Professional Documents
Culture Documents
A. JAHANBANI
R. AGUILERA
University of Calgary
Abstract
A tight gas reservoir is commonly defined as a reservoir having
less than 0.1 milliDarcies permeability. Because of the very low
permeability, hydraulic fracturing is usually carried out in efforts
to establish commercial production. There are several basic concepts and field cases of different well tests in tight gas reservoirs
in the literature. In this paper, we gather information and provide
a guide to some of the most important tests.
Generally because of low permeability, a well will not flow
initially at measurable rates and conventional well testing cannot
be applied. We review procedures for design of pre- and postfracture tests in single and dual porosity reservoirs. The prefracture test permits estimating preliminary values of reservoir
permeability and initial pressure. The post-fracture test provides
data for estimating fracture half length and conductivity. We also
review the application of convolution/deconvolution methods to
analyze well tests with significant wellbore storage.
Because of economic and environmental reasons, short duration procedures are preferred. However, although effective in
many instances, these methods also have their own limitations.
Introduction
Unconventional reservoirs (tight gas, coal bed methane, shales
gas and gas hydrates) will be an important pat of the global energy
mix for decades to come. Large reserves, long-term potential, costs
and gas prices and some other factors account for the great influence of these resources on the future of energy.
There is no formal definition for tight gas. A commonly used
definition, describes tight gas reservoirs as those having permeabilities smaller than 0.1 milliDarcies.
Well testing is generally done as an aid to estimate gas in-place
and recoverable volumes. Initial pressure (pi) is a critical parameter not only for estimating gas in-place, but also for determining
how much field development is required and whether or not the
field is overdeveloped. In addition to pi, well testing provides an
estimate of permeability and skin.
A problem associated with well testing in tight gas sands is
that usually long times are required to reach radial flow, due to
their extremely low permeabilities. Therefore, conventional well
tests cannot be applied to these reservoirs. Because of initial uneconomic rates, fracturing is usually required. Lee(2) has suggested
procedures for pre- and post-fracture tests design. In order to have
measurable gas rates for pre-frac testing, often a breakdown with
acid, KCl water or N2 is necessary.
There are many different types of pre-frac open-hole and casedhole tests. Some of the important open-hole tests include DST
Peer Reviewed Paper (Review and Publication Process can be found on our Website)
64
104
X eD
103
feeding into the fractures. This could be the result of damage induced during the hydraulic fracturing job.
102
101
100
10-2 -4
10
Infinite
10
5
2
1
CD =
10-1
01
0.0
10
0.0
0.1
10-3
10-2
10-1
100
101
102
103
104
PD (dpD/dtD)
101
100
10-1
10-2
10-3
10-4
10-2
100
102
104
106
tD
FIGURE 2: Dimensionless pressure and pressure derivative curves
for a single and dual porosity reservoirs dominated by linear flow
at early times, yet obscured by wellbore storage effects (CD = 0.01,
= 0.01 and f = 40,000). In this case it is nearly impossible to
distinguish the behaviour of the single and dual porosity reservoirs.
(Source: Shahamat and Aguilera, 2008)(2).
October 2009, Volume 48, No. 10
Pre-Closure Analysis
The pre-closure methods were proposed by Mayerhofer and
Economides(4, 5). They introduced a technique for estimation of
formation permeability along with fracture parameters from a
mini-frac test.
Their technique is based on an infinite-conductivity vertical
fracture solution to the diffusivity equation with a changing fracture face skin effect. They have decoupled the total pressure change
resulting from the transient flow from fracture to reservoir (during
fracture closure) into some components. Only the pressure change
of reservoir and that of filter cake are considered to be important. In this approach, the filter cake is treated as a changing skin
effect.
The first step in their method of analysis is to prepare a diagnostic log-log plot of the rate-normalized pressure (RNP) and its
derivative versus time. From there, the transient reservoir response
can be identified. The next step is to analyze a specialized plot that
permits estimating permeability to gas.
