SPE 77949

Determination of Fracture Orientation

by Multi-Well Interference Testing
*
*

Mohamed Cherifi/Sonatrach, Inc., Djebbar Tiab /University of Oklahoma, and Freddy H. Escobar /Universidad
Surcolombiana

* SPE MEMBERS
Copyright 2002, Society of Petroleum Engineers Inc.
This paper was prepared for presentation at the SPE Asia Pacific Oil and Gas Conference and
Exhibition held in Melbourne, Australia 8-10 October 2002.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes 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 where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract
It is often desirable to know patterns created after a number of
wells in reservoir have been fractured. Interference testing is
one of the best techniques that can be used for estimating
some of the important formation properties (transmissibility
and storativity) and fracture orientation.
Generalized correlations are presented relating (a) the
dimensionless pressure response, PD,inter, at the intersection
point of the pressure and the pressure derivative to the
compass orientation, , and (b) the quotient of the
dimensionless pressure and the ratio of dimensionless time to
the square of the dimensionless radial distance, (tD/rD2)inter, at
the intersection point to the compass orientation.
The correlations can be used for analyzing interference test
pressure response at an unfractured shut-in well caused by a
production of a vertically fractured well with a uniform flux,
infinite or finite conductivity fracture provided that the half
fracture length is known a priori.
Detailed procedures for the analysis of interference testing
of a vertically fractured well are presented and demostrated
through a simulated example.
Introduction
The data required to select the better reservoir management
method include inter-well properties, the degree of
communication between different wells, and information
about the reservoir heterogeneity. Pressure transient testing
techniques, such as pressure drawdown, build-up, injectivity,
falloff, and interference are some of the most frequently used
methods in reservoir and production engineering for obtaining
these data. Practical information obtainable from these

techniques includes amount of wellbore damage or

stimulation, formation permeability, porosity, reservoir, and
fluid discontinuities, and other related data.
In certain situations, pressure transient testing is
indispensable for reservoir analysis; for example, in definition
of near-wellbore and inter-well conditions. In other cases,
even with the most complex and through transient analysis, a
unique solution often is not possible without considering other
information such as production logs, stimulation history, core
descriptions. Interference test and pulse tests are best suited
for obtaining inter-well properties.
Interference comes from the fact that the pressure drop
caused by the producing well or wells at the shut-in
observation well interferes with the pressure at the observation
well or wells. Therefore, interference test is a multiple-welltest, which requires at least one active well and at least one
observation well. The active well is either a producer or
injector and the observation wells are shut-in wells in which
pressure effects caused by the active well are measured.
Multiple well tests are more sensitive to reservoir
heterogeneity than single well tests and give information on
reservoir properties than cannot be obtained from ordinary
well tests, for instance, the orientation of the hydraulically
vertical fractures, the orientation of the maximum and
minimum permeabilities.
Tongpenyai and Raghavan1 have presented solutions that
can be used to analyze interference test data influenced by
wellbore storage and skin effect at both the active and the
observation well. Several studies have considered the effect of
vertical fractures on transient pressure behavior of wells2. It
was recognized that single well tests and interference testing
could be used to determine properties of vertical fractures such
as fracture length and fracture orientation respectively.
In 1973, Gringarten and Ramey3 prepared tables of
instantaneous Greens and source functions, which can be
used with the Newmans product method to generate solutions
for a wide variety of reservoir flow problems. This approach
has been applied in the efforts to formulate analytical solutions
to flow equations in fractured reservoirs.
Gringarten, Ramey and Raghavan4 used this method to
derive analytical expressions for unsteady-state pressure
distribution created in a reservoir by vertically fractured well

M. CHERIFI, D. TIAB AND F.H. ESCOBAR

with an infinite conductivity fracture. In addition to infinite

conductivity case, Gringarten et al4 also provided another
solution called the uniform flux solution. The uniform flux
solution gives the appearance of high, but not infinite,
conductivity fracture. Thus, unlike infinite conductivity
boundary conditions, the pressure along the fracture plane
varies.
The analytical equations given in Ref. 4 contributed
significantly to the analysis of pressure response data for the
more complicated well test techniques such as interference
tests and pulse tests. This enabled Uriaet, Raghavan and
Thomas5 to apply uniform flux solution to the analysis of
interference test of a vertically fractured well. They showed
how the solution might be used to analyze pressure data in
adjacent observation wells to determine fracture orientation in
a manner analogous to standard interference tests.
Cinco-Ley et al9 developed finite-conductivity fracture
solution. They found that a match might not be obtained when
pressure interference data are matched with the uniform flux
type curves of Uriaet et al5, if finite-conductivity fractures
exist.
Tiab and Abolise7 presented a procedure for determining a
complex orientation of a uniform flux fracture and the
formation permeability from pulse testing. Generalized
correlations relating the quotient of dimensionless response
amplitude and dimensionless cycle period to dimensionless
time lag are presented. The correlations can be used for
analyzing pulse test response at an unfractured observation
well resulting from pulsing a vertically fractured active well.
In 1989, Meeha et al8 presented techniques for the design
and analysis of interference tests when both the active and
observation wells are intercepted by hydraulic fractures. These
techniques are based on a new mathematical solution. The
solution is general for any values of dimensionless fracture
conductivity including infinite conductivity. They generated
type curves of pressure and pressure derivative by using the
Laplace transform.
Theoretical Backgroud
The vertically fractured well models discussed in this paper
are the uniform flux and the infinite conductivity models
presented by Gringarten et al4 and the finite conductivity
model presented by Cinco-Ley et al9.
Uniform flux model
The basic assumptions made for this model are identical to
those proposed by Gringarten et al4.
Pressure Distribution
An analytical expression for the pressure distribution created
by uniform flux fracture was derived by Gringarten et al4 as
shown here:
2
(1 x D )
(1 + x D ) d (1)
yD
t
) erf
PD ( x D , y D , t D ) =
+ erf
exp(

