SPE 95241
The Effect of Fracture Relative Permeabilities and Capillary Pressures on the
Numerical Simulation of Naturally Fractured Reservoirs
J.J. de la Porte, SPE, and C.A. Kossack, SPE, Schlumberger, and R.W. Zimmerman, SPE, Imperial C.
Copyright 2005, Society of Petroleum Engineers
This paper was prepared for presentation at the 2005 SPE Annual Technical Conference and
Exhibition held in Dallas, Texas, U.S.A., 9 – 12 October 2005.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in a proposal 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 a proposal of not more than 300
words; illustrations may not be copied. The proposal 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
Engineers using numerical reservoir simulators to model
fractured reservoirs typically have used straight-line (corner-
to-corner) relative permeabilities and zero capillary pressure in
the fractures, without a clear understanding of how these two
parameters affect simulation accuracy and with no practical
method for selecting alternate values. The present study uses
the theoretical work of Rossen and Kumar (non-straight-line
fracture relative permeabilities)1,2 and of Firoozabadi and
Haugh (non-zero capillary pressure in rough-walled fractures)3
to quantify prediction errors and to demonstrate a method for
selecting the correct set of relative-permeability and capillary
curves for a particular simulation. Our results indicate that
using straight-line relative permeabilities can lead to predicted
oil-recovery errors as high as 70% in water-oil systems and to
underestimating oil production times in some gas-oil systems
by as much as a factor of three. In gas-oil systems where gas
flows into the fractures, oil recovery from the matrix blocks
could be underestimated by a factor of almost two when
fracture capillary pressures are set at zero.
Introduction
Using straight-line fracture relative permeabilities and zero
fracture capillary pressures, based on E. S. Romm’s
experiment4,5 in 1966, may not always be appropriate, and
could lead to large errors in reservoir definition and
performance prediction. Recent experimental research,
outlined below, clearly shows that relative permeabilities in a
certain range of fractures are not straight lines, but
implications for reservoir scale behavior have not yet been
systematically examined. The present study focuses on
quantifying prediction errors made in reservoir scale
simulations by comparing results from straight-line fracture
relative permeabilities and/or zero capillary pressures with
results from non-straight-line fracture relative permeabilities
and/or non-zero capillary pressures. A method for classifying
a reservoir system to select the correct set of relative
permeability and capillary pressure curves is given.
The study investigated differences in behavior of two-
phase (dead oil with water or gas injection) and three-phase
systems, using straight-line fracture relative permeabilities
and/or zero fracture capillary pressures (base case) and non-
straight-line fracture relative permeabilities and non-zero
fracture capillary pressures (sensitivity case). We used a
homogeneous reservoir for most of the simulations, although a
modified version of the reservoir model from the Sixth SPE
Comparative Solution Project: Dual Porosity Simulators
(Firoozabadi and Thomas6) was also used to assess the extent
to which our results were influenced by heterogeneity. To
determine sensitivities, we used different relative permeability
and capillary pressure curves and a wide range of values for
reservoir parameters, such as fracture spacing, fracture/matrix
permeability ratio, and fluid properties.
Background
Conventional straight-line relative permeability curves
originated with the 1966 publication by Romm.4,5 His
findings, based on experiments of flow between two parallel
glass plates, showed a linear dependency between phase
relative permeability and phase saturations, as well as zero
capillary pressures. The experiments did not examine the
effects of fracture aperture and roughness effects, or the
implications for reservoir scale behavior. Firoozabadi and
Thomas6 published the results on the Sixth SPE Comparative
Solution Project: Dual Porosity Simulators (referred to
hereafter as “SPE6”), which investigated aspects of the
physics of multiphase flow in fractured reservoirs. The use of
non-zero capillary pressures showed that predicted recoveries
are affected dramatically by combinations of wettability and
Enhanced Oil Recovery methods. In the same year,
Firoozabadi and Hauge3 published a phenomenological model
for calculating the capillary pressures of a system based on
fracture characteristics such as waviness, roughness, width,
and interfacial tensions.
In 1992, Rossen and Kumar1 introduced a method for
calculating non-zero relative permeabilities using the Effective
