SPE 101150

Two-Phase Relative Permeability Prediction: A Comparison of the Modified Brooks-

Corey Methodology with a New Carman-Kozeny Based Flow Formulation
P. Behrenbruch, SPE, and H.M. Goda, SPE, The U. of Adelaide

Copyright 2006, Society of Petroleum Engineers

parameters but involves relatively simple analytical
This paper was prepared for presentation at the 2006 SPE Asia Pacific Oil & Gas Conference expressions.
and Exhibition held in Adelaide, Australia, 11–13 September 2006.

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
Introduction
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to Unlike conventional core analysis, involving the measurement
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 of permeability and porosity, advanced (or special) core
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
analysis (SCAL) is more expensive and time consuming. For
for commercial purposes without the written consent of the Society of Petroleum Engineers is these reasons, the number of plugs used for SCAL for a
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 particular field tends to be limited, often not covering all
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.
existing depositional environments and flow zone units
adequately and leading to poor reservoir representation.
Abstract Prediction models provide a means to augment a data set
The prediction of relative permeability has been in the past and to validate the latter, allowing the characterization of a
and is currently a very active research area, with theoretical, reservoir. In a more abstract sense, capillary pressure and
experimental and empirical approaches under consideration. statistical models for relative permeability have their own
However, it is fair to say that the complexities of relative assumptions and rely purely on theory, making them less
permeability have to date eluded researchers and practitioners practical. Network models may be highly plug-specific. It is
alike, in that there is no universal formulation that is able to also difficult to translate results from experiments performed
predict two-phase relative permeability for the wide range of on standard core material, such as Berea sandstone, to other
rock and wettability characteristics observed. This paper types of reservoir rock. Purely empirical models, on the other
presents a new, generalised formulation, one that is truly hand, are perhaps the most widely used in the petroleum
predictive, and compares the same with the industry standard industry, but they may not be universally applicable. Finally,
– the modified Brooks-Corey (MBC) formulation. empirical models based on theoretical concepts tend to be
The MBC formulation is perhaps the most widely used, more satisfactory and may lead to greater universality(1).
practical method describing laboratory-derived relative Research in the area of single-phase flow and flow zone
permeability relationships in terms of simple power functions. unit identification have led to the conclusion that the
The shortcomings of this formulation are that it has no real theoretical Carman-Kozeny equation(2, 3) is an ideal
predictive capability and the relative contributions due to pore formulation for bridging the gap between the views of
structure components as compared to variation in wettability geologists and engineers. Good results obtained in previous
cannot be resolved. The new two-phase flow formulation research(4) utilizing the Carman-Kozeny (C-K) equation was
presented is based on a phenomenological approach related to then the impetus for investigating two-phase flow using a
the Carman-Kozeny equation, and is able to resolve the above modified C-K formulation. The modification required for a
mentioned shortcomings. two-phase formulation involved adaptation of certain
Included are several laboratory examples and the results of parameter groups, to allow the inclusion of the second phase.
a comparison of the two formulations is presented. It is shown The new model was tested using a large number of relative
how the new formulation is able to predict “the curvature” of permeability data sets which were generated by service
relative permeability curves when only the endpoints are laboratories, several hundred relationships, mainly for
known, duplicating observed behaviour from steady-state Australian fields.
relative permeability experiments. Alternatively, if the The purpose of the research described, part of a larger
endpoints can be derived, correlated or estimated with the use objective, is to compare the performance of a new mechanistic
of more fundamental data, the entire prediction of relative model, based on the Carman-Kozeny equation, with the
permeability is possible. Brooks and Corey model. Before presenting results from such
In conclusion, the formulation presented is able to predict comparison, a summary of respective formulations is given,
two-phase relative permeability under steady-state conditions, for the Modifed Brooks and Corey model and the new relative
not just merely fit data. The second advantage of this method permeability model.
is that it is theoretically based and does not involve any fitting
2 SPE 101150

