You are on page 1of 18

SPE

Society of Petroleum Engineers

SPE 16484
GASWAT-PC: A Microcomputer Program for Gas Material
Balance With Water Influx
by B. Wang and T.S. Teasdale, Texaco Inc.
SPE Members

Copyright 1987, Society of Petroleum Engineers


This paper was prepared for presentation at the Petroleum Industry Applications of Microcomputers held in Del Lago on Lake Conroe, Montgomery,
Texas, June 23-26, 1987.
This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the
author(s). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the
author(s). The material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Papers
presented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Permission to copy is
restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain conspicuous acknowledgment of
where and by whom the paper is presented. Write Publications Manager, SPE, P.O. Box 833836, Richardson, TX 75083-3836. Telex, 730989 SPEDAL.

incorrect OGIP. However, this problem may be avoided


by drawing the Cole plot, 3 where a flat line or
scattered distribution indicates natural depletion and
an increasing trend of the data points indicates water
influx.

ABSTRACT
GASWAT-PC is a microcomputer material balance program
that calculates original gas in place (OGIP) from
production history for gas reservoirs with or without
water influx.
Several well-known analytical techniques, and some extensions and improvements, are
applied to a single input data set. The output for
each method is given in both tabular and graphical
form.

Several widely accepted analytical methods include:


P/Z plots for natural depletion reservoirs, Ramagost
and Farshad4 (RF) P/Z plots for abnormally pressured
reservoirs, Cole and Havlena and OdehS (HO) methods
for water drive reservoirs, and pressure match methods6
for all types of reservoirs. For water drive reservoirs, five aquifer models are usually considered:
(1) small pot, 7 (2) Schilthuis steady state,S (3) Hurst
simplified,9 (4) van Everdingen and Hurst (EH) infinite
linear, and (5) EH finite/infinite radial aquifers.10

GASWAT-PC also has the capability to predict reservoir


pressure and recovery efficiency for a specified
production rate if OGIP and aquifer properties are
known.
An option is included to account for the
effect of a large pressure gradient in the waterinvaded region.

GASWAT, a Fortran program run on an IBM mainframe (IBM


3090) for two years, combines all the preceding techniques in a package where users input the required PVT
and production data.
We have found that the best
analysis is achieved by a series of interactive runs
where the user examines results in either graphical or
tabular form, and then individual data points are
selectively included in (or excluded from) each calculation procedure.

Case studies illustrate the application of GASWAT-PC


to an abnormally pressured reservoir and to two water
drive reservoirs.
Finally, we compare the pressure
performance predicted by a numerical simulator and by
GASWAT-PC for a water drive reservoir with a high dip.

INTRODUCTION
Material balance methods have been employed for many
years to calculate OGIP for gas reservoirs.
Since
reservoir data and aquifer properties are uncertain
for most reservoirs, it is important to analyze the
data using different techniques.
By comparing the
results of various methods, one can improve reservoir
performance interpretation and can predict OGIP more
accurately.

This kind of trial and error work can very economically


and easily be done on a microcomputer if computational
times are adequate. GASWAT-PC gives results identical
to the mainframe version with good response time. The
advantages gained by using the microcomputer include
portability, ease of operation, cost reduction, and
time savings.

For ~nstance, several authors1,2 reported that it is


difficult to distinguish natural depletion from partial
water drive by using only the P/Z plot. Thus, extrapolating an apparently straight P/Z line may yield an

GAS MATERIAL BALANCE EQUATION


The general form of the material balance equation for
a condensate gas reservoir is7
F = G(Eg + Efw) +We

References and illustrations at end of paper.


25

.(1)

GASWAT-PC:

A MICROCOMPUTER PROGRAM FOR GAS MATERIAL BALANCE WITH WATER INFLUX

where G is original wet gas in place, and


F = Gwgp Bg + Wp Bw
Eg

1.

(3)

Bg - Bgi
cw Swi + Cf

Erw = Bgi Ce (Pi-P)

Bgi

1 - Swi

(4)

(Pi-P).

(5)

We= U S(P,t)

2.

Note that for a dry gas reservoir, wet gas production


(Gwgp> equals dry gas production (Gp) since condensate
production (GLp) is zero. The condensate conversion
factor (Kc> is 132,790 rc/MLc and according to
Cragoe,11 the molecular weight of condensate is given
by MLc = 6084/(API gravity- 5.9).

Figure 1 demonstrates the G calculation for a volumetric reservoir using P/Z method 1 and method 2.
B.

OGIP CALCULATIONS

Rearrange equation (10)


similar to (9).
p

1.

Natural Deoletion Reservoirs


For natural depletion reservoirs with a normal
initial pressure gradient (e.g., less than 0.5
psi/ft [11.3 kPa/m]), We and Efw can be neglected
and equation (1) reduces to
F

= G Eg

2.

( 8)

Gwgp

[1-

-G-] ...........

P/Z Method 2:

RF Modified P/Z Method:


Ramagost and Farshad4 suggested plotting
(P/Z) [1 - ce (Pi-P)] versus Gwgp to yield a
straight line with intercept G at P = 0.
Note that this method is also valid for
normally pressured reservoirs, and Cf in
equation (11) does not need to be constant.
A variable pore volume compressibility table
of Cf versus P can be input in GASWAT-PC.
The calculated G and the normalized standard
deviation will be listed in the summary table.

Equation ( 6) can be converted to a more popular


form
p

(11)

Extrapolating the initial slope of the P/Z


plot to zero pressure would yield an incorrect
G.
However, carefully selecting the points
on the second slope for linear regression can
still give a reasonable G estimate.
One
disadvantage of the second slope approach is
that it may require a long production period
to develop the proper trend.

( 7)

n - 1

Y and Yc are observed and calculated values of Y


(dependent variable); Ymax and Ymin are maximum and
minimum observed Y, and n is the number of points.
A.

Gwgp

The conventional P/Z plot will be straight during


the early life, but will usually curve downward
during the later stages of depletion. This is the
so called dual-slope P/Z plot for abnormallypressured reservoirs.

( 6)

a(y) =

to obtain an equation

(1- - G - )

where
2<Ycj - Yj)2
j=1

. . (10)

F = G(Eg + Efw)

To compare different calculation methods, we use a


normal{zed standard deviation (%) defined as
a(Y)
(Ymax - Ymin>

Abnormally Pressured Reservoirs


For a natural depletion reservoir with an abnormal
initial pressure gradient (e.g., greater than 0.5
psi/ft [11. 3 kPa/m]), Efw becomes a significant
term and cannot be neglected in equation (1). For
this case we use

GASWAT-PC requires schedule data (time, pressure, and


gas, condensate, and water production) and PVT data.
GASWAT-PC calculates the Z-factor by interpolating
measured data, or alternatively, by the gas gravity
or the gas composition corrtHation.12 In the correlation approach, impurities (i.e., N2, C02 and so2 )13 and
the yield ratio are considered, and either Dranchuk
et al. 14 or Hall and Yarborough15 methods can be
selected for the Standing-Katz correlation16 calculation. Radial water drive reservoirs require additional
information, as discussed later.

<Y> = 1oo

P/Z Method 2:
This method is similar to P/Z method 1 except
that the regression line is not forced to
pass through the initial (Pi/Zi) point.
Usually, this method applies to those reser~
voirs with questionable initial pressures.

In the water influx term (We>, U is an aquifer constant


and S(P,t) is defined separately for different aquifer
types.
The theoretical values for U and S(P,t) are
given in Appendix A.