Medium Approach (EMA), which is based on the work of
Pruess and Tsang7 from 1990. EMA was used to illustrate the
effect of gravity and aperture distribution on Percolation
Theory relative permeabilities, which prohibits two-phase
2 SPE 95241
flow in fracture networks. In this study, Rossen and Kumar1
developed dimensionless parameters, PcD and HD, where HD is
a ratio between gravitational and capillary forces in the
fracture system. In the same year, Guzman and Aziz8 made a
valuable contribution to understanding the capillary/viscous
force ratio effect (captured in the capillary number) and the
effects of fracture relative permeabilities and capillary
pressure at the moment of water break-through in an oil-water
system.
In 1994, Rossen and Kumar2 published a refined model
of relative permeabilities in a single fracture, which included
effects of wetting-phase flow along the fracture walls and
asperities. Pieters and Graves10 used video-imaging to
measure saturations in fractures, and verified the non-linearity
of relative permeabilities. Persoff and Pruess11 performed an
experimental study to understand and quantify flow in
fractures by visualization and experimentally determined
relative permeabilities from flow rates and pressure-drops in
realistic rough-walled fractures. In 1985, Hughes and Blunt12
used CT-scanning to verify results from pore-scale network
modelling of multiphase flow that investigated imbibition in
fractures and matrix-fracture interaction. Their in situ
measurements confirmed non-straight-line relative
permeabilities.
In 2001, Bertels, DiCarlo and Blunt13 developed an
experimental technique that used CT-scanning to measure and
compute fracture aperture distribution, capillary pressure, and
relative permeability in rock fractures.
Theory
Fracture Relative Permeability Curves
The fracture relative permeability curves used in the present
study were obtained from the paper by Rossen and Kumar2,
and are shown in Figure 1. The main parameter used in
selecting the appropriate relative permeability curve for a
specific reservoir system, with gravity and capillary forces
acting, was defined by Rossen and Kumar as a dimensionless
fracture height, HD, which is essentially a ratio of the
gravitational force to the capillary forces in the system:
∆ρ gH
HD = , (1)
γ / bo
where ∆ρ is the density difference between phases, g is
gravitational acceleration, H is fracture height, γ is interfacial
tension, and bo is the mean half-aperture of the fracture.
The parameter HD quantifies the extent of gravitational
segregation. When HD is high (greater than 10), total phase
segregation allows use of straight-line relative permeabilities.
On the other hand, a value of HD = 0 indicates complete
domination of capillary forces, and percolation theory will
apply, with simultaneous two-phase flow impossible. The
latter case obviously holds severe implications in terms of
phase injectivity and mobilities.
Some examples of calculating the HD parameter, using
different fracture widths, are shown in Table 1. The
calculation of HD in field units is given in the Appendix.
In deriving the above curves, it is assumed that the half
aperture, bo, follows a lognormal distribution, with a mean of
50 µm and a log standard deviation of 0.43. Experimental
studies (Gale14, Hakami15, Pruess and Tsang7) have shown that
this is a valid assumption; however in systems where the
distributions are radically different, the curves will change as
described by Rossen and Kumar2.
Fracture Capillary Pressures
Fracture capillary pressure curves (as functions of wetting
phase saturation) used in the study were derived by
Firoozabadi and Hauge3, for various fracture apertures (t =
2bo) of 10, 20 and 100 µm. They developed a
phenomenological model, based on the Young-Laplace
equation of capillarity:
⎡⎛ 1 ⎞ ⎛ 1 ⎞⎤
Pc = γ ⎢⎜⎜ ⎟⎟ + ⎜⎜ ⎟⎟⎥ , (2)
⎣⎢⎝ r1 ⎠ ⎝ r2 ⎠⎦⎥
where Pc is pressure difference across the interface, r1 and r2
are radii of the curvature of the curved interface at any point,
and γ is surface tension. 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, as fully described in the
o
reference. A contact angle of 0 is assumed.
The derived fracture capillary pressure (assuming a
o
waviness angle α of 5 ), converted to a dimensionless form,
PcfD, is given as follows:
bo Pc
PcfD = , (3)
γ
where bo is the mean fracture half-width, γ is interfacial
tension, and Pc is capillary pressure. Numerical values of the
dimensionless capillary pressure PcfD,, as a function of wetting
phase saturation, are listed in Table 5. If bo is given in inches,
and γ is given in pounds/inch, Pc (in psi) would simply be
calculated as Pc = γ PcfD /bo. On the other hand, if bo is given in
microns and γ in dynes/cm, then Pc (in psi) would be given by
Pc = 0.145γ PcfD /bo. As a concrete example, the capillary