The Modified Brooks and Corey Model (MBC) indication for a plug’s degree of homogeneity) was found to
In a previous study by the authors(5), background and change over a narrow range of 2.3 – 3.7(8).
equations for the modified Brooks and Corey model were As can be seen, Corey’s original model and the Brooks and
presented and discussed. Corey’s orginal model(6), presented Corey model have some theoretical foundation. However, the
here as in equations 1 and 2 was primarly based on laboratory most utilized model by the petroleum industry is the so called
results. modified Brooks and Corey (MBC) model(9, 10, 11), or the
power law model(9). This model is explicitly a function of
k ro = (S oe ) relative permeability endpoints, i.e. endpoint relative
4
(1)
permeability to oil at irreducible water saturation, and
endpoint relative permeability to water at residual oil
k rg = (1 − S oe ) 1 − S oe2
2
( ) (2) saturation. The model may be presented as given by equations
7 and 8, where a linear relationship between normalized phase
saturation and normalized relative permeability is defined on a
S o − S or log-log plot, asuming valid relative permeability
where S oe = (3)
1 − S or measurements.
and
n
⎛ 1 − S w − S or ⎞ o
So: Oil saturation
'
k ro = k ro (S on )n o = k ro' ⎜⎜ ⎟⎟ (7)
Sor: Residual oil saturation ⎝ 1 − S wir − S or ⎠
kro: Relative permeability to oil
krg: Relative permeability to gas n
' ⎛⎜ S w − S wir ⎞ w
Later, Brooks and Corey(7, 8) defined a general relationship that
'
k rw = k rw (S wn )nw = k rw ⎜ ⎟⎟ (8)
could correlate normalized oil saturation (equation 3) to ⎝ 1 − S wir − S or ⎠
capillary pressure and entry pressure. where,

λ kro: Oil relative permeability normalized to absolute plug

S oe = ⎛⎜ d ⎞⎟
P air permeability
(4)
⎝ Pc ⎠
krw: Water relative permeability normalized to absolute
where plug air permeability

Pd: Entry pressure kro’: End point relative permeability to oil normalized to
P c: Capillary pressure absolute plug air permeability
λ: Pore size distribution index
krw’: End point relative permeability to water normalized
Using λ and normalised oil saturation from equation 3, Brooks to absolute plug air permeability
and Corey introduced another general formulation for relative
permeability determination, equations 5 and 6. Son: Normalized oil saturation

2+ 3λ Swn: Normalized water saturation

k rw = (S e ) λ (5)
S w: Water saturation
2+ λ
2⎛ ⎞
k rnw = (1 − S e ) ⎜⎜1 − S e λ ⎟⎟ (6) Swir: Irreducible water saturation
⎝ ⎠ Sor: Residual oil saturation
where no: Corey exponent to oil
krw: Wetting phase relative permeability n w: Corey exponent to water
krnw: Non-wetting phase relative permeability
One of the applications of the MBC model is to smooth,
However, it should be noted that, equation 5 and 6 are suitable modify and extend an existing relative permeability
for the determination of drainage relative permeability. In relationship. The model can also be used to estimate a “true”
other words, these formulations may not be able to define Sor value, one that gives the best linear relationship when
imbibition relative permeability. plotting log Son vs log (kro/kro’). Another application of the
Brooks and Corey examined their model using different model is to support rules of thumb for a qualitative measure of
materials. In general, except for the glass beads, λ (which is an the the wettability of a rock sample(5, 12), perhaps the most
SPE 101150 3