Vi 1
4 k h t
........................................................................... (1)
V 1
dp
i
d (ln t ) 4 k h t
............................................................................. (2)
Equation (2) shows that a unit-slope line is obtained if the pressure derivative is plotted against time on log-log coordinates. This
is useful in identifying the flow regime when there is no information about initial pressure.
A Cartesian plot of p versus 1/t yields a straight line for the
late-time data whose slope gives the value of permeability. Extrapolation of the straight line to 1/t = 0, gives the initial reservoir
pressure. The analysis could also be applied to the case with no
fractures, which is identical to an injection-type slug test.
Usually long shut-in times are required to apply the impulse solution. In 1997, Nolte et al.(8) further developed the application and
analysis techniques (also known in the industry as mini-fall-off
tests). The after-fracture closure period potentially consists of:
Pseudo-linear flow period: used to calculate closure time,
spurt loss and fracture length.
Pseudo-radial flow period: used to calculate reservoir pressure and transmissibility.
The late-time pressure decline data is analyzed for pseudo-radial flow, from which reservoir parameters can be calculated by
a technique similar to conventional Horner plot method. The following relationship exists:
()
( ) ........................................................................... (3)
p t pi = mR FR t , tc
where tc is the time to fracture closure with zero time at the start of
injection and FR is the radial flow time function defined as:
t
1
16
FR t , tc = ln 1 + c =
1.6
4 t tc
2
.................................................... (4)
( )
mR tc
.................................................................................... (5)
Convolution / Deconvolution
The application of superposition theorem (Duhamels principle)
in petroleum industry dates back to 1949 when van Everdingen
and Hurst(11) published their classic paper of the application of the
Laplace transform to flow problems. The terms superposition and
convolution and also the terms de-superposition and deconvolution are used interchangeably.
Convolution
The convolution theory is very useful in manipulating the constant rate and constant bottomhole pressure solutions of the diffusivity equation to obtain variable rate and variable bottomhole
pressure solutions. In 1949, van Everdingen and Hurst(11) proposed
the variable rate solution of the diffusivity equation using convolution of the constant rate solution with the variable rate under
consideration. Finally, they obtained a dimensionless convolution
integral that can be written as:
( )
tD
( ) d d
0
( )
qD D pDCR t D D d D
.............. (6)
( )
( )
n1
)
................. (7)
Deconvolution
The importance of the superposition theorem lies in determining
the constant rate solution when the variable rate and pressure data
are available. This approach is called deconvolution. From the previous equations, if pwD(tD) and qD(tD) are known, pDCR(tD) can be
determined through a de-convolution process. Once the constant
Journal of Canadian Petroleum Technology
wD =
( )
( )
( )
n1
( )
tD =
()
i w t
i wf
3.6 103 kt
ct Ac
)i
................................................................................ (14)
................................................................................... (15)
(8)
where
qD =
............................................................. (9)
( )
p =2
0
pdp
z
................................................................................... (10)
( )
DCR p t D = 4 t D
........................................................................ (11)
Substituting this solution into Equation (7) (in terms of pseudopressure) gives:
n1
( )
wD p t D = 4 qD ,1 t D + 4 qD , j1 qD , j
j =1
t D t D, j
.......... (12)
DCR
()
k Ac i w t
=
1.295 qT
.................................................................. (13)
()
1.295 q t T
k Ac i wf
) .......................................................................... (16)
i w t = C [q1 t + q j+1 q j
j =1
()
t tj
....................................... (17)
where
C=
0.55T
k Ac
( ct )i
.............................................................................. (18)
and Ac = 4xf.h. The factor C is determined through an optimization process to match the observed bottomhole pressure. Once C
is determined, permeability can be estimated. Viscosity and total
compressibility are functions of pressure but in this case they are
treated as constants evaluated at initial pressure. This will introduce
some errors in calculations as the pressure changes, particularly in
the analysis of afterflow period with large pressure variations. As
mentioned by Helmy(13), it is helpful to reconstruct the superposition based on average reservoir pressure instead of initial reservoir
pressure.