pressure curves for a water-oil system with interfacial tension
of 24 dynes/cm are shown in Figure 2, for three different
values of the half-aperture.
Matrix-Fracture Relative Permeabilities
The relative permeability function used for the fractures is
appropriate for the flow of fluids in the fractures. In cases
where HD decreases toward zero, the end points of the water
relative permeability curve (krw) and oil-to-water relative
permeability curve (krow) move away from the zero residual
saturations, and at some non-zero saturation of the wetting
phase the relative permeabilities of both the oil and water are
zero. In cases where HD is zero, the critical water saturation is
20%, as shown in Figure 1. This is especially important for the
water phase. In most reservoir simulators, the transmissibility
calculation for matrix-fracture transfer is upstream-weighted.
Thus, for the flow of water from the fracture to the matrix,
upstream weighting will look at the relative permeability of
water in the fractures. If the water relative permeability in the
fractures is zero, with non-zero water saturation, no water will
imbibe into the matrix, which most likely is not physically
correct. No flow of water in the fractures should not
necessarily equate to no imbibition into the matrix if there are
forces present in the matrix to cause imbibition.
SPE 95241
Zimmerman et al.16 suggested the use of the matrix
relative permeabilities, evaluated at the potential that exists at
the matrix-fracture boundary, to simulate the matrix-fracture
transfer. However, in the present study, straight-line relative
permeabilities have been used for fracture-to-matrix flow.
These straight-line functions equal the saturation of the phase
that is in contact with the matrix block, which conforms to
most engineers’ physical view of the situation.
Experimental Design
Two separate ECLIPSE Blackoil Dual Porosity simulation
studies were conducted: (1) a water-flooding scenario with
both live and dead oil, and (2) a gas injection study with a live
oil. The datasets used in all the experiments were based on the
SPE6 study data, extended to a sector model with a water or
gas injector and an oil producer in opposite corners of the
model. Specific modifications related to the gas-oil case are
described below.
Three cases were compared in the water-oil system: (1) a
base case with zero fracture capillary pressure and straight-
line relative permeabilities, (2) a case with non-straight-line
relative permeabilities and zero capillary pressures, and (3) a
case where the relative permeabilities were straight lines, and
the capillary pressures were non-zero. This was first
performed for a homogeneous, water-wet reservoir, then for a
heterogeneous system, for both a live oil (three-phase system
– oil, water and gas) and a dead oil (two-phase system).
Straight-line relative permeabilities were used for the gas-oil
and gas-water flows, as described in the gas-oil section.
Reservoir and production parameters (such as φ, σ, Kf, Pcf, and
∆P) were varied to simulate as many different reservoir
systems as possible, building a classification system of
sensitivities. De la Porte17 performed more than 800
simulations in a detailed analysis of the effect of non-straight-
line relative permeabilities and non-zero capillary pressures in
the fractures of the oil-water systems. A further investigation
into the gas-oil system, following the above-mentioned work,
is described in more detail in the present paper.
Results and Discussion
Gas-Oil Systems (Gas Injection): Fracture Relative
Permeabilities (krf )
With gas-oil systems, surface tension (γ) is relatively low and
the density difference (∆ρ) between phases is relatively high.
For example, ∆ρ will be in the range of 25 to 45 lbm/ft3 (the
lower value near the bubble point pressure), and γ will be in
the range of 0.1 to 7 dynes/cm (the lower value near the
bubble point pressure). Table 2 shows the values of HD at an
intermediate pressure, a typical value of H, and a range of bo
values from 10 microns to 1 cm. The values of HD are all high
enough that straight-line-relative permeabilities should be
used. As the pressure increases, ∆ρ and γ both decrease, and as
the pressure decreases, ∆ρ and γ both increase. Consequently,
the values of HD at various bo values remain relatively
constant and large in magnitude. Thus, for gas-oil systems,
one can conclude that straight-line relative permeabilities
should be used in the fractures.
3
Gas-Oil Systems (Gas Injection): Fracture Capillary
Pressures (Pcog-f)
In the traditional case of setting fracture capillary pressure
(Pcog-f) equal to zero, where the matrix is filled with oil and
connate water and the fracture filled with gas, the process in
which oil flows from the matrix into the fracture, and gas