influential parameter apart from saturation itself, affecting the Quality Index (RQI), Porosity Group (PG) and Flow Zone
shape of relative permeability curves(12). Based on a review of indicator (FZI).
an extensive data set for Australian basins, Goda and
k
Behrenbruch (2005) proposed a wettability matrix based on RQI = 0.0314 (11)
Corey exponents, see Table 1. φe
While the MBC model has many good features, it suffers
from a number of shortcomings. These shortcomings may be
summarized as follows: φe
PG = (12)
1 − φe
1. The model is a direct function of the endpoints, and
as such biased towards endpoint data, being less
influenced by the remaining data As such the method ⎛ 1 ⎞
is less suitable in actually validating a particular FZI = ⎜ ⎟ (13)
relative permeability data set. ⎜ FτS ⎟
⎝ s gv ⎠
2. In case one or both endpoints are not available, the
model fails to work properly. Substituting these definitions into equation 10, and taking
logarithms, a linar relationship with unit slope results, with a
3. The model cannot predict a relative permeability specific relationship defined for each separate geological
relationship, rather it is intended to smooth and entity, or flow zone unit, with a unique FZI index.
extend an existing relationship.
log RQI = log PG + log FZI (14)
The Carman-Kozeny Equation for Single-Phase Flow
and the Modified Carman-Kozeny Formulation for While this equation was initially postulated, the validity of
Two-Phase Flow (C – K Model) this equation could be confirmed by numerous studies,
The so called Carman-Kozeny equation is the result of demonstrating the applicability of the underlying assumptions
combining the early work conducted by Kozeny (1927)(2) and used in the establishment of the Carman-Kozeny equation. For
Carman (1937)(3), to study the problem of single phase flow. further information, the reader is referred to publication by
Kozeny (1927) studied the dependency between permeability Barr and Altunby (1992)(13) and Amaefule et al. (1993)(14).
and porosity. He originated the concept of a textural property, Recent applications may be found in Behrenbruch and
correlating factors such as shape factor, tortuosity, and Biniwale (2005)(15), and Biniwale and Behrenbruch (2004)(16).
specific surface area to permeability. Later, Carman (1937) Most recently, a semi-empirical model to predict relative
developed the direct relationship between permeability and permeability was established (Behrenbruch, 2006)(4). In
porosity. The Carman-Kozeny equation may be written as principle, the model can be considered as an extension of the
follows: Carman-Kozeny equation. In other words, parameters such as
φ e3 ⎛ 1 ⎞ reservoir quality index (equation 11) and porosity group
k= ⎜ ⎟ (9) (equation 12) were re-defined to allow the presence of a
(1 − φ e ) 2 ⎜ F τ 2S 2 ⎟
⎝ s gv ⎠ second phase. It was found that with extended definitions of
RQI and PG, relative permeability relationships take on the
where k is the absolute permeability in µm2 , φe is the effective form of a straight line when plotted in the two-phase Carman-
porosity, Fs is the shape factor (2 for circular openings), τ is Kozeny space. In final form, the model may be shown as
tortuosity and Sgv is the surface area per unit grain volume in follows:
µm-1.
2 3 2
While equation 9 may be used in estimating absolute 1014 m w φe S w ⎡ S w S wir ⎤
k rw = − (15)
permeability, the equation is often used to correlate properties ⎢ ⎥
for a fundamental geological unit (facies). The product Fs τ2
k ⎣1 − φ e S w 1 − φ e S wir ⎦
(known as Kozeny constant) is believed to be constant within
each single flow zone unit, and different from one zone to 2
1014 mo2 φ e3 (1 − S w ) ⎡ 1 − S w S or ⎤ (16)
another. This aspect was further studied by Barr and Altunby k ro = ⎢ − ⎥
(1992)(13) and later by Amaefule et al.(1993)(14), and equation k ⎣1 − φ e (1 − S w ) 1 − φ e S or ⎦
1 was proposed as a modified form, as follows:
In the above equations, mw and mo are the slopes of linear
relationships in the Carman-Kozeny space, generally thought
k φ e ⎛⎜ 1 ⎞
⎟ (10) to be representative of wettability.
0 .0314 =
φe (1 − φ e ) ⎜ Fs τ S gv ⎟ The validity of the model was tested by Behrenbruch
⎝ ⎠ (2006)(4) using a large number of data sets, mainly for
Australian fields. Two examples are given below (Figures 1
where 0.0314 is a conversion factor (millidarcy to µm2). and 2). As is the case with the MBC model, the new model
Furthermore, certain terms were grouped, defining Reservoir may also be used to calculate a new, more representative Sor
4 SPE 101150

for a relative permeability curve. An example that compares Simplifying and re-arranging equation 20:
both models in predicting a new (“true”) Sor is shown in Fig. 3
2
Comparison Studies 2 ⎛⎜1−φ ⎛⎜1−S ⎞⎟ ⎞⎟
krw ⎛ S −S ⎞ ⎝ or ⎠ ⎠ Sw
Due to the frequent employment of the modified Brooks and = ⎜⎜ w wir ⎟⎟ ⎝ (21)
Corey model (MBC) in smoothing and correcting relative '
krw ⎝ 1 − S − S
wir or ⎠ 1 − S or ⎛⎜1−φS ⎞⎟
2
permeability curves, this section presents a detailed ⎝ w ⎠
comparison study, carried out to investigate the relative
performance of both the new methodology to predict relative Taking a similar approach, the equation for oil may be
permeability (equations 15 and 16, also known as the new C-K derived:
relative permeability model) and the MBC model. Three
different approaches were explored. First of all, an attempt to 2
2 ⎛⎜1−φ⎛⎜1−S ⎞⎟⎞⎟
find an analytical expression that correlates both models is kro ⎛ 1− Sw − Sor ⎞ wir ⎠⎠ 1−Sw
presented. In other words, the aim was to investigate the = ⎜⎜ ⎟⎟ ⎝ ⎝ (22)
possibility of finding an analytical formulation connecting no
'
kro ⎝ 1− Swir − Sor ⎠ 1 −Swir (1−φ(1− Sw))2
(and/or nw) with mo (and/or mw). Secondly, a comparison was
made using synthetic data. Finally, a comparison was made
using real data. From equations 17 and 21,
In summary, the comparison studies investigated the
following approaches:
n −2
⎛ Sw − Swi ⎞ w
⎜⎜ ⎟⎟
(1−φ(1−Sor ))2 Sw
1. analytical equations = (23)
⎝ 1− Swir − Sor ⎠ 1−Sor (1−φSw)2
2. noise-free data