P/Z Method 1:
A plot of P/Z versus Gw~p should yield a
straight line, and G is obta~ned by extrapolating the line to P = 0. A linear regression
technique is used in GASWAT-PC to search for
a best fit line for all observed data points,
and this line is forced to pass through the
initial point (Pi/Zi).

(2)

(Gp + Kc GLp)Bg + Wp Bw

SPE 16484

(9)

26

B. Wang and T. S. Teasdale


3.

2.

HO Straight Line Method:

and equation (1) becomes


0

(14)

A plot of F/Eg versus (Pi-P)/Eg again yields


a straight line with intercept G at P = Pi.
The main advantage of this approach is that G
can be obtained without having to specify Cf
and U values.
C.

Eg

18

speaking, this method is most


very strong aquifers.
However,
often show the presence of water
more clearly than P/Z plots.

When water influx declines in reservoirs with


limited aquifers, Eg increases faster than
(We - Wp Bw> and the term (We - Wp Bw)/Eg in
equation (18) will decrease with time. Then
the Cole plot curve starts bending down and
this trend will cause a high and unrealistic
prediction of G. However, this error can be
minimized by excluding improper data points,
which exhibit a negative slope on the Cole
plot, from the linear regression.

We= U (Pi-P) . . . . . . . . (13)

G + (Bgi G ce + U)

Eg

Generally
useful for
Cole plots
influx much

This approach can be extended to an abnormally


pressured reservoir with some shale water or
with a small pot aquifer.
Then the water
influx (We> is defined as

Eg

We - Wp Bw

Cole3 proposed plotting GwgpBg/Eg versus Gwgp'


and then extrapolating this curve to Gwgp =
0. The intercept shows the true value of G,
since We and Wp are zero at zero Gwgp

(12)

A plot of F lEg versus (Pi-P) /Eg yields a


straight line with intercept G at P = Pi, and
slope (Bgi G ce>. Then Cf can be calculated
from equation (4).

Gwgp Bg

---- = G +

G ( 1 + Bgi ce

Cole Method:
Equation (15) can be written as

Havlena and Odeh5 proposed a straight line


method for water drive reservoirs. The same
technique can be used here by rearranging
equation (10).

Eg

3.

HO Straight Line Method:


The HO method is the most popular technique
used to estimate G for water drive reservoirs.
They rearranged equation (15) by dividing
both sides with Eg

Water Drive Reservoirs


_F_=G+U
Eg

For a water drive reservoir, Efw is neglected since


it is small compared to Eg and We. Then equation
(1) changes to

P/Z Methods:

pi

Zi

G - Gwgp
G- y

(16)

S(P,t) = Pi -

where
pi
y =

zi Psc

Tsc
T

(We - Wp Bw)

(19)

For instance, assuming a Schil thuis steady


state aquifer, then

Writing equation (15) in terms of P and Z


p

A plot of F /Eg versus S(P, t) /Eg yields a


straight line with intercept Gat S(P,t) = 0,
and slope U.
Note that the definition of U
and S(P,t) depends on the type of aquifer, as
given in Appendix A.

F = G Eg + We . . . . . . . . ( 1 5)

1.

S(P.t)
Eg

(20)

A plot of F/Eg versus [Pi- (Pj + Pj-1)/2]/Eg


will be made and, if this aqu~fer assumption
is correct, the points will approximate a
straight line with a small standard deviation
between the fitting line and the observed
points.

.(17)

The plot of P/Z versus Gwgp will not generally


be linear since We varies with production and
time.

Since the computation time required on each


aquifer calculation is small, GASWAT-PC makes
an HO calculation for each aquifer geometry
(small po't, Schil thuis, Hurst simplified,
infinite linear, finite/infinite radial
aquifers).
For each method, the G, U, and
standard deviations values are calculated.
The results are listed in a summary table for
comparison, and graphical output is also
available.

However, this method may sometimes be used if


we properly select the data points for regression.
Shagroni17 suggested using the early
portion of the P/Z curve in a linear regression since these points may not be strongly
influenced by water influx. This approach has
been tested in GASWAT-PC on many reservoirs,
and often gives reasonable results.

27

GASWAT-PC:

A MICROCOMPUTER PROGRAM FOR GAS MATERIAL BALANCE WITH WATER INFLUX

To analyze a radial aquifer situation, additional data for aquifer permeability, porosity,
water viscosity and compressibility, formation
compressibility, and gas reservoir radius
must be given to calculate the theoretical
dimensionless time coefficient [ equation
(A-10)]. Also, the aquifer radius is required.

The pressure match method is very accurate,6 and


it can used for all kinds of reservoirs. The major
drawback is that substantial computational time is
required for the pressure calculation.
In most
cases, we only use this procedure to compare the
results from other methods used in GASWAT-PC.

Because of uncertainties in the dimensionless


time coefficient and aquifer to reservoir
radius ratio ( ra/ rg) estimates, up to 14
adjustment factors {TDF) may be applied to
the theoretical dimensionless time, and 14
different ralrg values may be input.
Each
combination of TDF and ralrg is calculated,
and the "best fit" may be selected.

At the end of OGIP calculations, the calculated G


and U of each method (P/Z, Cole, HO, etc.) are
automatically input to the 'pressure match section.'
A set of pressure and water influx histories, and the standard deviation of the pressure
for each method, are calculated and listed in the
summary table for comparison. Since the comparison
criterion for all methods is the pressure standard
deviation, the m~n~mum value should give the best
estimates of G and U.

We use two selection criteria for the radial


aquifer case: (1) normalized standard deviation
O'(F/Eg), and (2) curvature CURV(F/Eg).
The
curvature is a measure of how much the best
parabola departs from a straight line over
the range of the data (see Figure 8), and the
calculation is described in Appendix B.

PRESSURE AND ULTIMATE RECOVERY PREDICTIONS


For many water drive reservoirs, production performance
is very sensitive to gas rate. Field cases18,19 have
shown that deliverabili ty and ultimate recovery can
be greatly increased by properly controlling gas
production.

Theoretically, if the aquifer in the HO


analysis is correctly specified, then the data
points will lie on a straight line (a = 0,
CURV = 0). A systematic upward (CURV > 0) or
downward curvature (CURV < 0) in the data
indicates that the specified aquifer is too
small or too large, or it may be that the
analysis is not appropriate for this data set. 5

GASWAT-PC has the capability to predict pressure and


ultimate recovery for a gas reservoir produced at a
specified rate if G and aquifer character are given.
Thus, GASWAT-PC can both help reservoir engineers
correctly estimate reserves, and also plan an optimum
operational strategy.

In our experience, the minimum a and the


minimum absolute curvature lcURVI will not
occur for the same ralrg and TDF pair. It is
recommended to pick the minimum a provided
ICURVI is small enough (less than about 20%)
so as to give credence to the validity of the
analysis.
D.

The calculation procedure for pressure prediction is


the same as that for the pressure match method mentioned above, except that the future production data
is manually generated based on a specified gas rate.
Cumulative water influx at each pressure point is
also calculated and listed in the output.
A.

Pressure Match Method for All Kinds of Reservoirs

Modify equation (16) by including the formation


compressibility term,

Reservoir Pressure Gradient


The pressure used in the aquifer equation (5),
called P2, refers to the pressure at the original
gas water contact.
In a conventional material
balance calculation, it is assumed that P2 equals
P1 which is the average reservoir pressure in
front of the current gas water contact.
This
assumption is reasonable early in the production
life since only a small amount of influx has
entered the reservoir.

Most material balance methods calculate G and U


directly from historical pressure, production,
and time data.
Another method is to assume
values for G and U, and then calculate pressure
as a function of time. One can find the G and U
which minimize the difference between the calculated and observed pressure data. This procedure
is described next.