flows in the opposite direction, is drainage with water as the
wetting phase. With the oil phase pressure in the matrix lower
than the oil pressure in the fracture, this capillary force tends
to keep the oil in the matrix rock. The other force present is a
gravity force given by (ρo-ρg)H, where H is the matrix block
height or stack height, defined as the vertical height in the
matrix blocks for which there is capillary continuity. We
assume the base case where the stack height is equal to the
matrix block height is equal to the fracture block height.
Gravity forces tend to push oil from the matrix into the
fractures. Many reservoir engineers have long believed that a
fractured reservoir with a high Pcog-m and a small matrix block
height will trap the oil in the matrix when gas invades the
fractures. The following analysis shows that this may not
always be true.
In the section above on theoretical background, we
presented the development of an equation for Pc for fractures.
A table of dimensionless capillary pressures is provided
(Table 5). Values of Pcog-f for three fracture widths (2bo = 10,
20 and 100 µm) were derived and used in the study. For the
gas-oil cases, we used a reference pressure of 5000 psia, at
which surface tension is approximately 0.35 dynes/cm. The
simulation case was similar to that used in SPE6. Several
changes to the input data were made, and will be explained.
The gas-oil surface tension of 0.35 dynes/cm, at a
reservoir pressure of approximately 5000 psia, gives the
fracture gas-oil capillary pressures shown in Figure 3. Here we
also see the matrix Pcog-m; which is different from the Pcog-m
used in SPE6. Four dual-porosity models were created with
these curves, with all cases having the Pcog-m shown assigned
in the matrix blocks. The Pcog-f variation created the four cases,
distinguished by the maximum pressures at minimum wetting
phase saturation: (1) Pcog-f = 0 psia, with fracture width greater
than 100 µm, (2) Pcog-f = 0.1735 psia with fracture width of 10
µm, (3) Pcog-f = 0.0723 psia with a fracture width of 20 µm,
and (4) Pcog-f = 0.0417 psia with a fracture width of 100 µm.
In SPE6, Pcog in both the matrix and fractures varied with
pressure (increasing with decreasing pressure) based on
changes in the densities. This was also done in the present
investigation, but this calculation did not greatly influence the
results, because pressure in the matrix decreases, as the
pressure in the fracture decreases, with the result that both Pcog
values increase at the same time. For the initial simulations,
the matrix-fracture transfer coefficient (σ) was set to 1.0 ft-2
and the stack height (H) was set to 3.0 ft, resulting in high
matrix-fracture transfer and low gravity forces. We discuss
the effect of these two parameters later in this section. Other
reservoir characteristics, such as permeability and porosity,
were given values similar to those in SPE6, and are not
significant factors in the comparative results.
In the dual-porosity simulations of the sector model, a
corner well produced oil at a constant rate, and 95% of the
separator gas was injected into a well in the opposite corner.
4 SPE 95241
In all cases, the fractures rapidly filled with gas. This fracture
gas saturation was the key to the difference in the comparative
results. In all cases, the gravity forces were small and nearly
constant, because the fracture (or stack) height (H) was small.
The rate of oil production from the matrix block was a
function of the balance in the gravity and capillary forces.
The effective Pcog tending to keep the oil in the matrix was the
difference between the Pcog-m and the Pcog-f. With Pcog-f = 0,
this force was at maximum. If we reduce the capillary force
that keeps the oil in the matrix by putting a non-zero Pcog in
the fractures, the oil production rate from the matrix increases.
The effect of a non-zero Pcog-f is most easily observed by
looking at the oil pressures in the matrix and fracture. When
the oil pressure is greater in the matrix, oil will flow from the
matrix into the fractures. When Pcog-f = 0, the oil pressure
difference (Pom-Pof), as seen in Figure 4, decreases with time,
thus impeding the flow of oil from the matrix to the fracture.
When a non-zero Pcog-f is used (10 µm case is shown in Figure
4) and the fracture’s gas saturation is approximately 1.0, Pcog-f
in the fracture is 0.1735 psia, which is higher than Pcog-m,
which would initially be in the order of 0.051. Thus, the
drainage capillary force direction is reversed. As a result, the
oil pressure gradient has a small negative value, and is even
positive at times; see Figure 4. When this small negative (or
positive) capillary force is combined with gravity forces, oil
production from the matrix is very high.
When we simulated these four cases with Dual Porosity
ECLIPSE, we compared oil production rates to see the effect