3. actual laboratory/field data

and for oil (equating equations 18 and 22):
1. Comparsion Study No. 1: Analytical Equations
In this approach, equations governing both models were 2
n −2 ⎛⎜1−φ⎛⎜1−S ⎞⎟⎞⎟
studied in attempt to define a linkage between them. The ⎛ 1− Sw − Sor ⎞ o ⎝ wir⎠⎠ 1−Sw
⎜⎜ ⎟⎟ =⎝ (24)
(1−φ(1− Sw))2
intention was to check the applicability of determining no (and 1−Swir
nw) as a function of mo (and mw) or vise versa. ⎝1− Swir − Sor ⎠
The MBC model may be summarised here for reference as
follows.
The fact that equations 23 and 24 do not contain mw and mo
confirms that mw and mo are specific constants for each plug.
n
k rw ⎛ S w − S wir ⎞ w Also, it is clear that nw and no are not constants and have
= ⎜⎜ ⎟⎟ (17) dependency on water saturation. In order to determine the
⎝ 1 − S wir − S or
'
k rw ⎠ range of variation for nw and no, a parametric study was
and similarly, for oil conducted. Table 2 indicates the range considered for values
of φ, Swir, and Sor. A number of combinations of variables was
n studied, as shown in Table 1. For each case, no and nw were
k ro ⎛ 1 − S w − S or ⎞ o then determined at the middle saturation values. Results
= ⎜⎜ ⎟⎟ (18) obtained for no and nw are shown in Figure 4 as a cross plot (no
⎝ 1 − S wir − S or
'
k ro ⎠ versus nw).
To further investigate the range for no and nw, two
For the new methodology, krw’ can be determined considering additional cases were considered, with their results presented
equation 15: in Figure 5:
2
' 1014mw φ (1- Sor ) ⎡ (1− Sor )
2 3
Swir ⎤ ƒ Porosity is equal to zero
krw = ⎢ − ⎥ (19)
ka ⎣1− φ(1− Sor ) 1− φSwir ⎦ ƒ Porosity is unity
Dividing equation 15 and 19, and re-arranging:
The approximate range observed for no and nw is 2.5 – 3.4,
increasing to 7 if porosity is assumed to be unity. The same
( )( ) ( )( )⎞⎟2⎛⎜ (1−φ(1− Sor))(1−φSwir) ⎞ (20)
2 range was also evident when other saturation values were
krw Sw ⎛ Sw 1−φSwir − Swir 1 −φSw
= ⎜ ⎟
'
krw (1− Sor ) ⎜
⎝ ( )(
1− φSw 1−φSwir ) ⎟ ⎜ (1− S )(1− φS ) − (S )(1 −φ(1 − S )) ⎟
⎠ ⎝ or wir wir or ⎠ considered.
It is noteworthy that this range also corresponds
approximately to the range of λ investigated by Brooks and
Corey in their original model(9), for the case of homogenous
SPE 101150 5

plugs. In their paper, λ was an indication of the degree of plug 3 Comparison Study No. 3: Actual Laboratory/Field Data
heterogeneity (sorting). Hence, no and nw may be considered In order to obtain a more comprehensive coverage, and as a
as measures of plug heterogeneity. In other words, higher final test, the comparison between the two formulations was
values for no and nw may be due to plugs being more extended to actual measured data on a variety of core plugs.
heterogenous (or faulty experimental procedure). Some Three samples, two from the Bonaparte Basin, and one
researchers have also indicated that the value of these from the Carnarvon basin are presented here. For such
exponents may be influenced by wettability(5, 17). comparisons, the intermediate saturation range was fitted in
Finally, no direct relationship could be established case of the C-K relative permeability formulation, with the
between no (and nw) and mo (and mw). aim to demonstrate the ability of the new model to predict
endpoints. For each sample, profiles are plotted on both
2. Comparison Study No. 2: Noise-Free Data normal and semi-log scale. It may be concluded that (see
To further show similarities and differences between the Figures 19 through 24):
models, a number of experimental data sets were studied. In
this approach the new methodology is used to predict relative
1. Good agreement exists between results obtained from
permeability curves from experimental measurements. The
the MBC and the new methodology models.
predicted curve is then believed to be accurate and may be
taken as a synthetic curve that is noise free. The Modified
2. Endpoint relative permeabilities predicted from the
Brooks and Corey model was then used to fit the data. Four
new methodology demonstrate slight deviation from
different profiles were compared:
measured values. This is not the case with MBC, as
the latter uses endpoints as input and does not predict
1. Profile from the new methodology (synthetic curve).
these.
2. Profile from the MBC model implementing no and nw
as constants (conventional model, no and nw are
obtained from a log-log plot). Conclusions
1. A comparison study was conducted between a new,
3. Profile from the MBC model implementing nw and no Carman-Kozeny based relative permeability model
as constants determined by averaging nw and no, and the modified Brooks and Corey model. Three
obtained from equations 23 and 24 at different comparisons were made: mathematical, using noise-
saturation values. free data and using actual lab/field data.