The pressure drop in the plane is derived by setting yD = 0 in

Eq. (1):

PD ( xD, o, tD ) =

SPE 77949

(1 x )
(1 + xD ) (1 xD ) (1 xD )2
D
Ei
+ erf
erf

4
4tD
2 tD
2 tD

(1 + xD ) (1 + xD )2
Ei

4
4tD

tD
2

(2)

The dimensionless pressure derivative expression is as

follows:
2
.t D
(1 + x D ) (3)
(1 x D )
yD
t D * PD ' =

exp(

4t D

) erf (
) + erf (
)
2 tD
2 t D

kh
( Pi P( x, y, t )) (4)
141.2qB
0 .0002637 kt
(5)
=
c t x 2f

Where: PD ( x D , y D , t D ) =
tD

xD =

x
xf

(6)

yD =

y
xf

(7)

The integral of the dimensionless pressure drop PD created by

uniform flux fracture presented by Eq. (1) was evaluated
numerically. The main problem in evaluating this integral
numerically is to determine a value for tD such that PD is
sufficiently small so that it can be neglected, and then replace
the lower limit of the integral by this value of tD. In doing so,
the PD for tD=0.01 is in the order of 10-2 or less which could be
ignored. The dimensionless pressure values are obtained by
the trapezoidal rule:
f (t D ) =

y D2
1
exp
4t
4
D

erf

1 x
D

2 t
D

+ erf 1 + x D

2 t
D

tD

(8)

m1
m1
h

PD (xD, yD, tD ) = f (a) + f (tD ) + 2 f (t2n ) + 4 f (t2n1)

(9)
3
n=1
n=1

t a
n tD
, tn =
Where: a=0.001, h = D
2m
2m
2 m is the number of tD segments and n is the increment of
time.

Infinite conductivity model

The assumptions made for this model are the same as that
made by Gringarten et al4.
Pressure Distribution:
The dimensionless pressure drop was given by Gringarten et
al4 as follows:
j

xD +
tD

y D2 n
n)
) q Dj ( ) erf (
PD ( x D , y D , t D ) = exp(
4 j =1

2
0

SPE 77949

DETERMINATION OF FRACTURE ORIENTATION BY MULTI-WELL INTERFERENCE TESTING

j 1
j
j 1
xD
xD
n ) erf (
n ) + erf (
n ) .d
erf (

2
2
2
4

(10)
2q j (t D )hx f
Where: q Dj (t D ) =
q
The dimensionless pressure derivative is as follows:
j
j 1

xD +
xD +

y D2 n
n
n )
t D * PD ' = exp( ) qDj ( ) erf (
) erf (
4 j =1

2
2

j
j 1
xD
xD
n ) + erf (
n ) .d
erf (
(11)

2
2
4

The infinite conductivity solution is obtained from the

uniform flux solution by dividing the half-fracture length into
30 segments each with uniform flux per unit area, and then
adjusting the unitary flow-rate in each segment so that the
resulting pressure distribution would be approximately
uniform along the fracture. In the following, we give the
procedure used to determine the unitary flow-rate distribution
at each time-level.
Step 1: The transient period is divided into m different
intervals and the dimensionless half-fracture length is divided
into n segments. Each of the segments produces a
dimensionless flow-rate equal qm per unit area.
Assuming that the flux entering each segment is uniform
and has a stepwise distribution in time and space: i.e. qm is
constant for segment, j, at time-level, m, and independent of
production history.
Step 2: Then, we solve the following constructed system of n
equations:
PD (m,2)q D (m,2) PD (m,1)q D (m,1) = 0
P (m,3)q (m,3) P (m,2)q (m,2) = 0
D
D
D
D
..................................................

(12)
PD (m, j )q D (m, j ) PD (m, j 1)q D (m, j 1) = 0
..................................................

PD (m, n)q D (m, n) PD (m, n 1)q D (m, n 1) = 0

q1 + q2 + q3 + ..................... + qn
xD +

Where the dimensionless pressure, PD, for each segment is

given as:

t D ( m )

+
erf (1 x D ( j )) + erf (1 x D ( j ))
PD ( m, j ) =

2
2 t D ( m)
2 t D (m )

2
(1 x D ( j )) (1 x D ( j )) (1 + x D ( j )) (1 + x D ( j )) 2

Ei
Ei
4
4t D (m )
4
4t D ( m)

(13)

Step 3: Solving the previous system of equation (12), the flux

distribution for each segment at each time-level is obtained
(fig. 1). Knowing the flux distribution, the dimensionless
pressure and the dimensionless pressure derivative are
calculated at any point in the reservoir.