of non-zero gas-oil capillary pressure in the fractures. Figure 5
shows the results of a comparison of the field oil production
rates.
The case with the narrowest fractures (10 µm) and the
highest Pcog-f has the highest oil production rates and continues
to produce oil many years after the other cases have stopped.
The incremental oil recoveries above the base case (Pcog-f = 0)
are as follows: 100 µm = 4%, 20 µm = 25%, and 10 µm =
76%.
Sensitivity simulations were run with a larger matrix
block height, H, and a smaller value of σ, since these are key
parameters in the comparison. When the matrix block height
was increased to 20 ft (from 3 ft), the gravity forces were
increased and, thus, dominated the oil flow from the matrix
blocks, reducing the differences between the four cases
(Figure 6). When the value of σ was reduced to 0.001 ft-2
(from 1.0 ft-2), the rate of flow between the matrix and fracture
was controlled by the matrix-fracture transmissibility. The
results show (Figure 7) that a difference between the cases
appears only after 3000 days of simulation with the 10 µm
fracture case having the highest oil production rate. When a
case with H = 20 ft and σ = 0.001 ft-2 was run, the differences
between the four cases (figure not shown) were even smaller.
Water-Oil System (Water Flooding): Fracture Relative
Permeabilities (krf )
In general, simulation results involving straight-line fracture
relative permeability curves (base case) and curves from cases
at low values of HD (sensitivity case) showed significant
differences (>10%) in one or more key reservoir production
results. De la Porte17 looked at differences in oil recovery
factors, water injection rates, reservoir pressures, water cuts,
gas-oil ratios, and oil production rates when comparing cases.
Often, differences were very large - as high as a 70% variation
in recovery factor between the base case and the sensitivity
case. Differences were greatest when HD was 0.5, and least
when HD = 5.0 or higher. This result was expected, with
water- and oil-phase mobilities in the fractures, at any
saturation, decreasing with decreasing HD (Figure 1).
Differences in results were least sensitive to the pressure
difference between the injection and production wells, and
most sensitive to the matrix-fracture transfer coefficient or
shape factor, σ. These conclusions were the same for the
heterogeneous case.
The general conclusion was that any simulation case
with HD < 5 should use non-straight-line relative
permeabilities in the fractures. This corresponds to reservoirs
with a high fracture density (small H) and very narrow
fractures (small t or bo), and/or low-density difference and
high interfacial tension among the phases. The use of straight-
line fracture relative permeabilities would most certainly result
in high errors in predictions. The main causes for the
differences were lower injectivity at the injector well, resulting
in less pressure support, lower phase mobilities at intermediate
saturations, and later water breakthrough at the producer.
Some cases showed a prolonged production lifetime due to
slower movement of the waterfront, resulting in higher oil
recovery factors.
To demonstrate the effect of using non-straight-line
fracture relative permeabilities (with Pcf = 0) on the results of
a typical simulation study, we compared five cases: a base
case with relative permeabilities taken to be straight lines (krf
= Sw) and capillary pressures set to zero (Pcf = 0), and four
sensitivity cases each using fracture relative permeabilities
associated with values of HD = 0.5, 1.0, 1.5 or 5.0. The case
shown here has medium-sized matrix blocks with σ = 0.01 ft-2,
a low effective fracture permeability of Kf = 20 mD, zero
fracture capillary pressure, Pcf = 0, and a pressure drop across
the flow field of ∆P = 2000 psi.
Figure 8 illustrates that in this case, total oil recovery
within a certain production period decreases with a decrease in
HD, whereas Figure 9 shows the decreasing effect on water
injectivity. The subsequent effect on the reservoir pressure
maintenance can be seen in Figure 10, and the effect on the
speed of the movement of the waterfront from the injector
well to the producer, is clear in Figure 11, with the latest
break-through in the HD = 0.5 case.
This example demonstrates clearly that the effect of the
fracture relative permeabilities is substantial in the case where
when HD < 5, and that the use of straight-line relative
permeabilities (HD = ∞) will give incorrect answers.
Water-Oil Systems (Water Flooding): Fracture Capillary
Pressures (Pcow-f)
Results from systematic simulation comparisons yield the
conclusion that the non-zero capillary pressure in fractures has
little effect on overall results, with differences in the results
indicators of less than 3% in all water flooding test cases.
