SPE 114594

Estimating Fracture Permeability and Shape Factor by Use of Image Log

Data in Welltest Analysis
Hassan Bahrami, Sharif University of Technology; Jamal Siavoshi, Husky Energy Canada; Soodabeh Esmaili and
Mohammad Hussein Karimi, POGC Iran; and Ramin Bahraie, Azad University Iran

Copyright 2008, Society of Petroleum Engineers

This paper was prepared for presentation at the 2008 SPE Asia Pacific Oil & Gas Conference and Exhibition held in Perth, Australia, 2022 October 2008.

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 initially assumed in the model and they are usually tuned
during history matching which can be time consuming and
The identification of fractures is essential during
also affect other history match parameters. Reservoir
exploration, drilling and well completion of naturally
simulation results and predictions might be inaccurate if
fractured reservoirs since they have a significant impact on
the values of fracture properties in the model are not
flow contribution. There are different methods to
reliable.
characterize these systems based on formation properties
This paper shows using image log data associated with
and fluid flow behaviour such as logging and testing.
welltest analysis in order to determine dynamic fracture
Pressure-transient testing has long been recognized as a
parameters such as fracture permeability and shape factor
reservoir characterization tool. Although welltest analysis
for reservoir simulation. In this study, sensitivity analysis
is a recommended technique for fracture evaluation, but its
has also been performed on fracture permeability, fracture
use is still not well understood.
porosity and matrix block size in real simulation models
Analysis of pressure transient data provides dynamic
in order to show importance of accurate determination of
reservoir properties such as average permeability, fracture
fracture parameters.
storativity and fracture conductivity. An infusion of
geological knowledge helps reducing uncertainty
Introduction
associated with any well-test interpretation. The static
Naturally fractured reservoirs differ from homogeneous
properties of naturally fractured reservoirs such as fracture
reservoirs from many points of view: geological,
distribution, fracture aperture, matrix block size and
petrophysical, production and economics. We may think of
fracture porosity can be obtained from processing of Image
fractured reservoirs as initially homogeneous systems
Log data.
whose physical properties have been deformed or altered
Simulation of naturally fractured reservoirs needs defining
during their deposition. As a consequence, it is not always
fracture permeability, shape factor and fracture porosity in
easy to match the behavior of these systems, specifically to
the fracture model. However, in most simulation studies,
forecast their production during simulation.
due to high uncertainties in estimating fracture
permeability and shape factor values, these parameters are
2 SPE 114594
The major challenges in fractured reservoirs are: 1) They
act as drains or barriers; 2) They significantly affect the
productivity at producers and the injectivity at injectors; 3)
They control sweep efficiency and water/gas breakthrough
within the reservoir; 4) They are difficult to simulate in
geological models.
A naturally fractured reservoir is a network of fractures,
vugs and matrix which are randomly distributed. The
porous networks have very different geometrical
characteristics, porosities, capacities and permeability. The
porosity associated with a fracture system is usually low.
However, fractures can have very high permeability that
results from the low tortuosity of the fluid flow path
through them.
The most common geometrical representations of fractured
reservoirs are the models introduced by Warren&Root and
Kazemi, assuming discrete matrix blocks separated by an
orthogonal system of continuous and uniform fractures.
Both models (Fig. 1) assume that fluids are mostly stored
in matrix blocks that have relatively high porosity and
relatively low permeability. These blocks are separated by
fractures that have relatively low porosity and relatively
high permeability. The fluid flow to the wellbore initiates
from the matrix blocks into the fractures under pseudo-
steady state conditions. The models have been introduced
to simulate flow through naturally fractured reservoirs.
Using dual-porosity and dual-permeability concept in
rectangular parallelepiped blocks, the idealized models
have been developed for the purpose of studying the
behavior of fractured reservoirs.
In reservoir simulation, fracture permeability and shape
factor are the main input parameters in the fracture model.
Although welltest analysis provides average permeability
(k), fracture storativity (), and fracture conductivity (),
but their use for determining fracture permeability and shape
factor is not well understood. Therefore in reservoir
simulation, the model initialization for fracture permeability
and shape factor is usually based on initial guess, not
reliable calculations, and the fracture properties are
considered as matching parameters during field history
matching. This can have negative impact on other matching
parameters and affect simulation results for predicting
reservoir dynamic behaviour.
Fracture Characterization using Image Log Data
Processing:
In Formation Micro-Imagers, the amount of the current
