Académique Documents
Professionnel Documents
Culture Documents
C H A P T E R 1 6
ANALYSIS OF DECLINE
AND TYPE CURVES
DECLINE-CURVE ANALYSIS
Decline curves are one of the most extensively used forms of data analy-
sis employed in evaluating gas reserves and predicting future production.
The decline-curve analysis technique is based on the assumption that past
production trends and their controlling factors will continue in the future
and, therefore, can be extrapolated and described by a mathematical
expression.
1235
Ch16.qxd 3/15/06 8:02 PM Page 1236
Exponential decline
Harmonic decline
Hyperbolic decline
Each type of decline curve has a different curvature, as shown in Fig-
ure 16-1. This figure depicts the characteristic shape of each type of
decline when the flow rate is plotted versus time or versus cumulative
production on Cartesian, semi-log, and log-log scales. The main char-
acteristics of these decline curves can be used to select the flow-rate-
decline model that is appropriate for describing the ratetime relationship
of the hydrocarbon system:
1000 I 1000 II
I
LOG RATE (q)
2000
RATE (q)
II III
I
III
II
1000
100 100
III
0
0 TIME (t) 120 0 TIME (t) 120 0 10 100 120
RATE - CUMULATIVE CURVES LOG TIME (t)
1000 1000 II
LOG RATE (q)
2000
RATE (q)
I III
II
1000 II
III
100 100
III
0
0 100000 0 100000 100 1000 10000 100000
CUMULATIVE CUMULATIVE
LOG CUMULATIVE
Figure 16-1. Classification of production decline curves. (After Arps, J.J. Estima-
tion of Primary Oil Reserves, Courtesy of Trans, AIME, vol. 207,1956 ).
Ch16.qxd 3/15/06 8:02 PM Page 1238
qi
qt = (16-1)
(1 + bD i t ) 1/ b
d (ln q ) 1 dq
D= = (16-2)
dt q dt
The minus sign has been added because dq and dt have opposite signs
and it is convenient to have D always positive. Notice that the decline-
rate equation, Equation 16-2, describes the instantaneous changes in the
slope of the curvature, dq/dt, with the change in the flow rate, q, over
time.
Ch16.qxd 3/15/06 8:02 PM Page 1239
The parameters determined from the classical fit of the historical data,
namely the decline rate, D, and the exponent, b, can be used to predict
future production. This type of decline-curve analysis can be applied to
individual wells or the entire reservoir. The accuracy of the entire-reser-
voir application is sometimes even better than for individual wells due to
smoothing of the rate data. Based on the type of rate-decline behavior of the
hydrocarbon system, the value of b ranges from 0 to 1, and, accordingly,
Arps equation can be conveniently expressed in the following three forms:
qi
qt =
Harmonic b=1 (1 + D i t ) (16-5)
Figure 16-2 illustrates the general shape of the three curves at different
possible values of b. These mathematical relations can be applied equally
for gas and oil reservoirs.
It should be pointed out that these three forms of decline-curve equa-
tions are applicable ONLY when the well/reservoir is under pseudo-
steady (semi-steady)-state flow conditions, that is, boundary-dominated
flow conditions. Arps equation has been often misused to model the per-
formance of oil and gas wells whose flow regimes are in a transient state.
As presented in Chapter 6, when a well is first open to flow, it is in a
transient (unsteady-state) condition. It remains in this condition until the
production from the well affects the total reservoir system by reaching its
drainage boundary, at which time the well is said to be flowing in a pseu-
do-steady-state or boundary-dominated flow condition. The following
is a list of inherent assumptions that must be satisfied before perfor-
mance of ratetime decline-curve analysis:
The well is draining a constant drainage area, that is, the well is in a
boundary-dominated flow condition
Ch16.qxd 3/15/06 8:02 PM Page 1240
qi
qt =
(1+ bDit)1/b
(1+ bDit)1/b
qi
Harmonic b=1 qt =
(1+ Dit)
Hyperbolic
b = 0.5
Harmonic
b=1
Time
G p = q t dt (16-6)
t1
Ch16.qxd 3/15/06 8:02 PM Page 1241
Replacing the flow rate, qt, in the above equation with the three individual
expressions that describe types of decline curves (Equations 16-3, 16-4,
and 16-5), and integrating gives the following:
G p( t ) =
(q i
qt )
Di
Exponential b = 0: (16-7)
(q i ) q t 1 b
G p( t ) = 1
Hyperbolic 0 < b < 1: D i (1 b) q i (16-8)
q q
G p ( t ) = i ln i
Di q t
Harmonic b = 1: (16-9)
All the expressions given by Equations 16-3 through 16-9 require con-
sistent units. Any convenient unit of time can be used, but, again, care
should be taken to make certain that the time unit of the gas flow rates, qi
and qt, matches the time unit of the decline rate, Di, for example, for flow
rate q in scf/month or STB/month with Di in month1.
Note that the traditional Arps decline-curve analysis, as given in Equa-
tions 16-7 through 16-9, gives a reasonable estimation of reserve but also
has its failings, the most important one being that it completely ignores
the flowing pressure data. As a result, it can underestimate or overesti-
mate the reserves. The practical applications of these three commonly
used decline curves for gas reservoirs are as follows:
Exponential Decline, b = 0
q t = q i exp( D i t )
Or linearly as
ln(q t ) = ln(q i ) D i t
Similarly,
qi qt
G p( t ) =
Di
Or linearly as
q t = q i Di G p( t )
This type of decline curve is perhaps the simplest to use and perhaps
the most conservative. It is widely used in the industry for the following
reasons:
Many wells follow a constant decline rate over a great portion of their
productive life and will deviate significantly from this trend toward the
end of this period
The mathematics involved, as described by the line expressions just
given, are easier to apply than those for the other line types
Assuming that the historical production from a well or field is recog-
nized by its exponential production-decline behavior, the following steps
summarize the procedure to predict the behavior of the well or the field
as a function of time.
Step 2. For both plots, draw the best straight line through the points.
Step 4. Calculate the initial decline rate, Di, by selecting a point on the
Cartesian straight line with a coordinate of (qt,Gpt) or on a semi-
log line with a coordinate of (qt,t) and solve for Di by applying
Equation 16-5 or Equation 16-7.
Ch16.qxd 3/15/06 8:02 PM Page 1243
ln (q i / q t )
Di = (16-10)
t
or equivalently as
qi qt
Di = (16-11)
G p( t )
t ln (q i
/ q t )
Di = t
t 2 (16-12)
t
or equivalently as
n (q t G p ( t ) ) q t G p ( t )
Di =
t t t
2 (16-13)
n (G p ( t ) )2 G p ( t )
t t
Step 5. Calculate the time it will take to reach the economic flow rate, qa
(or any rate) and corresponding cumulative gas production from
Equations 16-3 and 16-7.
ln (q i / q a )
ta =
Di
qi qa
G pa =
ta
Example 16-1
The following production data are available from a dry gas field:
Estimate
(a) The future cumulative gas production when the gas flow rate reaches
80 MMscf/day
(b) Extra time to reach 80 MMscf/day
Solution
Part a
Step 3. The intercept of the straight line with the y-axis gives a value of
qi = 344 MMscf/day.
qi qt 344 197
Di = = = 0.000418 day1
G p( t ) 352, 000
Ch16.qxd 3/15/06 8:02 PM Page 1245
400
350
300
Gas flow rate, scf
250
200
150
100
50
0
0 100 200 300 400 500 600 700 800
Cumulative gas producton, MMMscf
It should be pointed out that the monthly and yearly nominal decline. that
is, Dim and Diy, respectively, can be determined as
n (q t G p ( t ) ) [( q t )( G p ( t ) )]
Di = t t t
n (G p ( t ) ) ( G p ( t ) )2
2
t t
Part b
Step 1. Calculate the time to reach the last recorded flow rate, 184
MMscf, using Equation 16-10:
ln (q i / q a ) ln (344 / 184)
ta = = = 1560 days = 4.275 year
Di 0.000401
Step 2. Calculate the total time to reach a gas flow rate of 80 MMscf/day:
ln (344 / 80)
t= = 3637 days = 9.966 years
0.000401
Example 16-2
1-1-02 0 1,240
2-1-02 1 1,193
3-1-02 2 1,148
4-1-02 3 1,104
5-1-02 4 1,066
6-1-02 5 1,023
7-1-02 6 986
8-1-02 7 949
9-1-02 8 911
10-1-02 9 880
11-1-02 10 843
12-1-02 11 813
1-1-03 12 782
(a) Use the first six months of the production history data to determine
the coefficient of the decline-curve equation.
(b) Predict flow rates and cumulative gas production from August 1,
2002 through January 1, 2003.
(c) Assuming that the economic limit is 30 MMscf/month, estimate the
time to reach the economic limit and the corresponding cumulative
gas production.