4. Profile from the MBC model implementing nw and no 2. No direct relationship exists between no and nw
as variables determined from equations 23 and 24, (Corey exponents) and mo and mw (characteristic
corresponding to different saturation values. slopes for the new model). Also, no and nw are not
constants in equations comparing the formulations,
but a function of Sw. It was found that no and nw
Three examples from the Bonaparte Basin are presented. For
change over the range 2.6 – 3.5. This range is similar
reference, Figure 6 is a location map showing major
to the range across which λ (in the Brooks and Corey
Australian Hydrocarbon basins(18). The result for each example
model for relative permeability) changes. This would
is shown on a cartesian plot and a semi-log plot (Figures 7
tend to support the view that Corey exponents are a
through 18). In these figures, the “New Model” refers to the
measure of the degree of heterogeneity of plugs,
new C-K relative permeability methodology and “MBC”
rather than wettability.
signifies the modified Brooks and Corey model.
3. Other comparisons between the two models showed
From these figures and profiles the following may be very good performance of the C-K model, where the
observed: latter has the advantage of being a predictive tool,
1. A very good match is indicated between profiles rather than just fitting the data.
generated from the new methodology and the MBC
model using average no and nw obtained from Acknowledgments
equations. The authors would like to gratefully acknowledge the
sponsors: BHP Billiton, Chevron, Santos Ltd and Woodside
2. When using values of nw and no determined from Energy for their financial support and permission to publish
equations 23 and 24 at different saturations, the their data and the presented analysis results. The Australian
profile became identical to that of the new Postgraduate Award, APA, granted by the Australian
methodology, as could be expected. Government to the second author is also acknowledged.
6 SPE 101150

References 12. Stiles, J. H. and Hutfilz, J. M.: “The Use of Routine and Special
1. Honarpour, M., Koederitz, L. F., and Harvey, A. H.: Relative Core Analysis in Characterizing Brent Group Reservoirs, U. K.
Permeability of Petroleum Reservoir, CRC Press, Inc, Boca North Sea,” JPT, pp. 704, (June, 1992)
Raton (1986) 13. Barr, D.C., Altunbay, M.: “Identifying hydraulic units as an aid
2. Kozeny, J.: “Uber Kapillare Leitung des Wassers im Boden, to quantifying depositional environments and diagenetic facies.
Sitzungsberichte,” Royal Academy of Science, Vienna (1927) Geology of Malaysia, Symp. Res. Eval./Form. Damage, Kuala
3. Carman, P. C.: “Fluid Flow Through Granular Beds,” AIChE Lumpur, Malaysia, (1992)
(1937) 15, 150 14. Amaefule, J.O., Altunbay, M., Tiab, D., Kersey, D.G., Keelan,
4. Behrenbruch, P.: “Two-Phase Relative Permeability Prediction: D.K.:Enhanced reservoir description: using core and log data to
A New Semi-Empirical Model Based on a Modified Carman- identify hydraulic (flow) units and predict permeability in
Kozeny Equation.”, to be published uncored intervals/wells. SPE 26436, presented at the Annu.
5. Goda, H. and Behrenbruch, P.: ”Using a modified Brooks- Tech. Conf. Exhib., Houston, Tex. (1993)
Corey Model to Study Oil-Water Relative Permeability for 15. Behrenbruch, P. and Biniwale, S.: “Two-Phase Relative
Diverse Pore Structures,” SPE 88538, presented at the SPE Permeability Prediction: A New Semi-Empirical Model Based
Asia-Pacific Annual Technical Conference and Exhibition, on a Modified Carman-Kozeny Equation.” Journal of
Perth, Australia, 18 – 20 Oct, (2004) Petroleum Science andEngineering, Vol. (47), pp. 175 – 196,
6. Corey, A.T.: “The Interrelation Between Gas and Oil Relative (2005)
Permeability,” Producers Monthly (1954) 16. Biniwale, S. and Behrenbruch P.:”The Mapping of Hydraulic
7. Brooks, R. H., and Corey, A. T.: “Hydraulic Properties of Flow Zone Units and Characterization of Australian Geological
Porous Media,” Hydrology Papers, Colorado State University, Depositional Environments,” SPE 88521, to be presented at
Colorado (1964) The Asia Pacific Oil and Gas Conference and Exhibition, Perth,
8. Brooks, R. H., and Corey, A. T.: “Properties of Porous Media Australia, Oct. 18 –20, (2004)
Affecting Fluid Flow,” Journal of Irrigation and Drainage 17. Stiles, J.: “Using Special Core Analysis in Reservoir
Division, (June 1966) 61. Engineering – Relative Permeability & Capillary Pressure,”
9. Lake, L. W.: Enhanced Oil Recovery, Chapter 3, Prentice Hall, Course Notes, England, (1994)
Inc. (1989) 18. Ruth, P. V.: “Overpressure in the Cooper and Carnarvon
10. Embid, S. M.: “Modelling Capillary Pressure and Relative Basins, Australia,” PhD Dissertation, the University of
Permeability for Systems with Heterogeneous Wettability,” Adelaide, Australia, (2003).
Ph.D. Dissertation, The University of Texas at Austin (1997)
11. Alpak, F. O., Lake, L. W. and Embid, S. M.: “ Validation of a
Modified Carman-Kozeny Equation to Model Two-Phase
Relative Permeability,” SPE 56479, Presented at the Annual
Technical Conference and Exhibition, Texas, Oct. 3 – 6 (1999)