SPE 16484

But, later in the production life, a large pressure


gradient may be created in the water invaded
region between the original and the current water
gas contact, which results in a lower P1 value.
In other words, as production from a reservoir
increases, conventional material balance methods
may predict a higher reservoir pressure than the
observed pressure. This difference is especially
significant for reservoirs with large dip and
strong aquifers.

G- Gwgp
. . . . . (21)
G[1 - ce (Pi-P)] - Y

For a given G and U, the first pressure point


P(1) at the first schedule data point [t(1),
Gp(l), GLp(l) and Wp(l) 1 can be solved using an
iteration technique.
Then the second pressure
point P(2) can be found with the second schedule
data and P(l). The procedure continues until all
the pressure points are found.
Finally, the
standard deviation of the observed and the calculated pressure, and the water influx, are calculated and saved for comparison.

Lutes et al.19 proposed a modified material


balance method to account for the pressure gradient
effect in a radial water drive reservoir.
A
similar approach was included in GASWAT-PC. In
addition, the new option includes gravity20 and
28

SPE

I(Q4 <gl\

B. Wang and T. S. Teasdale

formation compaction effects, and is extended to


reservoirs with linear geometry. Figures 2.a and
2.b illustrate the schematic configuration for a
linear and a radial water drive reservoir, and
Figure 2.c shows their saturation profile.

where Ft is the hydrocarbon pore volume of


trapped gas, given by
Ft

The general material balance equation (1) can be


modified as
Gwgp Bgl

G(Egl + Efwl>
+ (We2- Wp Bwl)(l- A- B)

. . . . . (27)

(1 - Ev) +

Ev is the volumetric sweep efficiency,


Equation (26) can be rearranged as

. . . (22)

(1 -

Gwgpaw
G
)

and

(28)

where
A

Sgr

Bgl - Bgt

1 - Sgr - Swi

Bgt

Swi ow + Of
1 - Sgr - Swi

(Pt - P1)

Equation (28) is a line with intercept G at


(P/Z) = 0, and slope[- Pi/(Ft G Zi)].
By
plotting Equation (28) and P/Z versus Gwgp on
the same graph (Figure 3), we can determine
Gwgpaw and Paw from the intercept of the
curves .

. (23)

. (24)

If the pressure gradient in the water invaded


region is significant, the above relation is
not correct since the reservoir pressure is
not equal to the average pressure in the
trapped area. However, the cumulative water
influx at watered-out conditions is given by

Here, P1 is the average reservoir pressure in


front of the current gas water contact, P2 is the
reservoir pressure at the original gas water
contact, and Pt is the average reservoir pressure
in the water-invaded region. Subscripts 1, 2 and
t represent functions of P1, P2 and Pt.

Bgi G (1-Ft) + Wp Bw

Equation (22) is derived in Appendix C.


P1, P2
and Pt depend on reservoir geometry, and Appendix
D shows the relations for linear and radial
reservoirs.

Once the relations of Gwgp and P1 versus We


are generated by solving equation (22), one
can find Gwgpaw and Paw from Weaw

An iterative technique is used to solve the


modified material balance equation (22).
This
gives P1 and the current water gas contact (11 or
r1) at any time step t.

B.

Since it is difficult to distinguish a partial


water drive reservoir at early production
times, GASWAT-PC will first calculate Paw and
Gwgpaw for any water dri.ve reservoir. If the
calculated Paw is larger than the input Pa
which indicates that the reservoir will be
watered-out, then the ultimate recovery and
the abandonment pressure will be assigned to
Gwgpaw and Paw respectively. Otherwise, the
reservoir is a partial water drive type, and
Gwgp and Pa will be assigned to Gwgpa and Pa
respectively.

Ultimate Gas Recovery Prediction


1.

Natural Depletion and Partial Water Reservoirs


For a natural depletion or partial water drive
reservoir, the pressure drops rapidly as gas
is produced, and production stops at the
abandonment pressure (Pa), which depends on
operating line pressure, depth of the reservoir, and the productivity index. The recovery
efficiency (ER) is expressed as

CASE STUDIES

where Gwgpa is the ultimate gas recovery at Pa

To illustrate the various methods and their applications, four cases were analyzed with GASWAT-PC:
(1) an abnormally pressured reservoir with no influx,
( 2) a reservoir with a strong radial aquifer, ( 3) a
reservoir with an infinite linear aquifer, and (4) a
water drive reservoir with a large dip angle.

Watered-Out Reservoirs

Case 1:

For a strong water drive reservoir, water is


continuously entering the reservoir, and all
existing wells eventually water out. Therefore, the abandonment pressure (Paw> is
higher than for a volumetric or partial water
drive reservoir, and ER is lower since residual
gas in the water invaded zone (and in unswept
areas) is trapped at higher pressure. ER for
a watered-out reservoir is defined as

The Anderson "L" gas reservoir presented in the


Ramagost and Farshad paper4 is used here for illustration.
The reservoir has an initial pressure
gradient of 0. 843 psi/ft [19. 07 kPa/m] and a formation compressibility of 15 microsips (E-6 psi-1)
[2.18 x lo-6 kPa-1]. G = 69 BCF [1.98 x 109 m3] by
v ol umetrics.
Reservoir information and production
data are listed in Table 1.

Gwgpa
G

2.

. . . (29)

Gwgpaw
G

1 -

1 - Ft

. . . . . . (25)

An Abnormally Pressured Reservoir


With No Influx

The results are summarized in Table 1 and shown in


Figures 4. a to 4. d. All the data points are used in
the regression for P/Z method 1, and an incorrect

. . . . . (26)

29

GASWAT-PC:

A MICROCOMPUTER PROGRAM FOR GAS MATERIAL BALANCE WITH WATER INFLUX

G = 96.61 BCF [2.77 x 109m3] is obtained. Figure 4.a


shows the result of P/Z method 2.
Only the last
three points on the second slope are used in this
regression, and a more reasonable G = 79.86 BCF
[2.29 x 109 m3] is calculated.

aquifers.
The calculated G and U are 961.39 BCF and
15143.31 RB/psi/yr [27.53 x 109 m3 and 349.19
m3/kPa/yr] for the Schilthuis aquifer, and 892.683
BCF and 81380.50 RB/psi [25.56 x 109 m3 and 1876.57
m3/kPa] for the Hurst simplifiedaquifer, respectively.
The Hurst aquifer yields a a smaller deviation of
6.94%.

Figure 4. b demonstrates the result of the RF P/Z


method. All of the points are included in the regression and G = 74.04 BCF [2.12 x 109m3],

For the radial aquifer calculation, a range of ralrg


ratios (5 to 20) and TDF values (0 .S to 2. 0) are
input due to the uncertainties of aquifer size and
reservoir parameters.
Each pair of ralrg ratio and
TDF is used to calculate G and U.
So a total of 132
runs (12*11) are made, and then the two "best fit"
runs (minimum a and minimum absolute CURV) are automatically selected and listed in the summary table.

Figure 4.c shows the result of HO method. The calculated G is 77.05 BCF [2.22 x 109m3], which is close
to the RF P/Z value.
A Cf of 12 microsips [1. 74 x
10-6 kPa-1] is calculated from the slope using equation
(12), which is lower than the reported value of 15
microsips [2.18 x 10-6 kPa-1].