Ch16.qxd 3/15/06 8:02 PM Page 1247
Solution
Part a
Step 2. Determine the initial decline rate, Di, by selecting a point on the
straight line and substituting the coordinates of the point in
Equation 16-10 to give
Di =
(
ln q i / q t ) = ln (1240 / 986) = 0.0382 month 1
t 6
Alternatively, using the least-squares method as expressed by Equation
16-12 gives
t ln (q i
/ q t )
Di = t
t 2
10000
Rate, MMsct/month
1000
100
0 1 2 3 4 5 6 7
Time, months
Part b
Use Equations 16-3 and 16-7 to calculate qt and Gp(t), and tabulate the
results as follows:
(q i q t ) (q i q t )
G p( t ) = =
Di 0.0383
Part c
Apply Equations 16-10 and 16-11 to calculate the time, ta, to reach an
economic flow rate, qa, of 30 MMscf/month, and the corresponding
reserves, Gpa:
ta =
(
ln q i / q a ) = ln (1240 / 30) = 97 months = 8 year
Di 0.0383
qi qa (1240 30)106
G pa = == = 31.6 MMMscf
ta 0.0383
Ch16.qxd 3/15/06 8:02 PM Page 1249
Harmonic Decline, b = 1
qi
qt =
1+ Di t
q q
G p ( t ) = i ln i
Di q t
1 1 Di
= + t
q t q i q i (16 -14)
D
ln (q t ) = ln (q i ) i G p ( t ) (16 -15)
qi
The basic two plots for harmonic decline-curve analysis are based on
these two relationships. Equation 16-14 indicates that a plot of 1/qt ver-
sus t on a Cartesian scale will yield a straight line with a slope of (Di/qi)
and an intercept of 1/qi. Equation 16-15 suggests a plot of qt versus Gp(t)
on a semi-log scale and will yield a straight line with a negative slope of
(Di/qi) and an intercept of qi. The method of least squares can also be
used to calculate the decline rate, Di, to give
t qi
q t
t
t t
Di =
t 2
Other relationships that can be derived from Equations 16-14 and 16-15
include the time to reach the economic flow rate, qa (or any flow rate),
and the corresponding cumulative gas production, Gp(a):
Ch16.qxd 3/15/06 8:02 PM Page 1250
qi qa
ta = (16 -16)
q a Di
q q
G p ( a ) = i ln a
Di q t
qi
qt =
(1 + b D i t ) 1/ b
qi q t 1 b
G p( t ) = 1
D i (1 b) q i
Step 2. Extend the curve to intercept the y-axis at t = 0 and read qi.
Step 3. Select the other end point of the smooth curve, record the coordi-
nates of the point, and refer to it as (t2, q2).
q1 = q i q 2
(16 -17)
The corresponding value of t1 is read from the smooth curve at q1.
Ch16.qxd 3/15/06 8:02 PM Page 1251
q b q b
f ( b ) = t 2 i t1 i ( t 2 t1 ) = 0 (16 -18)
q1 q2
Starting with an initial value of b = 0.5, that is, bk = 0.5, the method
will usually converge after 45 iterations when the convergence criterion
is set at [bk+1 bk] 106.
Step 6. Solve for Di with Equation 16-4, by using the calculated value of
b from Step 5 and the coordinate of a point on the smooth graph,
for example, (t2, q2), to give
(q i / q 2 ) b 1
Di = (16 -21)
b t2
Example 16-3
The following production data were reported by Ikoku (1984) for a gas
well:
Solution
Step 1. Determine the type of decline that adequately represents the his-
torical data. This can be done by constructing the following two
plots:
Plot qt versus t on a semi-log scale, as shown in Figure 16-5.
The plot does not yield a straight line, and, thus, the decline is
not exponential.
Plot qt versus Gp(t) on a semi-log scale, as shown in Figure 16-6.
The plot again does not produce a straight line, and, therefore,
the decline is not harmonic.
The two generated plots indicate that the decline must be hyperbolic.
Step 2. From Figure 16-5, determine the initial flow rate, qi, by extend-
ing the smooth curve to intercept with the y-axis, at t = 0, to give
qi = 10 MMscf/day
Step 3. Select the coordinate of the other end point on the smooth curve
as (t2, q2), to give
1
10
2
q, MMscfd
1
0 1 2 3 4 5
t (years)
Figure 16-5. Rate-time plot for Example 16-3.
10
9
8
7
6
5
4
q, MMscfd
1
0 2 4 6 8 10
GPD (MMMscf)
Figure 16-6. Rate-cumulative plot for Example 16-3.
Ch16.qxd 3/15/06 8:02 PM Page 1254
With
f ( bk )
bk+1 = bk
f `( bk )
Step 7. Use Equations 16-4 and 16-8 to predict the future production
performance of the gas well. Notice that in Equation 16-4 the
denominator contains Dit and, therefore, the product must be
dimensionless, or
10 (106 ) (10)(106 )
qt = =
[ 1 + 0.5195 D i t ] (1/0.5195) [ 1 + 0.5195 (0.3668)( t )] (1/0.5195)
qi q 1 b
G p( t ) = 1 t
D i (1 b) qi
(10)(106 ) qt
1 0.5195
G p( t ) = 1
(0.001)(1 0.5195) (10)(106 )
20 20
19
18 18
17
16 16
15
14 14
12 12
11
10 10
9
8 8
7
6 6
5
4 4
3
2 2
1
0 0
0 5 10 15 20 25
Time, years
Figure 16-7. Decline curve data for Example 3-18.
0 10 10 0 0
0.5 8.4 8.392971 1.67 1.671857
1 7.12 7.147962 3.08 3.08535
1.5 6.16 6.163401 4.3 4.296641
2 5.36 5.37108 5.35 5.346644
2.5 4.72 4.723797 6.27 6.265881
3 4.18 4.188031 7.08 7.077596
3.5 3.72 3.739441 7.78 7.799804
4 3.36 3.36 8.44 8.44669
5 2.757413 9.557617
6 2.304959 10.477755
7 1.956406 11.252814
8 1.68208 11.914924
9 1.462215 12.487334
10 1.283229 12.987298
11 1.135536 13.427888
12 1.012209 13.819197
13 0.908144 14.169139
14 0.819508 14.484015
15 0.743381 14.768899
16 0.677503 15.027928
17 0.620105 15.264506
18 0.569783 15.481464
19 0.525414 15.681171
20 0.486091 15.86563
Ch16.qxd 3/15/06 8:02 PM Page 1257
100
90
80
0.2 0.1
0.4 0.3
5
0
70
0.6 0.
1.0 0. 7
0.8 0.
9
60
50
40
Values of "b"
30
qi /q at any time on decline
20
10
9
8
7
6
5
4
1
1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
Gp/tqi
100
90
80
0
0.1
70
0.2
3
60
0.
4
0.
50
5
0.
40 0.6
7
30 0.
0.8
qi/q at any time on decline
20 0.9
1.0
Values of "b"
10
9
8
7
6
5
4
1
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
D it
Figure 16-9. Relationship between production rate and time (After Gentry, 1972 ).
Example 16-4
Solution
Step 2. Enter Figure 16-9 with the values of 2.98 and 0.5 to give
Di t = 1.5
qi
Production rate
qr
Gpr
Gpr
A
A
B B
qa qa
Time Time
Figure 16-10. Estimation of the effect of restricting maximum production rate.
In many cases gas wells are not produced at their full capacity during
their early life for various reasons, such as limited capacity of flow lines,
transportation, low demands, or other types of restrictions. Figure 16-10
illustrates a model for estimating the time pattern of production where
the rate is restricted.
Figure 16-10 shows that the well produces at a restricted flow rate of
qr for a total time of tr with a cumulative production of Gpr. The pro-
posed methodology of estimating the restricted time, tr, is to set the total
cumulative production, Gp(tr), that would have occurred under normal
decline from the initial well capacity, qi, down to qr equal to Gpr. Even-
tually, the well will reach the time tr where it begins to decline with a
behavior similar to that of other wells in the area. The proposed method
for predicting the decline-rate behavior for a well under restricted flow
is based on the assumption that the following data are available and
applicable to the well:
Coefficients of Arps equation, that is, Di and b, by analogy with other
wells
Abandonment (economic) gas flow rate, qa
Ultimate recoverable reserves, Gpa
Allowable (restricted) flow rate, qr
Ch16.qxd 3/15/06 8:02 PM Page 1260
Step 1. Calculate the initial well flow capacity, qi, that would have
occurred with no restrictions, as follows:
D i G pa q
For Harmonic: q i = q r 1 + ln r (16-23)
qr q a
1/ b
D bG b (q r ) b q 1 b
For Hyperbolic: q i = (q r ) b + i 1 bpa 1 a
(q r ) 1 b qr
(16-24)
q q
For Harmonic: G pr = D ln q
i i
(16-26)
i r
qi q 1 b
For Hyperbolic: G pr = 1 r (16-27)
D i (1 b) qi
Step 3. Regardless of the type of decline, calculate the total time of the
restricted flow rate from
G pr
tr =
qr
(16-28)
Step 4. Generate the well-production performance as a function of time
by applying the appropriate decline relationships, as given by
Equations 16-3 through 16-14.
Ch16.qxd 3/15/06 8:02 PM Page 1261
Example 16-4
Solution
Step 1. Estimate the initial flow rate, qi, from Equation 16-22:
Step 3. Calculate the total time of the restricted flow from Equation 16-28:
G pr 16, 023
tr = = = 37.7 months = 3.14 years
qr 425
(q i q t ) 425 270
G p( t ) = = = 3523 MMscf
Di 0.044
Total production for the fourth year = 714 + 3,523 = 4237 MMscf
1 5100
2 5100
3 5100
4 4237
The flow rate at the end of the fourth year, 270 MMscf/month, is set
equal to the initial flow rate at the beginning of the fifth year. The flow
rate at the end of the fifth year, qend, is calculated from Equation 16-25 as
q end = q i exp[ D i (12)] = 270 exp[ 0.044 (12)] = 159 MMscf / month
159 94
Gp = = 1482 MMscf
0.044
Ch16.qxd 3/15/06 8:02 PM Page 1263
Reinitialization of data
Fetkovich (1971) points out that there are several obvious situations
where ratetime data must be reinitialized for reasons that include among
others,
The drive or production mechanism has changed
An abrupt change in the number of wells on a lease or a field due to
infill drilling
Changing the size of tubing would change qi and also the decline expo-
nent, b.
