SPE 167504

Analysis of Gas Production Data Using Flowing Material Balance Method

S. Mohammed, SPE, and G.S. Enty SPE, University of Mines and Technology
Copyright 2013, Society of Petroleum Engineers
This paper was prepared for presentation at the Nigeria Annual International Conference and Exhibition held in Lagos, Nigeria, 30th July 1st August 2013.
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 have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
This paper presents an analytical derivation that constitutes a sound theoretical background for gas flow equation. This gas
flow equation is based on a new pseudotime function which, unlike the computation of material balance pseudotime, is insensitive to time step-size and purely analytical. This new pseudotime function offers a simpler approach to handle viscositycompressibility variations since viscosity-compressibility ratio is a function of cumulative production. Unique to the proposed approach is that the flowing material balance method utilizes pseudocumulative, which is not a function of material
balance pseudotime.
Currently, the analytical derivation given in the literature for gas flow equation involving material balance pseudotime has
created the perception that material balance pseudotime is intuitive. An analytical derivation is given to show that material
balance pseudotime function has sound theoretical basis.
Presently, iterative scheme, algorithms and graphical techniques involving a number of plotting functions have been proposed to solve gas-in-place. This paper presents a direct approach to solve gas-in-place when early pseudosteady state line is
observed. Additionally, a technique is given to validate computed initial-gas-in-place; thus initial-gas-in-place computed by
any method can be verified with the proposed technique.
The proposed analysis yields initial-gas-in-place, pseudosteady state constant and drainage area. Two simulated and one field
published examples are presented to validate our proposed analysis.
Introduction
Material balance pseudotime has gained widespread acceptance due to its ability to handle general variable pressure/variable
rate scenarios. Further, this pseudotime function evaluates variations of gas properties (i.e., viscosity-compressibility values)
at average reservoir pressure, and thus, makes it suitable to handle longer-term boundary-dominated flow of real gas. While
material balance pseudotime is rigorous for gas wells, it is dependent, as regard its computation, on time step-size. Also, it is
not suitable for shut-in periods (Suabdi 2001). Thus, it gives misleading result during shut-in times. Normalized cumulative,
has been shown to be a function of material balance pseudotime (Argawal et al 1999; Mattar and Anderson 2003), and thus
render its computation time-step-dependent as well.
The computation of pseudotime functions requires a prior knowledge of average reservoir pressure, or indirectly original-gasin-place. Consequently, various methods have been proposed in the literature to solve gas-in-place. Unfortunately, some of
these methods are time-consuming; others utilize derivatives which may introduce noise; others are restricted to constant bottomhole flowing pressure and others involve a lot of plotting functions.
Blasingame and Lee (1988) proposed a modified gas flow equation based on material balance pseudotime to model the general variable pressure/variable rate case during boundary-dominated flow. The authors based their direction of thought on the
material balance equation for liquid case, and, then verified their equation using simulated data. Palacio and Blasingame
(1993) provided a theoretical development of Blasingame and Lees (1988) modified gas flow equation. In their derivation,
the authors developed an expression for gas flow rate and then substituted it into a predefined material balance pseudotime
function to yield gas material balance equation. Thus, material balance pseudotime function did not emanate from their derivation. Further, the authors introduced an algorithm, which require back calculations, to evaluate the average reservoir pres-

SPE 167504

sure profile.
Mattar and McNeil (1998) proposed a flowing gas material balance method for constant rate case without shut-in to compute
gas-in-place. Their analysis is based on the fact the pressure at any point in the reservoir declines at the same rate during constant rate boundary-dominated flow. Thus, the pressure drop measured at the wellbore is the same as the pressure drop that
would be observed anywhere in the reservoir for constant rate boundary-dominated flow. Consequently, the authors shifted
the straight line depicted by a plot of sandface or wellhead flowing pressure versus cumulative production to the initial reservoir or initial wellhead pressure to yield gas-in-place on the x-intercept. Mattar and Anderson (2003) proposed a flowing material balance method based on the modified version of Agarwal-Gardener rate/cumulative type curves. Their analysis involves a plot of pseudopressure drop normalized rate against pseudopressure drop normalized cumulative on a linear scale.
Their analysis yields initial-fluid-in-place on the x-intercept. The authors defined their normalized cumulative in terms of
material balance pseudotime. Mattar and Anderson (2005) presented a dynamic flowing material balance method to extend
Mattar and McNeil (1998) technique to variable rate case. Their method, however, requires a prior knowledge of pseudosteady constant. Further, the authors used an iterative scheme to solve gas-in-place.
Argawal (2010) demonstrated that a log-log plot of pseudopressure drop logarithmic derivative against actual flowing time
yields a unit slope during early time pseudosteady state but deviates during the late time pseudosteady state. He also noted a
similar observation with prime derivative: a horizontal line (constant value) is observed during early time pseudosteady state
and deviation from the horizontal line at late time during pseudosteady state. He determined the gas-in-place with the constant value. Unfortunately, derivatives introduce noise, which may pose difficulty in observing the correct horizontal line.
This noise may even be worse for production data which has lower frequency.
Purpose of this study
The purpose of this study is to present a flowing material balance equation based on pseudocumulative function to model gas
production data. Although, this pseudocumulative function has been reported in the literature (Callard and Schenewerk
1995), it has NEVER been applied to flowing material balance method before.
First, we will introduce a new material balance pseudotime function to model gas production data. Second, we will show that
this new material balance pseudotime function and the traditional material balance pseudotime have sound theoretical basis.
Third, we will propose a flowing material balance equation based on pseudocumulative function. Fourth, we will present a
new finding to compute gas-in-place. Fifth, we will present a universal method to validate computed gas-in-place. Thus, gasin-place computed by any technique can be validated by our method. Lastly, we will present a simple analytical technique to
compute pseudocumulative function.
The New Material Balance Pseudotime Function
We now present an analytical derivation for gas flow equation for general variable pressure/variable rate case involving the
new material balance pseudotime function. This new material balance pseudotime function is given by, Ta:

1
Ta
qg

Gp

( g Ct )i

g Ct

dG p .. (1)

We now proceed with the theoretical development of gas flow equation for general variable pressure/variable rate case involving the new material balance pseudotime function.
The general expression for

dp

q g Bg
V P S gi C t

d p is given by (Fraim and Wattenbarger 1987):

dt .. (2)

The relation between gas cumulative production and gas flow rate is given by:

qg

dG P
...................................................... (3)
dt

Combining Eqs. 2 and 3 yields:

dp

Bg
V P S gi C t

dG P ..

