SPE 132475

Using Second Derivative of Transient Pressure in Welltest Analysis of Low

Permeability Gas Reservoirs
Hassan Bahrami, M. Reza Rezaee, and Akim Kabir, Curtin University of Technology; Jamal Siavoshi, Husky Energy;
and Rochdi Jammazi, Oil and Gas Institute of Monastir

Copyright 2010, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Production and Operations Conference and Exhibition held in Tunis, Tunisia, 8–10 June 2010.

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
Welltest interpretation requires diagnosis of reservoir flow regimes in order to determine the basic reservoir characteristics such as
average reservoir permeability and skin factor. Due to wellbore storage effect, wellbore phenomena and complexity of reservoir
response from heterogeneous reservoir layers, detection of the reservoir flow regimes using standard welltest diagnostic plots
might be challenging and have some uncertainties.
In pressure transient testing, there are instances where the flow regimes might not clearly be revealed on diagnostic plots of
pressure build-up and its derivative, such as incomplete pressure build-up tests, low permeability reservoirs and multi-phase
producing wells. In such cases, the Semi-Log plot of first and second derivative of transient pressure versus time can be used to
reduce the uncertainties associated with welltest analysis.
This paper describes a new method for well test interpretation using second derivative of transient pressure. Two field examples
are shown in which a reliable radial flow regime on pressure build-up data could not be detected using standard plots. The second
derivative approach was used to predict radial flow regime trend and estimate the reservoir permeability and skin factor, which the
results were in good agreement with production data in these wells.

Introduction
Pressure-transient testing has long been recognized as a reservoir characterization tool. The welltest can characterize the ability of
the fluid to flow through the reservoir and to the well and provide a description of the reservoir in dynamic conditions, as opposed
to geological and log data. The well test interpretation is the process of obtaining information about a reservoir by analyzing a
pressure transient response caused by a change in production rate. Welltest analysis results are the overall response of reservoir to
dynamic disturbances made to the formation at the testing time.
A pressure transient test breaks into several flow regimes, each seeing deeper in reservoir than the last. Depending on well
completion type, completion configuration, reservoir geological and geometric attributes, different flow regime might be revealed
on pressure transient data. The early portion of the welltest data is controlled by the wellbore storage and skin effect. Wellbore
storage is due to decompression of fluid in the wellbore during pressure drawdown. The reverse happens during pressure build-up,
in which the reservoir continues to flow into the well after shut-in. A long enough test is often required to overcome effects of the
wellbore storage and wellbore phenomena, in order to have the reservoir response revealed on the pressure transient data [1].
The radial flow regime is the most common flow regime observed during transient testing. Diagnosing the radial-flow regime is
critical to quantitative welltest interpretation, since it provides values for permeability * thickness and skin factor for the formation
layers that contributed to the test [2, 3]. The reservoir flow regimes might be distorted or even masked by long wellbore storage in
low permeability reservoirs, complex heterogeneous reservoirs, or due to wellbore phenomena such as presence of multi-phase in
wellbore during the pressure build-up testing.

Pressure Derivative Approach

The diffusivity equation solution that describes the radial flow regime in a homogeneous porous medium for a constant flow rate
drawdown test is expressed in field units as follows [2, 3]:
2 SPE 132475

162.6qPB k
'P [log 't  log( )  3.23  0.87 S ] (1)
kh M .P .C t .rw 2
Equation (1) shows that the plot of pressure change versus logarithm of elapsed time, results in a straight line with the certain slope
of “m” as shown in Figure (1-a). The above equation can be simplified to the following general form:

'P = m Log(Time fuction) + b (2)