Finite conductivity model

The same assumptions made by Cinco-Ley et al9 are kept for
this model.
Dimensionless pressure distribution
The dimensionless pressure drop at any point in the reservoir
may be obtained by the following equation proposed by
Cinco-Ley et al9:
2
( x x)2 + yD
( D
)
4
e

tD 1

1
dxd
qD (x, )
(14)

4 0 1
The dimensionless pressure drop in the fracture is given by
the following equation6:
PD (xD , yD , t D ) =

tD
e

PFD ( x D , t D ) =

1

fD n = 1 0

fDf

2 n +1

q fD ( x , )

(xD 2n)2
4 fD

(xD x)
4 fD

2 n 1

(15)

dx d

The dimensionless pressure derivative equation is derived

from the equation governing the pressure within the reservoir
(eq. 14):

1
t D * PD ' =
4
Where:

( x, t D )e

( x D x ) 2 + y D2
)
4tD

dx

(16)

q D ( x, ) =
q fD ( x, ) =

2 q ( x, ) x f
q
2q f ( x , ) x f

q
kh
PfD ( x D , t D ) =
( Pi Pf ( x, t ))
141.2qB
w f c ft
C fDf =
x f ct

fD =

k f ct
k f c ft

(17)
(18)
(19)
(20)
(21)

The finite conductivity solution is obtained by considering

two flow regions: (1) the fracture, and (2) the reservoir.
Fracture flow model
The fracture has the those characteristics as stated by CincoLey et al9. Fluid enters the fracture at a rate q(x,) per unit of a
fracture half-length, and the flow across the edges of this
porous medium is negligible. As a result by applying these
assumptions, Cinco L. et al9 developed an equation governing
the pressure within the fracture (eq. 15).

M. CHERIFI, D. TIAB AND F.H. ESCOBAR

Reservoir flow model

The transient pressure behavior in the reservoir may be studied
by considering the fracture as a plane source of flux density
qf(x,t). The dimensionless pressure drop at any point in the
reservoir can be obtained by equation (14).
To solve Eqs. (14) and (15) simultaneously, continuity
between the two flow regions must be established. Thus:
PfD ( xD , t D ) = PD ( xD , y D = 0, t D )
(22)
And q fD ( xD , t D ) = q D ( x D , t D )

(23)

For 1 x D 1 and tD > 0.

After combining equations 14, 15, 22, and 23 using the
Poissons summation formula, and discretizing in space and
time so that the fracture is divided into 2N equal segments and
time is divided into K different intervals, one could end up
with the following system of equations:
N k 1

qD (1, k)z(1,1, k) + qD (2, k)z(2,1, k) + ...+ qD(N, k)z(N,1, k) = g(1) qD (i, l)z(i,1, l)
i =1 l =1

N k 1
qD (1, k)z(1,2, k) + qD (2, k)z(2,2, k) + ...+ qD (N, k)z(N,2, k) = g(2) qD(i, l)z(i,2, l)

i =1 l =1
.....................................................................................................................

N k 1
qD (1, k)z(1, j, k) + qD (2, k)z(2, j, k) + ...+ qD (N, k)z(N, j, k) = g( j) qD (i, l)z(i, j, l)

i =1 l =1
.....................................................................................................................
N k 1

qD (1, k)z(1, N, k) + qD (2, k)z(2, N, k) + ...+ qD (N, k)z(N, N, k) = g(N) qD (i, l)z(i, N, l)
i =1 l =1

(24)

SPE 77949

dimensionless fracture conductivity which are 10, 2, 0.5, to

generate the new correlations.
Using a computer program, the flux distribution graphs for
each fracture conductivity (10, 2, 0.5) are shown in figures 3,
4, and 5, respectively.
New Correlations
The equations in uniform flux model, infinite conductivity
model, and finite conductivity model sections provide the
theoretical basic for calculating the ratio of dimensionless time
to the sequare of dimensionless radial distance, (tD/rD2)inter, and
dimensionless pressure response at the intersection point
between the dimensionless pressure curve and dimensionless
derivative curve due to interference testing of a vertical
fracture well. Subsequently, there are new means of obtaining
PD,inter versus correlations and PD,inter/(tD/rD2)inter versus
correlations. A description of the computer program used in
generating these new correlations will be presented in this
section.
It was necessary to perform all computations in
dimensionless variable groups in order to reduce the number
of correlations to be generated. The following dimensionless
groups are defined in order to facilitate the discussion.
Dimensionless radial distance, rD =

r
=
xf

x+ y
xf

(25)

Where the g(j) and z(i,j,l) are defined in Appendix A.