(4)
Also, gas formation volume factor can be expressed as:

SPE 167504

Bg

zTPsc

. (5)

PTsc

Upon substituting Eq. 5 into Eq. 4, we have:

dp

zT
P

1

dG P . (6)
VP S gi C t

Psc

Tsc

Now, normalized pseudopressure is given by (Meunier et al 1987; and Palacio and Blasingame 1993):

m( P ) n

z
g

Pi

g z

d p ........ (7)

which is differentiated to give:

PZ

dm P n

P
d p .. (8)
g z

Now, when we substitute Eq. 6 into Eq. 8, we have:

Psc g i
T

dm P n
dG p .. (9)
P

VP S gi i Tsc g C t

zi

Initial-gas-in-place can be expressed as :

VP S gi

G
T

Pi

z i Tsc

Psc

....... (10)

When we combine Eqs. 9 and 10, we have:

dm P

g 1
dG p (11)
i
G
C

We can now integrate Eq. 11 from the initial normalized pseudopressure, which corresponds to zero cumulative, to any given
average reservoir pressure, which corresponds to a cumulative production of GP. In other words, we are integrating from initial time (i.e., t=0) to any given time, t,. That is:

n
1
dm P n
GCti
m Pi n
m P

which then becomes:

GP

C
g

t i

g Ct

dGP ... (12)

SPE 167504

mPi n

1
m P n
GCti

Gp

( g Ct ) i

g Ct

dG p ... (13)

Now, pseudocumulative is given by (Callard and Schenewerk 1995):

Gp

G pn

( g Ct )i

dG p ...... (14)

g Ct

Eq. 13 then becomes:

m P n mPi n

1
G pn .... (15)
GCti

Clealy, Eq. 15 has a form that is similar to the liquid material balance equation (Dake 1978):

P Pi

1
N P (16).
NCt

Therefore, using normalized pseudopressure instead of actual pressure, and pseudocumulative instead of actual cumulative
linearizes the single-phase gas flow equation. This allows us to solve single-phase gas problems with the liquid solution. To
our knowledge, Eq. 15 has not been reported in the literature.
Now, gas flow equation during pseudosteady state (pss) flow regime is given by (Al Hussainy et al 1966; Palacio and
Blasingame 1993):

m P n mPwf n q g ba , pss (17)

When we combine Eqs 13 and 17, we have:

mPi n mPwf n
qg

GCti

q g

Gp

( g Ct ) i

g Ct

dG p ba , pss ..... (18)

The expression in the bracket on the right hand side (RHS) of Eq. 18 becomes our basis for defining our new material balance pseudotime function, Ta, in Eq. 1:

mPi n mPwf

qg

1
Ta ba , pss ....... (19)
GCti

This constitutes a sound theoretical proof of our new pseudotime function. This new material balance pseudotime function,
Ta , shall be called rate normalized pseudocumulative. Eq. 19 is a gas flow equation based on the rate normalized pseudocumulative function, and, is strictly applicable to boundary-dominated flow. Eq. 19 will be extended to decline type curve analysis in our future papers. We shall express Eq. 18 later in this work to yield a form similar to the traditional flowing material
balance method.
Palacio and Blasingame (1993) presented a theoretical derivation of gas flow equation involving material balance pseudotime. To our knowledge, their derivation does NOT develop material balance pseudotime function. It rather shows that material balance pseudotime is suitable to model gas production data.
To extend Eq. 18 to the traditional material balance pseudotime function, we substitute Eq. 3 into the integral expression of
Eq. 18 to yield:

mPi n mPwf n
qg

GCti

( g Ct ) i

q g

dt
0 g C t ba, pss ... (20)

qg

Eq. 20 is exactly the gas flow equation based on the traditional material balance pseudotime, which was presented by

SPE 167504

Blasingame and Lee (1988) and later by Palacio and Blasingame (1993). Hence, material balance pseudotime has sound
theoretical basis.
Note that a comparison of Eq. 18 and Eq. 20 shows that rate normalized pseudocumulative (introduced in this paper) and
material balance pseudotime (introduced by Palacio and Blasingame 1993) are equivalent.
Proposed Flowing Material Balance Method
We now present a flowing material balance (FMB) equation based on Eq. 18 to analyze gas production data.
First, we express Eq. 18 as:

mPi n mPwf n
qg

1
GCti

G pn

ba , pss .... (21)

q g

Second, we rearrange Eq. 21 to yield:

qg

mP n

1
ba , pss GCti

G pn
1
.... (22)

mP n ba , pss

where

mP n mPi n mPwf t n ..... (23)

Eq. 22 describes the proposed flowing material balance for gas wells. Note that, as opposed to the traditional flowing material
balance equation (Mattar and Anderson 2003; Syah 1999), Eq. 22 does not involve material balance pseudotime. Thus, a different pseudocumulative function (i.e., Eq. 14) is utilized here. From Eq. 22, a plot of

qg

mP n

against

G pn

mP n

on linear

co-ordinates yields a straight line from which the initial-gas-in-place can be determined on the x-intercept. The pseudosteady
state constant can be determined from the y-intercept.
The viscosity-compressibility values required to compute pseudocumulative, just like pseudotime, ought to be evaluated at
the average reservoir pressure, which in turn are functions of initial-gas-in-place, G,. Consequently, a number of methods
have been proposed in the literature (Fraim and Wattenbarger 1987; Blasingame and Lee 1988; Palacio and Blasingame
1993; Keating et al 1994; Agarwal 2010) to solve for G value.
In the next section, we present a quick graphical method to solve for G.
Estimation of Gas-in-Place
Before the non-linear effect of viscosity-compressibility product dominates a rate/time data, the early portion of boundarydominated flow data can be modeled with actual cumulative production, instead of pseudocumulative. In other words, at early time during pss flow regime, both actual and pseudo cumulative will trace the same path. At late time however, the actual
cumulative will deviate due to the non-linear effect of viscosity-compressibility product. We therefore extrapolate the early
time pss line to the x-intercept to estimate G.
To take advantage of the above, we approximate Eq. 22 as:

qg

mP n

1
ba , pss GCti

Gp
1
.. (24)

mP n ba , pss

We found out (shown later in the example applications) that this approximate plot (i.e., Eq. 24) yields two straight lines during pss flow regime early time and late time pss lines. Extrapolation of the early time pss line to the x-intercept yields GCti.
This method provides a quicker approach than the ones proposed in the literature, and will be of enormous benefit as regard
the computation of pseudotime functions evaluated at the average reservoir pressure.