The time fuction in case of pressure drawdown test is “'t” and for pressure build-up test it can expressed as “(t p +'t)/'t”, where 't
is elapsed time and t p is production time prior to the pressure build-up test. In case of pressure drawdown test, by taking derivative
of Equation (2) with respect to logarithm of ǻt:
d ( ' P) (3)
m
d (log ' t)
The above equation can be written as follows:
d ( 'P)
Log [ ] 0 log [ 't ]  log[ m ] (4)
d (log 't)
Equation (4) indicates that in a constant flow rate drawdown test, for the pressure data related to radial flow regime, plot of
pressure derivative, P’: d('P)/d(Log't), versus elapsed time on a Log-Log scale results in a zero-slope line. Using the pressure
drivative curve as described in Figure (1-b), the intercept of the zero slope straight line with the vertical axis, “m”, can provide
permeability and skin values in field units as follows:
Q PB
k *h 162 .6 (5)
m
' P1hr k (6)
S 1 .1513 [  Log ( )  3 .2275 ]
m M .P .ct rw 2
In order to diagnose reservoir flow regimes, the plot of the downhole pressure derivative versus time functions on Log-Log scale
can be used. For the pressure transient test as described in Figure (1-b), after the wellbore storage effect, the pressure continues
building up while the derivative curve swings down and eventually flattens out as the transient moves far enough from wellbore
into the formation.
The pressure derivative plot is more informative than conventional Semi-Log plots, as the pressure derivative is much more
sensitive to reservoir characteristics than the pressure and also responds distinctly to the downhole changes. The permeability and
skin calculations depend on value of effective producing thickness. However skin is less sensitive since change in skin is related to
logarithm of the reservoir parameters values.
Furthermore, based on derivations of fluid flow and diffusivity equations for different flow regimes, on the pressure derivative
curve, the slope of +1 shows pure wellbore storage effect, and the slopes -0.5, +0.5, +0.25 and +0.36 indicate spherical, linear, bi-
linear and elliptical flow regimes respectively [2, 3 and 9].

Taking derivative of Equation (3) with respect to logarithm of ǻt, provides the equation for second derivative of pressure:
d 2 ( ' P) (7)
0
d (log ' t) 2
The above Equation can be written as follows

d 2 ( 'P)
 0 log[ 't ]  0 (8)
d (log't) 2
Equation 8 shows that for a pressure drawdown test, plot of second derivative of pressure, P’’: -w2['P]/w[Log't]2, versus Log (ǻt)
results in a zero-slope straight line with intercept of zero (For pressure build-up test, the time function is “(t p +'t)/'t"). A typical
second derivative curve for a pressure transient test has been shown in Figure 1-c. The second derivative can validates the
existence of radial-flow regime on first derivative, when there is uncertainty in radial flow regime identification using the standard
diagnostic plots.
SPE 132475 3

Radial Flow Regime Prediction

To calculate permeability from the derivative plot, reservoir response should be significant and the test should be long enough to
have the radial flow regime established in the formation and observe a reliable zero slope line on the pressure derivative curve
data. Compared with first derivative, the advantage of second derivative of pressure is that its intercept is certain (zero) and
therefore the second derivative curve trend might be predictable.
The radial flow regime prediction needs knowing approximate time at which the beginning of radial flow regime is expected. The
beginning time of the radial flow regime is approximately 1.5 log cycles after pure wellbore storage effect is ended [2]. Also the
second derivative curve has two extremum points, and approximately 1.5 cycles after the first extremum point of the second
derivative plot can be considered as radial flow regime beginning time, as shown in Figures (1-b) and (1-c).
The second derivative trend from its second extremum point to radial flow regime can be predicted by knowing the second
derivative intercept (zero) and having radial flow regime beginning time (1.5 log cycles after end of wellbore pure storage effect).
Once the second derivative curve is determined, the first derivative curve can be back calculated from the predicted second
derivative trend, which eventually provides permeability and skin values.
An important parameter in the radial flow regime prediction is estimation of the time at which radial flow regime is started. For
this purpose, sensitivity analysis was performed for several cases with different permeability, skin and wellbore storage constant
values as shown in Figure (2). The results indicated that reservoir parameters can have a significant influence on duration of
wellbore storage effect. The time period between pure wellbore storage effect and radial flow regime can differ in diverse cases,
and it is actually controlled by skin and permeability (also by wellbore storage constant C, but in terms of number of cycles it
appears to be the same). Therefore, the assumption of 1.5 log cycles duration might not always be valid.
The outputs of the sensitivity analysis suggest that the time period can vary from 1 cycle to 2.5 cycles depending on wellbore and
reservoir parameters. Therefore, for radial flow regime prediction based on second derivative curve, a sensitivity analysis needs to
be performed for effect of wellbore storage duration on calculated permeability and skin.

