average water saturation. In any case, if left with
a choice between arithmetic average permeability and geometric average
permeability, the choice should be geometric average permeability.

APPLICATION

Consider a stratified reservoir which has layers with the follow-


ing permeabilities:

subsea depth, feet permeability, md

3000 Gas-oil contact


3000 - 3040 10
3040 - 3100 200
3100 - 3143 50
3143 - 3200 500
3200 Water-oil contact

i
SPWLA TWELFTH ANNUAL LOGGING SYMPOSIUM, MAY 2-5, 1971

Assume that the capillary pressure data in Figure 2 applies to this reser-
voir and that core data show that the free water level is ten feet below
the water-oil contact. The average water saturation of the reservoir is
determined as follows:

1. Determine geometric average permeability:

log Zg = 40 log 10 + 60 log 200 + 43 log 50 + 57 log 500


40+60+43+57

~g = 90.6

2. Replot the capillary pressure data in Figure 2 as logarithm of permea-


bility against water saturation with parameters of height above free
water level. Results are given in Figure 3.

3. Enter Figure 3 with ~ = 90.6 and determine water saturation and height
values that correspon% to this average permeability.

Swg% H, feet

61.5 10
51 20
47 30
44 40
42 60
37.5 170

Plot these Swg vs. H values as in Figure 4.

8
SPWLA TWELFTH ANNUAL LOGGING SYMPOSIUM, MAY 2

4. Determine average water saturation from Figure 4 by solving Eq. 2


graphically. The area under the curve between the water-oil contact
and the gas-oil contact, shown as the shaded area in Figure 5, is
divided by the thickness between the water-oil contact and the gas-
oil contact to obtain the average water saturation.

Woc
SwgdH
[ GOC
Fwg = = 7960
200
Woc - Goc

Fwg = 39.8%

The same problem can be solved using Eq. 3 to calculate the true average
water saturation. For the purpose of illustration and comparison the
same problem is solved using Eq. 3 as follows:

1. Eq. 3 is best solved in steps because average water saturation for


each layer requires graphical integration of the capillary pressure
curve for that layer. Also, because water saturation depends on
height above free water level, the average water saturation of each
layer must be determined at the layers position with respect to
the water-oil contact. Therefore, it is convenient to number the
layers from bottom to top in the reservoir, which will place laYer
one nearest the water-oil contact. Figure 6 shows how the average
water saturation should be obtained for each layer in the reservoir
from capillary pressure data.

0 10 20 30 40 50 60 70 m 90 ,m

wATm MATURATION,% WATER 3ATuRATWM, %

FIQURE 3 GRAPHICAL INTEGRATION OF THE WATE8 DISTRIBUTION FIGIRE 6 CORnELATlON OF WATER SATURATION WITH MEIQHT ABOVE
CURVE AT AvERAGE RESERVOIR PERMEABILITY, FREE WATER LEVEL SMOWING ORAPHICAL lNTEQRAnON OF
THE WATER DISTRIBUTION CURVE FM EACH LAwIR IN A
RESJZRVO!R

9
SPWLA TWELFTH ANNUAL LOGGING SYMPOSIUM, MAY 2.5, 1971

2. For layer one, the average water saturation is:

67
SwdH
J 10
Fwl = = 1588
57
67 - 10

Fwl = 27.8%

Similarly for layers two, three, and four:

ZW2 = 46.8%

:W3 = 30.5%

FW4 = 57.5%

3. After the individual average water saturations have been determined for
each layer, Eq. 3 is used directly

Zwt . 127.8)(57) + (46.8)(43)+ (30.5)(60)+ (57.5)(40)


57+43+60+40

FWt= 7750
200

Zwt = 38.8%

The value of w~ter saturation obtained using the geometrically averaged


permeability~ Swg = 39.8%, is within 2.5% of the true average water
saturation, Swt_= 38.8%. It is obvious that the calculations involved
in determining Swt would become very tedious ~or a reservoir with many
strata; however, the calculations leading to S would not be complicated
by an increase in the number of strata. Wg

1
SPWLA TWELFTH ANNUAL LOGGING SYMPOSIUM, MAY 2-5, 19

NOMENCLATURE

error in average water saturation, %

WC = gas-oil contact

h= thickness of individual bed, ft

H= height above free water level, ft

k= absolute permeability of individual bed, md

E= average absolute permeability for reservoir, md

capillary pressure, psi

water saturation, %

average water saturation, %

Woc = water-oil contact

density of fluid, lb/ft3

SUBSCRIPTS

123 = individual beds numbered from bottom to top


99s...

a= quantity obtained using arithmetic average permeability

g = quantity obtained using geometric average permeability

o = oil

t = true value

w= water
I
SPWLA TWELFTH ANNUAL LOGGING SYMPOSIUM, MAY 2-5, 1971
1
I

REFERENCES

1. Amyx, J. W., Bass, D. M., Jr. and Whiting, R. L. : Petroleun~ Reservoir


Engineering, McGraw-Hill Book Company, Inc., New York (1960) 546.

2. Luffel, D. L., and Randall, R. V. :Core Handling and Measurement


Techniques for Obtaining Reliable Reservoir Characteristics, Formation
Evaluation Symposium, AIME, Houston, Texas, 1960, 1-21.

3. Wright, H. T., Jr. and Woody, L. D., Jr. : Formation Evaluation of


the Borregos and Seeligson Field, Brooks and Jim Wells County, Texas,
Formation Evaluation Symposium, AIME, New Orleans, Louisiana, October 1955.

ABOUT THE AUTHORS

Paul Westbrook is a Petroleum Engineer


with Shell Oil Company in New Orleans,
Louisiana. He received his B.S. degree
in Petroleum Engineering from Mississippi
State University in 1970.

(Please refer to Paper E for the biography


John Lee.)
and photo of the co-author, !IJ.

L. Paul Westbrook

-12-

You might also like