SPE 167504

In appendix A, we have proposed a universal technique to validate computed gas-in-place. We shall utilize this method in
some of our example applications.
Computation of Pseudocumulative
As shown in appendix B, the equation required to compute pseudocumulative function will have this polynomial form:
Gp

G pn

where

( g Ct )i

g Ct

dG p G p

C1 2 C2 3 C3 4 C4 5
G p G p G p G p ... (25)
2
3
4
5

C1 , C2 , C3 ... , in Eq. 25 are constants.

Procedure for computing Pseudocumulative

1. Make a linear plot of viscosity-compressibility ratio against actual cumulative production.
2. Fit a polynomial curve to the data at least to the third degree and set the intercept to 1.
3. Integrate the resulting polynomial equation with respect to the actual cumulative production to yield a form similar to Eq.
25.
4. Substitute each actual cumulative production data point into the integral expression to obtain the corresponding pseudocumulative.
In the next section, we present the procedure for generating the proposed flowing material balance plot.
Procedure for Generating the Proposed FMB Plot
1. Make an approximate plot using Eq. 24.
2. If two lines are observed (i.e., early and late time pss lines), extrapolate the early time pss line to the x-axis and solve for
G. This yields the correct G at one time. However, if the early time pss line is absent, any known method can be used to solve
for G.
3. Use the material balance equation to compute the average reservoir pressure profile, and, subsequently, the viscositycompressibility values.
4. Compute for pseudocumulative using the procedure outlined previously.
5. Now, make the rigorous plot using Eq. 22 on a linear graph.
Table 1 summarizes the flowing material balance method introduced in this work.
Table 1: Flowing material balance method introduced in this work.
Rate and Pressure
scenario
General
(rigorous)
General
(approximate)

y-axis

x-axis

qg
mP n
qg
mP n

G pn
mP n
Gp

mP n

Linearization
(pss flow regime)

Computation of GCti on
x-intercept

(Almost) complete
(one straight line)

Extrapolates to GCti

Partial
(two straight lines are
observed)

1. Early straight line

extrapolates to GCti.
2. Late straight line
overestimates GCti

Applications
Example 1: Constant pressure/variable rate (simulated) case
The gas well data in table 2 were taken from Fraim and Wattenbarger (1987, case 1). The reader may resort to Fraim and
Wattenbarger (1987) for the PVT data. Cumulative production data in table 2 were computed using trapezoidal rule.
Our first goal is to solve for G in order to determine the average reservoir pressure profile. As shown in Fig. 1, an approximate plot (see Eq. 24) is made. It can be seen that two lines are noted during the pss flow regime the early time and late time
pss lines. The early time pss line is then extrapolated to the x-axis to yield:

GCti 850 Mscf / psia

from which the initial-gas-in-place can be estimated as:

850
61.95Bscf
0.00001372

which compares well with the simulated data of 61.51 Bscf (see Fraim and Wattenbarger, 1987, Case 1).

SPE 167504

Clearly, the proposed approach provides a direct and quick method to compute G and hence, average reservoir pressure profile. Note that Fraim and Wattenbarger (1987) converged on G after three iterations. Our approach, however, yields G at one
time. The computed G was then used to determine the average reservoir pressure profile, which in turn was used to evaluate
the viscosity-compressibility values. Note that if the late time pss line had been used, the G computed would have been optimistic.
Next, we plotted gas viscosity-compressibility ratio against actual cumulative production as shown in Fig. 2, and a polynomial curve was then fitted to the data points to yield polynomial equation (see Fig. 2), which was integrated with respect to
Gp to yield an equation similar in form to Eq. 25 as:
Gp

G pn

( g Ct ) i

g Ct

dG p GP

3 *10 G
8

2
p

7 *10 G
16

3
p

3 *10 G
23

4
p

6 *10 G
31

5
p

...

. (26)
Pseudocumulative can then be computed by substituting each cumulative production data point into Eq. 26. Table 3 shows
the computed plotting variables.
Fig. 3 shows the rigorous material balance plot. As opposed to the approximate plot, the rigorous plot linearizes the entire pss
data. However, both the rigorous plot (i.e., fig. 3) and the extrapolated early time pss line of the approximate plot (i.e., fig. 1)
intercept the x-axis at the same point. Thus, both plots yield the same G. Therefore, the extrapolated early time pss line of the
approximate plot, if present, will give us the correct G even if we dont utilize the rigorous plot. It is, however, advised to use
the rigorous plot. Fig. 4 is a composite graph of the rigorous and approximate plots. It is observed that during the transient
state, both plots trace the same path implying that the pseudocumulative function has no influence on the transient data.
Fraim and Wattenbarger (1987) and Callard and Schenewerk (1995) made a similar observation with normalized time and
pseudocumulative respectively using type curves. During the pss flow regime at early time, both plots continue to trace the
same path. However, the approximate plot deviates at late time while the rigorous plot continues to track the extrapolated
early time pss line and intercepts the same point on the x-axis as the extrapolated early time pss line. Therefore, instead of a
guess and time-consuming iterative scheme, we can use the early time pss line of the approximate plot to our advantage: it
gives us G at one time. The pss constant is 1.16 Mscf/D/psia.