Provision of a well is not limited by tubing or equipment; the effects of
stimulation will result in a change in deliverability, qi, and possibly the
remaining recoverable gas. However, the decline exponent, b, normally
can be assumed constant. Fetkovich et al. (1996) suggested a rule-of-
thumb equation to approximate an increase in rate due to stimulation:
7 + s old
(q i ) new = (q t )old
7 + s new
where (qt)old = producing rate just prior to stimulation
s = skin factor
(q i ) new
qt =
(1 + b t ( D i ) new )1/ b
Ch16.qxd 3/15/06 8:02 PM Page 1264
with
(q i ) new
( D i ) new =
(1 b) G
where G = gas in place, scf
TYPE-CURVE ANALYSIS
recall Darcys equation (Chapter 6), which describes the radial, incom-
pressible, steady-state flow as expressed by
kh
Q= p (16-29)
141.2 B [ln( re / rwa ) 0.5 ]
where rwa is the apparent (effective) wellbore radius, as defined by Equa-
tion 6-152 in terms of the skin factors by
rwa = rw e s
Group A can be then defined by rearranging Darcys equation as:
r 1 kh
ln e = p
rwa 2 141.2 Q B
kh
pD = p (16-30)
141.2 Q B
Taking the logarithm of both sides of the above equation gives
kh
log ( p D ) = log ( p) + log
141.2 Q B (16-31)
kh
log
141.2 Q B
Ch16.qxd 3/15/06 8:02 PM Page 1266
t
0.1 1 10 102 103
1 102
kh
141. 2QBm
1 10 p
PD 10 0.0002637k /fmct r 2 w
102 1
10 102 103 104 105
tD
Figure 16-11. Concept of type curves.
Ch16.qxd 3/15/06 8:02 PM Page 1267
Not only do these two curves have the same shape, but if they are
moved relative to each other until they coincide or match, the vertical
and horizontal displacements required to achieve the match are related to
these constants in Equations 16-31 and 16-32. Once these constants are
determined from the vertical and horizontal displacements, it is possible
to estimate reservoir properties such as permeability and porosity. This
process of matching two curves through the vertical and horizontal dis-
placements and determining the reservoir or well properties is called
type-curve matching.
Consider the Ei-function solution to the diffusivity equations, as given
in Chapter 6 by Equation 6-78:
70.6 Q B 948 c t r 2
p ( r , t ) = pi + Ei
kh kt
1 rD
2
p D = Ei (16-33)
2 4 tD
It should be noted that when tD /rD2 >25 , Equation 16-33 can be approxi-
mated by
1 t
p D = ln D2 + 0.080907
2 rD
Notice that
tD 0.0002637 k
= t
c t r
2 2
rD
Ch16.qxd 3/15/06 8:02 PM Page 1268
t 0.0002637 k
log D2 = log + log ( t ) (16-34)
c t r
2
rD
Equations 16-31 and 16-34 indicate that a graph of log(p) versus log(t)
will have an identical shape (i.e., will be parallel) to a graph of log(pD)
versus log(tD/rD2 ), although the curve will be shifted by log(kh141.2/QB)
vertically in pressure and log(0.0002637k/ ct r2) horizontally in time.
When these two curves are moved relative to each other until they coin-
cide or match, the vertical and horizontal movements, in mathematical
terms, are given by
pD kh
p = 141.2 Q B (16-35)
MP
and
t D / rD2 0.0002637 k
= (16-36)
t MP c t r 2
Step 2. Place a tracing paper over Figure 16-12 and construct a log-log
scale that has the same dimensions as those of the type curve.
This can be achieved by tracing the major and minor grid lines
from the type curve to the tracing paper.
Ch16.qxd 3/15/06 8:02 PM Page 1269
1 1 1
9
8 9
7 8
6 7
5 6
PD
5
4
4
3
3
2
2
101 1
9 1
8 9
7 8
6 7
5 6
5
4 4
3 3
2 2
102 1 1
2 3 4 5 6 7 8 91 2 3 4 5 6 7 89 1 2 3 4 5 6 7 89 1 2 3 4 5 6 7 89 1 2 3 4 5 6 7 89 1
102 10 102 103 104
tD /r2D
Figure 16-12. Dimensionless pressure for a single well in an infinite system, no well-
bore storage, no skin. Exponential-integral solution (After Earlougher, R. Advances in
Well Test Analysis) (Permission to publish by the SPE, coyright SPE, 1977 ).
Step 3. Plot the well-test data in terms of p versus t on the tracing paper.
Step 4. Overlay the tracing paper on the type curve and slide the actual
data plot, keeping the x- and y-axes of both graphs parallel, until
the actual data point curve coincides with or matches the type
curve.
Step 6. Using the match point, calculate the properties of the reservoir.
Example 16-51
0 pi= 0 0
4.3 22 22
21.6 82 82
28.2 95 95
45.0 119 119
48.0 injection ends
51.0 109 109
69.0 55 55
73.0 47 47
93.0 32 32
142.0 16 16
148.0 15 15
Solution
Step 1. Figure 16-13 shows a plot of the well-test data during the injec-
tion period (48 hours) in terms of p versus t on a tracing paper
with the same scale dimensions as in Figure 16-12. Using the
overlay technique with the vertical and horizontal movements,
find the segment of the type curve that matches the actual data.
1This example problem and the solution procedure were given by Earlougher, R.,
Advanced Well Test Analysis, SPE Monograph Series, SPE, Dallas, TX (1977).
Ch16.qxd 3/15/06 8:02 PM Page 1271
5
4
3
1 9 MATCH POINT
8
7 10
6
5
4 t = 100 HR
PD
M
3
(t /r 2) = 50
D D
2
pM = 10 PSI
(pD)M = 0.80
101 9
8
7 1
6 1 10 10
2
103
5
TESTING TIME,t, HR
4
3
2 1
10
1 2 3 4 5 6 7 891 2 3 4 5 6 7 8 91 2 3 4 5 6 7 8 91 2 3 4 5 6 7 8 91 2 3 4 5 67 89
101 1 10 102 103
tD/rD2
Figure 16-13. Illustration of type curve matching for an interference test using the
type curve (After Earlougher, R. Advances in Well Test Analysis) (Permission to pub-
lish by the SPE, copyright SPE, 1977 ).
Step 3. Using Equations 16-35 and 16-36, solve for the permeability and
porosity:
and:
Example 16-6
(a) Calculate the pressure at radii of 10 ft and 100 ft for the flowing times
0.1, 0.5, 1.0, 2.0, 5.0, 10, 20, 50 and 100 hours. Plot [pi-p(r,t)] versus
(t/r2) on a log-log scale.
(b) Present the data from part a in terms of [pi-p(r,t)] versus (t/r2) on a
log-log scale.
Solution
70.6 Q B 948 c t r 2
p ( r , t ) = pi + Ei
kh kt
or
70.6 (360) (1.444) (1.3) 948 (0.15) (1.3) (17.6 x 106 ) r 2
pi p ( r , t ) = Ei
( 22.9) ( 20) ( 22.9) t
Ch16.qxd 3/15/06 8:02 PM Page 1273
r2
pi p ( r , t ) = 104 Ei 0.0001418
t
Assumed
t, hours r = 10 feet r = 100 feet
Ei[0.0001418 Ei[0.0001418
t/r2 r2/t] pi-p(r,t) t/r2 r2/t] pi-p(r,t)
100
90
et
r = 10 fe
80
70
70.6QB 948fct r 2
p(r, t) = pi + Ei
kh kt
60
Pi p(r,t)
50
40
feet
r = 100
30
20
10
0
0 20 40 60 80 100 120
Time, hrs
Figure 16-14. Pressure profile at 10 feet and 100 feet as a function of time.
Ch16.qxd 3/15/06 8:02 PM Page 1274
Figure 16-14 shows different curves for the two radii. Obviously, the
same calculations can be repeated for any number of radii and, conse-
quently, the same number of curves will be generated. However, the solu-
tion can be greatly simplified by examining Figure 16-15. This plot
shows that when the pressure difference pi p(r,t) is plotted versus t/r2,
the data for both radii form a common curve. In fact, the pressure differ-
ence for any reservoir radius will plot on this exact same curve.
1000
900
800
700
600 r2
pi p(r,t) = 104Ei 0.0001418
pi p(r,t)
t
500
400
300
200
100
0
0 0.2 0.4 0.6 0.8 1 1.2
t/r2
Figure 16-15. Pressure profile at 10 feet and 100 feet as a function of t/r2.
Ch16.qxd 3/15/06 8:02 PM Page 1275
qt 1
Hyperbolic: = 1/ b
qi 1 + b D i t
In a dimensionless form:
1
q Dd =
1 + b t Dd
1/ b (16-37)
Ch16.qxd 3/15/06 8:02 PM Page 1276
where the decline-curve dimensionless variables qDd and tDd are defined by
qt
q Dd =
qi (16-38)
t Dd = D i t (16-39)
qt 1
Exponential: =
qi exp D i t
1
Similarly, q Dd = (16-40)
exp t Dd
qt 1
Harmonic: =
qi 1+ Di t
1
or q Dd = (16-41)
1 + t Dd
141.2 B 1
qi =
k h p ln rD
1
2
( )
It is obvious that the right-hand side of the previous equation is dimen-
sionless, which indicates that the left-hand side of the equation is also
dimensionless. This relationship thus defines the dimensionless rate qD as
follows:
141.2 B q i 1
qD = =
k h p 1 (16-43)
ln( rD )
2
Recall the dimensionless form of the diffusivity equation from Chapter 6,
Equation 6-90:
2 pD 1 pD pD
+ =
rD2 rD rD tD
qDd and tDd in terms of the transient dimensionless rate qD and time tD.