SPE 95241
Conclusions
Our conclusions for the proper simulation of fractured
reservoirs are as follows:
• Water-oil systems with water injection into the fractures:
When HD is less than 5 (very narrow fractures, low
fracture heights, low fluid-density difference and/or high
interfacial tension between phases), it is necessary to use
non-straight-line fracture relative permeability curves.
Use of straight-line fracture relative permeabilities in
these cases could result in high errors in all results and
economic predictions.
• Water-oil systems with water injection into the fractures:
It is acceptable to use zero fracture capillary pressures.
• Gas-oil systems with gas injection into the fractures: Pcog-f
should not be set to zero for dual-porosity reservoirs with
narrow fractures (less than 100 microns). This is
especially important for reservoirs with small matrix
blocks (small H and large σ).
• Gas-oil systems with gas injection into the fractures: Use
of straight-line fracture relative permeabilities is
acceptable.
• Engineering guidelines in the next section provide the
simulation engineer with a formula for determining HD
and Pcf, and criteria for when they must be used in
simulations (or when the straight-line facture relative
permeabilities and/or Pcf = 0 may be used).
Engineering Guidelines or System Classification
Required data for a water-oil System
• Water-oil density difference at reservoir conditions (∆ρ)
• Average fracture half-width (bo)
• Estimate of water-oil interfacial tension (γ)
• Estimate of matrix block height or fracture height (H)
Required data for a gas-oil System
• Gas-oil interfacial tension at a range of pressures (γ)
• Average fracture half-width (bo)
Method to classify reservoir system: water-oil systems
1) Fracture Relative Permeabilities: Calculate the
dimensionless fracture height, HD, (see Appendix for details)
to get an indication of which forces are dominant in a specific
system.
If the value of HD is less than 7, one should use the
calculated HD value rounded to the nearest value of the set 0.5,
1.0, 1.5 or 5 (for which non-straight line relative
permeabilities are available) to select the appropriate set of
fracture relative permeability curves. For example, if the
calculated value of HD = 1.7, then the tabulated values for HD
= 1.5 should be used. Given the uncertainties in many of the
parameters used to calculate HD, this approximation is
justified.
2) Fracture Capillary Pressure: In water-oil systems, the
fracture capillary pressure can be set at zero.
5
Method to classify reservoir system: gas-oil systems
1) Fracture Relative Permeabilities: In gas-oil systems,
straight-line relative permeabilities (krf = So or Sg) can be used
in the fractures.
2) Fracture Capillary Pressure: In gas-oil systems with gas
invading the fractures; use non-zero gas-oil capillary pressure
in narrow fractures, that is, fractures that are 100 microns or
less in width. To generate the appropriate Pcog-f, use Table 5 to
get the PcfD as a function of the wetting phase saturation and
Equation (3) to convert the dimensionless capillary pressure
into psia. If bo is given in microns and γ in dynes/cm, then Pc
(in psia) is given by Pc = 0.145γ PcfD /bo. The gas-oil
interfacial tension (γ) used in the above calculation should be
the value at the average reservoir pressure or at the reference
pressure used in the “gas-oil capillary pressures, adjusted
according to the ratio of the surface tension”18 calculation.
Appendix
HD Calculation in Field Units
The expression for HD in any consistent set of units is given by
equation (1). However, the terms on the right side of this
equation are more traditionally measured in the following
units:
HD = dimensionless fracture height
∆ρ = density difference between phases [lbm/ft3]
=
H Fracture height [ft]
G = gravitational acceleration = 32.17405 ft/sec2
gc = conversion constant = 32.17405 lbm·ft·lbf-1·sec-2
g´ = G/gc = 0.0069444 lbf ft2 / lbm in2
γ = surface or interfacial tension [dynes/cm]
bo = mean half-aperture or width of the fracture [cm]
If these units are used, HD would be calculated as follows:
⎡ ∆ρH ⎤
H D = 4.7892 × 10 2 ⎢ ⎥. (A-1)
⎣ γ / bo ⎦
Matrix-Fracture transfer coefficient or shape factor, σ
To calculate σ for any specific reservoir, an estimate of the
fracture height is required. This could be estimated from
downhole imaging tools, and the associated x and y
dimensions of an average matrix block could be calculated
from correlations.
For a rectangular matrix block of lengths (Lx,Ly,Lz),
ECLIPSE uses Kazemi’s expression:
⎛ ⎞
⎜ 1 1 1 ⎟
σ = 4⎜ + + ⎟. (A-2)
⎜ L2x L2y L2z ⎟
⎝ ⎠
Zimmerman and Bodvarsson16 used Fourier methods to
analyze the asymptotic behavior of the matrix block pressure,
and found that the coefficient 4 should be replaced by π2.
ECLIPSE uses expression (A-2).
6 SPE 95241