penetrated into the formation from each of the electrodes
is measured, generating conductivity curves, which is
then converted into an image for geological and
petrophysical interpretation. Since resistivity of a matrix
rock is higher than a matrix saturated with mud, when a
button electrode approaches a fracture filled with mud or
other fluid of resistivity (Rm), an increase in current
begins to flow due to the presence of this low resistivity
anomaly [7].
The fracture aperture is proportional to the sum of the
increased current flow; and mathematically can be
expressed as [7]:
C2 1 C 2
b = C1. A.Rm Rxo [1]
Where b is fracture aperture, C1 and C2 are constant from
tool modeling, A is excess current divided by voltage and
integrated along a line perpendicular to the fracture trace,
Rm is mud resistivity and R XO is flushed zone resistivity.
Fracture porosity is another output of Image Log
processing [7]:
FMI = matrix [LLS . Ci ]1/ m [2]
Where Ci is conductivity of each electrode, LLS is shallow
Laterolog resistivity, m is cementation factor, FMI is
secondary porosity and Matrix is open hole log porosity.
Processing of image log data can give reliable values for
fracture spacing, fracture porosity and fracture aperture,
and provides a real picture of the fracture network in a
reservoir.
SPE 114594
Fracture Characterization Using Pressure
Transient Analysis:
Pressure-transient testing has long been recognized as a
reservoir characterization tool. The main results of welltest
analysis in naturally fractured reservoirs are
permeability*thickness, fracture storativity and fracture
conductivity (interporosity flow coefficient).
In analysis of pressure transient data, average reservoir
permeability is determined from slope of straight line on
semi-log Horner plot as shown in figure 2:
K Average = [162.6 q.o .Bo ] /[m.h] [3]
It is necessary to know the effective producing thickness
of reservoir in order to determine reservoir permeability.
In order to describe matrix and interconnecting fracture
network, the main dynamic parameters commonly used are
fracture conductivity, , and fracture storativity, .
The fracture storativity is the fraction of fluid stored in the
fracture system, and is related to the porosities and
compressibilities of matrix and fracture. The storativity
ratio represents in percentage, the contribution of the
fracture system to reservoir fluid storage. The fracture
storativity is defined as follows [2]:
(VC t ) f
= [4]
[(VC t ) f + (VC t ) m ]
The smaller the value, the more is the fracture effect on
pressure variations. The value of 1 means that the
reservoir behaves like a homogeneous reservoir made up
solely of fractures.
The fracture conductivity characterizes the ability of the
matrix blocks to flow into the fracture system and is related
to the permeability contrast between the two matrix and
fracture media. The smaller the , the later the start of
total system flow (delays the beginning of the transition
period). Fracture conductivity is defined as follows [2]:
K
= m rw2 [5]
Kf
3
Where the shape factor () is [5]:
= 4 (1/ LX 2 + 1/ LY 2 + 1/ LZ 2 ) [6]
In the standard Semi-Log plot of the buildup-pressure
response vs. shut-in time or drawdown-pressure response
vs. flowing time, neglecting the wellbore storage effects,
dual porosity response is characterized by two parallel
straight lines of the early and late time with a transition
period in between (Figure 2). The displacement of these
straight lines is function of and . The dual porosity
response on Log-Log plot is observed as a valley on
pressure derivative curve followed by a platue.
Using equations 5 and 6, having fracture conductivity
from welltest analysis, matrix permeability from core
data, and knowing fracture permeability (which is still
unknown in the calculations), the parameters shape factor
and average matrix block size for unidimentional matrix
blocks can be obtained using as follows:
= [.Kf ] /[K m .rw 2 ] [7]
L = rw 12 K m / (.K f ) [8]
Note that matrix block size (fracture spacing) along the
logged interval can directly be obtained from processing
of Image Log data. Using the log results, accuracy of the
estimated average block size from welltest can be
checked.
Using equation 4, and having fracture storativity value
from welltest, the fracture porosity can be estimated as
follows:
Cm
= m [9]
C f 1
f
Fracture compressibility might be different from matrix
compressibility by an order of magnitude. Fracture
compressibility might be in the range from 4E-04 to 4E-
05. In other words, Cf values are 10 to 100 folds higher
than Cm [1].
4 SPE 114594
The fracture compressibility can be estimated from the
following equation [1]:
1 (k / k i ) 2 / 3
Cf = [10]
P
Where K i is welltest permeability at the initial reservoir
pressure; and K is welltest permeability at the current
reservoir pressure.
Fracture porosity is another output of Image Log data
processing, and it can be used to check accuracy of average
fracture porosity estimated from welltest analysis.
Challenges Regarding Fracture Permeability in
Naturally Fractured Reservoirs:
Determination of permeability value is one of the biggest
challenges in reservoir/production engineering. Pressure
transient analysis yields a value for permeability*thickness,
and in order to calculate reservoir permeability, defining an
accurate value as effective producing interval (thickness) is