Combining Equation 16-38 with Equation 16-43 gives
qt qt
q Dd = =
qi k h ( p i p)
1
( )
141.2 B ln rD
2
or
1
( )
q Dd = q D ln rD
2
1 0.006328 t
t Dd =
1 ( c t ) rwa
2
1 2
( )
rD 1 ln rD (16-46)
2 2
1 0.006328
Di = 2
1 ( c t ) rwa
1 2
( )
r 1 ln rD
2 D 2
(16-47)
10
re /
rw =
10 Empirical Type Curve Solution
20 1
50 q(t) ; for b > 0
100 q Dd = =
200 TRANSIENT DEPLETION qi [ 1 + bDit ]1/b
10000 2000
q(t) 1
1.0 q Dd = = ; for b > 0
re /rw = 10000 qi 2
e Dit
t Dd = Dit
qDd Exponential
common to analytical
0.1 and empirical solutions
b=0
2 rw rw 2 2 rw rw 2
0.7
0.
0.6
0.5
0.4
0.
0.2
0.3
0.1
9
0.001
104 103 102 101 1 10 102 103
tDd
Figure 16-16. Fetkovich type curves (After Fetkovich, 1980, JPT June 1980, copyright
SPE 1980).
tDd and the solution to Equation 16-37 as a function of tDd for several val-
ues of b ranging from 0 to 1.
Notice for Figure 16-16 that all curves coincide and become indistin-
guishable at tDt 0.3. Any data existing before a tDt of 0.3 will appear to
represent exponential decline regardless of the true value of b and, thus,
will plot as a straight line on a semi-log scale. With regard to the initial
rate qi, it is not the actual producing rate at early time; it is very specifical-
ly a pseudo-steady-state rate at the surface. This pseudo-state rate can be
substantially less than the actual early time transient flow rates that would
be produced from low-permeability wells with large negative skins.
The basic steps used in Fetkovich type-curve matching of declining
ratetime data are as follows:
Step 1. Plot the historical flow rate, qt, versus time, t, in any convenient
units on log-log paper or tracing paper with the same logarithmic
cycles as in the Fetkovich type curve.
Step 2. Place the tracing-paper data plot over the type curve and slide
the tracing paper with plotted data, keeping the coordinate axes
parallel, until the actual data points match one of the type curves
with a specific value of b.
Ch16.qxd 3/15/06 8:02 PM Page 1280
Step 3. From the match of the particular type curve of Step 2, record values
of the reservoir dimensionless radius re/rwa and the parameter b.
Step 4. Select any convenient match point on the actual data plot (qt and
t)mp and the corresponding values lying beneath that point on the
type-curve grid (qDd, tDd)mp.
Step 5. Calculate the initial surface gas flow rate, qi, at t = 0 from the
rate match point:
q
qi = t
q D t mp (16-48)
Step 6. Calculate the initial decline rate, Di, from the time match point:
t Dd
Di = (16-49)
t mp
Step 7. Using the value of re/rwa from Step 3 and the calculated value of
qi, calculate the formation permeability, k, by applying Darcys
equation in one of the following three forms:
Pseudo-pressure form:
Pressure-squared form:
1422 T ( g Z)avg [ ln ( re / rwa ) 0.5 ] q i
k= (16-51)
h ( pi2 p2wf )
Pressure-approximation form:
where k = permeability, md
pi = initial pressure, psia
pwf = bottom-hole flowing pressure, psia
m(P) = pseudo-pressure, psi2/cp
qi = initial gas flow rate, Mscf/day
T = temperature, R
h = thickness, ft
g = gas viscosity, cp
Z = gas deviation factor
Bg = gas formation volume factor, bbl/scf
Step 8. Determine the reservoir pore volume (PV) of the well drainage
area at the beginning of the boundary-dominated flow from the
following expression:
56.54 T qi
PV =
( g c t )i [ m ( pi ) m ( pwf )] D i (16-53)
28.27 T ( g Z) avg q i
PV = (16-54)
( g c t )i [ p2i p2wf ] D i
with
( PV )
re = (16-55)
h
Ch16.qxd 3/15/06 8:02 PM Page 1282
re 2
A= (16-56)
43, 560
where PV = pore volume, ft3
= porosity, fraction
g = gas viscosity, cp
ct = total compressibility coefficient, psi1
qi = initial gas rate, Mscf/day
Di = decline rate, day1
re = drainage radius of the well, ft
A = drainage area, acres
subscripts
i = initial
avg = average
Step 9. Calculate the skin factor, s, from the re/rwa matching parameter
and the calculated values of A and re from Step 8.
r r
s = ln e w (16-57)
rwa mp re
( PV )[1 Sw ]
G= (16-58)
5.615 Bgi
The initial gas in place can also be estimated from the following
relationship:
qi
G= (16-59)
D i (1 b)
more the b value curves approach one another, which leads to the difficulty
of obtaining a unique match. Arguably, applying the type-curve approach
with only three years of production history may not be possible for some
pools. Unfortunately, since time is plotted on a log scale, the production
history becomes compressed so that even when incremental history is
added, it may still be difficult to differentiate and clearly identify the
appropriate decline exponent b.
The following example illustrates the use of the type-curve approach
to determine reserves and other reservoir properties.
Example 16-7
The ratetime data from the past 8 years were plotted and matched to an ED: 8 or
eight?
re/rwa stem of 20 and b of 0.5, as shown in Figure 16-17. The resulting
match point has the following coordinates:
qt = 1000 Mscf/day
t = 100 days
qDd = 0.58
tDd = 0.126
Ch16.qxd 3/15/06 8:02 PM Page 1284
56.54T qi
PV =
( gc t)i [m(pi ) m(p wf)] D i
1000 b = .5
MATCH POINT
q = 1000 Mcf/d: qDd = 0.58
qD = 2400 Mcf/d t = 100 DAYS: tDd = 0.126
ONSET OF q
DEPLETION tW = 0.0700 md qi =
rW1 = 02.1 lt qD i MP
qD = 1400 Mcf/d
r = 6.17
qg, MCF/D
ro = 1242 lt tD
i
Di =
100 b=0 t MP
MATCH POINT
q = 2000 Mcf/d: qa = 0.68
t = 100 DAYS: tW = 0.128
tW = 0.006 md 8 YEARS
rW1 = 77.4 ft
r = 6.38
re = 1647 ft
b=0 b = 0.5
10
10 100 1000 10000
TIME, days
Figure 16-17. West Virginia gas well A type curve fit (Copyright SPE 1987 ).
(a) Permeability, k
(b) Drainage area, A
(c) Skin factor, s
(d) Gas in place, G
Solution
q
qi = t
q D t mp
1000
qi = = 1724.1 Mscf /day
0.58
Ch16.qxd 3/15/06 8:02 PM Page 1285
and:
t Dd
Di =
t mp
0.126
Di = = 0.00126 day1
100
Step 3. Calculate the reservoir pore volume of the well drainage area
using Equation 16-53:
56.54 T qi
PV =
( g c t )i [ m ( pi ) m ( pwf )] D i
( PV )
re =
h
( 20.36) 106
re = = 1242 ft
(70)(0.06)
Ch16.qxd 3/15/06 8:02 PM Page 1286
and:
re 2
A=
43, 560
(1, 242)2
A= = 111 acres
43, 560
r r
s = ln e w
rwa mp re
0.35
s = ln ( 20) = 5.18
1242
( PV )[1 Sw ]
G=
5.615 Bgi
The initial gas G can also be estimated from Equation 16-59, to give
qi
G=
D i (1 b)
1.7241(106 )
G= = 2.737 Bscf
0.00126 (1 0.5)
Ch16.qxd 3/15/06 8:02 PM Page 1287
q g = C ( pr2 pwf
2
)n
Fetkovich (1996) suggested that the Arps decline exponent b and the
decline rate can be expressed in terms of the exponent n as follows:
1 pwf
2
b=
2n
( 2 n 1)
pi (16-60)
q
Di = 2n i (16-61)
G
pressure (i.e., pwf = 0 or pwf < pi), the equation can be reduced to the fol-
lowing expression:
1
b =1 (16-62)
2n
The exponent n from a gas well back-pressure performance curve can
therefore be used to calculate or estimate b and Di. Equation 16-61 pro-
vides the physical limits of b, which is between 0 and 0.5, over the accept-
ed theoretical range of n, which is between 0.5 and 1.0 for a single-layer
homogeneous system, as shown in the following table:
n b
Layered No-Crossflow
Behavior, 0.5 < b < 1
0.1
St
q, Rate
im
ul
Single Layer Behavior, at
io
n
0.5 b < 1 O
pp
or
0.01 tu
ni
ty
b
=
b=0
b=
1
5
0.001
0.01 0.1 1 10 100 1000
t, Time
Figure 16-18. Depletion decline curves (After Fetkovich, 1997, copyright SPE 1997 ).
Ch16.qxd 3/15/06 8:02 PM Page 1290
qi
qt = 2n
q i 2 n 1 (16-63)
1 + ( 2 n 1) t
G
1
q i 1 2 n
G p( t ) = G 1 1 + ( 2 n 1) t (16-64)
G
For n = 0.5, b = 0:
q
q t = q i exp i t (16-65)
G
Ch16.qxd 3/15/06 8:02 PM Page 1291
q
G p ( t ) = G 1 exp i t
G (16-66)
For n = 1, b = 0.5:
qi
qt =
qi 2
1+ t (16-67)
G
G
G p( t ) = G
q t (16-68)
1+ i
G
These relationships are based on Pwf = 0, which implies that qi = qimax,
as given by
k h p2i
q i = q i max =
1422 T ( g Z) avg ln ( re / rw ) 0.75 + s (16-69)
where
qimax = stabilized absolute open-flow potential, i.e., at Pwf = 0, Mscf/day
G = initial gas in place, Mscf
qt = gas flow rate at time t, Mscf/day
t = time
Gp(t) = cumulative gas production at time t, Mscf
For a commingled well producing from two layers at a constant Pwf,
the total flow rate (qt)total is essentially the sum of the flow rates from all
layers, or
(q t ) total = (q t )1 + (q t )2
where the subscripts 1 and 2 represents the more permeable layer and less
permeable layer, respectively. For a hyperbolic exponent of b = 0.5, Equa-
tion 16-67 can be substituted into the above expression to give
Ch16.qxd 3/15/06 8:02 PM Page 1292
(q )max total
=
(q max )1
+
(q max )2
2 2 2
q q q
1 + t max 1 + t max 1 + t max (16-70)
G total G 1 G 2
q max q max
G 1 = G 2
Equation 16-70 indicates that only if will the
value of b = 0.5 for each layer yield a composite ratetime value of b = 0.5.