The minimum a run (ralrg=10 and TDF=0.6) shows a


small CURV of -11.82%, so we choose it as the best
representation for a radial aquifer. G = 821.88 BCF
[23 .54 x 109 m3], and U = 6330.95 RB/psi [145.99
m3 /kPa] which is smaller than the theoretical value
of 10740 RB/psi [247.66 m3/kPa]. The correct dimensionless time coefficient (Ktn> can be calculated by

Note that six data points (in the first 10% gas
production) are excluded from the HO calculation due
to their scattered distribution.
However, these
points are included in P/Z method 1 and RF P/Z method,
since P/Z plots are not as sensitive as HO calculations
to the early life data.6
The G and Cf values from the RF P/Z and HO calculations
are automatically input to the 'pressure match section, ' and then the reservoir pressure at each time
step is calculated and plotted in Figure 4.d. Comparing the observed data and the calculated results, it
is seen that both cases give excellent matches which
indicate the correct G value is within the range of
74 to 78 BCF [2.12 x 109 to 2.22 x 109m3],

KtD

= KtD

TDF

(12.78) (0.6)

7. 668 yr-1 (30)

Figure S .f demonstrates the HO plot of the radial


aquifer of ralrg = 10 and TDF = 0.6. Note that the
points moves up from Y value of 1210 to 1630, then
drop back to 1340 along a straight line.
This indicates that the assumed radial geometry is correct.

The calculation time required for this run on an IBM


AT is 42 seconds.
Case 2:

SPE 16484

After the OGIP calculations, the pressure match is


automatically checked for each method as shown in
Figure S.g.
In general, the match is good for all
methods up to 400 BCF [11.45 x 109 m3] production.
However, the radial aquifer gives a better pressure
match during the 400 to 600 BCF [11.45 x 109 to 17.18
x 109 m3] production period.

A Gas Reservoir With A Strong Radial Aquifer

This is a typical Louisiana offshore gas reservoir


which has a strong radial water drive.
G = 806 BCF
[23.08 x 109 m3] is estimated by volumetrics.
The
reservoir data and the output summary are listed in
Table 2.
Figures S.a-S.h demonstrate the results of
the various methods,

Now we can predict the future pressure performance


and ultimate gas recovery by using the select~d G and
U values. G = 821.88 BCF [23.54 X 109m3], U = 6330.95
RB/psi [145.99 m3/kPa], ralrg = 10, and tn/t = 7.668
1/yr are input in the pressure prediction section. A
constant rate of 30 BCF/yr is assumed, and the pressure
gradient in the water invaded region is neglected
since the dip angle is small. Sgr and Ev are assigned
0.25 and 0.8, respectively.

G = 1074.75 and 947.54 BCF [30.78 x 109 and 27.13 x


109 m3] are calculated by P/Z method 1 and method 2.
Only the first four points, which are considered to
be less influenced by the influx, are included in the
regression for both methods.
Figure S.a shows that
the rest of the points depart from the regression line
indicating the influx entering the reservoir.

The result is shown in Figure S.h. Under the assumed


conditions, the reservoir will water out at an abandonment pressure of 2138 psi a [14, 741 kPa] and an ultimate
wet gas recovery of 631.68 BCF [18.09 x 109m3],

Figure S.b is the Cole plot, which gives G = 916.65


BCF [26.25 x 109 m3] and U = 1.74.
The first two
points, and those points after 400 BCF [11.45 x
109 m3] gas production (which have negative slope) are
excluded from the calculation.

The calculation time required for this study is 333


seconds.

Six data points are excluded from the HO regressions


(see 'data select' Table 2).
Except for the small
pot aquifer, these excluded points are not shown on
HO plots since they change the value of the S(P,t)
term [see S(P,t) calculation in Appendix A].

Case 3:

A Gas Reservoir With An Infinite


Linear Aquifer

This is an infinite linear water drive reservoir


illustrated in reference 21.
The reservoir has an
initial pressure of 2333 psia and G = 337.7 BCF by
volumetrics.
The reservoir information and the
output summary are listed in Table 3. Figures 6.a to
6.d show the results of various methods.
All the
data points are included in regression for all methods.

Figure S.c shows the HO plot of small pot aquifer.


The calculated G = 1430.23 [40.96 x 109 m3] is incorrect due to the large deviation. Figures S.d and S.e
are the HO plots of Schil thuis and Hurst Simplified
30

SPE

I {Q4cg4-

B. Wang and T. S. Teasdale


The CPU time required for Run1 was 10 seconds.

Figure 6. a is the result of P/Z method 1, which


yields a G of 383.27 BCF [10.98 x 109m3] and a small
a of 0.91%. Figure 6.b shows the result of the Cole
plot, and a lower G = 347.83 BCF [9.96 x 109 m3] is
found.
The increasing trend of the data points
reveals influx entering the reservoir even though the
P/Z plot (Figure 6.a) appears quite linear.

CONCLUSIONS
GASWAT-PC is a powerful, low cost microcomputer
program for material balance analysis of gas reservoirs.
Several commonly accepted techniques and
aquifer models (and extensions) are included for
interactive examination of condensate, volumetric,
abnormally pressured, and water drive situations.

The result of the HO method for small pot aquifer is


incorrect since the slope of the regression line is
negative.
The calculated G and U for Schilthuis,
Hurst simplified and infinite linear aquifers are
339.69 BCF and 5960 RB/psi/yr [9. 73 x 109 m3 and
137.43 m3/kPa/yr], 332.13 BCF and 34315.29 RB/psi
[9.51 x 109 m3 and 791.28 m3/kPa], and 309.03 and
28424.93 RB/psi/yr**O.S [8.85 x 109 m3 and 655.46
m3/kPa/yr**0.5], respectively.
The infinite linear
aquifer yields the best straight line with a 3.09%
deviation (see Figure 6.c).

This program also can predict reservoir pressure


performance and ultimate gas recovery.
Some other
important features of this system include:

Just comparing the P/Z and the linear aquifer plots,


it is difficult to decide which is the better representation.
Here, the pressure match method can help
assist in this decision.
Figure 6. d illustrates the matches of the observed
P/Z and the calculated P/Z values for both methods.
It is evident that the linear aquifer gives the
better pressure (or P/Z) match.
Note that this straight line behavior of the P/Z plot
for water drive gas reservoirs has been reported in
many references.1,2 Bruns et al. indicated that the
inflection of this curve must be considered as a sign
of water drive.
The calculation time required in this case is 40
seconds.
Case 4:

1.

Gas gravity and gas composition correlations for


Z-factor calculation,

2.

User selection of the data points included in


each method,

3.

Automatic selection for the best radial aquifer,

4.

Calculation of pressure gradient in the water


invaded region,

S.

Pressure match checking for each method,

6.

Input panel to assist users in creating or editing


dataset, and

7.

Summary table and graphic output to help users in


analyzing results.