Nomenclature 13. Bertels, S. P. DiCarlo, D. A. and Blunt, M. J.: “Measurement of

Aperture Distribution, Capillary Pressure, Relative Permeability,
bo = mean half-width of the fracture [microns] and in situ Saturation in a Rock Fracture using Computed
Tomography Scanning”, Water Resour. Res. 37, 649-662, 2001.
b = fracture width [microns] 14. Gale, J. E.; “Comparison of Coupled Fracture Deformation and
t = fracture width (aperture)3 [microns] Fluid Flow Models With Direct Measurements of Fracture Pore
g = gravitational acceleration [Nm/s2] Structure and Stress-flow Properties, Proc. 28th U. S. Symp. on
Rock Mechanics, Tucson, Ariz., A. A. Balkema, Rotterdam, pp.
H = fracture height or matrix block height [ft] 1213-1222, 1987.
HD = dimensionless fracture height 15. Hakami, E.: “Water Flow in Single Rock Joints”, Licentiate
Kf = fracture effective permeability [mD] thesis, Luleå Univ. Tech., Luleå, Sweden, 1988.
16. Zimmerman, R. W., Hadgu, T., Bodvarsson, G. S.: “A New
Pcf = fracture capillary pressure [psi] Lumped-parameter Model for Flow in Unsaturated Dual-
PcfD = dimensionless fracture capillary pressure porosity Media”, Adv. Water Resour., 19, 317-327, 1996.
17. De la Porte J. J.: “Effect of Fracture Relative Permeabilities and
∆P = pressure drop between injector/producer [psi]
Capillary Pressures on the Numerical Simulation of Fractured
Sw = water saturation Oil and Gas Reservoirs”, MSc thesis, Imperial College, London,
∆ρ = density difference between the fluids [lb/ft2] 2004.
18. ECLIPSE Reference Manual, 2004a. Schlumberger, Houston,
σ = matrix-fracture transfer coefficient [ft-2] Texas, 2004.
φ = porosity
γ = interfacial tension between fluids [dynes/cm]

Acknowledgments
This work was performed as an Imperial College MSc
research project, sponsored by Schlumberger Information
Solutions (SIS). We thank the staff at SIS Abingdon
Technology Centre for their support with the interpretation of
the results.

References
1. Rossen, W. R. and Kumar, A. T. A.: “Single and Two-Phase
Flow in Natural Fractures”, paper SPE 24915, 1992.
2. Rossen, W.R. and Kumar, A. T. A.: “Effect of Fractures
Relative Permeabilities on Performance of Naturally Fractured
Reservoirs”, paper SPE 28700, 1994.
3. Firoozabadi, A. and Hauge, J.: “Capillary Pressure in Fractured
Porous Media”, J. Petrol. Tech., pp. 784-791, June 1990.
4. Romm, E. S.: “Fluid Flow in Fractured Rocks” (in Russian),
Nedra, Moscow, 1966.
5. Romm, E. S.: “Fluid Flow in Fractured Rocks” (English
translation) W. R. Blake (transl.), Phillips Petroleum Co.,
Bartlesville, Okla., 1972.
6. Firoozabadi, A. and Thomas, L.K.: “Sixth SPE Comparative
Solution Project: Dual Porosity Simulators”, paper SPE 18741,
1990.
7. Pruess, K. and Tsang, Y.W.: “On Two-Phase Relative
Permeability and Capillary Pressures of Rough-Walled
Fractures” Water Resour. Res., 26, 1915-1926, 1990.
8. Guzman, R. E. and Aziz, K.: “Fine Grid Simulation of Two-
Phase Flow in Fractured Porous Media”, paper SPE 24916,
1992.
9. Horie, T., Firoozabadi, A. and Ishimoto, K.: “Laboratory Studies
of Capillary Interaction in Fractured/Matrix Systems”, paper
SPE 18282, 1990.
10. Pieters, D. A. and Graves, R. M.: “Fracture Relative
Permeability: Linear or Non-linear Function?”, paper SPE
28701, 1994.
11. Persoff, P. and Pruess, K.: “Two-phase Flow Visualization and
Relative Permeability Measurement in Natural Rough-walled
Rock Fractures”, Water Resour. Res., 31, 1175-1186, 1995.
12. Hughes, R. G. and Blunt, M. J.: “Pore-Scale Modelling of
Multiphase Flow in Fractures and Matrix/Fracture Transfer”,
paper SPE 56411, 1999.
SPE 95241 7

Table 1. Examples of HD calculations for typical Table 4. Fracture relative permeabilities vs. water
water-oil system. saturation for HD = 1.0, 0.5 and 0.0 (2).
HD = 1.0 HD = 0.5 HD = 0
∆ρ H γ bo bo HD
Sw Krw Krow Krw Krow Krw Krow
[lbm/ft3] [ft] [dynes/cm] [cm] [microns] [ ] 0.00 0.00 1.00 0.00 1.00 0.00 1.0

30 3 25 0.0001 1 0.172 0.05 0.00 0.87 0.00 0.885 0.00 0.9

0.10 0.00 0.53 0.00 0.60 0.00 0.675
30 3 25 0.001 10 1.724
0.15 0.01 0.30 0.001 0.295 0.00 0.29
30 3 25 0.01 100 17.24 0.20 0.03 0.17 0.02 0.12 0.01 0.00