Table 2-Simulated Gas well production data for example 1. (Fraim and Wattenbarger, 1987, Case 1)
qg,
(Mscf/D)

t, days

Gp, Mscf

qg,
(Mscf/D)

t, days

Gp, Mscf

30400

0.052

1580.8

11300

520

7944571.6

27300

0.18

5273.6

9650

668

9494871.6

24600

0.729

19520.15

7990

880

11364712

22200

3.38

81553.55

6330

1150

13297912

20,100

17.7

384421.55

4750

1490

15181512

18100

83.2

1635471.6

3250

1950

17021512

16400

174

3201771.6

1800

2690

18890012

14700

276

4787871.6

921

3550

20060042

13000

389

6352921.6

472

4410

20659032

SPE 167504

Fig. 1-An approximate FMB plot for example 1

Fig 2: A linear plot of viscosity-compressibility ratio against raw cumulative production for example 1

Fig. 3: A rigorous FMB plot for example 1

SPE 167504

Fig. 4-A composite graph of rigorous and approximate FMB plots for example 1

Table 3: Computed Plotting variables for example 1

Gp, Mscf

Gpn, Mscf

Gp/m(P)n,
Mscf/psia

Gpn/m(P)n,
Mscf/psia

qg/m(p)n,
Mscf/D/psia

1580.8

1580.7625

0.0894215

0.0894194

1.719645

5273.6

5273.1829

0.2983131

0.2982895

1.544286

19520.15

19514.436

1.1042013

1.1038782

1.391555

81553.55

81453.912

4.6132603

4.6076248

1.255793

384421.55

382217.94

21.745671

21.621023

1.137002

1635471.6

1596318.5

92.514135

90.299369

1.023868

3201771.6

3054912.2

181.11542

172.80802

0.927703

4787871.6

4465986.2

270.83674

252.62862

0.831539

6352921.6

5796379.4

359.36732

327.88532

0.735374

7944571.6

7088749.3

449.40259

400.99115

0.63921

9494871.6

8290617.3

537.0988

468.97753

0.545874

11364712

9667493.5

642.87053

546.86365

0.451972

13297912

11009456

752.22635

622.77482

0.358071

15181512

12239144

858.77642

692.33489

0.268695

17021512

13368137

962.86017

756.19893

0.183844

18890012

14443983

1068.5561

817.05658

0.101821

20060042

15082815

1134.7415

853.19357

0.052098

20659032

15399858

1168.6247

871.12783

0.0267

Example 2: Field case (West Virginia Gas Well data)-Low Permeability fractured gas well
This field data (see tables 4 and 5; we computed cumulative production using the trapezoidal rule) were first analyzed by
Fetkovich et al (1987), who used decline curves with ordinary time. Fraim and Wattenbarger (1987) later analyzed this data
with normalized time. Blasingame and Lee (1988) also analyzed this data with material balance pseudotime on linear coordinates. Ansah et al (1996) used a semi-analytic approach to analyze this same data.
We intend to validate our approach by comparing our results with that of the above authors. As was shown in the previous
example, an approximate plot was used to identify the early and late time pss line. This is shown in fig 5. The early time pss
line is extrapolated to the x-axis.
Results from x-intercept using fig. 5 (extrapolated early time pss line):

10

SPE 167504

GCti (x-intercept) = 520 Mscf/psia

or

520
2.7807 Bscf
0.000187

Table 7 compares our result with that of other investigators. Our result compares well with those of Blasingame and Lee
(1998) and Ansah et al (1996). Note that Blasingame and Lee (1988) used an iterative scheme to converge on G while Ansah
et al (1996) used a semi-analytic gas flow model. Clearly, the proposed approach provides a direct and quick method to compute G and hence, average reservoir pressure profile. We will later validate the result obtained by the various investigators
using Eq. A-2.
The average reservoir pressure, and hence the viscosity-compressibility values were then evaluated using the computed G.
The polynomial equation shown in Fig. 6 was used to compute pseudocumulative as was shown in the previous example. The
computed plotting variables are given in table 6.
Fig. 7 shows the rigorous material balance plot. It is seen that the entire data is linearized during the pss flow regime. Again,
a comparison of the approximate plot (i.e., fig. 5) and the rigorous plot (i.e., fig. 7) shows that both the rigorous plot and the
extrapolated early time pss line of the approximate plot intercept the x-axis at the same point. Fig. 8 is a composite graph of
rigorous and approximate plots. Similar trend as shown in the previous example is observed: it can be seen that the rigorous
plot continues to track the extrapolated early time pss line of the approximate plot. Note that the same point on the x-intercept
for the early time pss line of the approximate is also observed for the rigorous plot. Therefore, the rigorous plot yields the
same G. From fig. 7, the pss constant is 780 scf/D/psia.
We now use the technique proposed in this paper (see appendix A, Eq. A-2) to determine the correct G from among the various investigators. Fig. 9 shows such a plot. It can be seen from fig. 9 that our plot passes through the anchor point (i.e., x =
0.159). The results obtained by Fetkovich et al (1987) and Fraim and Wattenbarger (1987) undershoot the anchor point implying that their results overestimate the initial-gas-in-place. Note that Fetkovich et al (1987) used ordinary time, which is not
rigorous for gas well data. Fraim and Wattenbarger (1987) used normalized time, which though, is rigorous for gas well data,
they used decline type curve analysis: late-time data on type curves tend to be somewhat compressed due to the logarithmic
nature of the plot, and may affect the analysis. Blasingame and Lee (1988) used material balance pseudotime on linear plot
and had good result: their plot is very close to the anchor point. It has however, long been recognized that plots involving
time as independent variable are not as smooth as cumulative production. This may have an influence on Blasingame and
Lees (1988) analysis. Note that Ansah et al (1996) result almost touched the anchor point - we wish to say that Ansah et al
(1996) approach gives very good result. Nevertheless, our result seems the best.
We wish to emphasize that our proposed approach, that of Fraim and Wattenbarger (1987), Blasingame and Lees (1988) and
Ansah et al (1996) are all rigorous for single-phase gas flow.

Fig. 5-An approximate FMB plot for example 2

SPE 167504

11

Fig. 6-A linear plot of viscosity-compressibility ratio against raw cumulative for example 2

Fig. 7: A rigorous FMB plot for example 2

12

SPE 167504

Fig. 8: -A composite graph of rigorous and approximate FMB plot for example 2

Fig. 9: Determination of the correct G from among various investigators (example 2).

Table 4: System properties for example 2 (Fetkovich et al and Fraim Wattenbarger.)

