You are on page 1of 12

SPE 132838

Effects of Fracture Properties on Numerical Simulation of a Naturally


Fractured Reservoir
M.M. Noroozi and B. Moradi, SPE, Iranian Central Oil Fields Company, and G. Bashiri, Research Institute of
Petroleum Industry
Copyright 2010, Society of Petroleum Engineers
This paper was prepared for presentation at the Trinidad and Tobago Energy Resources Conference held in Port of Spain, Trinidad, 2730 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
Effects of Some of the fracture properties on numerical simulation of fractured reservoirs are usually not taken into consideration.
Two important fracture properties, which are used in simulation, are fracture capillary pressure and fracture relative permeabilities.
Engineers usually assume the first one to be zero and the second one to be a straight line without paying attention to their
significant role in the simulation of fractured reservoirs. In addition, effects on these parameters on the behavior of fractured
reservoir model when Dual Porosity Dual Permeability (DSDP) concept is used are not investigated yet. The present study
investigates the effects of above mentioned properties through other fracture properties like fracture permeability and matrixfracture transfer coefficient(shape factor) to analyze completely effects of the whole fracture characteristics on numerical
simulation of a reservoir. The oil field under study is a highly fractured carbonate reservoir located in Southwest of Iran. This
paper indicates when straight-line fracture relative permeabilities and zero fracture capillary pressure can be used in the simulation.
Effect of reservoir heterogeneity was investigated on numerical simulation too. Sensitivity analysis also has been done to clearly
indicate the behavior of the DSDP model by assuming /nonassuming zero fracture capillary pressure and straight line fracture
relative permeabilities. Sensitivity analysis was done in three main production scenarios: natural depletion, water injection, gas
injection. This sensitivity study will also show the magnitude of the error which simulation engineers are making if they use
straight-line relative permeabilities and zero capillary pressure in the fractures.
Introduction
Simulation of naturally fractured reservoirs is a challenging task from both a reservoir description and a numerical standpoint:
Flow of fluids through the reservoir primarily is through the high permeability, low effective porosity fractures surrounding
individual matrix blocks.The matrix blocks contain the majority of the reservoir pore volume and act as source or sink terms to the
fractures. The rate of recovery of oil and gas from a fractured reservoir is a function of several variables, including:a) Size and
properties of matrix block b) Size and properties of fracture blocks. The study of naturally fractured reservoirs has been the
subject of numerous papers over the last four decades. but there are some fracture properties which usually their effects is ignored
by researchers and engineers .Two important fracture properties which are usually neglected are: a) Fracture relative
permeabilities .b) Fracture capillary pressure. Currently engineers simulating fractures reservoirs by following assumptions: a)
straight line fracture relative permeabilities. b)zero capillary pressure in the fractures.They do not know when non-straight line
relative permeabilities and non-zero capillary pressure are important and when they are not important, and they do not know the
magnitude of the error they are making by applying these simplified assumptions. De la Porte [1] investigated the effect of fracture
properties on the synthetic model while Dual Porosity concept is used; however, the effects of fracture properties on the behavior
of the real fractured reservoir model when DSDP concept is used, is not investigated yet. In this study, one of the fractured
reservoirs of Iran has been used as a real model. ECLIPSE, which is one of the best worldwide simulators, has been used in this
project to perform the simulation.simulation results with economical impact, which is analyzed, Are final oil Recovery, Oil rate,
Water injection rate, Gas oil ratio and Gas injection rate.