Table 1: Corey Exponents Matrix for Wettability – after Goda and Behrenbruch (2004)

Corey exponent Corey exponent

Wettability Class
to water to oil

Oil wet 2-3 6-8

Slightly oil wet 2-4 2-6

Slightly water wet 4-6 2-6

Water wet 6-8 2-4

Table 2: Parametric Study: Ranges for φ, Swir and Sor

Parameters Minimum (%) Maximum (%)

Porosity (φ) 5 35

Irreducible water saturation (Swir) 10 60

Residual oil saturation (Sor) 10 40

SPE 101150 7

1.0
kro (measured) kabs = 5.07 md
0.9
krw (measured) φ = 15.1%
0.8 kro (predicted) Swir = 45.1%
0.7 krw (predicted) kom = 2.72 md
Sor = 14.1%
0.6
0.5
kr

0.4
0.3
0.2
0.1
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Sw

Fig. 1: Measured Versus Predicted Relative Permeability Curve – Sample A

1.0
kro (measured) kabs = 481 md
0.9
krw (measured) φ = 17.4%
0.8
kro (predicted) Swir = 12.2%
0.7
krw (predicted) kom = 384 md
0.6
Sor = 22.7%
0.5
kr

0.4
0.3
0.2
0.1
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Sw

Fig. 2: Measured Versus Predicted Relative Permeability Curve – Sample B

8 SPE 101150

Sw
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
1

0.1
kr

0.01

krw (MBC)
0.001 kro (MBC)
krw (New Model)
kro (New Model)
krw (Experimental Data)
kro (Experimental Data)
0.0001

Fig. 3: “True” Sor Estimated from MBC and the New Model

4
no

1
1 2 3 4 5 6 7
nw

Fig. 4: Range of Distribution for no and nw as Obtained from Equations 23 and 24

SPE 101150 9

7
Results obtained for the range specified
Results for porosity = 0.0
6 Results for porosity = 1.0

4
no

1
1 2 3 4 5 6 7 8 9
nw

Fig. 5: Range of Distribution for no and nw as Obtained from Equations 23 and 24 for Porosity = 0.0 and
Porosity = 1.0
ga
an
om
Er

COOPER
BASIN

Fig. 6 – Major Australian Hydrocarbon Basins - Modified After Ruth, 2003

10 SPE 101150

1.0
kabs = 1411 md krw (new Model)
0.9 kro (new Model)
φ = 15.7%
krw (MBC)
0.8 Swir = 10.2% kro (MBC)
kom = 1091 md krw (Average nw )(Eq. 4.34))
0.7 kro (Average no )(Eq. 4.35))
Sor = 22.6%
0.6
kr

0.5

0.4

0.3

0.2

0.1

0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Sw

Fig. 7: Synthetic Data: Comparison between Different Profiles for Sample 1 – Bonaparte Basin (Normal Scale, Average
no and nw)

Sw
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
1.0E +00

1.0E -01
kabs = 1411 md
φ = 15.7%
1.0E -02 Swir = 10.2%
kr

kom = 1091 md
Sor = 22.6%
1.0E -03
krw (new Model)
krok(new
rw (new methodology)
Model)
k
krw ro (new
(MBC) methodology)
1.0E -04 k (MBC)
kro (MBC)
rw

krwkro (MBC) n )(Eq. 4.34))