Field Examples
In pressure transient testing, there are instances where the radial flow regime might not clearly be reached on diagnostic plots of
pressure build-up and its derivative, for example: incomplete pressure build-up tests, low permeability reservoirs and multi-phase
producing wells. If the reservoir permeability is very low such as in tight gas formations, the wellbore storage effect can be very
long, masking the reservoir flow regimes. In such cases, the second derivative approach might help to extrapolate the second
derivative curve to the time related to start of radial flow regime. Then first derivative curve can then be determined in order to
estimate permeability and skin. It should be noted that since the second derivative is more sensitive to the downhole pressure
changes, data smoothing should also be applied on the second derivative curve. A couple of field examples where the above
developed theory can be applied are described below.

- Incomplete Welltest in a Low Permeability Gas Reservoir

A pressure build-up test was performed in a gas well in order to evaluate well productivity. Figure (3-a) shows the Log-Log
diagnostic plot of pressure and pressure derivative for the main build-up period. The derivative curve shows dominant effect of
wellbore storage on pressure data. It indicates that the test duration in the main build-up period is not long enough to detect radial
flow regime. Therefore the radial flow regime was predicted using by the second derivative approach.
The first and second derivative data were plotted on a Semi-Log scale as illustrated in Figure (3-b). As a rule of thumb, beginning
of radial flow regime was considered approximately 1.5 time log cycles after the end of pure wellbore storage region. Using this
assumption, beginning of radial flow regime was approximated at 12 hours.
Using available second derivative data, curve fit was performed from the second extremum point on second derivative to zero
value point at 12 hours where beginning of radial flow regime is expected. Then, the first derivative of pressure data was
determined from the fitted curve on second derivative points. The value of pressure derivative in radial flow region was estimated
as 4.1E+07 psi2/cp, which corresponds to permeability of 0.2 mD and skin of +9.
A consistency check of the results was also performed by considering the beginning of the infinite acting radial flow at 9 hrs and
25 hrs after the start of the build-up test which resulted in permeability of 0.1 mD and 0.3 mD respectively for the different cases.
The results highlighted an excellent convergence of the permeability values considering different times for beginning of radial
flow regime, compared with the 0.2 mD obtained by considering the beginning of the infinite acting radial flow at 12 hr (1.5 cycles
after end of pure wellbore storage). According to this welltest analysis results, the low well productivity is mainly due to the low
permeability.
The petrophysical evaluations in this well showed average reservoir porosity of 0.03. There was also no evidence of natural
fractures contribution on production. These additional data confirm the low value of reservoir permeability estimated from the
welltest analysis.
4 SPE 132475

- Multi-Phase Producing Gas Well

In a gas-water producing well, pressure build-up test was performed to identify the well productivity problems. Figure (4-a) shows
the Log-Log diagnostic plot of pressure and pressure derivative for the main build-up period. The derivative curve shows that after
wellbore storage and during build-up of pressure, the data have been affected due to phase segregation in wellbore. This caused
uncertainty in identifying reservoir flow regimes on pressure derivative.
First, the data that were affected by phase segregation were removed, and then first and second derivatives of pressure were plotted
for the unaffected data on a Semi-Log scale as shown in Figure (4-b). As a rule of thumb, beginning of radial flow regime was
considered approximately 1.5 cycles in time after the end of pure wellbore storage region. Using available second derivative data,
curve fit was performed from the second extremum point on second derivative to zero value point at 1 hour where beginning of
radial flow regime is expected, as show in Figure (4-c). Then the first derivative of pressure data was determined from the fitted
curve on second derivative points. The value of pressure derivative in radial flow region was estimated to be 2.8E+06 psi2/cp,
which corresponds to permeability of 4.1 mD and a skin of +86.