required. The effective producing thickness can be
determined based on producing (perforated) interval from
petrophysical evaluations and/or production logging
results.
For naturally fractured reservoirs, since cumulative
thickness of fractures is not easy to determine, reservoir
thickness or perforated thickness might be used as
producing thickness. This generally results
underestimating fracture permeability as high value of
fracture permeability will be distributed over matrix
thickness.
It has also been suggested that fracture permeability can be
estimated by assuming welltest permeability as geometric
mean value of fracture permeability and matrix
permeability [1]:
k f = k / .km
2
[11]
But note that fracture permeability mainly depends on
fracture aperture, while it is not considered in the
permeability averaging. Also in using this averaging, the
results for fracture permeability will be sensitive to matrix
permeability value, and uncertainty in value of matrix
permeability, increases uncertainties in determining value
of fracture permeability.
However using welltest results and cumulative fracture
thickness from Image Log data, more reliable values for
fracture permeability and shape factor can be estimated.
Estimating Fracture Permeability by Use of
Image Log data in Welltest Analysis:
For layered reservoirs, average permeability of layers are
calculated based on thickness of each layer [5], and the
concept can also be used in naturally fractured reservoirs.
By considering matrix and fractures as layered system,
average permeability*thickness based on thickness of
matrix and fracture layers can be written as follows:
m=1....i f =1.... j
khAverage =
Matrix
k m hm + k
fracture
f hf [12]
Where h f is fracture aperture, hm is matrix block
thickness, k is average permeability, and h is reservoir
thickness. Average permeability is obtained from welltest
analysis, and matrix permeability can be estimated from
core analysis.
Since in the permeability averaging based on thickness,
fracture aperture is also taken into account, more realistic
results for fracture permeability can be obtained
In Kazemi simplified fracture model, matrix and fracture
in systems are considered as parallel layers in which, matrix
blocks have similar dimensions, and fracture layers have
similar aperture values. In other words, all matrix blocks
have the same values for average matrix permeability and
block size, and all fractures have similar fracture
permeability and aperture values. Therefore equation 12
can be simplified by considering Kazemi model as
follows:
k f =(k.h) / hf + (h/ hf 1).km [13]
Calculating fracture permeability using this technique
needs knowing total fracture thickness (cumulative fracture
aperture value) as well, which can be obtained from Image
Log processing results. Therefore in addition to
SPE 114594
petrophysical evaluation results which are important for
welltest analysis of conventional reservoirs, Image Log
data are needed for welltest data analysis of Naturally
Fractured Reservoirs. Use of both welltest and Image Log
data can provide more reliable fracture characterization as
shown in the field example below.
Field Example: Fracture Characterization
In a Naturally Fracture Reservoir in Middle East, results of
welltest data, core analysis, and Image Log data were
studied in order to characterize fracture parameters.
Image log data of wells X-Y1, X-Y2, X-Y3, X-Y4 and X-
Y5 were studied in order to estimate average static fracture
properties. Figure 3 shows processed Image Log data for
fracture dip, azimuth, porosity and aperture along the
logged interval in well X-Y5. Similarly, the Image Log
data of the other 4 wells were processed. Summary of
image log data processing results for the 5 wells have been
shown in figure 4, in which total number of fractures, total
fracture height (cumulative fracture aperture value),
average fracture spacing and average fracture porosity in
each well have been reported. The results showed average
matrix block size of 7 ft, average fracture aperture of 1
mm, average cumulative fracture aperture value of 3ft, and
average fracture porosity of 0.50%, as reported in table 1.
Analysis of core data of well X-Y7 assuming 3% porosity
cut-off, showed average matrix permeability of 1.5 md and
average matrix porosity of 0.08 as can be seen in figure 5.
The core permeability values are in the range of 0.01 to 10
md, which shows very heterogeneous matrix rock. In other
words, there is a high uncertainty in the parameters which
are calculated based on the matrix permeability value. Also
ratio of fracture compressibility to matrix compressibility
has been assumed as 10.
Pressure build-up testing data in well X-Y6, which has
been completed in the same reservoir layer as other wells,
was analyzed as shown in figure 6. The results showed
average permeability value of 295 md (k.h value of 542
Darcy.ft), value of 0.32 and value of 1.56E-7.
Use of value in Eq. 9 and considering matrix porosity
5
value of 8% from the core data, resulted in fracture
porosity of 0.007 which was in good agreement with
porosity value from Image Log results (0.005).
Using equation 13, and considering welltest k.h value (542
Darcy.ft), core matrix permeability value (1.5 md), and
average cumulative fracture aperture value from Image log
data (3 ft), resulted in the fracture permeability value of
136,411 md.
Using equations 7 and 8, and considering the fracture