A computer program was written to determine the flux
distribution along the fracture at each step time level where
the qD(i,l=1) equal zero at the first step time level, and also
dimensionless pressure drop, dimensionless pressure
derivatives are evaluated at any point in the reservoir. A
sensitivity analysis was conducted to obtain accurate results.
We found that solutions do not change appreciably when more
than 20 segments are taken per fracture half-length, xf, and 10
intervals in each log cycle of dimensionless time are taken.
Therefore, in all cases studied here, the fracture half-length
was divided into 20 equal segments and 10-time intervals were
taken in each log cycle of dimensionless time.

Hence PD(xD,yD,tD) and PD(xD,yD,tD) can be expressed as

PD(rD,,tD) and PD(rD, ,tD) respectively.
A flow chart for the computation is shown in appendix A
(fig. 6). For a given dimensionless radial distance, rD, fracture
orientation, , and using Table Curve Software to obtain
fitting equations for portions of dimensionless pressure and
dimensionless pressure derivative responses, we can find the
dimensionless (tD/rD2)inter and dimensionless pressure, PD,inter at
the intersection point such that PD t D PD' 0.0001 .

The flux distribution

The study of the combined effects of CfDf and fD on the
pressure behavior was fist examined by Cinco-Ley et al9. They
found that as long as most of the fluid produced at the
wellbore comes from the formation and the expansion of the
fracture system is negligible, solutions could be correlated by
one parameter. This correlating parameter was found to be
CfDffD = kf w/( k xf) and they call it the dimensionless
fracture conductivity.
The solutions obtained in this study were compared with
solutions published in the literature. Results for a highly
conductive fracture (CfDf=10-3, fD=107) show excellent
agreement with the infinite conductivity solution of Gringarten
et al4 (Fig. 2).
For the sake of simplicity, and based on the published
papers6,8,9, we have chosen three distinct values of the

Discussion
We generate for each conductivity model the PD,inter versus
correlations, the PD,inter/(tD/rD2)inter versus correlations, and
the PD,inter versus (tD/rD2)inter correlations.
Investigating the effect of the fracture conductivity models
on the fracture orientation, we conclude that there is a range
where we can apply the uniform flux model most of the times.
Whereas in the same correlations there is a range where one
could read different fracture orientation values depending on
whether we use the correlations that based on the infinite
conductivity, finite conductivity or the uniform flux models.
Although the PD,inter versus correlations and the
PD,inter/(tD/rD2)inter versus correlations obtained from these
models are slightly analogous, in other words, they have the
same shape, the PD,inter and PD,inter/(tD/rD2)inter values are
different for the three models.
For a given fracture orientation, , and dimensionless
radial distance, rD, the PD,inter obtained by using the uniform

y
Fracture compass orientation, = tan 1
x

(26)

SPE 77949

DETERMINATION OF FRACTURE ORIENTATION BY MULTI-WELL INTERFERENCE TESTING

flux conductivity model is greater than that obtained by using

the infinite and finite conductivity models. Other than the
PD,inter/(tD/rD2)inter values from the uniform flux model is less
than that obtained by using the infinite and finite conductivity
models.
As a result, where we can apply the three models, the
fracture orientation from the two correlations (the PD,inter
versus correlations and the PD,inter/(tD/rD2)inter versus
correlations) obtained by using the uniform flux model is less
than that obtained by using the infinite and finite conductivity
models.
Procedure analysis
PD,inter versus correlations:
This correlation is used when we can calculate the
transmissibility from the horizontal derivative portion of the
curve.
1. Decide which model you want to use depending on the
fracture conductivity.
r
xf

2.

Calculate the dimensionless radial distance: rD =

3.

Calculate the transmissibility: kh = 70.6 qB'

4.

Calculate PD,inter by: PD ,int er =

5.

Using the calculated value of PD,inter, rD, and PD,inter versus

correlations, the compass orientation of the fracture can
be read.
Using PD,inter versus (tD/rD2)inter correlations, we can obtain
the formation storativity by:

6.

ct h =

0 . 0002637
r 2 ( t D / r D2 ) int er

(27)

(t P ) r

kh Pint er
141.2qB

(28)

kh

t int er

(29)

3.

Calculate the dimensionless radial distance: rD =

4.

Calculate thePD,inter/(tD/rD2)inter by using:

5.

6.

(31)

Where Pinter, tinter are the coordinates of the intersection point

of the pressure and pressure derivative curves in a loglog plot.
For the two new correlations, the fracture half-length should
be obtained from a post-test of the active well.
The PD,inter versus (tD/rD2)inter correlation is used in
conjunction with the PD,inter versus correlation to obtain the
formation storativity.
Simulated example
The following example is designed to illustrate how the
preceding procedures may be applied in the analysis of
interference tests of vertically fractured wells. Assume that the
reservoir has the approximate properties shown in table 1.
Table 1. Reservoir and well data for simulated example
Reservoir and well data
0.26
Viscosity, , cp
0.08
Porosity, , fraction
Thickness, h, ft
30
Flow rate, q, STB/D
900
Formation volume factor ,Bo, RB/STB
1.6
System compressibility, Ct, psi-1
1.24x10-6
Distance between fractured well A and
400
responding well B, rB, ft
Distance between fractured well A and
responding well C, rc, ft
600
Fracture half-length, xf, ft
400
Dimensionless fracture conductivity,
10
CfDf*fD
Table 2. Pressure data

PD,inter/(tD/rD2)inter versus correlations:

This correlation provides a direct means of estimating the
fracture compass orientation, , even if we do not know the
formation permeability. This means that when the horizontal
portion of the derivative has not been established yet, the
correlation can be applied. But this correlation is applicable
when the storativity is known apriori. Therefore, the following
procedure is recommended:
1. Obtain the best estimate of the formation storativity.
2. Choose the adequate model.