Gas gravity, g

0.57 (air=1.0)

Initial
pressure

Porosity,

0.06, fraction

Initial gas FVF, Bgi

0.70942 rb/Mscf

Water saturation, Sw

0.35, fraction

Initial gas viscosity

0.02167 cp

Reservoir net pay, h

70.0 ft

Initial gas
pressibility,

1.870*104 psi-1

Wellbore radius, rw

0.354, ft

Reservoir
ture, T

1600 F

tempera-

reservoir
4175 psia

com-

Bottomhole presure

710 psia

SPE 167504

13

Table 5 (Example 2): Gas production data from Fraim and Wattenbarger (1987).
t, days

qg,
Mscf/D

Gp, Mscf

t, days

qg,
Mscf/D

Gp, Mscf

t, days

qg,
Mscf/D

Gp, Mscf

428.66

1,057.50

657012.74

1,453.01

575.4

1385700.1

9.36

2,382.80

22303.008

466.85

1,040.50

697074.05

1,559.88

410.8

1438397.7

31.57

2,339.80

74747.481

502.28

1,132.80

735574.06

1,662.55

353.1

1477612.5

58.12

1,755.30

129109.93

554.18

915.6

788730.04

1,799.91

359.4

1526547

89.97

1,832.20

186240.87

635.06

927

863244.78

1,892.29

409.2

1562048.6

119.02

1,542.60

235259.84

689.14

712.1

907566.04

2,110.90

346

1644595.8

148.51

1,580.60

281311.43

728.47

834

937970.1

2,152.70

254

1657135.8

167.35

1,555.70

310855.37

759.19

692.9

961423.29

2,331.22

249.9

1702113.9

206.37

1,676.50

373915.59

789.24

775.7

983489

2,465.40

276.6

1737436.8

230.26

1,386.00

410497.16

865.26

682.1

1038900

2,498.23

220.9

1745603.2

260.95

1,364.20

452698.97

969.14

702.2

1110800.5

2,583.15

239.1

1765134.8

287.52

1,297.90

488064.97

1,049.65

679.3

1166412.8

2,692.33

196.4

1788908.8

315.07

1,207.30

522574.1

1,056.52

614

1170855.3

2,818.95

184.7

1813036.2

349.02

1,194.80

563349.75

1,144.50

580.8

1223414.5

2,930.14

210.3

1834996.2

375.68

1,229.60

595667

1,276.13

519.4

1295824.2

3,104.16

189.1

1869748.0

406.99

1,156.60

633022.96

1,367.30

469.7

1340912.3

Table 6 (Example 2): Computed Plotting variables

Gp,Mscf
0
22303.00
8
74747.48
1
129109.9
3
186240.8
7
235259.8
4
281311.4
3
310855.3
7

Gp,Mscf

Gp/m(P)n
,Mscf/psia

Gpn,Mscf

Gpn/m(P)n
,Mscf/psia

qg/m(P)n,
Mscf/psia

Gp,Mscf

Gp/m(P)n
,Mscf/psia

Gpn,Mscf

Gpn/m(P),
Mscf/psia

qg/m(P)n,
Mscf/psia

657012.7
4

259.00

572090.7
9

225.11607

0.4161232

1385700.
1

545.00

1019754

401.27024

0.2264182

0.9376249

697074.0
5

274.00

601610.9
3

236.73216

0.4094337

1438397.
7

566.00

1045222

411.29182

0.1616486

0.9207045

735574.0
6

289.00

629418.7
6

0.4457535

1477612.
5

581.00

1063615.
9

418.52979

0.1389438

0.6907055

788730.0
4

310.00

666914.0
8

427.30326

0.1414229

0.7209654

863244.7
8

340.00

717739.0
2

0.6070086

907566.0
4

357.00

747018.7
3

Gp/m(P)n
,Mscf/psia

Gpn,Mscf

Gpn/m(P)n
,Mscf/psia

qg/m(Pn),
Mscf/psia

8.78

22203.56
1

8.7370369

29.40

73631.51
4

50.80

125783.9
3

73.30

179328.2
8

92.60

224241.4
6

111.00

265574.0
6

28.973788
49.495612
70.565159
88.238367
104.50263

0.6219615

937970.1

0.6121635

961423.2
9

369.00

766698.3

378.00

781654.5
8

247.67446
262.42876
282.42823
293.94971
301.69356

0.360286

1526547

601.00

1085912.
1

0.3647718

1562048.
6

615.00

1101638

433.49138

0.161019

0.2802093

1644595.
8

647.00

1136772.
3

447.3166

0.13615

0.3281766

1657135.
8

652.00

1141937.
5

449.3491

0.0999483

0.2726541

1702113.
9

670.00

1160097.
7

456.49508

0.0983349

122.00

291652.6
3

373915.5
9

147.00

346176.1
5

136.21931

0.6596979

983489

387.00

795548.7
5

313.04613

0.3052357

1737436.
8

684.00

1173963.
2

461.95113

0.1088413

410497.1
6

162.00

377097.1
3

148.38663

0.545387

1038900

409.00

829687.3

326.47955

0.2684044

1745603.
2

687.00

1177119.
9

463.19327

0.0869235

452698.9
7

178.00

412125.9
5

162.17037

0.5368088

1110800.
5

437.00

872399.7
4

343.28677

0.2763137

1765134.
8

695.00

1184595.
7

466.13499

0.0940852

192.00

440952.8
8

0.5107199

1166412.
8

459.00

904226.3
2

0.2673026

1788908.
8

704.00

1193555.
8

469.66076

0.0772828

206.00

468619.4
9

0.4750691

1170855.
3

461.00

906723.7
2

0.2416072

1813036.
2

713.00

1202494.
2

473.17797

0.0726789

222.00

500724.9
1

0.4701503

1223414.
5

0.2285431

1834996.
2

722.00

1210495.
3

476.32642

0.0827524

234.00

525722.4
2

0.483844

1295824.
2

1869748

736.00

1222899.
3

481.20735

0.0744103

249.00

554126.3
3

0.4551188

1340912.
3

488064.9
7
522574.1
563349.7
5
595667
633022.9
6

114.76447

173.51368
184.40041
197.03381
206.87025
218.04711

307.57882

355.81044
356.79316

481.00

935770.3

510.00

974294.0
7

368.22291
383.3819

0.2043824

528.00

997422.3
1

392.4828

0.1848256

Table 7 (Example 2): Comparison of results.

Parameters

Fetkovich
et al
(1987)

Fraim and
Wattenbarger
(1987)

Blasingame
and Lee
(1988)

Ansah et
al
(1996)

This work

G, Bscf

3.36

3.3045

2.6281

2.849

2.7807

A, acres

112.6

101.6

88.03

95.54

92.6

1/ba,pss,
scf/D/psia

1086.8

764

756

780

14

SPE 167504

Example 3: Variable pressure/variable rate (simulated) case.