SPE 132838

Theory
Fracture Relative Permeabilities
Relative permeability functions are usually taken to be dependent on phase saturation. The two most commonly used expression
for relative permeability for homogeneous porous media are the X-curve and Corey curve (Corey, 1954). The X-curve describes
relative permeability as a linear function of saturation.[2] Earliest is Romms (1966) experiment with kerosene and water through an
artificial parallel-plate fracture lined with strips of polyethylene or waxed paper. Romm found a linear relationship between
permeability and saturation.[3] His experimemits did not examine the effects of fracture aperture and roughness effects, or the
implications for reservoir scale behavior. Pruess and Tsang (1990) conducted numerical simulation for flow through rough-walled
fractures. They modeled fractures as two-dimensional porous media with apertures varying with position. Their study shows that
residual saturation of the nonwetting phase is large and phase interference is greatly dependent on the presence or absence of
spatial correlation of aperture in the direction of flow.[4]Persoff et al. (1991) did experiments on gas and water flow through roughwalled fractures using transparent casts of natural fractured rocks. The experiment showed strong phase interference similar to the
flow in porous media. [5]Persoff and pruess did other experiments on multiphase flow in rough walled fracture. [6]Their results are
compared with commonly used relative permeability relations for porous media, the X-curve and Corey curve as shown in Figure
1. In 1992, Rossen and Kumar introduced a method for calculating non straight-line relative permeabilities usig the Effective
Mediumn Approach (EMA), which is based on the work of Pruess and Tsang as discussed above. EMA was used to illustrate the
effect of gravity and aperture distribution on relative permeabilities. [7] The fracture relative permeability curves used in this study
were obtained from the research by Rossen and Kumar. According to their study the main parameter used in selecting the
appropriate relative permeability curve for a specific reservoir system, with gravity and capillary forces acting, is dimensionless
fracture height, HD. That is essentially a ratio of the gravitational force to the capillary forces in the system: [1]
* g * H
(1)
H D =
/ bo
Where is the density difference between the fluids, is the interfacial or surface tension between the fluids,g is gravitational
acceleration,H is fracture height and bo is mean half-aperture of the fracture.
HD quantifies the extend of gravity segregation. HD of zero indicates complete domination of capillary forces, so simultaneous
two-phase flow is impossible. When HD is infinity, total phase segregation will allow the use of straight-line relative permeabilities
and zero capillary pressures.
Figure 3 shows the fracture relative permeability curves associated with the various HD parameters.
According to their data, it can be shown that HD is straightly proportional to mean half-aperture of the fracture (Figure 2).
Their experiments shows that when HD is high (greater than 10). Total phase segregation allows use of straight-line relative
permeabihities.according to above result in oil-gas systems straight-line relative permeabihities can be used (It will be discussed in
details in gas injection section)
Fracture Capillary Pressure
Capillary pressure in the fracture of a naturally fractured reservoir would exist due to two reasons: [8]
In vertical fractures due to gravitational segregation of the fluids and capillary rise.Due to Wettability, where fracture walls and
very narrow parts of the fractures would be covered or filled with the wetting phase, thus causing an interface between the phase to
be present.
The fracture capillary pressures used in the study were derived by Firoozabadi and Hauge [9], varying per fracture aperture of sizes
10, 20 and 100 m. They developed a phenomenological model, based on the Young-Laplace equation of capillarity.The model is
based on assumptions regarding fracture properties,Including roughness, shape of the asperities and number of asperities in contact
with each other at opposite fracture faces. They calculate fracture capillary pressure for different waviness angels (Fig 3) and
various fracture width (Fig 4). [9]
Based on the above results Rossen and Kumar converted fracture capillary pressure to dimensionless form(PcfD) and tabulated the
results(Table 2.3).They defined PcfD as follows:[7]
(2)
p
p

cfD

/b

Where bo is the fracture half aperture (t= 2bo).


If bo is given in inches, and is given in pounds/inch; Pc (in psi) would simply be calculated as p cfD =
if bo is given in microns and in dines/cm, then Pc (in psi) would be given by:[1]
pcfD
(3)
pc = 0.145*
bo