PD ,int er ( t D / rD2 ) int er = 26 . 85685

PD ,int er
kh
= 141.2qB
Pint er

c t hr 2 Pint er
qB
t int er

r
xf

(30)
2

By using the calculated value of PD,inter/(tD/rD )inter and the

PD,inter/(tD/rD2)inter versus correlations, the fracture
orientation can be determined.
By using the PD,inter versus correlations and the obtained
compass orientation, the transmissibility can be estimated
from the following formula:

time (hours)
0.039079
0.078158
0.117236
0.156315
0.195394
0.234473
0.273552
0.31263
0.351709
0.390788
0.781576
1.172364
1.563152
1.953939
2.344727
2.735515
3.126303
3.517091
3.907879
7.815758
11.72364

Well B

0.000735
0.160182
0.819175
2.172343
4.153475
6.622738
9.447539
12.52367
15.77332
19.13933
53.04256
84.76397
112.7853
137.6511
159.9375
180.1048
198.5093
215.4277
231.0781
351.4479
428.1732

psi

time (hours)
0.117236
0.156315
0.195394
0.234473
0.273552
0.31263
0.351709
0.390788
0.781576
1.172364
1.563152
2.344727
2.735515
3.126303
3.907879
7.815758
11.72364
15.63152
23.44727
31.26303
39.07879
78.15758
156.3152
234.4727
312.6303
390.7879

Well C P
psi
0.036074
0.15104
0.396106
0.794578
1.349113
2.050899
2.886039
3.839172
16.60071
32.45876
48.56409
79.02098
93.11478
106.4592
131.0838
228.5671
296.4368
347.8578
423.884
479.7863
524.019
677.0512
829.34
918.0843
981.0795
1029.979

M. CHERIFI, D. TIAB AND F.H. ESCOBAR

The pertinent interference test data tabulated in table 2 will

now be used with the reservoir well data in analyzing the test
results. From the pressure and the pressure derivative plots
(figures 7, 8) we obtain:
Well B: tinter = 1.367 hrs, Pinter= 97.7 psi.
Well C: (t*P)r= 220 psi, tinter = 2.93 hrs, Pinter= 100 psi.
Analysis of well B data:
1. Assuming porosity 0.08, thickness 30 ft, and total
compressibility 1.24 10-6 psi-1. rD=400/400=1.
2. By using Eq. 30 , we have:
PD,inter

3.
4.

(t

rD2

int er

5.

0.081.2410 30 400 97.7

= 26.85685
= 0.6347
9001.6
1.367

By using PD,inter/(tD/rD2)inter=0.6347 and PD,inter/(tD/rD2)inter

versus correlations (fig. 9 ), the orientation equals 50.
By using PD,inter versus correlations and = 50 (fig.
10), one could obtain PD,inter equal to 0.218, and by using
eq. 31, the transmissibility is estimated as follows:
0.218
kh
= 141.2 900 1.6
= 453.69
97.7

By using Eq. 27:

6.

md-ft/cp.
7.

Analysis of well C data:

1. rD=600/400=1.5
2.

4.

900 1.6
kh
= 70.6
= 462.109 md-ft/cp.
220

Therefore k = 462.109 0.26 / 30 = 4 md.

3.
4.
5.

By using eq. 28, we get: PD ,int er = 462.109

100
141.2 900 1.6

PD,inter=0.227
By using PD,inter=0.227 and PD,inter versus correlations
(fig. 11) the orientation ( rD=1.5, PD,inter=0.227) =30.
By using PD,inter versus (tD/rD2)inter correlations (fig.12), we
find a value of (tD/rD2)inter equal 0.336 corresponding to
PD,inter=0.227. By using Eq. 29, the storativity is:
ct h =

0 . 0002637
0 . 336 600

462 . 109 2 . 93

ft-psi-1, ct=1.229 10-6 psi-1.

The transmissibility and the storativity obtained from
analysis of the well pair A B and A C are approximately
the same. Therefore, the formation transmissibility and
storativity can be assumed uniform for all practical purposes.
The orientation of the fracture, which results in one
consistent direction, is the true compass orientation as shown
in figure 13.
c t h = 2 . 951 10

Conclusions
1. The interference test of a vertically fractured well may be
analyzed to determine the compass orientation of the
fracture, the formation permeability and the storativity of
the reservoir zone influenced by the test.
2. The uniform flux, infinite, and finite conductivities of
hydraulic fractures significantly affect interference test
response. Most significant results are for rD 2.
3. At late times, all derivatives reach a value of 0.5
indicating the presence of a pseudo-radial flow regime for

8.