permeability value of 136411 md, the core matrix
permeability of 1.5 md, wellbore radius value of 0.35 ft,
and value of 1.56E-7, resulted in shape factor value of
0.12 (matrix block size of 10.2 ft).
Also values of fracture permeability from different
techniques were compared, which showed significantly
different results for shape factor (Equation 7), as shown in
tables 2, 3 and 4:
- Fracture permeability by use of Image Log data in
welltest analysis (equation 13), resulted in Kf of
136411 md, and block size of 10.2 ft (shape factor:
0.12).
- Considering geometric mean of fracture
permeability and matrix permeability as average
reservoir permeability (equation 12), resulted in Kf
value of 58017 md and block size of 15.6 ft (shape
factor: 0.05).
- By considering reservoir thickness as producing
thickness with Kf of 295 md, block size of 218 ft
was estimated (shape factor: 0.0003).
The results showed that use of fracture aperture data from
Image log in welltest analysis, provides more reliable
fracture permeability calculation compared to other
techniques, since calculated matrix block size from the
technique (10.2 ft) was in better agreement with reality of
fracture distribution from Image log processing (table 1).
Also note that in permeability calculation from welltest
and Image log (table 2), the calculated fracture
permeability is not very sensitive to matrix permeability
value and it mostly depends on fracture aperture
(calculated Kf is in the range of 132,000-137,000 md).
6 SPE 114594
But in use of geometric mean of Kf and Km as welltest
permeability (table 3), the results are very sensitive to
matrix permeability value and as can be seen, use of
minimum and maximum core permeability values, causes
significant changes in fracture permeability (calculated Kf
is in the range of 8,700-1,740,000 md). Also in
considering welltest permeability as fracture permeability
(table 4), matrix block size of 218 ft is estimated, which is
very far from reality of reservoir based on Image Log
data.
As can be seen, this technique has provided more realistic
results, since the calculated fracture parameters in general
are in a better agreement with Image Log processing
results. However in case that no image log data is
available, geometric mean of fracture permeability and
matrix permeability as effective reservoir permeability,
has provided more reasonable results for fracture
parameters compared to considering welltest permeability
as fracture permeability.
Numerical Simulation for Sensitivity Analysis:
In an oil producing Naturally Fractured Reservoir in
Middle-East, numerical simulation for sensitivity analysis
was performed on fracture permeability, fracture porosity
and matrix block size in order to show importance of
accurate determination of fracture parameters, and effect
of different fracture parameters on dynamic behavior of
the reservoir.
In the reservoir model, a well (P1) was defined with
constant oil production rate of 5000 STBD. The matrix
permeability of 5 md, matrix porosity of 0.1, fracture
permeability of 10,000 md, fracture porosity of 0.005, and
sigma value of 0.1 (equivalent to matrix block size of 11
ft) were input in the base model.
Sensitivity analysis needs changing fracture properties in
the base model, and then comparing the results. Therefore
in addition to the base model simulation run, the model
was run for fracture permeability values of 1000md and
50000md, and fracture porosity values of 0.001 and 0.01.
Then the model was run for shape factor values of 0.001
and 1, and also non-uniform distribution of sigma values
among the matrix layers (shape factor in the range of 0.001
to 1). The reservoir 3-D model and simulation results (well
bottom-hole flowing pressure and well gas production rate)
for sensitivity analysis on fracture permeability, shape
factor, non-uniform shape factor distribution, and fracture
porosity have been shown in figures 7, 8 and 9. As can be
seen from the results, since each fracture parameter can
differently affect reservoir dynamic behavior, having more
accurate input values for fracture parameters can result in
better prediction results. It was also observed that sigma
value has small effect on bottomhole pressure, but more
significant effect on production rates of different phases.
Conclusions
1. Welltesting in naturally fractured reservoirs can
provide average values for reservoir
permeability*thickness, fracture storativity and
fracture conductivity, which are not the direct inputs
of a reservoir simulation model.
2. Reliable fracture permeability and shape factor
values can be determined by use of Image Log data
in welltest analysis. Reservoir model initialization
based on the estimated fracture parameter values can
provide more realistic prediction of reservoir
dynamic behavior.
3. For Kazemi simplified fracture model, permeability
averaging can be performed based on total thickness
of fractures, and total thickness of matrix. The
permeability averaging can provide more reliable
determination of fracture permeability, compared to
other averaging techniques.
4. Knowing values of total fracture aperture and total
matrix thickness from image log data, and
considering permeability averaging based on the
thickness values of fracture and matrix, can provide
reliable estimation of fracture permeability value
from welltest average permeability*thickness value.
SPE 114594 7