p
. On the other hand,
/ bo

SPE 132838

For example in graphical form, for a water-oil system with interfacial tension of 0.35 dynes/cm, the fracture capillary pressure
curves for different values of fracture apertures would be represented by Figure 5.
ECLIPSE models for fractured reservoirs
Dual-Porosity Model
The most commonly used flow model for practical simulations of fractured systems is the dual-porosity model. Here the basic idea
is to dissociate the flow inside the fracture network and the matrix and to model the exchange between these two media using a
transfer function. This concept was first introduced by Barenblatt and Zheltov (1960). [10]A simplified dual-porosity version of the
Barenblatt and Zheltov flow model was used, in which the block-to-block flow takes place only through the fracture network, with
the matrix feeding the fractures through a transfer function.
Dual Porosity Dual Permeability Model
The dual-porosity models discussed above do not represent inter block matrix-matrix flow. This approximation is reasonable when
large-scale flow is solely through the fractures. When matrix-matrix interblock flow is significant and must therefore be included
in the model, we require a DSDP representation.
Blaskovich et al. (1983) first introduced models of this type [11]. By adding the matrix-to-matrix connections, the matrix blocks are
no longer isolated, and contribute to the overall fluid flow. Being more general than the dual-porosity model, which is limited to
strongly connected fractured reservoirs, the Dual porosity Dual permeability model is capable of simulating a wide variety of
problems ranging from slightly fractured to highly fractured systems. [12]
In this study, DSDP model is used to describe flow in fractures. This can be done in ECLIPSE by using the DUALPERM
keyword.
Simulation Design
The main objective of the sensitivity study was to determine in which situations it would make a significant difference to the
simulated reservoir behavior to use HD, associated fracture relative permeabilities and non-zero fracture capillary pressures.
Significant difference was defined where the difference between the simulation with straight line fracture relative permeabilities
and zero capillary pressure and the non-straight line fracture relative permeabilities and non-zero capillary pressure is greater than
typical uncertainties in the input data .these is done by take in to account other fracture properties. So different combination of
fracture properties is used to prepare an engineering guide lines in fractured reservoir simulation.
The objective was achieved by running simulations, varying the fracture parameters mainly influencing the flow directly, i.e.
effective fracture permeability, sigma. Each combination was run with each of the different sets of relative permeabilities, as
associated with the HD parameter, using a zero capillary pressure. The results were compared to the base case, which had straightline fracture relative permeabilities and zero capillary pressures. This was performed on three different cases:
Reservoir with primary Depletion
Reservoir with water injection
Reservoir with gas injection
Water injection scenario
1-Effect of Fracture Relative Permeabiliy
As can be seen in Figure 6 differences decrease with an increase of HD, with the differences for HD=5 less than about
10%.Differences in ultimate oil recovery are all positive, which indicates that the recovery will always be less (slower) if nonstraight-line relative permeability curves have been used. Figure 7 shows that water injection rates in cases with a low effective
fracture permeability (for example, Kf=l00mD), are always much lower than in straight-line cases. At high fracture permeabilities,
with Kf around 500 to 1000mD and HD>=1, the injectivity is the same. Figure 8 shows the differences in the output vectors that
were observed in a HD=0.5 case and Figure 9 shows how the differences reduced in the HD=5 case.As can be seen from Figure 8
there are significant difference between main output vectors and base case.
Summary of cases where straight-line fracture relative permeability would make no difference in any of the results parameters are
listed in Table 1.
2-Effect of Capillary Pressure
The effect of fracture capillary pressure in the three-phase system, combined with non- straight-line fracture relative permeability,
is minimal and can be ignored. A combination of HD=1 and the capillary pressure for a fracture width of 10 microns, were tested.
Figure 10 shows the differences between i) test case with HD=1 and Pc=0 and ii) Test case with HD=1 and Pc=High

SPE 132838

Gas injection scenario


1-Effect of Fracture Relative Permeabiliy
Straight-line relative permeabilities is used for gas oil case in this study.this is based on the results of gas-oil density differences,
as calculated by the ECLIPSE. Figure 11 shows the density differences with pressure change. This figure clearly indicates that
minimum difference between gas and oil density is 27.3 lb/ft3. Using the minimum differences above and calculating the HD; the
results can be seen in Table 2 for various aperture widths. For this minimum difference, it was proven that HD is always greater
than 5 so straight line relative permeabilities can be used .for the other gas-oil density differences which is more than 27, HD even
will be more than HD related to this minimum differences.Therefore gas-oil relative permeabilities in the fractures should always
be set to straight lines.
It is good to be mentioned that the expression for HD in any consistent set of units is given by Eq. (1) However; the terms on the
right side of this equation in this study are more traditionally measured in the following units:
: lbm/ft3, H: ft, G: ft/sec2, gc: (lb.ft)/(lbf.sec2), : dynes/cm, bo: cm, If these units are used, HD could be calculated as follows:

H D = 4.7892 *10 2 [

* H
]
/ bo

(3)