9.

SPE 77949

any value of dimensionless radial distance. This

characteristic allows us to calculate the transmissibility, as
given by Eq. 27.
For a line source vertical well, Ouandlous10 showed that a
value of PD,inter equal to 0.3237 is reached at an (tD/rD2)inter
equal to 0.5749. It is shown in this study that for
vertically fractured wells, this PD,inter value and the
(tD/rD2)inter at which it occurs both depend on the
observation well location. But for large values of rD, these
values are converging to those obtained by Ouandlous10.
Direct methods for analyzing the vertically fractured well
interference test data are developed based on PD,inter,
(tD/rD2)inter and the observation well location. These
methods lead to more accurate results than the tedious
type curve matching technique.
The PD,inter versus correlations gives a more accurate
estimate of the fracture orientation when the
dimensionless radial distance, rD, is greater than 0.5. For
rD ranging from 1 to 2 the observation well is less affected
by fracture orientation when the range of the fracture
orientation, , is 70 to 90.
The PD,inter/(tD/rD2)inter versus correlations gives a more
accurate estimation of the fracture orientation when the
angle between the fracture and the line joining the active
and the responding wells is less than approximately 50.
An important advantage of these correlations is that it
provides a direct means of estimating fracture orientation
without knowing the formation transmissibility.
Investigating the effect of the conductivity fracture
models, we conclude that the fracture orientation from the
tow correlations obtained by using the uniform flux model
is less than that obtained by using the infinite and finite
conductivity models.
For a given dimensionless radial distance, if the
calculated value of PD,inter or PD,inter/(tD/rD2)inter falls outside
the range of tabulated values, then it could be that either
the observation well is in the fracture plane, or the model
of the fracture is not applicable, depending on whether the
calculated value is greater or less than the tabulated
values.

Acknowledgement
We thank Sonatrach for providing financial support for The
University of Oklahoma Graduate Program in Petroleum
Engineering in Algeria. Special thanks to the technical staff
of Sonatrach-PED, CRD, and Exploration for their valuable
suggestions and input. Thanks are extended to IAP and INH
for permission to use their facilities and to Sonatrach-DRH for
assisting with the management of the project. We also extend
our thanks to Sonatrach for permission to publish this research
work.
Nomenclature
Symbols
Bo
CfDf

= Formation volume factor, RB/STB

= Dimensionless fracture storage capacity

SPE 77949

DETERMINATION OF FRACTURE ORIENTATION BY MULTI-WELL INTERFERENCE TESTING

= Dimensionless fracture conductivity

= Total system compressibility, psi-1
= Formation thickness, ft
= Effective formation permeability, md
= Pressure, psi
= Dimensionless pressure drop
= Initial pressure, psi
= Pressure change, psi
= Total flow rate, STB/D
= Radial distance between active and
responding wells, ft
= Dimensionless radial distance, based on
the fracture half-length, xf
= Time, hours
= Dimensionless time based on the
fracture half-length
= Coordinates in well system, ft
= Time rate of change of pressure psi/hr
= Dimensionless pressure derivative
= dimensionless segment s length

CfDf*fD
ct
h
k
P
PD(xD,yD,tD)
Pi
P
q
r
rD
t
tD
x, y
P
tD*PD
x
Special functions:
Ei( x ) =

erf ( x) =

fD
Subscripts
D
f
i
inter

u
e du
0

Greek symbols

6.

Abobise, E.O: Analysis of pulse tests of vertically

fractured wells, Master thesis in 1981, the University of
Oklahoma.
7. Tiab, D., and Abolise, E.O.: Determination of fracture
orientation from pulse testing, Paper SPE 11027,
presented at the 57th Annual Meeting of Society of
Petroleum Engineers, held in New Orleans, LA. Sep 1982
8. Meehan, D. N., Horne R. N., and Ramey, H., Jr.:
Interference testing of finite conductivity hydraulically
fractured wells, Paper SPE 19784, Oct. 1989.
9. Cinco-Ley, H., Samaniego, V. F. and Rodriguez, F.:
Determination of the orientation of a finite conductivity
vertical fracture by transient pressure analysis, Paper
SPE6750, presented at the 52nd Annual fall Technical
conference and Exhibition of the Society of Petroleum
Engineers of AIME, held in Denver, Colorado, Oct 9-12,
1977.
10. Ouandlous, A.: Interpretation of interference tests by
Tiab's Direct Synthesis Technique, M.S. Thesis. The
University of Oklahoma. 1999.
Appendix A
Equations used in the finite conductivity model
The equations that are used in the finite conductivity system:

e u

u du
x
2

= Viscosity, cp
= Porosity, fraction
= Compass orientation of the fracture,
degree
= Dimensionless hydraulic diffusivity of
the fracture
= Dimensionless
= Fracture
= Initial
= Intersection point

References
1. Tongpenyai, Y., and Raghavan, R.: The effect of
wellbore storage and skin on interference test data, JPT
(Jan 1981) 151-160, SPE 8863.
2. Russel, D. G., and Truitt, N. E.: Transient pressure
behavior in vertically fractured reservoirs, JPT (Oct
1964) 1159-1169.
3. Gringarten, A. C., and Ramey, H. J, Jr.: The use of point
source solution and Green's functions for solving flow
problems in reservoirs,. SPEJ (Oct. 1973) 285-296.
4. Gringarten, A. C, Ramey, H., Jr., and Raghavan, R.:
Unsteady-state pressure distributions created by a well
with a single infinite-conductivity vertical fracture,
Paper SPE 4051. August 1974.
5. Uraiet A., Raghavan R., and Thomas G. W.:
Determination of the orientation of a vertical fracture by
interference tests, Paper SPEc5845, January 1977.