Mattar and Anderson (2003) presented an excellent review of methods
that are available for analyzing production data using traditional and
modern type curves. Basically, modern type-curve analysis methods
incorporate the flowing pressure data along with production rates, and
they use the analytical solutions to calculate hydrocarbon in place. Two
important features of modern decline analysis that improve upon the tra-
ditional techniques are as follows:
Normalization of rates using flowing pressure drop: plotting a nor-
malized rate (q/p) enables the effects of back-pressure changes to be
accommodated in the reservoir analysis.
Handling the change in gas compressibility with pressure: using
pseudo-time as the time function, instead of real time, enables the gas
material balance to be handled rigorously as the reservoir pressure
declines with time.
Fetkovich originally developed his type curves for gas and oil wells
that are producing at constant pressures. Carter (1985) presented a new
set of type curves developed exclusively for the analysis of gas rate
data. Carter noted that the changes in fluid properties with pressure sig-
nificantly affect reservoir performance predictions. Of utmost impor-
tance is the variation in the gas viscositycompressibility product, gcg
which was ignored by Fetkovich. Carter developed another set of
decline curves for boundary-dominated flow that uses a new correlating
parameter, , to represent the changes in gcg during depletion. The
parameter, called the dimensionless drawdown correlating parameter,
Ch16.qxd 3/15/06 8:02 PM Page 1293
( g c g )i
= (16-71)
( g c g )avg
or, equivalently,
( g c g )i m ( pi ) m ( pwf )
=
2 p p (16-72)
i
wf
Zi Zwf
where
cg = gas compressibility coefficient, psi1
m(p) = real gas pseudo-pressure, psi2/cp
pwf = bottom-hole flowing pressure, psi
pi = initial pressure, psi
g = gas viscosity, cp
Z = gas deviation factor
10
= 1.234
= 1.184
= 1.045
= 1.01
= 1.004
=1
= 0.75
qD
0.1
= 0.55
1424qT (1/B1)
qD =
skh [m(p i) m(p wf)]
4
2. 364 10 24kt
tD = (a1)2
f (pi)cg(pj)r2w
0.01
0.001
0.01 0.1 1 10 100
tD
Figure 16-19. Radial-linear gas reservoir type curves (After Carter, SPEJ 1985,
copyright SPE 1985).
( g c g )i
=
( g c g )avg
Or
( g c g )i m ( pi ) m ( pwf )
=
2 p p
i
wf
Zi Zwf
Ch16.qxd 3/15/06 8:02 PM Page 1295
Step 3. Match the rate data to a type curve corresponding to the comput-
ed value of in Step 1. If the computed value of is not one of
the values for which a type curve is shown, the needed curve can
be obtained by interpolation and graphical construction.
Step 5. Estimate the gas that would be recoverable if the average reser-
voir pressure were reduced from its initial value to Pwf from the
following expression:
(q t )mp
G = G i G Pwf = (16-73)
(q D t D )mp
pi
Zi
Gi = G
p p (16-74)
i wf
Zi Zwf
Ch16.qxd 3/15/06 8:02 PM Page 1296
Bgi G i
A= (16-75)
43, 560 h (1 s w i )
Example 16-7
p, psia g, cp Z
1 0.0143 1.0000
601 0.0149 0.9641
1201 0.0157 0.9378
1801 0.0170 0.9231
2401 0.0188 0.9207
3001 0.0208 0.9298
3601 0.0230 0.9486
4201 0.0252 0.9747
4801 0.0275 1.0063
5401 0.0298 1.0418
pi = 5400 psia
T = 726 R
= 0.070
= 0.55
pwf = 500 psi
h = 50 ft
Swi = 0.50
1.27 8.300
10.20 3.400
20.50 2.630
40.90 2.090
Ch16.qxd 3/15/06 8:02 PM Page 1297
81.90 1.700
163.80 1.410
400.00 1.070
800.00 0.791
1,600.00 0.493
2,000.00 0.402
3,000.00 0.258
5,000.00 0.127
10,000.00 0.036
Solution
Step 2. Plot the production data, as shown in Figure 16-20, on the same
log-log scale as Figure 16-16 and determine the match points of
the following:
(q)mp = 1.0 MMscf/day
(t)mp = 1000 days
(qD)mp = 0.605
(tD)mp = 1.1
= 1.045
(q t )mp
G = G i G Pwf =
(q D t D )mp
10
= 1.234
= 1.184
= 1.045
= 1.01
= 1.004
1
Dimensionless rate qD
=1
= .75
= .55
0.1 1424qT(1/B1)
qD =
kh [m(pi) m(pwf)]
2.364 104 24kt
tD = (a1)2
f(pi)cg(pj) r 2m
0.01
0.01 0.1 1 10 100
Dimensionless time tD
pi
Zi
Gi = G
p p
i wf
Zi Zwf
5400
1.0418
Gi = 2860 = 3176 MMscf
5, 400 500
1.0418 0.970
Zi T (1.0418) (726)
Bgi = 0.0287 = 0.02827 = 0.000396 ft 3 /scf
pi 5400
Ch16.qxd 3/15/06 8:02 PM Page 1299
Bgi G i
A=
43, 560 h (1 s w i )
0.23396 Q B t 162.6 QB 4A
pwf = pi log
Ah c t kh 2
1.781CA rw
where
k = permeability, md
A = drainage area, ft2
CA = shape factor
Q = flow rate, STB/day
t = time, hrs
ct = total compressibility coefficient, psi1
Ch16.qxd 3/15/06 8:02 PM Page 1300
Expressing the time t in days and converting from log to a natural log-
arithm, ln, the above relation can be written as follows:
pi pwf p B 4A 5.615 B
= = 70.6 ln 2
+ t
q q k h 1.781CA rwa A h C t (16-76)
or more conveniently as
p
= b pss + m t (16-77)
q
B 4A
Intercept: b pss = 70.6 kh ln 2 (16-78)
1.781CA rwa
5.615 B
Slope: m = (16-79)
A h c t
m ( p)
= b pss + mt (16-81)
q
711 T 4A
Intercept: b pss = ln 2
kh 1.781CA rwa
56.54 T 56.54 T
Slope: m = =
( g c t )i ( hA ) ( g c t )i ( pore volume)
The linkage that allows for the conversion of gas-production data into
equivalent constant-rate liquid data is based on the use of a new time
function called pseudo-equivalent time or normalized material bal-
ance pseudo-time, defined as follows:
( g c g )i t q ( g c g )i Z G
ta = o cg dt = q 2ip [m ( pi ) m ( p)]
t
qt (16-82)
g t i
where
ta = pseudo-equivalent (normalized material balance) time, days
t = time, days
G = original gas-in-place, Mscf
qt = gas flow rate at time t, Mscf/day
p = average pressure, psi
g = gas viscosity at p , cp
c = gas compressibility at p , psi1
g
(p) = normalized gas pseudo-pressure, psi2/cp
m
Ch16.qxd 3/15/06 8:02 PM Page 1302
qg 1
b pss =
m ( pi ) m ( pwf ) m
1+ (16-83)
ta
b pss
(p) is the normalized pseudo-pressure as defined by
where m
gi Zi Pi
p
m ( pi ) =
pi 0 Z dp (16-84)
g
gi Zi P p
m ( p) =
pi 0 Z dp (16-85)
g
and
1
m= (16-86)
G c ti
70.6 gi Bgi 4A
b pss = ln 2
kg h 1.781 CA rwa (16-87)
pi = initial pressure
rwa = effective (apparent) wellbore radius, ft
Bgi = gas formation volume factor at pi, bbl/Mscf
1
q Dd =
1 + ( t a ) Dd (16-88)
with
qg
q Dd = b pss
m ( pi ) m ( pwf ) (16-89)
m
( t a ) Dd = ta (16-90)
b pss
m ( p i ) m ( p)
= b pss + m t a (16-91)
qg
0 pi Zi i pi/Zi pi/(Z )i
gi Zi P p
m ( p) =
pi Z dp
0 g
The required calculations of this step can be performed in the following
tabulated form:
P
p gi Z i
area = dp m ( p ) = (area)
p 0 g Z pi
0 0 0
pi
p pi G P
= 1
Z Zi G
Ch16.qxd 3/15/06 8:02 PM Page 1305
Step 6. From the plot generated in Step 3, enter the graph of p versus p/Z
with each value of the ratio p Z and determine the value of the
corresponding average reservoir pressure p . For each value of
(p
the average reservoir pressure p , determine the values m ) for
each p.
( c )
ta =
g g i
qt
ZiG
2 pi
( ) ()
m pi m p
( g c g )i Z i G
ta = [m( p i ) m( p )]
m( p ) qt 2p i
T qt GP p
Step 9. Recalculate the initial gas in place, G, by using the value m from
Step 8 and applying Equation 16-86 to give
1
G=
c ti m
Step 10. The new value of G from Step 9 is used for the next iteration,
i.e., starting from Step 5, and this process could continue until
some convergence tolerance for G is met.