5. Shape factor and matrix block size can be estimated L Fracture spacing
R Resistivity
from fracture permeability and fracture conductivity
NFR Naturally Fractured Reservoirs
values. More accurate determination of fracture R Resistivity
permeability results in more accurate estimation of Shape factor
Fracture conductivity
shape factor. Fracture porosity can be estimated from Fracture storativity
welltest fracture storativity value. Viscosity
B Formation volume factor
6. Image log data processing can provide fracture V Pore volume
spacing (matrix block size) and fracture porosity
Subscripts
values, which can be used to check accuracy of the
f fracture
estimated average matrix block size and average m matrix
x x direction
fracture porosity from welltest.
y y direction
7. In addition to petrophysical evaluation results which z z direction
are important for welltest analysis of conventional
References
reservoirs, Image Log data are needed in welltest
analysis for Naturally Fractured Reservoirs. 1. Tiab, D., Restrepo, D.P. and Lgbokoyi, A., Fracture
Porosity of Naturally Fractured Reservoirs, SPE 104056,
8. Using sensitivity analysis to know effect of each
Mexico, 2006
fracture parameter on dynamic behavior of naturally
2. Van Golf, T.D., Fundamentals of fractured reservoir
fractured reservoirs helps better matching of reservoir
engineering, 1st edition, 1982
history, and get more realistic predictions. 3. Warren, Root, The behavior of naturally fractured
reservoirs, SPE Journal, 1963
Acknowledgement 4. Saeidi, Ali, Reservoir engineering of fractured reservoirs,
The authors would like to appreciate Dr. Ali Saeidi, Dr. 1987
Seyyed Reza Shadizadeh (Petroleum University of 5. Horne, R., Modern Well test Analysis, 1995
Technology), Dr. Reza Azin (Persian Gulf University), 6. Reiss, Louis, The reservoir engineering aspects of fractured
Dr. Mohamed Tchambaz, Adil Gurkan Ceyhan, Zohreh formations, 1980
Movahhed, Meisam Ashraf, Rouhollah Dashti, Elaheh 7. Luthi, S.M, Fracture apertures from electrical borehole