q=
(24) Vw cw dp
Bg
dt
.............................................................................. (19)
( )
( )
pwD t D = qD ,1 pDCR t D +
n1
. ....................................... (20)
The pressure solution for infinite acting radial flow at the wellbore is approximated by:
Equation (23) states that a Cartesian plot of wellbore normalized pseudo-pressure (nw) versus the rate convolution time function, results in a straight line. From the slope of this line, formation
flow capacity or permeability can be calculated by Equation (25).
Equation (23) also states that a value of skin is required before performing the analysis. Therefore, different preliminary
values of skin must be assumed. On the Cartesian plot at very low
rates, curves with different values of skin form a single line. The
most correct line is the one that demonstrates the best straight line
through most of the curve. The skin value corresponding to this
line is therefore considered as the skin.
............................................................................. (26)
where
1
pDCR t D = ln t D + 0.80907 + s
2
..................................................... (21)
( )
( )
0 =
3.6 103 k
................................................................................... (27)
n =
i Zi
2 pi
.......................................................................................... (22)
( ) dtc
ta t =
........................................................................................ (28)
Tariq(17)
( ) s
q t
ni nw = m log t + 0 +
q1
q1
()
.................................................... (23)
The term in the bracket is called rate convolution time function and:
n1 q
q
) (
) (
) (
) (
) (
) +
. . (24)
where
m=
68
q Bg
0.535 kh
......................................................................................... (25)
The boundary conditions include having an instantaneous pressure drop from pi to wellbore pressure at the time of perforation.
The use of Laplace transformation with some approximations results in straight forward equations for handling early and late-time
analysis.
Different flow regimes are identified. The early-time data are
dominated by wellbore storage, from which skin can be calculated.
Late-time data are reservoir dominated and can be analyzed to provide estimates of reservoir pressure and permeability. In order to
obtain reasonable estimates of reservoir parameters, pressure data
must contain at least some part of the reservoir dominated flow. To
determine if this is the case, a special derivative known as impulse
derivative (IDER) can be used(17, 18):
( )
IDER = ta
d
w
dta ............................................................................ (29)
of 1. However, the two graphs are very similar to each other. The
part of the graph where derivative becomes flat (slope = 0) represents reservoir dominated flow and gives the approximate start of
radial flow. Sometimes the reservoir dominated flow is reached
after a very long time, which is not practical. This problem can be
reduced by making the wellbore volume smaller and hence getting
faster influx from reservoir.
In conventional well-test analysis, we start from early-time and
progress to late-time, but in PITA late-time data is analyzed first.
The working equation for late-time analysis is(14, 15):
w = i
.......................................... (30)
w = w0 +
kh i w0 ta
( )(
24 1.842 103 Vw s
period compared to the shut-in time (instantaneous source solution). Furthermore, unlike the Soliman et al.(10) method, the
impulse method cannot determine the proper flow regime and
assumes radial flow. Therefore, the former method seems to
be more reliable.
7. Convolution/deconvolution of wellbore-storage-dominated
data by simultaneously measuring and analyzing the pressure
and rate data offers a solution to the problem of long duration
pre-frac testing in tight formations. However, deconvolution
methods become unstable when there is severe rate change.
8. Rate convolution analysis reduces the entire data set to a
single linear trend by discretizing variable flow rates. However, in its derivation, radial flow was assumed, which seems
not to be as effective as linear flow for tight gas reservoirs.
........................................................ (31)
Acknowledgements
Parts of this work were funded by the Natural Sciences and Engineering Research Council of Canada (NSERC agreement 34782506), ConocoPhillips (agreement 4204638) and the Alberta Energy
Research Institute (AERI agreement 1711). Their contributions are
gratefully acknowledged. The authors express their gratitude to the
Department of Chemical & Petroleum Engineering of University
of Calgary for support during the course of this work.