Table 8: Reservoir and production data for example 3 (simulated case): Callard (1994) and Rodgers et al
(1983, case 1).
G

4.855 Bcf

Permeability

0.3 md

Height

80 ft

Temperature

636 R

Gas gravity

0.7

Gas saturation

75%

Pi

2500 psia

Porosity

10%

ba,pss

201,359 psia /cp/Mcf/D

4,840,000 ft

Rwa

2.8346 ft

CA

30.8822; well in the centre of square

time, yrs

qg, Mcf/D

GP, MMcf

Pwf, psia

m(P),
2
psia /cp

2500

4.77E+08

1000

365

1604

2.11E+08

1000

730

1361

1.54E+08

800

1022

1352

1.52E+08

800

1314

1153

1.12E+08

600

1533

1216

1.24E+08

600

1752

1071

9.80E+07

400

1898

1197

1.20E+08

400

2044

1107

1.03E+08

Reservoir properties and production data in Table 8 were taken from Callard (1994) and Rodgers et al (1983). Fig. 10 shows
the rate profile.
Fig. 11 shows the approximate plot. It can be seen that a single line is observed: this indicates that both the transient and early time pss data were not recorded. In this case, extrapolation of the single line (i.e., the late time pss line) will yield an optimistic result (see result in column 4 of table 10). Consequently, the method (i.e., the ease of determining G) breaks down
when the early time pss line is absent. Any known method can be used to solve for G in order to evaluate the average reservoir pressure. The average reservoir pressure profile was, however, given (see Rodgers et al 1983, case 1), and hence, the
values of the gas properties were determined from correlations. The viscosity-compressibility ratio was then plotted against
raw cumulative production shown in Fig. 12, and, the pseudocumulative profile was then computed as shown in column 2 of
table 9. The plotting variables are also shown in columns 4 and 5 in table 8.
Fig. 13 shows the rigorous material balance plot. Clearly, the entire data points were recorded during the late pss flow regime. Fig. 14 is a composite graph of the rigorous and approximate plots. Note that the plots diverge immediately after the
first data point. This confirms that the early time pss line is indeed absent. Also note, as expected, that both lines intercept the
x-axis at different points, and, hence would yield different G. See results in table 10.
The initial-gas-in-place-compressibility product can be determined from the x-intercept on fig. 13 as:

G pn
GCti
1.56MMcf / psia

m p n x int ercept
Now, assuming Cti C gi , the initial-gas-in-place can be estimated as:

SPE 167504

15

G pn

m p n x int ercept
1

*1.56 4.875Bcf
0.00032

1
Cgi

This value compares well with the simulated value of 4.855 Bcf
The pseudosteady state constant can be determined from the y-intercept as:

ba , pss

1
psia / Mcf / D
1.56

Or

ba , pss

1
1

psia / Mcf / D

psia / Mcf / D
1.56
1.56
2 * Pi *
2 * 2500 *
198,209.6 psia 2 / cp / Mcf / D
gi * zi
0.019024 * 0.85

which compares well with the simulated value of 201,359 psia2/cp/Mcf/D.

The drainage area may be estimated as:

5.615 * Bgi * G

h1 S wi

5.615 * 0.001089 * 4,875,000,000

4,968,222 ft 2
0.1 * 80 * 1 0.25

which compares well with the simulated value of 4,840,000 ft2.

Further, to validate Eq. A-2, we used three values of G: simulated or correct G (i.e., 4.855 Bscf), 30% short of simulated G
and 30% more of simulated G. This is shown in fig. 15. It can be seen that the correct or simulated G passes through the anchor point (i.e., 0.159) on the x-intercept. The lower and higher values (or incorrect Gs) overshoot and undershoot the anchor point respectively.

Fig. 10: Rate profile for example 3.

16

SPE 167504

Fig. 11: A linear graph of approximate FMB plot for example 3

Fig. 12: A linear plot of viscosity-compressibility ratio against cumulative production for example 3.

Fig. 13: A linear graph of rigorous FMB plot for example 3

SPE 167504

17

Fig. 14: A composite graph of rigorous and approximate FMB plot for example 3

Fig. 15-A linear plot of rate normalized pseudopressure against Qpn function to validate G for example 3.

Table 9: Computed plotting variables for example 3

Gp, MMcf

Gpn,
MMcf

Gp/m(P)n,
MMcf/psia

Gpn/m(P)n,
MMcf/psia

qg/m(P)n,
Mcf/D/psia

365

336.16

0.424448

0.390908

1.16

730

629.29

0.699046

0.602604

0.96

1022

843.08

0.972935

0.802604

0.76

1314

1042.15

1.112483

0.882610

0.67

1533

1181.48

1.343193

1.035197

0.52

1752

1310.41

1.429072

1.068876

0.49

1898

1389.15

1.645299

1.204194

0.35

2044

1460.75

1.692151

1.209296

0.33

18

SPE 167504

Table 10: Comparison of results for example 3

Parameter

True value
(simulated)

Pseudo,
Gpn/m(P)n

Raw,
Gp/m(P)n

G, Bcf

4.855

4.875

6.750

4,840,000

4,968,222

6,879,077

201,359

198,209.6

217,751.4

A, ft

ba, pss

Summary
This work has presented an analytical derivation of gas flow equation that constitutes a sound theoretical background for both
the new pseudotime function (i.e., rate normalized pseudocumulative) and the traditional material balance pseudotime. This
will give confidence to analysts to model gas production data. Further, the new pseudotime function introduced in this work
has been shown to be equivalent to the traditional material balance pseudotime. The benefit is that both pseudotime functions
have identical applications and one can be substituted for the other. However, unlike the traditional material balance pseudotime, the computation of this new pseudotime function is purely analytical and insensitive to time-step size.
For a complete data (i.e., both transient and pss data), early time pss line is likely to be present since early time pss data follow immediately the transient or transition state. The implication is that early time pss line will likely be present for a complete data. Therefore the approach presented in this paper to solve for G will almost always be feasible. However, for a highly
depleted reservoir having only pss data (such as example 3 in this paper), the early time pss line may not be present.
Conclusion
1. A new flowing material balance equation, which incorporates pseudocumulative instead of material balance pseudotime,
has been presented. This new version provides a convenient approach to handle variations in viscosity-compressibility values
with pressure since viscosity-compressibility ratio is a function of cumulative production.
2. The utilization of cumulative production and the independence of pseudocumulative on time-step size help yield accurate
results.
3. The approximate flowing material balance plot presented in this paper provides a useful and direct approach to compute
initial-gas-in-place at one time when early time pss line is observed. The late time pss line yields an optimistic result.
4. A universal technique has been developed to validate computed initial-gas-in-place. This will provide a level of certainty
as regard the accuracy of the computed gas-in-place, particularly for field data.
5. We have successfully applied our approach to both simulated and field data, and, good results were obtained.
6. The flow equations utilized in this paper are applicable to general pressure and rate scenario, all reservoir geometries and,
are strictly valid for boundary-dominated flow.
Acknowledgement
The authors would like to thank the University of Mines and Technology, Tarkwa, Ghana, for the financial support and the
permission to use its facilities. The authors would like to thank Dr. P.A. Eshun, the Head of Petroleum Engineering Department at University of Mines and Technology, Ghana, for the encouragement to prepare and publish this paper.
Nomenclature
A
= Drainage area, ft2
ba,pss = pseudosteady state constant, psia/scf/D