Palacio and Blasingame developed a modified Fetkovich-Carter type
curve, as shown in Figure 16-21, to allow the performance of constant-rate
Ch16.qxd 3/15/06 8:02 PM Page 1306
1
ta qg
(q Dd )i =
ta
0
dt
m ( p ) m ( p ) a (16-92)
i wf
1 d ta qg
( )
q Dd = 1
0 m p m p dt a
id
t a dt a ta
i( ) wf ( )
(16-93)
800
101 101
b =1
0.9
0.8
102 0.7 102
0.6
0.5
0.4
0.3
0.2
0.1
b =0
103 10 3
104 103 102 101 100 101 102 103
tDd
Figure 16-21. Palacio-Blasingame type curve.
Ch16.qxd 3/15/06 8:02 PM Page 1307
qg
m( p wf ) m ( p i ) m ( p wf )
t qg ta pwf
(p )m
Plot qg/[m (p)] versus t on a Cartesian scale.
i a
Step 3. Using the well production data as tabulated and plotted in Step 2,
compute the two complementary plotting functions, as given by
Equations (16-92) and (16-93) as a function of ta.
1
ta qg
(q Dd )i =
ta
0
dt
m ( p ) m ( p ) a
i wf
1 d 1
ta
qg
(q Dd )id = ( )
t a dt a t a dt a
m ( p ) m ( p )
0 i wf
Step 4. Plot both functions, i.e., (qDd)i and (qDd)id, versus ta on a tracing
paper so it can be laid over the type curve of Figure 16-22 for
matching.
1 t a (q Dd )i
G= (16-94)
c ti t Dd mp q Dd mp
5.615 G Bgi
A=
h (1 Swi )
Ch16.qxd 3/15/06 8:02 PM Page 1308
A
re =
re
rwa =
reD
rwa
s = ln r
w
141.2 Bgi gi re 1 (q Dd )i
k= ln (16-95)
h rw 2 q Dd mp
qg/Pp, Mscf/D/psia
Analysis for Gas Well A: 100
reD = re/r 1w = 4 (qg/Pp)fun/qDd = 0.78 Mscf/D/psia
ta,bart/tDd = 115 days Liquid Loading
12 7
G = 0.245 bscf 101
28 18 Gas Well A
80 48
160 Low Demand in Rate
800
4
100 110 102 100
100 101 102 103
qDdi and qDdid
q q 1 1
= =( ) QN + \
m ( pi ) m ( pwf ) m ( p) \
G b pss b pss
2 q t pi t a
QN =
(c t i Zi ) m ( p)
( g c g )i ZiG
ta = [m ( pi ) m ( p)]
qt 2 pi
Step 4. Using the assumed value of G and for each production data point
of GP at time t, calculate p/Z from the gas material balance equa-
tion, Equation 16-21:
p pi G P
= 1
Z Zi G
Step 5. For each production data point of qt and t, calculate ta and the
normalized cumulative production QN:
( g c g ) i ZiG
ta =
qt 2 pi i ()
m (p ) m p
2 q t pi t a
QN =
(c t i Zi ) m ( p)
Ch16.qxd 3/15/06 8:02 PM Page 1311
Step 6. Plot q/p versus QN on a Cartesian scale and obtain the best line
through the date points. Extrapolate the line to the x-axis and
read the OGIP.
Step 7. The new value of G from Step 6 is used for the next iteration,
i.e., Step 3, and this process could continue until some conver-
gence tolerance for G is met.
p pi G P
= 1
Z Zi G
p D = (1 G pD )
where
p/Z
pD = (16-96)
pi / Zi
GP
G PD =
G
Anash and co-authors indicated that the product ( g c t ) can be
expressed in a dimensionless ratio of (g ct)i/g ct) as a function of the
dimensionless pressure, pD, by one of the following three forms:
a) First-order polynomial
The first form is a first-degree polynomial that is adequate in describing
the product, gct, as a function of pressure at gas-reservoir pressures below
Ch16.qxd 3/15/06 8:02 PM Page 1312
5000 psi, that is, pi < 5000. The polynomial is expressed in a dimensionless
form as
i c ti
= pD (16-97)
ct
where
cti = total system compressibility at pi, psi1
i = gas viscosity at pi, cp
b) Exponential model
The second form is adequate in describing the product, gct,for high-
pressure gas reservoirs, that is, pi > 8000 psi
ic ti
= o exp(1 p D ) (16-98)
ct
i c ti
= a o + a 1p D + a 2 p2D + a 3p3D + a 4 p4D (16-99)
ct
The coefficient in Equations 16-98 and 16-99, 0, 1, ao, a1, etc., can
be determined by plotting the dimensionless ratio [i cti/ct] versus pD on
a Cartesian scale, as shown in Figure 16-23, and using the least-squares
regression model to determine the coefficients.
The authors also developed the following fundamental form of the sta-
bilized gas flow equation:
dG P Jg pD i c ti
dt
= qg =
c ti pwD
c t
dp D
pwf / Zwf
pwD =
pi / Zi
Ch16.qxd 3/15/06 8:02 PM Page 1313
1.0
Legend: Data Trends
0.9 (g = 0.55 and T =150F)
pi =1000 psi
pi =3000 psi
0.8
pi =5000 psi
pi =8000 psi
pi =1000 psi
0.7 pi =10,000 psi
pi =12000 psi 3000 psi
5000 psi
0.6
(0/cti)/(ci)
0.5
8000 psi
0.4
10,000 psi
pi = 12,000 psi
0.3
0.2
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
pD = (p/z)/(pi /z)
Figure 16-23. Typical distribution of the viscosity-compressibility function (After
Anash et al., 2000).
101
reD = re/r' w = 4 Legend: Type Curves for the Assumption of a Linear Relation of uct versus p/z
(i.e., the First-Order Polynomial Model). This is the Pressure-Squared
12 7 Flow Solution, and is valid for Low Pressure Reservoirs (p1 < 5000 psia)
28 18
80 48
800 160
1 104
100 Transient "Stems" Depletion Stems
(Transient Radial Flow Region- (Boundary-Dominated Flow Region-
Analytical Solutions Volumetric Reservoir Behavior)
Legend: "Knowles" Type Curves
Dd
q
(Semi-Analytical Solutions)
= 0.50
= 0.625
= 0.75
101 = 0.875
= 0.998
Fluid Parameters: = 0.998
g = 0.55; T= 150F; pi= 4000 psia 0.875
Solutions Legend: Boundary-Dominated Flow Region 0.75
Solid Lines ="Carter" Type Curves (Numerical Simulation) 0.625
Dased Lines = Semi-Analytical Gas Flow Solutions 0.5
102
103 102 101 100 101 102
tDd
Figure 16-24. First-order polynominal solution for real-gas flow under boundary-
dominated flow conditions. Solution assumes a ct profile that is linear with PD
(Permission to copy by the SPE, 2000 ).
101
Transient "Stems"
reD = re/r' w = 4 (Transiant Radial Flow Region-
Legend: Type Curves for the Assumption of an Exponential Relation of
ct versus p/z (i.e., the "Exponential" Model). This Model
Analytical Solutions is Only Valid for High Pressure Cases (pi <8000 psia)
12 7
28 18
80 48
800 160
1 104
100 Depletion "Stems"
(Boundary-Dominated Flow Region-
Volumetric Reservoir Behavior)
Legend: "Exponential" Model
qDd
(Semi-Analytical Solutions)
= 0.50
= 0.625
= 0.75
101 = 0.875
= 0.998
Cases for pi = 8000
Fluid Parameters:
g = 0.55; T= 150F; pi= 8000 psia and = 0.998 and 12,000 psia
12,000 psia 0.875 overlay one another
102
103 102 101 100 101 102
t
Dd
101
Transient "Stems"
reD = re/rw=4 Legend: Type Curves for the Assumption of an General Polynomial
(Transient Radial Flow Region-
Relation of ct versus plz (g = 0.55;T = 150F;pi = 4000,
Analytical Solutions)
12 7 8000, and 12,000 psia)
28 18
80 48
800 160
1 104
100
Depletion "Stems"
(Boundary-Dominated Flow Region-
Legend: "General Polynomial" Model Volumetric Reservoir Behavior)
qDd
(Semi-Analytical Solutions)
= 0.50
= 0.625 Cases for pi = 8000
= 0.75 and 12,000 psia
= 0.875 overlay one another
101
= 0.998
Fluid Parameters:
g =0.55; T = 150E; pi = 4000,8000 and = 0.998 = 0.5 (pi = 4000 psia
12,000 psia
0.875
Solutions Legend: Boundary-Dominated Flow Region 0.75
Solid Lines = "Carter" Type Curves (Numerical Simulation) 0.625
Dashed Lines = Semi-Analytical Gas Flow Solutions 0.5 0.4
102 3
10 102 101 100 101 102
tDd
Figure 16-26, General polydomial solution for real-gas flow under boundary-
dominated flow conditions (Permission to copy by the SPE, 2000).
Step 1. Using the available gas properties, prepare a plot of (i cti/ ct)
versus pD, where
p/Z
pD =
pi / Zi
Step 2. From the generated plot; select the appropriate functional form
that describes the resulting curve:
First-order polynomial
i c ti
= pD
ct
Ch16.qxd 3/15/06 8:02 PM Page 1316
Exponential model
ic ti
= o exp(1 p D )
ct
i c ti
= a o + a 1p D + a 2 p2D + a 3p3D + a 4 p4D
ct
Step 3. Plot the historical flow rate, qg, versus time, t on a log-log scale
with the same logarithmic cycles as the one given by the selected
type curves (i.e., Figures 16-24 through 16-26).