Nomenclature
= area perpendicular to flow, m2
= gas formation volume factor, m3/ m3
= dimensionless fracture conductivity
= total compressibility, kPa1
= gas compressibility in wellbore, kPa1
= factor defined by Equation (18)
= dimensionless wellbore storage as a function of xf
(rather than rw)
FL
= linear-flow time function
FR
= radial-flow time function
h
= net pay thickness, m
IDER = impulse derivative, (kPa)3.hr/(Pa.s)2
k
= permeability, mD
m
= slope defined by Equation (25)
mR
= slope of pseudo-radial ACA
p
= pressure, kPa
pD
= dimensionless pressure
pDCR = constant rate solution (unit rate response)
pi
= initial pressure, kPa
pwD = variable rate solution
q
= flow rate, m3/d
qD
= dimensionless flow rate
qD,1 = dimensionless initial flow rate
rw
= wellbore radius, m
s
= skin factor
t
= time, hr
ta
= pseudo-time, hr.kPa/Pa.s
tc
= fracture closure time, hr
tp
= producing time, hr
twbs = duration of wellbore storage distortion, hr
t
= shut-in time, hr
T
= temperature, K
V i
= injected volume, m3
Vw
= wellbore (chamber) volume, m3
xf
= fracture half length, m
Z
= gas deviation factor
= interporosity flow coefficient for dual porosity
reservoir
0
= defined by Equation (27)
j+1 = weighting parameter
f
= interporosity flow coefficient for hydraulically fractured well in dual porosity reservoir
= viscosity, Pa.s
= pi number
Ac
Bg
cfd
ct
cw
C
CD
69
D
DCR
i
ni
nw
w
wD
wf
w0
=
=
=
=
=
=
=
=
=
=
=
=
integration variable
porosity, fraction
defined in Equation (4)
pseudo-pressure, (kPa)2/Pa.s
constant rate solution (unit rate response)
initial pseudo-pressure, (kPa)2/Pa.s
initial normalized pseudo-pressure, kPa
wellbore normalized pseudo-pressure, kPa
wellbore pseudo-pressure, (kPa)2/Pa.s
variable rate solution
pseudo-pressure at flowing BHP, (kPa)2/Pa.s
initial cushion pseudo-pressure or wellbore pseudopressure at ta= 0, (kPa)2/Pa.s
1.589873 E01 = m3
1.0 E03 = Pa.s
3.048 E01 = m
2.831685 E02 = m3
1.450377 E04 = Pa1
6.894757 E+00 = kPa
9.87 E+14 = m2
References
1. Lee, W.J. 1987. Pressure-Transient Test Design in Tight Gas Formations. J. Pet Tech 39 (10): 11851195. SPE-17088-PA. doi:
10.2118/17088-PA.
2. Shahamat, M.S. and Aguilera, R. 2008. Pressure-Transient Test Design in Dual-Porosity Tight Gas Formations. Paper SPE 115001 at the
CIPC/SPE Gas Technology Symposium 2008 Joint Conference, Calgary, 1619 June. doi: 10.2118/115001-MS.
3. Nolte, K.G. 1979. Determination of Fracture Parameters from Fracturing Pressure Decline. Paper SPE 8341 presented at the SPE Annual
Technical Conference and Exhibition, Las Vegas, Nevada, USA,
2326 September. doi: 10.2118/8341-MS.
4. Mayerhofer, M.J. and Economides, M.J. 1993. Permeability Estimation From Fracture Calibration Treatments. Paper SPE 26039
presented at the SPE Western Regional Meeting, Anchorage, 2628
May. doi: 10.2118/26039-MS.
5. Mayerhofer, M.J. and Economides, M.J. 1996. Field Cases for Permeability Determination from Minifracs. SPE Advanced Technology
Series 4 (1): 111117. SPE-26999-PA. doi: 10.2118/26999-PA.