NOMENCLATURE
gas formation volume factor, RB/MCF
[res m3/stock-tank m3]
water formation volume factor, RB/STB
[res m3/stock-tank m3]
effective compressibility, psi-1 [kPa-1]
(cwSwi + Cf)/(1-Swi>
formation compressibility, psi-1 [kPa-1]
water compressibility, psi-1 [kPa-1]
cw
CGWC
current gas water contact
curvature, defined in equation (B-4)
CURV
expansion of water and reduction in pore
volume, RB/STB [res m3/stock-tank m3],
defined in Equation (4)
expansion of gas, RB/MCF [res m3 /stocktank m3], defined in Equation (3)
recovery efficiency, fraction
volumetric sweep efficiency, fraction
underground withdrawal, RB [res m3], defined
in Equation (1)
total trapped gas volume in HCPV, fraction
original wet gas in place, BCF [m3]
cumulative condensate produced, MMSTB [stocktank m3]
cumulative dry gas produced, BCF [m3]
trapped wet gas, BCF [m3]
cumulative wet gas produced, BCF [m3]
net thickness, ft. [m]
HCPV = hydrocarbon pore volume, MMRB [res m3]
Kc = condensate conversion factor, MCF/STB [m3Jm3]
~tD
dimensionless time coefficient, yr-1
KtD = theoretical dimensionless time coefficient,
yr-1, defined in Equation (A-10)

A Water Drive Reservoir With A High Dip Angle

To illustrate the effect of a pressure gradient in


the water invaded region, a hypothetical water drive
reservoir with a high dip angle was constructed. A
radial aquifer with ralrg = S was assigned, and
The dip angle is 30
G = 20 BCF [0. 57 x 109 m3 J.
degrees and Sgr is 0.25. The other reservoir parameters are listed in Table 4.
GASWAT-PC is used to predict the pressure performance
at constant gas production rate, with and without the
pressure gradient option ( Run1 and Run2, respectively).
The results are listed in Table 4 and shown in
Figure 7.
Note that P1 in Table 4 indicates the average pressure
in front of the current water gas contact.
It is
found that the pressure difference between the two
runs is 742 psi [5,116 kPa] at the end of production.
This shows that the pressure gradient effect is
significant in a strong or medium water drive reservoir
with a high dip angle.
The reservoir was modeled with a commercial numerical
simulator.
RZ coordinates of 134*6 grid blocks were
used in this study, and the results are listed in
Table 4 and shown in Figure 7.
GASWAT-PC and the
simulator give .very comparable pressure performance
predictions.
31

GASWAT-PC:

A MICROCOMPUTER PROGRAM FOR GAS MATERIAL BALANCE WITH WATER INFLUX

effective permeability to water in the


aquifer, md
effective permeability to water at residual
kwrg
gas saturation, md
distance of linear gas reservoir at current
L1
gas water contact, ft [m]
distance of linear gas reservoir at original
L2
gas water contact, ft [m]
molecular weight of condensate
MLc
original gas water contact
OGWC
p
pressure, psia [kPa]
average pressure in front of current gas
p1
water contact, psia [kPa]
pressure at original gas water contact, psia
p2
[kPa]
average pressure in the water invaded region,
Pt
psia [kPa]
water influx rate, RB/D [res m3/D]
qw
dimensionless water influx
Qd
aquifer radius, ft [m]
ra
radius of gas reservoir at original gas water
rg
contact, ft [m]
radius of gas reservoir at current gas water
r1
contact, ft [m]
rg f't [m]
r2
residual gas saturation, fraction
Sgr
initial water saturation, fraction
Swi
S(P,t)
aquifer function, defined in Appendix A
time, years
t
dimensionless time
tD
TDF
dimensionless time adjusting factor
reservoir temperature, F [C]
T
u aquifer constant
theoretical aquifer constant, defined in
Appendix A
pore volume of aquifer, MMRB [res m3]
vaa
width of linear reservoir, ft [m]
cumulative water influx, MMRB [res m3]
We
cumulative water produced, MMSTB [stockWp
tank m3]
y
function defined in equation (17)
z gas deviation factor
porosity, fraction
~
water viscosity, Cp [Pas]
1-lw
e dip angle, degree
'II= influx enchroachment angle, degree
specific gravity of condensate
Yc
specific gravity of formation water
Yw
a
standard deviation, defined in Equation (7)
0'
normalized standard deviation, defined in
Equation ( 6)

REFERENCES

kw

1.

J. R., Fetkovich, M. J. and Meitzen,


"The Effect of Water Influx on P/ZCumulative Gas Production Curves," J. Pet. Tech.
(March 1965) 287-291.

2.

Chierici, G. L., Pizzi, G. and Ciucci, G. M.:


"Water Drive Gas Reservoirs: Uncertainty in
Reserves Evaluation From Past History," J. Pet.
Tech. (February 1967) 237-244.

3.

Cole F. W.: Reservoir Engineering Manual, Houton,


Gulf Publishing Co., 1969,

4.

Ramagost, B. P. and Farshad, F. F.: "P/Z Abnormally Pressured Gas Reservoirs," paper SPE 10125,
presented at the 1981 SPE Annual Technical
Conference and Exhibition, San Antonio, Oct. 1981.

S.

Havlena, D. and Odeh, A. S.: "The Material Balance


as an Equation of a Straight Line," Trans., AIME.
Part 1:228 (1963). Part 2:231 (1964) I-815.

6.

Tehrani, D. H.: "An Analysis of Volumetric Balance


Equation for Calculation of Oil in Place and Water
Influx," J. Pet. Tech. (September 1985) 1664-1670.

Dake, L. P. : Fundamentals of" Reservoir Engineering, Elservier Scientific Pubishing Company, 1978.

8.

Schilthuis, R. J.: "Active Oil and Reservoir


Energy," Trans., AIME, 118:37.

9.

Hurst, W.: "Water Influx Into a Reservoir and Its


Application to the Equation of Volumetric
Balance," Trans., AIME, 151. 1943, 57.

10.

van Everdingen, A. F. and Hurst, W.: "Application


of the Laplace Transform to Flow Problems in
Reservoirs," Trans., AIME, 186. 1949. 305-324B.

11.

Cragoe, C. S.: "Thermodynamic Properties of Petroleum Product," Bureau of Standards, U. S. Department of Commerce Misc. Pub., No. 7 (1929) 26.

12.

Standing M. B.: "Volumetric and Phase Behavior of


Oil Field Hydrocarbon Systems," SPE AIME, Dallas,
1977.

13.

Wichert, E. and Aziz, K.: "Calculation of Z 's for


Sour Gases," Hydrocarbon Processing, 51(5) 1972,
119-122.

14.

Dranchuk, P. M., Purvis, R. A. and Robinson,


D. B.:
"Computer Calculation of Natural Gas
Compressibility Factors Using the Standing and
Katz Correlation," Institute of Petroleum, IP
74-008 (1974).

15.

Hall, K. R. and Yarborough, L.: "A New Equation of


State for Z-factor Calculations," Oil and Gas J.
(June 1973) 82-92.

16.

Standing M. B. and Katz, D. L.: "Density of


Natural Gases," Trans., AIME, 146. 1942, 64-66.

17.

Shagroni, M. A.: "Effect of Formation Compressibility and Edge Water on Gas Field Performance,"
master thesis, Colorado School of Mines, 1977.

Subscriots
a
aw
i
j
sc
1
2
t

=
=

SPE 16484

minimum abandonment pressure condition


watered-out abandonment condition
initial condition
index of loops
standard condition
location at current gas water contact
location at original gas water contact
trapped gas in water invaded region

ACKNOWLEDGMENTS
The authors thank Texaco Inc. for permission to
publish this paper.
They also thank Dr. Joe P. Vogt
for his assistance in making the numerical simulation
run.

32

Bruns,

v. c.:

SPE Jv4~~-----------------------------B_._w_a_n~g_a_n~d_T_.__s_._T_e_a_sd_a_l_e_______________________________
9-.
18.

Agarwal. R. G., Al-Hussainy, R. and Ramey, H. J.,


Jr.: "The Importance of Water Influx in Gas
Reservoirs," J. Pet. Tech. (November 196S) 13361342.

and
~ kw <cw + Cf)

u = 0.30S2

h W[

llw

O.S
] . (A-8)

19.

Lutes, J. L. et al: "Accelerated Blowdown of a


Strong Water-Drive Gas Reservoir," J. Pet. Tech.
(December 1977) 1S33-1S38.

20.