( reD )mp
pwf / Zwf
pwD = (16-100)
pi / Zi
1
= ( 1 p2w ) (16-101)
2 D
Ch16.qxd 3/15/06 8:02 PM Page 1317
= A o + A1 + A 2 + A 3 + A 4 (16-103)
where
A o = ( A1pwD + A 2 pwD
2
+ A 3p3wD + A 4 pwD
4
)
A1 = a o
a1
A2 =
2
a2
A3 =
3
a3
A4 = s
4
C ti q g
Jg =
q Dd (16-104)
mp
Step 8. Estimate the OGIP, in scf, from the time match point:
Jg t
G=
C ti t Dd mp (16-105)
Ch16.qxd 3/15/06 8:02 PM Page 1318
Step 9. Calculate the reservoir drainage area, A, in ft2, from the follow-
ing expression:
5.615 Bgi G
A= (16-106)
h (1 Swi )
where A = drainage area, ft2
Bgi = gas formation volume factor at pi, bbl/scf
Swi = connate water saturation
Step 10. Calculate the permeability, k, in md, from the match curve of
the dimensionless drainage radius, reD:
141.2 i Bgi J g 1
k= ln[reD ] mp (16-107)
h 2
Step 11. Calculate the skin factor from the following relationships:
A
Drainage radius re = (16-108)
r
Apparent wellbore radius rwa = ( r )
e
(16-109)
eD mp
r
Skin factor s = ln
wa
(16-110)
w
r
Example 16-8
The West Virginia gas well, A, is a vertical gas well that has been hydrauli-
cally fractured and is undergoing depletion. The production data were pre-
sented by Fetkovich (1980) and used in Example 16-6. A summary of the
reservoir and fluid properties is given below:
rw = 0.354 ft
h = 70 ft
= 0.06
Ch16.qxd 3/15/06 8:02 PM Page 1319
T = 160 F
s = 5.17
k = 0.07 md
g = 0.57
Bgi = 0.00071 bbl/scf
gi = 0.0225 cp
cti = 0.000184 psi1
pi = 4175 psia
pwf = 710 psia
= 0.4855 (first-order polynomial)
Swi = 0.35
Solution
Step 1. Figure 16-27 shows the type-curve match of the production data
with that of Figure 16-24, to give:
(q Dd )mp = 1.0
( t Dd )mp = 1.0
( reD )mp = 28
101
Transient "Stems" Legend: West Virginia Gas Well A."from SPE 14238:
reD = re/r1w = 4
(Transient Radial Flow Regoin- General Polynomial Type Curve Approah.
12 7 Analytical Solution) (Low Pressure case pi = 4000 psia
28 18
80 48
800 160
1 x 104
100
Depletion "Stems"
Data for West Virginia Gas Well:
pi = 4175 psia (r = 0.9498) (Boundary-Dominated Flow Region-
g = 0.57 (sin = 1) T = 160F Volumetric Reservoir Behaviour)
qDd
rW = 0.354 ft
h = 70 ft
= 0.4701 (dimensionless)
PWi = 710 psia (zWt = 0.9488)
102
103 102 101 100 101 102
tDd
Figure 16-27. Type curve analysis of West Virginia gas well A (SPE 14238). Gen-
eral polynomial type curve analysis approach (Permission to copy by the SPE, 2000).
Jg t
G=
C ti t Dd mp
743.758 695
G= = 2.834 scf
0.0001824 1.0
Step 5. Compute the permeability from the match on the reD = 28 tran-
sient stem by using Equation 16-107:
Ch16.qxd 3/15/06 8:02 PM Page 1321
Step 6. Calculate the skin factor by applying Equations 16-108 and 16-109:
A 4.1398 x 106
re = = = 1147.9 ft
r 1147.9
rwa = ( r ) = 28 = 40.997 ft
e
eD mp
r 40.997
s = ln r = ln 0.354 = 4.7522
wa
Uniform-Flux Fractures
kf w f FC
Conductivity group: FCD = =
k xf k xf
re
Fracture group: reD =
xf
1 A 1 2.2458
p D = 2 t DA + ln 2 + ln + s
2 rw 2 CA
with the dimensionless time based on the wellbore radius, tD, or drainage
area, tDA, as given by Equations 6-87 and 6-87a:
0.0002637 kt
tD =
c t rw2
0.0002637 kt r2
t DA = = tA w
c t A A
The authors adopted the last form and suggested that, for a well pro-
ducing under pseudo-steady-state at a constant rate with a finite-conduc-
tivity fracture in a circular reservoir, the dimensionless pressure drop can
be expressed as follows:
pD = 2 tDA + bDpss
or
bDpss = pD 2tDA
where the term bDpss is the dimensionless pseudo-steady-state constant
that is independent of time; however, bDpss is a function of
the dimensionless radius, reD, and
the dimensionless fracture conductivity, FCD
The authors note that, during pseudo-steady flow, the equation describ-
ing the flow during this period yields constant values for given values of
reD and FCD that are closely given by the following relationship:
0.43464 a 1 + a 2 u + a 3 u + a 4 u + a 5 u
2 3 4
with
u = ln(FCD)
where a1 = 0.93626800 b1 = 0.38553900
a2 = 1.0048900 b2 = 0.06988650
a3 = 0.31973300 b3 = 0.04846530
a4 = 0.0423532 b4 = 0.00813558
a5 = 0.00221799
Based on the above equations, Pratikno et al. (2003) used Palacio and
Blasingames previously defined functions (i.e., ta, (qDd)i, and (qDd)id)
and the parameters reD and FCD to generate a set of decline curves for a
sequence of 13 values for FCD with a sampling of reD = 2, 3, 4, 5, 10, 20,
30, 40, 50, 100, 200, 300, 400, 500, and 1000. Type curves for FCD of
0.1, 1, 10, 100, and 1000 are shown in Figures 16-28 through 16-32.
The authors recommend the following type-curve matching procedure,
which is similar to the methodology used in applying Palacio and
Blasingames type curve:
Deplelion "Stems"
100 (Boundary-Dominated Flow
100
3
reD= 2 qDdid Region-Volumetric
10
40
20 Reservoir Behavior)
50
30
100 200
400 300 300
102 102
104 103 102 101 100 101 102 103
tDd, bar =NpDd /qDd
qDb
qDb
1 Depletion"Stems"
2 3
100 10 4 (Boundary-Dominated Flow) 100
30 F cD = 2 Region-Volumetric
40 20
Reservoir Behavior
100 50
400 200 300
300
101 101
Fractured Well for a F cD = 1000
bounded circular
reservoir
qDb/d
102 102
104 103 102 101 100 101 102 103
tDd' bar =NpDd /qDd
101 101
Fractured Well in a reD =1000
Bounded Circular
Reservoir
qDdid
102 102
104 103 102 101 100 101 102 103
tDd, bar = NpDd/qDd
101
Fractured Well in a reD =1000 101
Bounded Circular
Reservoir qDdid
102
102
104 103 102 101 100 101 102 103
tDd, bar=NpDd /qDd
100 100
30
Depletion "Stems" 100
20 3 (Boundary-Dominated Flow
10
200 50 40 4 tcD= 2 RegionVolumetric
400 5 Reservoir Behavior)
500 300
101 101
Fractured Well in a reD =1000
Bounded Circular
Reservoir qDdid
102 102
104 103 102 101 100 101 102 103
tDd, bar=NpDd /qDd
( g c g )i ZiG
For gas: t a = [m ( pi ) m ( p)]
qt 2 pi
(p ) and m
where m (p) are the normalized pseudo-pressures, as defined
i
by Equations 16-84 and 16-85:
gi Zi Pi
p
m ( pi ) =
pi Z dp
0 g
gi Zi P p
m ( p) =
pi Z dp
0 g
Notice that the GOIP must be calculated iteratively, as illustrated previ-
ously by Palacio and Blasingame (1993).
For gas:
qg
q Dd =
m ( pi ) m ( pwf )
1
ta qg
(q Dd )i =
ta m ( p ) m ( p dt a
0 i wf
)
Ch16.qxd 3/15/06 8:02 PM Page 1329
1 d 1
ta
qg
(q Dd )id = ( )
t a dt a t a dt a
m ( p ) m ( p )
0 i wf
For oil:
qo
q Dd =
pi pwf
1
ta
qo
(q Dd )i =
ta
0
p p dt a
i wf
1 d 1
ta
qo
(q Dd )id = ( ) p p dt a
wf
t a dt a t a 0 i
Step 4. Plot the three gas or oil functions, qDd, (qDd)i, and (qDd)id, versus
ta on a tracing paper so that it can be laid over the type curve
with the appropriate value of FCD.