ba , pss

141.2 gi Bgi 1 4
A

ln
2
Kgh
2 e C A r wa

Bgi
Bg
Cg
Cgi
CA
Cgi

=
=
=
=
=
=

Gas formation volume factor at initial pressure, RB/Mscf

Gas formation volume factor, RB/Mscf
Gas compressibility, psi-1
Gas compressibility at initial pressure, psi-1
Shape factor, dimensionless
Gas compressibility at initial pressure, psi-1

Ct

= Total compressibility at average reservoir pressure, psi-1

SPE 167504

19

Cti
G
Gp
Gpn
h
Kg
m(P)n

=
=
=
=
=
=
=

Total compressibility at initial pressure, psi-1

Initial-gas-in-place, scf
(Actual) Cumulative gas production, scf
Pseudocumulative, scf
Reservoir thickness, ft
Gas permeability, md
Normalized pseudopressure, psia

=
=
=
=
=
=
=
=
=
=
=
=
=
=
=

Average reservoir pressure, psia

Initial pressure, psia
Pressure at standard condition, psia
Bottom-hole flowing pressure, psia
Gas flow rate, scf/D
Initial gas flow rate, scf/D
Drainage radius
Effective wellbore radius, ft
Initial gas saturation, dimensionless
Initial water saturation, dimensionless
Production time, days
Reservoir temperature, 0R
Rate normalized pseudocumulative, days
Temperature at standard condition, 0R
Pore volume, Mcf

=
=
=
=

Gas compressibility factor evaluated at the average reservoir pressure

Gas compressibility factor at initial pressure
Porosity, fraction
Gas viscosity at initial pressure, cp

= Gas viscosity at average reservoir pressure, cp

Pi
Psc
Pwf
qg
qgi
re
rwa
Sgi
Swi
t
T
Ta
Tsc
Vp
zi

gi

= Eulers constant, 0.5772

Reference
Agarwal, R. G. David C. G. Stanley W. Kleinsteiber et al. 1999. Analyzing Well Production Data Using Combined-Type-Curves and Decline-Curve Analysis Concepts. Paper SPE 57916 presented at the Annual Technical Conference and Exhibition, New Orleans. 27-30 September.
Agarwal, R. G. 2010. Direct Method of Estimating Average Reservoir Pressure for Flowing Oil and Gas Wells. Paper SPE 135804 presented at the Annual Technical Conference and Exhibition, Florence.19-22 September.
Agarwal, R. G. 1979. Real Gas Pseudo Time A New Pressure Buildup Analysis of MHF Gas Wells. Paper SPE 8729 presented at the
SPE Annual Technical Conference and Exhibition, Las Vegas. 23-26 September.
Al-Hussainy R., Ramey, H.J. Jr. and Crawford P.B. 1966. The Flow of Real Gases through Porous Media. JPT. Trans, AIME, 237. 624636. May.
Al-Hussainy R. and Ramey, H.J. Jr. 1966 Application of Real Gas Flow to Well Testing and Deliverability Forecasting. JPT. Trans, AIME,
237. 637-642. May.
Anderson, D.M., Stotts, G.W.J., Mattar, L., Ilk, D., and Blasingame, T.A. 2006. Production Data Analysis Challenges, Pitfalls, Diagnostics. Paper SPE 102048 presented at the Annual Technical Conference and Exhibition held in San Antonio, Texas, U.S.A, 24-27 September.
Ansah J., Knowles, R.S. and Blasingame, T.A. 1996. A Semi-analytical (P/Z) Rate-Time Relation for the Analysis and Prediction of Gas
WellPerforamance. Paper SPE 35268 presented at the SPE Mid-Continent Gas Symposium, Amarillo, TX, 28-30 April.
Arps, J.J. 1945. Analysis of Decline Curves. In Transactions of the Society of Petroleum Engineers, AIME 160, 228-247.
Arps, J.J. 1956. Estimation of Primary Oil Reservoirs. In Transactions of the Society of Petroleum Engineers, AIME 207, 182-191.
Blasingame, T.A., and Lee W.J. 1986. The Variable-Rate Reservoir Limits Testing. Paper SPE 15028 presented at the SPE Permium Basin
Oil & Gas Recovery Conference, Midland. 13-14 March.
Blasingame, T.A., and Lee W.J. 1988. The Variable-Rate Reservoir Limits Testing of Gas Wells. Paper SPE 17708 presented at the SPE

20

SPE 167504

Gas Technology Symposium, Dallas, TX. 13-15 June.