Dumore, J. M. : "Material Balance for a BottomWater Drive Gas Reservoir," Soc. Pet. Eng. J.
(December 1973) 328-334.

S(P,t) =

21.

Ikoku, C. U.: "Natural Gas Engineering," PennWell


Publishing Co. (1980).

APj is still defined by equations (A-6) and (A-7),


and Qd is dimensionless water influx which is a
function of dimensionless time t 0 and the ratio of
aquifer to gas reservoir radius, ralrg.

S.

van Everdingen-Hurst finite/infinite


aquiferS 8
n-1

APPENDIX A

The ore tic ally,


coefficient

DEFINITIONS OF AQUIFER FUNCTIONS AND AQUIFER CONSTANTS

Reference 8 provides tables and work curves of


Qd(to,ralrg), for infinite ralrg and for ralrg
ratios ranging from 1.S to 10.
In many field
cases, it was round that the aquifer ralrg = 10
was too small, but the infinite radial aquifer
was too large, so additional tables for ralrg
ratios between 10 and 100 were computed and are
included in GASWAT-PC.

. (A-1)

<cw + Cf) Vaq . . . . . . . . (A-2)

Vaq is the pore volume of aquifer.


2.

. . . . (A-10)

to get dimensionless time.


In radial aquifer
calculations, GASWAT-PC multiplies each dimensionless time value by the dimensionless time adjustment factor (TDF) to account for uncertainties in
Equation (A-10).

where subscr~pt n refers to time step n, and the


theoretical U is defined as

real time is multiplied by the

KtD

Small Pot AquiferS


S(P,t) = (Pi - Pn)

ra
- - ) . (A-9)
rg

2.309 kw

The following equations define the aquifer functions


[S(P, t)] for various aquifers, and the theoretical
aquifer constants (U) for the small pot, infinite
linear, and finite/infinite radial aquifers.
Note
that the correct U for various aquifers are determined
from matching production history, using HO or other
material balance methods.
1.

l APj Qd(tDj
j=1

radial

for the radial aquifer is

U = 1.119 ~ h(cw + Cf) rg2 ~


(A- 11 >
360

Schilthuis Steady State Aquifer6

where

~is

encroachment angle.

APPENDIX B
3,

Hurst Simplified Aquifer? ,9


n

S(P,t) =

L[Pij=1

CURVATURE CALCULATION

(Pj + Pj-1)] tj - tj-1


2
ln(12 tj)

In GASWAT-PC, the curvature of a set of data points


(x1Y1), (x2Y2), . (xnYn> is defined as follows:

(A-4)

1.
4.

Normalize the data set (xiYi) by setting

van Everdingen-Hurst infinite linear aquiferS,8


xi - Xmin

n-1

S(P,t)

. (A-S)

j=1

Yi - y(xmin>

where
for j

where max and mix represent the maximum and minimum


values, and the range of Xi is 0 to 1.

(A-6)
2.

Pj-1 - Pj+1
APj

. (B-2)

Xmax - Xmin
Pi-P1

APj

. (B-1)

Xmax - Xmin

for j

2 to n-1

Find the best quadratic fit (2nd order polynomial


regression) to the normal;ized data points (Xi, Yi).

(A-7)
(B-3)
33

GASWAT-PC:

10
3.

The curvature CURV (%) is then defined as


CURV = 100

SPE 16484

A MICROCOMPUTER PROGRAM FOR GAS MATERIAL BALANCE WITH WATER INFLUX

If the pressure gradient is negligible, P1 equals to


P2 and A and B will be zero.
Then the modified
material balance equation will be identical to the
conventional equation (1).

(C-4)

CURV represents a measure of how much the parabola


(B-3) departs from a straight line over the range
X = 0 to 1 (Y = bo + b1 X is not necessarily the
"best" straight line for the data set).

APPENDIX D
P1 AND Pt CALCULATIONS FOR LIBEAR AND RADIAL RESERVOIRS

Figl,lre 8 shows 100(Y - bo) /b1 for different


values of CURV.
Note that CURV > 0 yields an
upward curvature while CURV < 0 gives a downward
curvature.

Linear Reservoir:
(D-1)

APPENDIX C
MODIFIED MATERIAL BALANCE EQUATION

P2 - ( .001127 kwrg h W

As water influx enters into the reservoir, the conventional material balance equation has to be modified
by considering the pressure gradient in the water
invaded region (see Figure 2).

+ 0.433 rw tane) (L2-L1).

.(D-2)
.(D-3)

original HCPV = current HCPV in front of CGWC


+ current HCPV between CGWC and OGWC
+ HCPV reduction in front of CGWC
+ HCPV reduction between CGWC and OGWC
+ water influx
- underground water withdrawal

Radial Reservoir:
5.6146 x 106 (We2-WpBw1> 0.5
(1-Sgr-Swi> ~nh W/360) ]
. . (D- 4 )

th\,lS,

qwuwln(r2/r1)
P2(G - Gwgp - Gt)Bg1

G Bgi

- 0.433rwtane(r2-r1) (D-5)

0.00708kwrgh

+ Gt Bgt
1 - Swi
+ ce (Pi-P1) [aBgi- Gt Bgt
1 - Swi
+ Ce (Pi-Pt) [at Bgt

Sgr

Sgr

P2 -

365.25
(We2 - Wp Bw1> Sgr
1 - Sgr - Swi

. . . . (C-2)

OAPI

G(Eg1 + Efw1>
+ (We2 - Wp Bw1> (1 - A - B)

Sgr
1 - Sgr - Swi

Swi cw + Cf
B= 1 Sgr - Swi

r1+r2

] . (D-6)

Bg1 - Bgt

. . (C-3)

(C-4)

Bgt
(Pt-P1)

We2- Wp Bw1
tn - tn-1

. . . (D-7)

141.5/(131.5 +

= g/cm3

API)
E-01

m3

cp x 1.0*

E-03

Pas

ft

3.048*

E-01 = m

ft3

2.831 685

bbl

where
A

SI METRIC CONVERSION FACTORS

equation (C-1) can be rearranged as


Gwgp Bg1

r22+r1r2+r12

. . . . . (C-1)

and since

Gt Bgt =

The qw in equations (D-2), (D-5) and (D-6) is defined


as

+ We2
- Wp Bw1

qwllw
[
0 5 0.00708kwrgh

1.589 873

oc

(F-32) /1.8
psi x 6.894 757

(C-5)

* Conversion factor is exact.

(C-6)
34

E+OO

kPa

SP'E 1 6 4 8 4,

TABLE 1

INPUT_AND OUTPUT OF CASE 1


This Example Chosen from SPE paper 10125:4
Anderson "L" Abnormally Pressured Reservoir. South Texas.
water compressibility
formation compressibility
porosity
initial water saturation
reservoir temperature
initial pressure
initial gas formation factor
depth
original gas in place
Data Selection
PZ1

PZ2

1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1

0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1

0
1

RF

HO

1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1

1
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1

3 microsips
15 microsips
0.2
0.35
267 F
9507 psia
0.5537 RB/MCF
-11167 ft
69 BCF by volumetrics

. . . . . . . . . . . . . . Production Data . . . . . . . . . . . . . . .
p
t
Gp
GLp
{MSTB}
z
{BCF~
{J2sia}
{da:i:sl
0.0
69.00
182.00
280.00
340.00
372.00
455.00
507.00
583.00
628.00
663.00
804.00
987.00
1183.00
1373.00
1556.00

9507.0
9292.0
8970.0
8595.0
8332.0
8009.0
7603.0
7406.0
7002.0
6721.0
6535.0
5764.0
4766.0
4295.0
3750.0
3247.0

1.4400
1.4180
1.3870
1.3440
1.3160
1.2820
1.2390
1.2180
1.1760
1.1470
1.1270
1.0480
0.9770
0.9280
0.8910
0.8540

o.o

o.o

0.393
1. 642
3.226
4.260
5.504
7.538
8.749
10.509
11.758
12.789
17.262
22.890
28.144
32.567
36.820

29.90
122.90
240.90
317.10
406.90
561.20
650.80
776.70
864.30
939.50
1255.30
1615.80
1913.40
2136.00
2307.80

excluded in regression
included in regression
. . . . . . . . . . . . . . . . Output Summary . . . . . . . . . . . . . . . . .