1
C fDf

g ( j) =

Where: x Dj
z (i , j , l ) =

2
t D ( k ) + 2

fD
j 0.5
=
N

1
1 exp( ( n ) 2 fD t D ( k ) cos( n x Dj )
2
n
n =1

4
x

v (i, j , l )
m (i, j , l ) +
t l ,l 1 + 3
4
C fDf
fD C fDf

Where: t l ,l 1 = t D (l ) t D (l 1) , x =
v (i , j , l ) =

n =1

1
exp( ( n ) 2
n3

fD

xf
N

t k , l ) exp( ( n ) 2

fD

t k ,l 1 )

n
cos( n x Dj ) cos( n x Di ) sin

2N

m(i, j , l ) = X ik, ,jl 1 X ik, ,jl + Yi ,kj,l 1 Yi ,kj,l

Where:

i, j

X ik, ,jl = 2 t k ,l erf

t k ,l

Yi,kj,l =

erf i , j

t
k ,l

+ erf i , j

t
k ,l

erf i , j

t
k ,l

i2, j
i2, j
i2, j
i2,j
2
i, j Ei
+ i , j Ei
i , j Ei

i, j Ei
t k ,l
t k ,l
t k ,l
t k ,l

The arguments of the erf and Ei functions are defined as:

i, j =

j i + 0.5
2N

, i, j =

j i 0.5
2N

, i, j =

j + i 0.5
2N

, i, j =

j + i 1.5
2N

M. CHERIFI, D. TIAB AND F.H. ESCOBAR

SPE 77949

Appendix B:

Flux distribution for CfDf*fD =0.5

1.6
1.5
1.4
1.3
1.2
1.1
1.0
0.9
0.8

3
2.5

tD =0.0001

tD =0.001

tD =0.011

tD >0.115

qD(x D,t D)

2qmhx f/qf

flux distribution of infinite conductivity

tD =0.001

tD =0.01

tD =0.1

tD =1

1.5
1
0.5
0
0

0.0

0.2

0.4

x D 0.6

0.8

0.2

0.4 x 0.6
D

0.8

Fig. 5

1.0

Fig. 1

Input : rD, ,

Start

fluxdistribution CfDf=1E-3, fD =1E+7

3

qD

2.5
2

This study
tD > 5

Do the fitting of the

pressure values.

Import pressure values

from Excel

Do the fitting of
the pressure
derivative values.

Import pressure
derivative values from
Excel

G ringarten etal(infinite
conductivity)

1.5
1
0.5
0
0

0.2

0.4

xD

0.6

0.8

tD=tD+step

Fig. 2

Yes

No
PD t D PD' 0.0001

Flux distribution for CfDf*fd =10

3

qD(x D,t D)

2.5
2

tD =1E-5

tD =1E-3

tD =0.1

tD =1

Send:
(tD/rD2)inter,
PD,inter to Excel

1.5

stop

1
0.5

Figure 6: flow chart of the computer program

0
0

0.2

0.4

x D 0.6

0.8

1
1000

Fig. 3

P, t*P' (psi)

Flux distribution for CfDf*fD =2

3

qD(x D,t D)

Pressure and pressure derivative of well B

2.5

tD =7E-5

tD =1E-3

tD =1E-1

tD =1

100
Pressure curve
Derivative curve

1.5

10

0.1

0.5

t(hours)

Fig. 7

0
0

0.2

0.4 x 0.6
D

Fig. 4

0.8

10

100

SPE 77949

DETERMINATION OF FRACTURE ORIENTATION BY MULTI-WELL INTERFERENCE TESTING

1000

100

P D,inter = 0.227

0.18

(t D/rD2 )inter = 0.336

0.08

10
0.1

1 t(hours) 10

100

rD= 1.2

0.13

Pressure curve
Derivative curve

1000

0.1

0.2
0.3
(t D/rD2)inter

(P D/(t D/rD2 ))inter

0.5

5
4

(P D/(t D/rD2 ))inter versus

correlations
rD = 0.9 till 1.1, CfDf*h fD=10

rD=0.9

=50

Fracture plane

Well B

Well A

2
1

Well C

0
0

10

20

30

40
50
(degree)

60

70

80

=30
Fig. 13

90

Appendix C
Uniform flux model:

Fig. 9
0.3
P D,inter=0.218

0.25

P D,inter versus correlations

0.35

rD = 1

rD=0.1 till 0.6

0.3

0.2

0.25

rD =0.7

0.15
0.1

P D,inter versus correlations

0.05

rD = 0.7 till 1, CfDf* fD=10

P D,inter

P D,inter

0.4

Fig. 12

Fig. 8

rD=0.6

0.2
0.15
0.1
0.05

0
0

20

40

60
(degree)

rD=0.1

80

20

Fig. 10
0.75

r D= 2

0.29

P D,inter /(t D/rD2)inter

0.27
rD=1.5

0.25
0.23
0.21

P D,inter versus correlations

0.19

rD = 1.5 till 2, CfDf* fD=10

0.17
0.15
0

10

20

30

40

50 60
(degree)

Fig. 11

40 degree 60
(
)

80

100

Fig. 14

0.31

P D,inter

rD= 1.5

rD=1.2 till 1.5, CfDf* fD=10

0.23
P D,inter

P, t*P' (psi)

P D,inter versus (t D/rD2 )inter correlations

0.28

Pressure and pressure derivative of well C

70

80

90

rD= 1.7

0.7

P D,inter/(t D/rD2 )inter versus

correlations, r D=1.7 till 2

0.65
0.6

rD= 2

0.55
0.5
0.45
0

20

40
60
(degree)

Fig. 15

80

100

10

M. CHERIFI, D. TIAB AND F.H. ESCOBAR

Finite conductivity model:

P D,inter versus (t D/rD2 )inter correlations r D=0.1 till 0.4

0.25

0.3
0.25

rD=0.6

0.2

rD= 0.4

0.2

0.15

PD,inter

P D,inter

SPE 77949

0.15
0.1
0.05
0
0.5 (t /r )
D D2 inter

P D,inter versus correlations

rD = 0.3 till 0.6, CfDf* fD = 10

0.05

rD= 0.1
0

rD=0.3

0.1

1.5

20

40

(degree)

Fig. 16
0.3

P D,inter

P D,inter/(t D/rD2 )inter

rD=1.5

0.2

rD=1.1

0.15
P D,inter versus correlations

0.1

rD=1.1 till 1.5

0.05
0

20

40
60
(degree)

80

0.9
0.85
0.8
0.75
0.7
0.65
0.6
0.55
0.5

P D,inter/(t D/rD2 )inter versus correlations

rD = 1.5 till 2, CfDf* fD = 10

rD=2

100

20

40

60

100

Fig. 21
0.25

2
P D,inter/(t D/rD2 )inter versus correlations

1.5

rD= 0.1, 0.2, 0.3

rD=0.3

0.15

0.05

20

40
60
(degree)

80

rD= 0.4

0.1

0.5

rD= 0.7

0.2
P D,inter

P D,inter /(t D/rD2)inter

80

(degree)

Fig. 17

P D,inter versus (t D/rD2 )inter correlations

rD = 0.4 till 0.7, CfDf* fD = 10
0

100

0.1

0.2

0.3
(t D/rD2 )inter

0.4

0.5

0.6

Fig. 22

Fig. 18
0.25

0.25

0.2

0.2

rD=0.7

rD=0.6

0.15

P D,inter

P D,inter

80

Fig. 20

Infinite conductivity model:

0.25

60

0.15
P D,inter versus (t D/rD2 )inter correlations

0.1

rD=0.3

0.1
P D,inter versus correlations
rD = 0.3 till 0.6, CfDf* fD = 2

0.05

rD=0.4 till 0.7

0.05

0.2

0.4
(t D/rD2)inter

Fig. 19

0.6

0.8

20

40

(degree)

Fig. 23

60

80

SPE 77949

DETERMINATION OF FRACTURE ORIENTATION BY MULTI-WELL INTERFERENCE TESTING

P D,inter/(t D/rD^2)inter versus correlations

rD = 1.2 till 1.4, CfDf* fD = 2

P D,inter/(t D/rD2 )inter

P D,inter/(t D/rD^2)inter

1.2

0.8
rD=1.4

0.6
0.4
0

20

40

60

80

2
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4

P D,inter/(t D/rD2 )inter versus correlations

rD = 0.5 till 0.8, CfDf* fD = 0.5

rD=0.8

(degree)

20

40

(degree)

Fig. 24
0.26

80

Fig. 27

0.32

rD= 0.8

0.18

rD= 1.5

0.3

0.14

0.28

0.1

P D,inter

P D,inter

60

rD= 1.1

0.22

P D,inter versus (t D/rD^2)inter correlations

0.06

rD = 0.8 till 1.1, CfDf* fD = 2

0.02
0

0.1

0.2

0.3

0.4

(t D/rD2 )inter

0.5

0.2
rD=0.3

0.15

P D,inter versus correlations

0.05

rD =0.3 till 0.6, CfDf* fD = 0.5

0
0

20

0.22

P D,inter versus (t D/rD2 )inter correlations

0.2

rD = 1.2 till 1.5, CfDf* fD = 0.5

0.25

0.3

0.35
0.4
(t D/rD2 )inter

Fig. 28

rD=0.6

0.1

0.24

0.2

0.3
0.25

rD= 1.2

0.26

0.18

Fig. 25

P D,inter

11

40

(degree)

Fig. 26

60

80

0.45

0.5

0.55