6. Gu, H., Eibel, J.L., Nolte, K.G., Cheng, A.H-D., and Abousleiman,
Y. 1993. Formation Permeability Determination Using ImpulseFracture Injection. Paper SPE 25425 presented at the Production
Operations Symposium, Oklahoma City, Oklahoma, USA, 2123
March. doi: 10.2118/25425-MS.
7. Abousleiman, Y., Cheng, A.H-D., and Gu, H. 1994. Formation Permeability Determination by Micro or Mini-Hydraulic Fracturing.
Journal of Energy Resources Technology 116 (2): 104114.
8. Nolte, K.G., Maniere, J.L., and Owens, K.A. 1997. After-Closure
Analysis of Fracture Calibration Tests. Paper SPE 38676 presented
at SPE Annual Technical Conference and Exhibition, San Antonio,
Texas, USA, 58 October. doi: 10.2118/38676-MS.
9. Talley, G.R., Swindell, T.M., Waters, G.A., and Nolte, K.G. 1999.
Field Application of After-Closure Analysis of Fracture Calibration
Tests. Paper SPE 52220 presented at the SPE Mid-Continent Operations Symposium, Oklahoma City, Oklahoma, USA, 2831 March.
doi: 10.2118/52220-MS.
10. Soliman, M.Y., Craig, D., Bartko, K., Rahim, Z., and Adams, D.
2005. Post-Closure Analysis to Determine Formation Permeability,
Reservoir Pressure, Residual Fracture Properties. Paper SPE 93419
presented at the SPE Middle East Oil and Gas Show and Conference,
Bahrain, 1215 March. doi: 10.2118/93419-MS.
11. van Everdingen, A.F. and Hurst, W. 1949. The Application of the
Laplace Transformation to Flow Problems in Reservoirs. Trans.,
AIME, 186: 305324.
12. Raghavan, R. 1993. Well Test Analysis, 268335. Englewood Cliffs,
New Jersey: Prentice-Hall.
13. Helmy, M.W. 1999. Analysis of Well Performance with Multiple
Shut-in Periods. PhD dissertation, Texas A&M University, College
Station, Texas.
14. Rahman, N.M.A., Pooladi-Darvish, M., and Mattar, L. 2005. Development of Equations and Procedure for Perforation Inflow Test
Analysis (PITA). Paper SPE 95510 presented at the SPE Annual
70
Authors Biographies
Ashkan Jahanbani was awarded the university prize for being ranked as the top student of the B.Sc. and M.Sc. programs from
the Petroleum University of Technology.
He was the leader of the student scientific
association at the Petroleum University of
Technology, Ahwaz (2005 2006). His interests include fluid flow through porous
media, well-test analysis, naturally fractured reservoirs and tight gas sands. He
holds a B.Sc. degree in petroleum reservoir
engineering from the Petroleum University of Technology, Iran,
and M.Sc. and M.Eng. degrees from the Petroleum University of
Technology and the University of Calgary in a dual-degree program. He is also a member of SPE.
Roberto Aguilera is Professor and ConocoPhillips-NSERC-AERI Chair in the
Schulich School of Engineering at the University of Calgary, Chairman of the Journal
of Canadian Petroleum Technology (JCPT)
Editorial Review Board, a principal of
Servipetrol Ltd., and a director of Junex
in Quebec. He is a petroleum engineering
graduate from the Universidad de America
at Bogota, Colombia, and holds Masters
and Ph.D. degrees in petroleum engineering
from the Colorado School of Mines. He is a Distinguished Author
of JCPT (1993 and 1999), a recipient of the Outstanding Service
Award (1994) and the Distinguished Service Medal (2006) from
the Petroleum Society of Canada, and a Society of Petroleum Engineers (SPE) Distinguished Lecturer on the subject of naturally
fractured reservoirs for the 2000 2001 season.
Journal of Canadian Petroleum Technology