2-Effect of Capillary Pressure


The base case, with zero fracture capillary pressures, was compared to three cases in which the fracture capillary pressure was set
to correspond to three fracture widths, as described before. The results show a clear increase in field oil recovery (FOE), higher oil
production rates (FOPR) with increasing capillary pressure because as fracture capillary pressure decrease later and more gradual
gas breakthrough will occur . It is obvious that Gas- oil ratio is inversely proportional to fracture capillary pressure (Figure 12).
Primary depletion scenario
1-Effect of Fracture Relative Permeabiliy
The differences between ultimate recoveries were small between base case with straight-line fracture relative permeabilities and
test cases. As can be observed on Figure 13 all differences are less than 10% so the effect of fracture relative permeabilities can be
ignored.
2-Effect of Capillary Pressure
As the main driving mechanism in this case is gravity drainage (like gas injection case), fracture capillary pressure have a
significant effect on the simulation results.Fig 14 shows the difference in ultimate oil recovery in base case(pcf=0) and the test
cases (with nonzero fracture capillary pressures ) are exactly like the gas injection case as described before.

Conclusion
1) This research has shown that the use of fracture relative permeabilities in numerical simulations of this fractured reservoir
sometimes make a substantial difference in the simulated reservoir behavior and subsequently the prediction results. The
difference up to 70 % in prediction oil recovery has been shown.
2) The use of straight-line fracture relative permeabilities should be limited to the following water injection cases: HD
factor of 1.5: low sigma (around .001 to .0001 ft-2) and high Kf (more than 1000 mD). All cases with an HD of 5 and
higher.
3) In gas injection case, the difference between gas and oil densities are high enough for HD to be greater than 5; so straight
line fracture relative permeabilities could be used.
4) Results from the primary depletion case approved that effect of fracture relative permeabilities are insignificant on the
simulation results.
5) Fracture capillary pressures showed significant differences in the simulation results of gas injection case, where
recoveries were up to 60% higher. Gas-oil systems with: a) Low fracture permeability (kf=100), b) small fractures widths
(<100 micros) should be always simulated with non-zero fracture capillary pressure.

SPE 132838

Acknowledgments
This work was supported by the National Iranian Oil Company (NIOC). The authors acknowledge the Iranian Central Oil Fields
Company (ICOFC), especially the R&D department, for their help in preparing this paper.
References
1) De la Porte ,J.J.; Kossack,C.A.:"The Effect of Fracture Relative Permeabilities and Capillary Pressures on the Numerical Simulation of
Naturally Fractured Reservoirs, paper SPE No.95241

2) Corey, A.T.: The interrelationship between gas and oil relative permeabilities, Prod. Mon., Vol. 19, pp. 38-41, 1954
3) Serhat,A.:Estimation of fracture relative permeabilities from unsteady state core floods, Journal of Petroleum Science
and Engineering 30 ,2001
4) Pruess, K.; Tsang ,Y.W.: Two-Phase Relative Permeability and Capillary Pressure of Rough-Walled Fracture,Water
Resource. Res. 1915-1926., Sept. 1990
5) Persoff, P. K.; Pruess, K.; Myer, L.: Two-Phase Flow Visualization and Relative Permeability Measurement in
Transparent Replicas of Rough-Walled Rock Fractures, Proceedings 6th Workshop on Geothermal Reservoir
Engineering, Stanford University, Stanford, California, January 23-2 5, 1991
6) Persoff, P.;Pruess,K.:Two-Phase Flow Visualization and Relative Permeability Measurement in Natural Rough-Walled
Rock Fractures ,Water Resources Research Vol. 31, No. 5, pp. 1175-1186, May, 1995
7) Rossen, W.R.; Kumar, A.T.A.: Single and Two-Phase Flow in Natural Fractures, paper SPE No.24915 ,1992
8) Saidi,A.M.:Reservoir Engineering of Fractured Reservoirs(Fundamental and Practical Aspects),Total edition
presse,1987
9) Firoozabadi, A.; Hauge, J.: Capillary pressure in fractured porous media. JPT, 784 791. 1990
10) Barenblatt, G. I.; Zheltov, Y. P.:Fundamental equations of filtration of homogeneous liquids in fissured rocks,1960
11) Blaskovich,F.T.;Cain,G.M.; Sonier,F.; Waidren,D.;Webb,S.J.:A multicomponent isothermal system for efficient
reservoir simulation, paper SPE NO.11480 presented at the Middle East Oil Technical Conference, Bahrain,1983
12) Gong.B.:Effective Models of Fracture Reservoirs ,Stanford University,september 2007
Table 1: Summary of cases where straight-line fracture relative permeability would make no difference in any of the results
parameters.
HD

0
0.5
1
1.5
1.5
1.5
1.5
5

Matrix-Fracture
Coefficient, Sigma
(0.0001 to 1 ft-2)
None
None
None
1
0.01
0.001
0.0001
All

Effective fracture
permeability ,Kf
(100 to 1000 mD)
None
None
None
None
None
High only
High only
All

SPE 132838

Table 2: Gas HD calculation.