Step 5. Establish a match point for each of the three functions (qDd,
(qDd)i, and (qDd)id). Once a match is obtained, record the time and
rate match points as well as the dimensionless radius value, reD:
a) Rate-axis match point Any (q/p)MP (qDd)MP pair
b) Time-axis match point Any (t )MP (tDd)MP pair
c) Transient flow stem Select the (q/p), (q/p)i
and (q/p)id functions that best match the transient data stem and
record reD
Step 6. Solve for bDpss by using the values of FCD and reD:
u = ln(FCD)
0.43464 a 1 + a 2 u + a 3 u + a 4 u + a 5 u
2 3 4
Step 7. Using the results of the match point, estimate the following
reservoir properties:
Ch16.qxd 3/15/06 8:02 PM Page 1330
1 ta (q / m ( p)
For gas: G = g
c ti t Dd mp q Dd mp
141.2 Bgi gi (q g / m ( p) MP
kg = b Dpss
h (q Dd ) MP
5.615 G Bgi
A=
h (1 Swi )
A
re =
1 t a (q o / p)i
For oil: N =
c t t Dd mp q Dd mp
5.615 N Boi
A=
h (1 Swi )
A
re =
Step 8. Calculate the fracture half-length, xf, and compare with Step 1:
re
xf =
reD
Ch16.qxd 3/15/06 8:03 PM Page 1331
Example 16-9
A Texas field vertical gas well has been hydraulically fractured and is
undergoing depletion. A summary of the reservoir and fluid properties is
as follows:
rw = 0.333 ft
h = 170 ft
= 0.088
T = 300 oF
g = 0.70
Bgi = 0.5498 bbl/Mscf
gi = 0.0361 cp
cti = 5.1032 180 105 psi1
pi = 9330 psia
pwf = 710 psia
Swi = 0.131
FCD = 5.0
Fetkovich-McCray Type Curve for a Vertical Well with a Finite Conductivity Vertical Fracture (FcD = 5)
[Example 1-Low Permeability/High Pressure Gas Reservoir (Texas)]
104 103 102 101 100 101 102 103
102 102
Legend: Data Functions Results Legend: Example
Transient "Stems" (qDg/,\Pp) Data
(Transient Flow Region (qDg/,\Pg) Data
Analytial Solution: FcD=5)
(qDg/,\Pp)id Data
101 101
qDd
qDdi
qDd, qDdi, and qDdid
Depletion "Stems"
100 100
10 2
3
(Boundary-Dominated Flow
100
30 4
40 20 5 fcD= 2 RegionVolumetric
200
400 Reservoir Behavior)
50
300
500
reD =1000
101 101
Fractured Well in a qDdid
Bounded Circular
Reservoir Model Legend: Fetkovich-McCray Type Curve-Fractured
Well Centered In a Bounded Circular Reservoir
(Finite Conductivity: FcD=5)
102 102
104 103 102 101 100 101 102 103
Material Balance Pseudotime Function, days [G=1.0 BSCF (forced)]
Figure 16-33. Match of production data for Example 1 on the Fetkovich-McCray
decline type curve (pseudopressure drop normalized rate versus material balance
time format) for a well with a finite conductivity vertical fracture (FcD=5) (Permission to
copy the SPE, 2003 ).
Ch16.qxd 3/15/06 8:03 PM Page 1332
Figure 16-33 shows the type-curve match for FCD = 5, with the matching
points:
(qDd)mp = 1.0
[(qg / m(p )]mp = 0.89Mscf/psi
(tDd)mp = 1.0
(ta)mp = 58 days
(reD)mp = 2.0
Solution
Step 1. Solve for bDpss by using the values of FCD and reD:
0.43464 a 1 + a 2 u + a 3 u + a 4 u + a 5 u
2 3 4
0.43464 a 1 + a 2 u + a 3 u + a 4 u + a 5 u
2 3 4
Step 2. Using the results of the match point, estimate the following
reservoir properties:
1 ta (q / m ( p)
G= mp g
c ti t Dd q Dd mp
1 58 0.89
G= mp = 1.012 x10 MMscf
6
5.1032 x105
1 . 0 1.
141.2 Bgi gi (q g / m ( p) MP
kg = b Dpss
h (q Dd ) MP
Ch16.qxd 3/15/06 8:03 PM Page 1333
5.615 G Bgi
A=
h (1 Swi )
A 2401195
re = = = 277 ft
Step 3. Calculate the fracture half-length, xf, and compare with Step 1:
re 277
xf = = = 138 ft
reD 2
PROBLEMS
1/1/2000 0 1,017
2/1/2000 1 978
3/1/2000 2 941
4/1/2000 3 905
5/1/2000 4 874
6/1/2000 5 839
7/1/2000 6 809
8/1/2000 7 778
9/1/2000 8 747
10/1/2000 9 722
11/1/2000 10 691
12/1/2000 11 667
1/1/2001 12 641
Ch16.qxd 3/15/06 8:03 PM Page 1334
a) the first six months of the production history data to determine the
coefficient of the decline-curve equation.
b) predict flow rates and cumulative gas production from August 1,
2000 through January 1, 2001.
c) Assuming that the economic limit is 20 MMscf/month, estimate the
time to reach the economic limit and the corresponding cumulative
gas production.
2. The volumetric calculations on a gas well show that the ultimate
recoverable Reserves, Gpa, are 18 MMMscf of gas. By analogy with
other wells in the area, the following data are assigned to the well:
Exponential decline
Allowable (restricted) production rate = 425 MMscf/month
Economic limit = 20 MMscf/month
Nominal decline rate = 0.034 month1
Calculate the yearly production performance of the well.
3. The following data are available on a gas wells production:
pi = 4100 psia
= 0.10
pwf = 400 psi
Swi = 0.30
T = 600 R
g = 0.65
h = 40 ft
Time, days qt, MMscf/day
0.7874 5.146
6.324 2.108
12.71 1.6306
25.358 1.2958
50.778 1.054
101.556 0.8742
248 0.6634
496 0.49042
992 0.30566
1240 0.24924
1860 0.15996
3100 0.07874
6200 0.02232
REFERENCES
13. Chen, H.-Y., and Teufel, L. W., A New Rate-Time Type Curve for Analy-
sis of Tight-Gas Linear and Radial Flows, SPE 63094, 2000 SPE Annual
Technical Conference & Exhibition, Dallas, TX, October 2000.
14. Cox, S. A., Gilbert, J. V., Sutton, R. P., and Stoltz, R. P., Reserve Analysis
for Tight Gas, SPE 78695, Eastern Regional Meeting, Lexington, KY,
October 2002.
15. Doublet, L. E., Blasingame, T. A., Pande, P. K., and McCollum, T. J.,
Decline Curve Analysis Using Type CurvesAnalysis of Oil Well Pro-
duction Data Using Material Balance Time: Application to Field Cases,
SPE 28688, 1994 Petroleum Conference & Exhibition of Mexico, Veracruz,
Mexico, October 1994.
16. Earlougher, R. C., Jr., Advances in Well Test Analysis. SPE of AIME 5,
1977. In Theory and Practice of the Testing of Gas Wells, 3 ed., Calgary:
Energy Resources Conservation Board, 1975.
17. Engler, T. W., A New Approach to Gas Material Balance in Tight Gas
Reservoirs, SPE 62883, 2000 APE Annual Tech Conference and Exhibi-
tion, Dallas, TX, October 2000.
18. Fetkovich, M. J., Multipoint Testing of Gas Wells, SPE Mid-Continent
Section Continuing Education Course of Well Test Analysis, March 17,
1975.
19. Fetkovich, M. J., Decline Curve Analysis Using Type Curves, SPE 4629,
SPE Journal, June 1980.
20. Fetkovich, M. J., Vienot, M. E., Johnson, R. D., and Bowman, B. A., Case
Study of a Low-Permeability Volatile Oil Field Using Individual-Well
Advanced Decline Curve Analysis, SPE 14237, SPE 60th Annual Tech
Conference & Exhibition, Las Vegas, NV, September 1985.
21. Fetkovich, E. J., Fetkovich, M. J., and Fetkovich, M. D., Useful Concepts
for Decline Curve Forecasting, Reserve Estimation, and Analysis, SPE
Reservoir Engineering, February 1996.
22. Fetkovich, M. J., Vienot, M. E., Bradley, M. D., and Kiesow, U. G.,
Decline Curve Analysis Using Type Curves: Case Histories, SPE 13169,
SPE Formation Evaluation, December 1987.
23. Fetkovich, M. D., Fetkovich, M. J., Thomas, L. K., and Guerrero, E. T.,
Oil and Gas Relative Permeabilities Determined from Rate-Time Perfor-
mance Data, SPE 15431, SPE 61st Annual Tech Conference & Exhibition,
New Orleans, LA, October 1985.
24. Gentry, R. W. Decline Curve Analysis. JPT, January 1972, p. 3842.
25. Ikoku, C., Natural Gas Reservoir Engineering. John Wiley & Sons, Inc.,
1984.
Ch16.qxd 3/15/06 8:03 PM Page 1337
26. Jeimert, T. A., Theory and Application of Polynomial Type Curves, SPE
25876, SPE Rocky Mountain Regional/Low Permeability Reservoirs Sym-
posium, Denver, CO, April 1993.
27. Lee, J., Well Testing, Society of Petroleum Engineers of AIME, Dallas,
TX, 1982.
28. Lord, M. E., and Collins, R. E., Detecting Compartmented Gas Reservoirs
Through Production Performance, SPE 22941, SPE 66th Annual Technical
Conference & Exhibition, Dallas, TX, October 1991.
29. McCain, W. D., Voneiff, G. W., Semmelbeck, M. E., and Hunt, E. R.,
A Tight Gas Field Study: Carthage (Cotton Valley) Field, SPE 26141,
SPE Gas Technology Symposium, Calgary, Alberta, Canada, June 1993.
30. Palacio, C. and Blasingame, T. Decline-Curve Analysis Using Type-Curve,
SPE paper 25909, Presented at the 1993 SPE Rocky Mountain regional
Meeting, Denver, CO April 26-28; 1993.
31. Pratikno, H., Rushing, J., and Blasingame, T. A., Decline Curve Analysis
Using Type Curves: Fractured Wells, SPE 84287, SPE Annual Technical
Conference, Denver, CO, October 2003.
32. Spivey, J. P., Gatens, J. M., Semmelbeck, M. E., and Lee, W. J., Integral
Type Curves for Advanced Decline Curve Analysis, SPE 24301, SPE Mid-
Continent Gas Symposium, Amarillo, TX, April 1992.
33. van Everdingen, A. F., and Hurst, W., Application of Laplace Transform to
Flow Problems in Reservoirs, Trans. AIME 186, 1949, pp. 305-324B.
34. Marhaendrajana, T., and Blasingame, T. A., Decline Curve Analysis Using
Type Curves: Evaluation of Well Performance Behavior in a Multiwell
Reservoir System, SPE 71517, 2001 SPE Annual Technical Conference
& Exhibition, New Orleans, LA, September 2001.
35. Valko, P. P., Doublet, L. E., and Blasingame, T. A., Development and
Application of the Multiwell Productivity Index (MPI), SPE Journal 5(1),
March 2000.