Shams and Hadi Parvizi (Schlumbeger) for many helpful scans, journal of GEOPHISICS, Vol.55, NO.7, July-1990
8. Bourbiaux, B., Scaling up matrix-fracture transfers in dual-
discussions on this work.
porosity models, SPE 56557, 1999

Nomenclature 9. Ahriche, The effect of conductivity and storativity in dual

porosity reservoir, SPE 71088, 2001
P Pressure
K Permeability 10. Anguiano, Mapping of permeability structure in a naturally
Q Flow rate fractured reservoir, SPE 68833, 2001.
C Compressibility
t Time
11. C. Ehlig-Economides, Use of the Pressure Derivative for
h Thickness Diagnosing Pressure-Transient behavior, SPE 18594.
b Fracture aperture
r Radius
Porosity
8 SPE 114594

Warren-Root Kazemi model

Actual fracture model
network

Figure 1: Dual porosity dual permeability system (Warren-Root and Kazemi simplified fracture models)

Log-Log Plot of pressure

Semi-Log Plot of pressure
and pressure derivative

P f
Pressure
(DP)
Average
Permeability
Pressure
derivative
Shape
P factor
(DP)

Figure 2: Pressure Transient behavior in Naturally Fractured Reservoirs

SPE 114594 9

Figure 3: Image Log processed data in well X-Y5

10 SPE 114594

Figure 4: Fracture distribution in the wells having Image Log data

Table 1: Image Log processing results

Figure 5: Core data for matrix permeability and matrix porosity

SPE 114594 11

C=8.2 bbl/psi
K=295 md
=0.32
=1.56E-7
Skin= 3.2

Figure 6: Log-Log and Semi-Log plots of pressure and pressure derivative of the build-up test in well X-Y6

Table 2: Fracture permeability calculation by use of Image Log data in welltest analysis

Table 3: Fracture permeability calculation by considering welltest permeability as geometric mean of

fracture and matrix permeabilities

Table 4: Fracture permeability calculation by considering welltest permeability as fracture permeability

12 SPE 114594

Figure 7: Reservoir model, 3-Dimentional view

Figure 8: Sensitivity analysis results for well bottom hole flowing pressure on fracture permeability, shape factor
and fracture porosity in well P1 (perforations in lower layer)
SPE 114594 13

Figure 9: Sensitivity analysis results for well gas production rate on fracture permeability, shape factor
and fracture porosity in well P1 (perforations in upper layer)