30 3 25 0.1 1000 172.4 0.25 0.06 0.08 0.04 0.04 0.025 0.00
0.30 0.10 0.04 0.07 0.02 0.05 0.00
0.35 0.15 0.02 0.11 0.005 0.08 0.00
0.40 0.21 0.01 0.16 0.00 0.12 0.00
0.45 0.27 0.00 0.22 0.00 0.18 0.00
Table 2. Gas HD calculation. 0.50 0.325 0.00 0.28 0.00 0.24 0.00
Pressure ∆ρ H γ bo bo HD 0.6 0.45 0.00 0.42 0.00 0.4 0.00
0.7 0.58 0.00 0.58 0.00 0.58 0.00
3
[psia] [lbm/ft ] [ft] [dynes/cm] [cm] [microns] [ ]
0.8 0.72 0.00 0.75 0.00 0.75 0.00
2200 36 5 3.6 0.001 10 23.9462 0.9 0.86 0.00 0.91 0.00 0.91 0.00
2200 36 5 3.6 0.01 100 239.462 1.0 1.00 0.00 1.00 0.00 1.00 0.00
2200 36 5 3.6 0.1 1000 2394.62
2200 36 5 3.6 1 10000 23946.2

Table 5. Dimensionless Fracture Capillary Pressure,

Table 3. Fracture relative permeabilities vs. water where bo is the half-aperture in inches, Pc is the
saturation for HD = ∞, 5.0 and 1.5 (2). capillary pressure in psia, and γ is the surface
HD = ∞ HD = 5 HD = 1.5 tension in lbf/in (or any other consistent set of
Sw Krw Krow Krw Krow Krw Krow units).
0.00 0.00 1.00 0.00 1.00 0.00 1.00
0.05 0.05 0.95 0.00 0.82 0.00 0.87 SLiquid PcfD = bo Pc /γ
0.10 0.10 0.90 0.04 0.74 0.00 0.53
0.00 17.240
0.15 0.15 0.85 0.09 0.67 0.02 0.38
0.01 4.7400
0.20 0.20 0.80 0.14 0.62 0.05 0.29
0.02 3.8070
0.30 0.30 0.70 0.24 0.52 0.14 0.18
0.40 0.40 0.60 0.34 0.42 0.24 0.20
0.04 3.0160
0.50 0.50 0.50 0.44 0.32 0.36 0.04 0.06 2.5850
0.60 0.60 0.40 0.54 0.23 0.48 0.00 0.08 2.4420
0.70 0.70 0.30 0.65 0.13 0.59 0.00 0.10 2.2980
0.80 0.80 0.20 0.75 0.05 0.72 0.00 0.12 2.1540
0.85 0.85 0.15 0.81 0.02 0.78 0.00 0.14 2.0110
0.90 0.90 0.10 0.87 0.00 0.86 0.00 0.16 1.8670
1.00 1.00 0.00 1.00 0.00 1.00 0.00 1.00 0.8619
8 SPE 95241

1.0
0.2

Pressure Difference (psi)

0.9
10 micron Fracture Width
0.8 0
Krw and Krow

0.7

-0.2
0.6 HD=Infinity Pc = 0
0.5 HD=5
HD=1.5
-0.4
0.4
HD=1
0.3
HD=0 -0.6
0.2

0.1
-0.8
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 -1
Wetting (water) Phase Saturation 0 1000 2000 3000 4000 5000
Time (days)
Figure 4. Oil pressure difference (M-F) of base case (Pc = 0) and
case with t = 10 microns.
Figure 1. Fracture relative permeabilities, for different values of
the dimensionless height, HD; numerical values are provided in
the Appendix.

Figure 5. Field oil production rates comparison, using different

Pcog curves.
Figure 2. Capillary pressure curves for different fracture half-
widths, with γ = 24 dynes/cm.

0.351 Pc(t=100micro m) Pc(t=20micro m)

Gas-Oil Capillary Pressure (psia)

0.301 Pc(t=10 micro m) Pc matrix

0.251

0.201

0.151

0.101

0.051

0.001
0 0.2 0.4 0.6 0.8 1

Saturation of Gas

Figure 6. Oil production rate, with H = 20 ft.

Figure 3. Matrix capillary pressures and fracture capillary

pressures for three fracture widths.
SPE 95241 9

-2
Figure 7. Oil production rate, with σ = 0.001 ft . Figure 10. Field pressure as a function of time.

Figure 8. Oil recovery factor (FOE = RF) as a function of time. Figure 11. Well water-cut as a function of time.

Figure 9. Water injection rate as a function of time.