Callard, J.G. 1994. Reservoir Performance History Matching Using Type-Curves. PhD Dissertation, Louisiana State University, Baton
Rouge, LA (December 1994).
Callard, J.G. and Schenewerk, P.A. 1995. Reservoir Performance History Matching Using Rate/Cumulative Type-Curves. Paper SPE
30793 presented at the Annual Technical Conference and Exhibition held at Dallas, U.S.A., 22-25 October.
Carter, R.D. 1985. Type Curves for Finite Radial and Linear Gas Flow Syetems: Constant Terminal Pressure Case. SPEJ. October, 719728.
Fetkovich, M.J. 1980. Decline Curve Analysis Using Type curves. SPE 4629, June.
Fetkovich, M.J. Vienot M.E. Bradley M.D. et al. 1987. Decline Curve Analysis Using Type curves Case Histories. SPE 13169, December.
Fraim, M.L. and Wattenbarger, R.A. 1987. Gas Reservoir Decline Curve Analysis Using Type Curves With Real Gas Pseudopressure and
Normalized Time. SPE, December.
Johnson, N.L., Currie, S.M., Ilk D. and Blasingame, T.A. 2009. A Simple Methodology for Direct Estimation of Gas-in-Place and Reserves
Using Rate-Time Data. paper SPE 123298 presented at the Rocky Mountain Petroleum Technology Conference, Denver Colorado U.S.A.
April 14-16.
Keating, J.F., Chen, H.Y. and Wattenbarger R.A. 1994. Original Gas in Place and Decline Curves from Early Stabilized Rate-time Data.
paper SPE 27666 presented at the Permian Basin Oil and Gas Recovery Conference, Midland, Texas. 16-18 March.
Lee J., and Wattenbarger, R.A. 1996. Gas Reservoir Engineering. Vol. 5, 214-226, Texas: Textbook Series, SPE.
Mattar, L. and Anderson D.M. 2003. A Systematic and Comprehensive Methodology for Advanced Analysis of Production Data. Paper
SPE 84472 presented at the SPE Annual Technical Conference and Exhibition held in Denver Colorado, U.S.A. 5-8 October.
Mattar, L. and Anderson, D. 2005. Dynamic Material Balance (Oil or Gas-in-Place without shut-ins). Paper SPE presented at the 6th Canadian International Petroleum Conference, Calgary, Alberta, Canada, June, 7-9.
Mattar, L. and McNeil, R. 1998. The Flowing Gas Material Balance. Journal of JCPT. Vol. 37 #2.
Menuier, D.F., Kabir, C.S. and Wittman, M.J. 1987. Gas Well Test Analysis: Use of Normalized Pseudovariables SPEFE. 629-636. December.
Palacio, J.C. and Blasingame T.A. 1993. Decline Curve Analysis Using Type curves-Analysis of Gas Well Production Data. SPE paper
25909 presented at the 1993 Joint Rocky Mountain Regional and Low Permeability Reservoirs Symposium, Denver, Colorado, 26-28
April.
Poston S.W. and Poe B.D. Jr. 2008. Analysis of Production Decline Curves: Monograph SPE. pp 132.
Ramagost, B.P. and Farshad, F.F. 1981. P/Z Abnormally Pressured Gas Reservoirs. paper SPE 10125 presented at the Annual Technical
Conference and Exhibition, San Antonio. Oct. 5-7.
Rodgers, J.S. Boykin, R.S. and Coble L.E. 1983. Nonstatic Pressure History Analysis for Gas Reservoirs. SPEJ. April. 209-18.
Suabdi, I.N. 2001. Analysis of Layered Gas Reservoir Performance Using a Semi-Analytical Solution for Rate and Pressure Behaviour.
PhD Dissertation, Texas A&M University. (December 2001).
Syah, I. 1999. The Analysis and Interpretation of Well Interference Effects on Well Performance of Gas-Condensate Reservoirs: A Case
Study of Arun Field. PhD Dissertation, Vol. 1, Texas A&M University. (May 1999).

APPENDIX A
VERIFICATION OF INITIAL-GAS-IN-PLACE
Eq. 22 in the main text can be expressed as:

qg

mP n

2
ba , pss

G pn

1
... (A-1)

2 GCti mP n ba , pss

SPE 167504

21

or

qg

mP n

2
1
...... (A-2)
Q pn
ba , pss
ba , pss

where

G pn
....(A-3)
2GCti mP n
qg
against Q pn on linear coordinates will pass through the anchor point
Therefore, according to Eq. A-2, a plot of
mP n
1

0.159 on the x-intercept during the boundary-dominated flow as long as G (or GCti) is correct. Hence, ini i.e.,
2

Qpn

tial-gas-in-place computed by any method can be validated using Eq. A-2.

APPENDIX B
COMPUTATION OF PSEUDOCUMULATIVE
Ansah et al (1996) showed that the viscosity-compressibility ratio is related to the fractional recovery, Gp/G, as:

C t i

g Ct

Gp
a 0 a1 1
G

G
a 2 1 p
G

G
a 3 1 p
G

G
a 4 1 p
G

...

. (B-1)

where

a 0 , a1 , a 2 , a 3 , a 4 ...... are constants.

Eq. B-1 can be expressed as:

C
g

t i

g Ct

C0 C1G p C2G p C3G p C4G p ... .. (B-2)

where

C o ao a1 a 2 a3 a 4 ... 1 ... (B-3)

Before we proceed, we wish to explain why we have set the constant in Eq. B-3 to 1.
At the initial condition, i.e., t 0 , cumulative production is equal to zero G p 0 . From Eq. B-2, the y-intercept occurs

at G p

0 . This implies that the y-intercept occurs at the initial condition. Therefore, at the initial condition, which is actual-

ly G p

0 , Eq. A-2 reduces to:

g Ct i
C0 . (B-4)

t
g
t 0G p 0

Also note that at the initial condition, the average reservoir pressure is equal to the initial reservoir pressure. Therefore at the
initial condition, the viscosity-compressibility values are evaluated at the initial reservoir pressure. Hence, Eq. B-4 becomes:

C
C
g

t i

t i

1 C0 ....... (B-5)

Eq. B-5 is the basis of setting the constant in Eq. B-3 to 1. Microsoft Excel gives a value of the intercept very close to 1 when
the analyst does not even set the intercept to 1, but accurate result is obtained when the intercept is set to 1.
Now, we proceed with the definition of the other constants in Eq. B-2:

22

SPE 167504

a 2a2 3a3 4a4 ...

C1 1
..... (B-6)
G

a 3a3 6a4 ...

C2 2
............ (B-7)
G2
a 4a4 ...
C3 3
.... (B-8)
G3

a4
C4 4 ..... (B-9)
G
g Ct i
against G p is made on a linear graph, the constants can be determined
Now, according to Eq. B-2, when a plot of
g Ct
by fitting a polynomial curve to the data and setting the intercept to 1. It is recommended that the polynomial expression be at
least to the third degree.
Now, integrating Eq. B-2 with respect to G p yields:
Gp

G pn

( g Ct ) i

dG p

0 g Ct
C
C
C
C
2
3
4
5
G p 1 G p 2 G p 3 G p 4 G p ...
2
3
4
5

(B-10)

Therefore, to compute pseudocumulative, we substitute each gas cumulative production data point into the polynomial expression (i.e., Eq. B-10). Note that since Eq. B-1 is general (Ansah et al 1996), it follows that Eq. B-10 is general, and, can be
used to compute pseudocumulative regardless of the initial pressure.