(Average w
krw (from equation)
kro (Average no )(Eq. 4.35))
kro (from equation)
1.0E -05

Fig. 8: Synthetic Data: Comparison between Different Profiles for Sample 1 – Bonaparte Basin (Semi-log Scale,
Average no and nw)
SPE 101150 11

1.0
kabs = 1411 md krw (new Model)
0.9 kro (new Model)
φ = 15.7%
krw (MBC)
0.8 Swir = 10.2%
kro (MBC)
kom = 1091 md krw (Varying nw )(Eq. 4.34))
0.7
Sor = 22.6% kro (Varying no )(Eq. 4.35))
0.6
kr

0.5

0.4

0.3

0.2

0.1

0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Sw
Fig. 9: Synthetic Data: Comparison between Different Profiles for Sample 1 – Bonaparte Basin (Normal Scale, Varying no
and nw)

Sw

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
1.0E +00

1.0E -01

kabs = 1411 md
1.0E -02
φ = 15.7%
kr

Swir = 10.2%
kom = 1091 md
1.0E -03
Sor = 22.6%
krw (new Model)
kro (new Model)
1.0E -04 krw (MBC)
kro (MBC)
krw (Varying nw)(Eq. 4.34))
kro (Varying no )(Eq. 4.35))
1.0E -05
Fig. 10: Synthetic Data: Comparison between Different Profiles for Sample 1 – Bonaparte Basin (Semi-log Scale, Varying
no and nw)
12 SPE 101150

1.0
kabs = 1345 md krw (new Model)
0.9 kro (new Model)
φ = 15.3%
krw (MBC)
0.8 Swir = 9.6%
kro (MBC)
kom = 719 md krw (Average nw )(Eq. 4.34))
0.7
Sor = 19.8% kro (Average no )(Eq. 4.35))
0.6
0.5
kr

0.4
0.3
0.2
0.1
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Sw
Fig. 11: Synthetic Data: Comparison between Different Profiles for Sample 2 – Bonaparte Basin (Normal Scale, Average no
and nw)

Sw
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
1.0E+00

1.0E-01

1.0E-02

1.0E-03
kr

1.0E-04

1.0E-05
kabs = 1345 md krw (new Model)
φ = 15.3% kro (new Model)
1.0E-06 krw (MBC)
Swir = 9.6%
kro (MBC)
1.0E-07 kom = 719 md krw (Average nw)(Eq. 4.34))
Sor = 19.8% kro (Average no )(Eq. 4.35))
1.0E-08
Fig. 12: Synthetic Data: Comparison between Different Profiles for Sample 2 – Bonaparte Basin (Semi-log Scale,
Average no and nw)
SPE 101150 13

1.0
kabs = 1345 md krw (new Model)
0.9 kro (new Model)
φ = 15.3%
krw (MBC)
0.8 Swir = 9.6%
kro (MBC)
0.7 kom = 719 md krw (Varying nw) (Eq. 4.34))
Sor = 19.8% kro (Varying no)(Eq. 4.35))
0.6
0.5
kr

0.4
0.3
0.2
0.1
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Sw
Fig. 13: Synthetic Data: Comparison between Different Profiles for Sample 2 – Bonaparte Basin (Normal Scale, Varying
no and nw)

Sw
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
1.0E+00

1.0E-01

1.0E-02

1.0E-03
kr

1.0E-04
kabs = 1345 md krw (new Model)
1.0E-05 φ = 15.3% kro (new Model)
Swir = 9.6% krw (MBC)
kro (MBC)
1.0E-06 kom = 719 md
krw (Varying nw)(Eq. 4.34))
Sor = 19.8% kro (Varying no )(Eq. 4.35))
1.0E-07
Fig. 14: Synthetic Data: Comparison between Different Profiles for Sample 2 – Bonaparte Basin (Semi-log Scale,
Varying no and nw)
14 SPE 101150

1.0
kabs = 557 md krw (new Model)
0.9 φ = 16.6% kro (new Model)
Swir = 18.7% krw (MBC)
0.8 kro (MBC)
kom = 477 md krw (Average nw )(Eq. 4.34))
0.7
Sor = 18.9% kro (Average no)(Eq. 4.35))
0.6

0.5
kr

0.4

0.3

0.2

0.1

0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Sw
Fig. 15: Synthetic Data: Comparison between Different Profiles for Sample 3 – Bonaparte Basin (Normal Scale,
Average no and nw)