Method Used

Gl

BCF

P/Z method 1
P/Z method 2
Ramagost-Farshad
Havlena-Odeh *

96.61
79.86
74.04
77.05

~~.

3.18
0.22
2.12
9.94

Formation compressibility of 12 microsips is calculated from the


slope

35

SPE 1 b 4 84.

TABLE 2
INPUT AND OUTPUT OF CASE 2

A Typical Louisana Offshore Gas Reservoir with Strong water Influx


water compressibility
formation compressibility
water viscosity
porosity
initial water saturation
reservoir temperature
initial pressure
initial gas formation factor
gas radius
aquifer permeability
aquifer thickness
water enchroachment angle
original gas in place
theoretical dimensionless time
coefficient KtD
theoretical radial aquifer constant

1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1

c..QI&

1
0
0
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0

12.78 yr-1
10740 RB/psi

.... Production Data ...............


p
t
Gp
Wp
GLp

Data Select
~

3.1 microsips
3.2 microsips
0.33 cp
0.306
0.139
201 F
4243 psia
0.7673 RB/MCF
13000 ft
571 md
59 ft
180 degree
806 BCF by volumetrics

tlQ

__z_

1
0
0
0
0
1
0
1
1
0
1
1
1
1
1
1
1
1
1
1
1
1
1

o.o

4243.0
4165.0
4108.0
3935.0
3760.0
3538.0
3433.0
3408.0
3331.0
3178.0
3138.0
3100.0
3054.0
3078.0
2933.0
2917.0
2897.0
2875.0
2807.0
2496.0
2355.0
2300.0
2291.0

0.9718
0.9676
0.9651
0.9566
0.9486
0.9395
0.9357
0.9348
0.9322
0.9275
0. 92 64
0.9253
0.9242
0.9248
0.9214
0.9211
0.9207
0.9202
0.9190
0.9146
0.9142
0.9143
0.9143

283.0
452,0
799.0
1296.0
2172.0
2607.0
2966.0
3298.0
3663.0
4028.0
4390.0
4767.0
5156.0
5414.0
5504.0
5579.0
5809.0
5940,0
62 9 3 . 0
6501.0
6807.0
6875.0

CMSTB)

_.J.('1..S_'rn.l

o.o

o.o

23.451
29.324
57.343
102.318
165.984
190.122
226.208
263.241
297.909
327.622
352.843
370.186
387.236
417.052
424.861
433.398
449.844
465.379
521.950
549.323
575.786
581.104

129.58
150.52
214.96
338.86
534.18
603.17
698.68
808.73
902.81
1001.46
1070.02
1114.92
1165.00
1251.62
1272.07
1291.01
1330.60
1359.84
1457.16
1493.04
1524.69
1531.62

o.o
o.o
0.72
9.42
27.54
56,95
62.65
78.22
93.05
118.55
173.58
201.13
207.46
240.77
307.01
317.04
343.42
439.21
457.99
667.32
861.46
1153.86
1245.00

excluded in regression
included in regression

TABLE 2 (Cont.)
INPUT AND OUTPUT OF

CAS.IL2

...................... Output Summary ........................


___M_et,\}~q .. P.~-~g____
P/Z method 1
P/Z method 2
Cole
Havlena-Odeh:
Small Pot
Schilthuis
Hurst Simplified
Radial (best fit)
ralrg=10 TDF=0.6
ralrg=18 TDF=0.6

G. BCF

_L_%

1074.75
947.54
916.65

1. 74

9.19
7.85
6.97

1430.23
961.39
892.68

7681.15
15143.31
81380.50

26.69
10.31
6.94

821.88
862.67

6330.95
5270.78

3. 04
5.70

units of aquifer constant U for different types:


Cole method
Small Pot
Schilthuis
Hurst simplified
Radial

dimensionless
RB/psi
RB/psi/yr
RB/psi
RB/psi

minimal absolute normalized stanoa.rcl deviation a(F/EG)


minimal absolute curvature ICURVI

36

QQ.RY .....%

-11.82
- 0,08

TABLE 4

TABLE 3
INPUT AND

OUTPJ!~ .. Of .. CASE

INPUT AND OUTPUT OF CASE 4

A Gas Reservoir with Infinite Linear Water Influx


This example chosen from "Natural Gas Engineering" by C. U. Ikoku21
3 microsips
4 microsips
0. 308
0.425
155 F
2333 psia
1.172 RB/MCF
100 md
32.5 ft
337.7 BCF by volumetric

water compressibility
formation compressibility
porosity
initial water saturation
reservoir temperature
initial pressure
initial gas formation factor
aquifer permeability
aquifer thickness
original gas in place

............ Production Data ........... .

-...j

U?~r:..tl

_ _z_

___rn_&:L_

0.0
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
18.00
20.00

2333.0
2321.0
2203.0
2028.0
1854.0
1711.0
1531.0
1418.0
1306.0
1227.0
1153.0

0.8829
0.8843
0.8849
0.8893
0.8956
0. 9 005
0.8997
0.9134
0.9223
0.9231
0.9289

0.0
2.305
20.257
49.719
80.134
105.930
135.350
157.110
178.300
192.089
205.744

. . . . . . . . . . . . . . . . . . . Output Summary . . . . . . . . . . . . . . . . . . . .
Method Used
G. BCF
U *
~
383.27
347.83

0.21

0.91
9.77

429.56
339.69
332.13
309.03

-44339.40
5960.92
34315.29
28424.93

18.37
10.17
9.36
3. 09

Units of aquifer constant U for different types:


Cole method
= dimensionless
Small Pot
= RB/psi
Schilthuis
= RB/psi/yr
Hurst simplified = RB/psi
Infinite Linear = RB/psi/yr0.5

formation compressibility
water viscosity
porosity
initial water saturation
residual gas saturation
reservoir temperature
gas gravity
gas radius
aquifer permeability, kw
effective permeability to water at
residual gas saturation, kwrg
aquifer thickness
water enchroachment angle
dip angle
original gas in place
aquifer radius/gas radius ratio, ralrg
dimensionless time coefficient KtD
radial aquifer constant u

4 microsips
0.25 cp
0.2
0.25
0.25
250 F
0.6
3786 ft
500 md

125 md
100 ft
30 degree
30 degree
20 BCF
5

230.1 yr- 1
187.2 RB/psi

. . . . . . . . . . . . . . . . . . . . . . Output Summary . . . . . . . . . . . . . . . . . . . .
Calculated P1, psia
t
Gp
.lQgy_.l
..._...R]m!_
Simulation
~
~

All data points are included in regression for all methods.

P/Z method 1
Cole
Havlena-Odeh:
Small Pot
Schilthuis
Hurst Simplified
Infinite Linear

Pressure Prediction for a Hypothetical Water Drive Reservoir With


A High Dip Angle.