Conclusions
x Radial flow regime can be indicated by zero slope line with a certain intercept on first derivative curve, and zero slope line
with zero intercept on second derivative curve.
x End of wellbore storage can be detected using second derivative plot. The first extremum point on the second derivative plot
can approximate the time at which pure wellbore storage effect is ended.
x There are instances where the flow regimes might not clearly be revealed in the diagnostic pressure build-up plots, e.g.,
incomplete pressure build-up tests, low permeability reservoirs and multi-phase producing wells. In such cases, the Semi-Log
plot of first and second derivative of transient pressure versus time can be used to reduce the uncertainties associated with
welltest analysis.
x Radial flow regime can be predicted using curve fitting on the second derivative points from the second extremum point on
second derivative to the zero value point at 1.5 cycles after pure wellbore storage effect.
x Once the second derivative curve is determined, the first derivative curve can be back calculated from the predicted second
derivative trend, and therefore permeability and skin can be estimated.
x As a rule of thumb, radial flow regime is assumed to be started from 1.5 time log cycles after pure wellbore storage effect.
However depending on well and reservoir parameters, it can vary from 1.0 to 2.5 log cycles. Therefore for radial flow regime
prediction based on second derivative curve, sensitivity analysis needs to be performed for effect of skin and permeability on
wellbore storage duration.

Acknowledgement
The authors would like to appreciate Dr. Seyyed Reza Shadizadeh (Petroleum University of Technology, Iran) for many helpful
discussions on this work.

Nomenclatures:
P Pressure
Q Flow rate
B Formation Volume Factor
μ Viscosity
t Time
K Permeability
S Skin
rw Wellbore radius
h Thickness
C Wellbore Storage Constant
ct Total Compressibility
M Porosity
RF Radial Flow
WBS Wellbore Storage
P’ First Derivative of Pressure
P’’ Second Derivative of Pressure
m(P) Pseudo Pressure
SPE 132475 5

References
1. Ahmed, Tarek, « Handbook of reservoir engineering », 2nd edition, 2000
2. Bahrami H., Siavoshi J., Second derivative yields new insights to well test analyses, Oil and Gas Journal, Canada, Dec 2005
3. KAPPA engineering team, 2005—Production Logging Technical Reference, Emeraude 2 (42), Section E-04, 1–8
4. Earlougher, R.C. Jr. « Advances in Well Test Analysis » Monograph Series, SPE ,1977.
5. Bourdarot G., « Well Testing : Interpretation Methods », Edition Technip, Paris 1998.
6. Guo, B. and Ghalambor, A., 2005—Natural Gas Engineering Handbook. Houston: Gulf Publishing Company.
7. Lakovlev, S.V. , « Multi-Phase Flow in Several Layers Limits the Applicability of Conventional Buildup Analysis »,
SPE/AAPG Western Regional Meeting, 19-22 June 2000, Long Beach, California
8. Ayan, C., « Multiphase Pressure Buildup Analysis: Field Examples», SPE 17412, SPE California Regional Meeting, 23-25
March 1988, Long Beach, California
9. Badazhkov, D., « Analysis of Production Data with Elliptical Flow Regime in Tight Gas Reservoirs », SPE 117023, 2008
6 SPE 132475

(a)

(b)

+1 slope
m 0 slope

(c)

Figure 1- Pressure Transient Analysis Diagnostic Plots; a: Semi Log Horner Plot, b: Log-Log Standard Plot, c: Semi-Log Plot of
2nd Derivative curve (In the example, the first extremum point of second derivative curve is at 0.9 hrs and the second extremum
point at 7 hrs)
SPE 132475 7

(a)

(b)

(c)

Figure 2- Sensitivity Analysis Plots for a pressure drawdown test, to see effect of reservoir and well parameters on wellbore
storage duration and beginning of radial flow regime, a: Effect of wellbore storage constant, b: Effect of Skin, c: Effect of
Permeability
8 SPE 132475

(a)

(b)

K=0.2 md
S=+9

Figure 3- Pressure build up analysis for an incomplete test in a low permeability gas reservoir, a: Log-Log standard diagnostic plot,
b: Semi-Log 2nd Derivative plot [Time function: (t p +'t)/'t]
SPE 132475 9

(a)

Phase segregation
effect

(b)

(c)

(d)

K=5.3 md
S=+85
Actual data

Figure 4- Pressure build up analysis for a water-gas producing well in a medium permeability gas reservoir; a: Log-Log standard
diagnostic plot, b: Semi-Log 2nd Derivative plot for reliable data, c: Semi-Log 2nd Derivative plot showing curve for on 2nd
derivative cure, d: Log-Log standard diagnostic plot with predicted pressure 1st derivative [Time function: (t p +'t)/'t]