Sw
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
1.00E+00

1.00E-01
kabs = 557 md
1.00E-02 φ = 16.6%
Swir = 18.7%
1.00E-03 kom = 477 md
kr

Sor = 18.9%
1.00E-04
krw (new Model)
1.00E-05 kro (new Model)
krw (MBC)
1.00E-06 kro (MBC)
krw (Average nw )(Eq. 4.34))
kro (Average no )(Eq. 4.35))
1.00E-07

Fig. 16: Synthetic Data: Comparison between Different Profiles for Sample 3 – Bonaparte Basin (Semi-Log Scale,
Average no and nw)
SPE 101150 15

1.0
kabs = 557 md krw (new Model)
0.9 φ = 16.6% kro (new Model)
krw (MBC)
0.8 Swir = 18.7%
kro (MBC)
kom = 477 md krw (Varying nw)(Eq. 4.34))
0.7
Sor = 18.9% kro (Varying no)(Eq. 4.35))
0.6

0.5
kr

0.4

0.3

0.2

0.1

0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Sw
Fig. 17: Synthetic Data: Comparison between Different Profiles for Sample 3 – Bonaparte Basin (Normal Scale, Varying no
and nw)

Sw
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
1.00E+00

1.00E-01
kabs = 557 md
φ = 16.6%
1.00E-02
Swir = 18.7%
kom = 477 md
1.00E-03
kr

Sor = 18.9%

1.00E-04 krw (new Model)

kro (new Model)
krw (MBC)
1.00E-05 kro (MBC)
krw (Varying nw )(Eq. 4.34))
kro (Varying no )(Eq. 4.35))
1.00E-06

Fig. 18: Synthetic Data: Comparison between Different Profiles for Sample 3 – Bonaparte Basin (Semi-Log Scale, Varying
no and nw)
16 SPE 101150

1
kabs = 1345 md kro (measured)
0.9 krw (measured)
φ = 15.3%
0.8 Swir = 9.6%
kro (MBC)
krw (MBC)
0.7 kom = 719 md
kro (New Model)
Sor = 19.8%
0.6 krw (New Model)

0.5
kr

0.4

0.3
0.2
0.1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Sw
Fig. 19: Real Data: Comparison between New Model and MBC for Sample 1 – Bonaparte Basin (Normal Scale)

Sw
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
1.00E+00

1.00E-01
kr

1.00E-02
kro (measured)
kabs = 1345 md krw (measured)
1.00E-03 φ = 15.3% kro (MBC)
Swir = 9.6% krw (MBC)
kom = 719 md kro (New Model)
Sor = 19.8% krw (New Model)
1.00E-04

Fig. 20: Real Data: Comparison between New Model and MBC for Sample 1 – Bonaparte Basin (Semi-Log Scale)
SPE 101150 17

1
kro (measured) kabs = 1411 md
0.9
krw (measured) φ = 15.7%
0.8 kro (MBC) Swir = 10.2%
0.7 krw (MBC) kom = 1091 md
kro (New Model)
0.6 Sor = 22.6%
krw (New Model)
0.5
kr

0.4
0.3
0.2
0.1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Sw
Fig. 21: Real Data: Comparison between New Model and MBC for Sample 2 – Bonaparte Basin (Normal Scale)

Sw
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
1.00E+00

1.00E-01
kr

1.00E-02
kro (measured)
kabs = 1411 md
krw (measured)
kro (MBC) φ = 15.7%
1.00E-03
krw (MBC) Swir = 10.2%
kro (New Model) kom = 1091 md
krw (New Model) Sor = 22.6%
1.00E-04

Fig. 22: Real Data: Comparison between New Model and MBC for Sample 2 – Bonaparte Basin (Semi-Log Scale)
18 SPE 101150

1
kabs = 2.92 md kro (measured)
0.9
φ = 26.0% krw (measured)
0.8 Swir = 66.1%
kro (MBC)
krw (MBC)
0.7 kom = 1.2 md
kro (New Model)
Sor = 12.6%
0.6 krw (New Model)
kr

0.5
0.4

0.3
0.2

0.1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Sw
Fig. 23: Real Data: Comparison between New Model and MBC for Sample 1 – Carnarvon Basin (Normal Scale)

Sw

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

1.00E+00

1.00E-01
kr

1.00E-02

kro (measured) kabs = 2.92 md

krw (measured)
φ = 26.0%
1.00E-03 kro (MBC)
Swir = 66.1%
krw (MBC)
kro (New Model) kom = 1.2 md

krw (New Model) Sor = 12.6%

1.00E-04
Fig. 24: Real Data: Comparison between New Model and MBC for Sample 1 – Carnarvon Basin (Semi-Log Scale)