0.0
182.5
3 65.0
547.5
730.0
912.5
1095.0
1277.5
1460.0
1642.5
1825.0
2007.5
2190.0
2372.5
2555.0
2737.5
2920.0
3102.5
3285.0
3467.5

o.o
0.730
1.460
2.190
2.920
3.650
4.380
5.110
5.840
6.570
7.300
8.030
8.760
9.490
10.220
10.950
11.680
12.410
13 .140
13.870

10173.0
10009. 0
9866.0
9709.0
9562.0
9403.0
9236.0
9068.0
8908.0
8747.0
8567.0
8392.0
8221.0
8030.0
7840.0
7643.0
7435.0
7214.0
6972.0
6684.0

10173.0
10010.9
9859.4
9704.6
9547.8
9388.5
9226.4
9061.5
8893.9
8723.4
8549.0
8370.2
8186.4
7996.6
7801.2
7597.8
7384.3
7156.9
6909.9
6629.0

10173.0
10027.4
9893.8
9757.9
9620.9
9482.7
9343.2
9202.1
9059.5
8915.7
8770.8
8623.9
8475.3
8324.4
8171.5
8016.1
7859.2
7699.6
7537.1
7371.3

Simulation = numerical simulation result


Run1
Run2

GASWAT-PC result, included pressure gradient effect in


water invaded region
GASWAT-PC result, excluded pressure gradient effect

en

-o
rn
a---a
0
~

ex;

....

SPE 1 6 4 8~

5000

.....a:S

4000

en

p..

N'

P:;-

3000

2000

OBS
P-~!_

____ _

PZ2
1000~----~--~-----r----~----~--~-----r--~

20

40

80

60

Gwgp,BCF
Fig. 1-Sample plots of P/Z Method 1 and Method 2.

WATER DRIVE \
PERFORMANCE \

0.8

0.6

\\WATER-OUT LINE

',

\\
\

(Gwgp aw, Paw)


\
G
Zaw

',

0.4

0.2

0~------~----~--~-------r------~~----~
0
0.4
0.6
0.8
1
0.2

vr
z

Gwgp/G, fraction

1.o

~ 1-Sor

oC

ffi
....
oC

INFLUX

GAS

Ill

AQUIFER

Fig. 3-Uitimate gas recovery calculations for depletion .and water-drive reservoirs.

Swl
I

0
,. 1

Lz

La

~"z

~"a

LENGTH OR RADIUS
(C)

Fig. 2-Schematic of water-drive reservoirs: (a) linear flow model, (b) radial flow
model, and (c) saturation profiles.

38

SP'E 1 6 4 84.

(a) P/Z METHOD 2. G = .79.86 BCF

(b) RF P/Z' PLOT I G = 74:.04: B CF

: ....I

8000T-----------------------------~

USED IN FIT

0.

NOT USED

~ 8000

..,.

v
0

0)

0.

N
.........
Q.

'"""-.,..,.. ............_

"0

i:,
~

4000

40 0 0

.,.

.'""'

2000+-------~------~------~----~

2000,+-------~------~------~------~

40

20

Gwgp, BCF

(c) HO METHOD. G = 77.05 BCF

10000T-----------------------------~

-~~

.............

USED IN FIT
NOT USED

(.)

m
100

v--.............

....

~'

40

(4) COJIP.ARISON OF PRESSURE JI.ATCBIS

150~----~----------~----------~

lL

20
Gwgp,BCF

111.

1000

...~.....

.,.

HO

--RF

---...

..............

----.

................

0+-------~------T-------~----~

50 ~o----~---2-oTo,-o-o--~---4o-oro-o---?---8-0~ooo

(PI-P)/Eg, psi-MCF/RB

Fig. 4-Piots of Case 1.

39

20

Gwgp, BCF

40

SPE 1 6 4 84;

(a) P/Z METHOD 1, G = 107-&.'75 BCF

(b) COLE PLOT, G

2000r----------------------------,

SOOOr----------------------------,

~!!~P.l!tf.l.~

~~m~
o+---~----r-------~r-------~
0

800

40 0

100

400

Gwgp,BCf

(o) SIIALL POT AQUIFIR, G '" U80.88 BCF

(4) SCBILTBUIS AQUIFIR, G '" tel.at BCF

,.

1800

100

Gwgp, BCf

2000r----------------------------,

2000r----------------------------,

...

NOT USED

20oo+---~----r---~--~r-------~
20 0

....

1000 ........ . -

3000

-~

.-"

~~

::

4000

~
0..

.
. ..........

....0

NOT USED
"

= 916.85 BCF

1--------:::'""''--,._
,....

w"'
-;;:. 1000

~!!~~.l!4.f.l.~

-~~

1100

_.,...,..
~1000/'

NOT USED

soo+-------2 o-o~o-o------.-o~oo-o------,-4oooo

soo+-----~--~--~------r-~---4
2000

4000

1000

8000

S(P,T)/Eg, pii-YR-MCf/RB

(PI-P)/Eg, psi-MCf/RB

(f) ~A.~~"\1~ \u,~v n/ra=10 TDF=O.II

11

(e) BURST AQUIFER, G::: 8112.118 BCF


2000

....

1800

2000

...
.
___.___....
:~:

.... uoo
0

.;.

-;;:. 1000

~--

~.......~

-;;:. 1000

sao

500
0

5000

10000

S(P,T)/Eg, pei-MCf/RB

SOOOr----------------------------,

:."

a: 3000

- - - - SCHIL THUIS
-----HURST

..... ..........,,...____

"a.

4000

'ii

---RADIAL

"'~

~
Q.

..............

1000

200

Gwgp,BCf

40 0

.....\ ;,

---RADIAL

~..;:~

'''

,.~

2000

150000

8000~---------------------------,

-----COLE

........
.......

100000

(b) ULTIIIATI GAB RICOYIRY CALCULATION


Pa'" 11811 pla OWapa'" 881,811 BCF

(a) COMPARISON OF PRISSURIIIATCBIS

4000

50000

S(P,T)/Eg, pal-MCf/RB

----WATER-OUT
LINE

""'''

eoo

500

Gwgp,BCf

Fig. 5-Piots of Case 2.

40

''

'

1000

SPE 1 6 4 84,

(a) P/Z METHOD 2, G::: !183.2'7 BCF

(b) COLE PLOT, G

3000,.---"'---'~---------,

= 3f'7.83 BCF

400,.--~~----------,

~.

~.A~~

CD

.;.
w

;;

~ J!IO ~~
If
~

": 2000

0.

0>

~
IOOO+-----,--~---r--------1

100

200

300~-~-~--~--r--~-~

300

(o) llffiMIT. LIMIAa .lCIUIJIII,

/.

.;.

J!IO

Q "

{4) COMPARISON

ad.OII ICr

/ ..t<"'

or PRISSURI 114TCBIS

40

....

-PZI

'iii
0.

20

ui

/,/'

JOO

300

Cwgp, BCF

400

II...

zoo

100

Gwgp, BCF

.0
0

0..

cu

a..

1000
toOO
S(P,l')/Eg, psi-MCF/RB

-20

3000

100

100

300

Gwgp, BCF

Fig. 6-Piots of Case 3.

PRESSURE PREDICTION

10000

<

9000

(/]

Il-l

8000

7000

SIMUL
~-V.~.L .. .

8000~------~------~------~~~----~------~--._--~

10

Gwgp,BCF
Fig. 7-Piots of Case 4.

41

15

R~~.&.. ...

SPE 1 b 4 8 4~

CURV=%

1.5------------------------ 60
40
30
20
10
1 ... ..... ........ ......................... a

,.....

-10

-20
-30

,.0

-40
~~..........- , -60

0~--------~--------+-----------------~
0
0.5
1

42

You might also like