Pressure
4194

[lbm/ft3]
27.3
27.3
27.3
27.3

H
[feet]
5
5
5
5

[dynes/cm]
0.66
0.66
0.66
0.66

bo
HD
[micro meters]
10
99.0493
100
990.493
1000
9904.93
10000
99049.3

Figure 1:Measurement of air-water relative permeabilities in rough-walleded fractures (graph from Horne et al. 2000)

Figure 2:Dimensionless fracture height parameter, HD, vs mean half-aperture of the fracture.

SPE 132838

Figure 3:Computed capillary pressure for t=100m and different waviness angels

Figure 4:Computed capillary pressure for =5 and various fracture width

p c(p si)

Capillary pressure vs.saturation


0.2
0.18
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0

bo=5 microns
bo=10 microns
bo=50 microns

0.2

0.4

0.6

0.8

1.2

liquid saturation

Figure 5: Fracture Capillary Pressures per fracture width.

SPE 132838

(4,3,,1)

(4,2,,1)

(4,1,,1)

(3,3,,1)

(3,2,,1)

(3,1,,1)

(2,3,,1)

(2,2,,1)

(2,1,,1)

(1,3,,1)

(1,2,,1)

(1,1,,1)

(0,3,,1)

(0,2,,1)

80
70
60
50
40
30
20
10
0
(0,1,,1)

% Difference

Ultimate Recovery (FOE) Differences Results

Experiment No.
sigma=1

sigma=0.0001

Figure 6:Differences in Ultimate Oil Recovery (FOE) with an increase in HD towards the right.

Water Injection Rate after 1 year vs K (Sigma=1)


80
70
%Difference

60
HD=0.5

50

HD=1

40

HD=1.5

30

HD=5

20
10
0
k=100

k=500

k=1000

Figure 7:Difference in water injection rates (injectivity) after 1 year of production(sigma=1).

Figure 8:Results from Case with HD=0.5 and HD=: Gas-oil ratio ,Oil Recovery Factor, oil production rate , water cut. The green
curves represent the base case, using straight line relative permeabilites.

SPE 132838

Figure 9:Results from Case with HD=5 and HD=: Gas-oil ratio ,Oil Recovery Factor, oil production rate , water cut. The green
curves represent the base case, using straight line relative permeabilites.

10

SPE 132838

Figure 10:The combined effect of non-zero capillary pressure and non-straight-line relative permeabilities. Blue represents the test
case with Pc=0, and red the test case with Pc=High

Gas-Oil Density Difference vs Pressure


45
40
35
30
25
20
1000

1500

2000

2500

3000

3500

4000

4500

Figure 11:Gas-Oil density difference with pressure change, as determined by ECLIPSE

SPE 132838

11

Figure 12:Comparison of results, using different capillary pressure tables based on fracture aperture . green
curves(b0=0),pink curves(b0=5), light blue curves(b0=10), dark blue curves(b0=50).

LIVE OIL PRIMARY RECOVERY: Difference in Oil Recovery (FOE)


9
8
7
6
5
FOE Difference (%)
4
3
2
1
0

kf="100"
kf="500"

kf="100"
(4,kf,4,1)

(3,kf,4,1)

(0,kf,4,1)

Experiment number (HD, Kf, sigma,deltaP)

(2,kf,4,1)

(4,kf,1,1)

(2,kf,1,1)

kf="1000"
(3,kf,1,1)

(0,kf,1,1)

kf="1000"

Figure 13: summary of all differences in the ultimate oil.

12

SPE 132838

50 microns fracture aperture

10 microns fracture aperture

5 microns fracture
aperture

Fracture aperture=0

Figure 14:Comparison of field oil recovery in base case and test cases.