SPE-188260-MS

Innovative Workflow for Grouping, Averaging, End-Point Scaling and

Assessing Uncertainty of Water-Oil Relative-Permeability Curves,
Considering Corresponding Normalized Water-Fractional-Flow Curves,
Reservoir-Rock Types and Wettability Indexes

F. C. Ferreira, F. Pairoys, D. V. Klemin, B. Baser-Langenau, S. Safonov, O. Dinariev, N. Evseev, M. D. Silva, and

G. Da Fraga Carneiro, Schlumberger

Copyright 2017, Society of Petroleum Engineers

This paper was prepared for presentation at the Abu Dhabi International Petroleum Exhibition & Conference held in Abu Dhabi, UAE, 13-16 November 2017.

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
Relative permeability is a key input for multiphase reservoir simulations. Challenges related to the validation
and interpretation of the laboratory core measurements are associated with the restoration processes and
resulting wettability states, the heterogeneities and multi-scale aspects of complex rocks, as well as the
limitations of core flooding experiments. Moreover, the relative permeability curves from several samples
can be scattered and their correlation with wettability and reservoir-rock types not directly apparent.
Grouping, averaging, end-point scaling and assessing the data uncertainty are crucial steps in relative
permeability data processing. To improve these processes, a new workflow is proposed, based on the water
fractional flow concept (fw), which is an effective representation of the behavior of oil displacement by water
injection, combining both relative permeability curves to oil and water into a single, equivalent curve.
First, the water fractional flow curve, obtained from a relative permeability core flooding test, is
normalized according to saturation end-points at a constant viscosity ratio equal to one. Such normalization
allows the separation of the fw plot area into two regions according to the wettability state of the samples.
Fractional flow curves for the same sample but at opposite wettability conditions, i.e. strongly oil-wet or
water-wet, present a remarkable symmetry, from which a wettability index is calculated. This proposed
new wettability index may then be compared to other indexes like Amott-Harvey or USBM for validation.
Additionally, the shape of the normalized water fractional flow curves is influenced by rock-pore sizes.
Subsequently, the normalized fractional flow curves are grouped by wettability and by reservoir-rock
type, supporting the validation of the relative permeability data and identifying associated trends and
uncertainties. The average, lower and upper-bound normalized fractional flow curves are obtained for each
group. Likewise, relative-permeability and saturation end-points are correlated with reservoir-rock-type
index or other rock properties. Finally, average, lower and upper-bound normalized relative permeability
sets of curves and corresponding end-points may be used for reservoir simulation. Alternatively, de-
normalized relative permeability curves can be obtained.
2 SPE-188260-MS

By varying wettability, a controlled relative permeability dataset is obtained using direct-hydrodynamic

(DHD) simulations on 3D digital rock model of a carbonate core sample. The proposed workflow is applied
to such a dataset. The results confirm the ability of the method to correctly identify the different wetting
states and to group the fractional flow curves accordingly.
The proposed wettability index, directly obtained from relative permeability data, may be complementary
to other industry wettability indexes and better represent the expected displacement behavior. The proposed
workflow, although simple and widely applicable, considerably improves the relative permeability analysis
process. It can be integrated with other core analysis, well-log analysis and digital-rock analysis workflows.

Introduction
Capillary effects in multiphase systems result from the combination of cohesion forces inside each phase and
adhesion forces between the different phases (Abdallah et al., 2007). In particular, when immiscible fluids
like water and oil flow through small pores of reservoir rocks at low flow rates, capillary effects determine
multiphase flow behavior. Cohesion forces result in interfacial tension acting through a meniscus separating
the immiscible fluid phases. Adhesion forces are present between each fluid and the surrounding pore walls.
The balance of these cohesion and adhesion forces will define the wettability of reservoir rock towards one
of the fluids, and the corresponding contact angle between the meniscus and the pore walls. The smaller
the pores, the stronger are such capillary effects. Capillary effects will assist or hinder the movement of one
fluid phase compared to the other. Capillary effects are represented in reservoir flow simulation by relative
permeability and capillary pressure curves. Besides wettability, such experimental data are also significantly
influenced by pore-space geometry, rock heterogeneity, experimental conditions and procedures; they are
therefore subject to large uncertainties.
Multiple challenges are related to the validation, analysis and processing of relative permeability data, e.g.
wettability preservation or restoration processes and resulting wettability states, heterogeneities, multi-scale
aspects of complex rocks, and limitations of waterflood experiments. After quality checking and processing
the experimental data, the relative permeability results are usually mapped into a reservoir simulation
model. Frequently, several realizations of the simulation model are prepared for sensitivity studies. For
these reasons, upscaling, grouping, averaging, normalizing, end-point scaling and assessing the uncertainty
envelope of relative permeability data are key strategies to build a reservoir simulation model.
Absolute permeability k is defined by Darcy’s equation for a single phase fluid completely saturating
a porous media. When two or more fluids share the porous space, each phase will present a different
saturation and a different effective permeability. Relative permeability is defined for each fluid as the ratio
between the effective and chosen reference permeability, e.g. Klinkenberg-corrected absolute permeability
or effective permeability to oil at irreducible water saturation condition. During the displacement of a fluid
phase by another, saturation changes. The relative permeability of each fluid phase will therefore also
change, resulting in a relative permeability versus saturation curve. Due to capillary effects, the impact
of saturation change on relative permeability is different for wetting and non-wetting fluids. Each pair of
curves presents saturation end-points, e.g. the irreducible water saturation and the residual oil saturation.
Each curve also presents relative permeability end-points, e.g. the oil initial relative permeability and the
water final relative permeability. Relative permeability curves are different if the displacement fluid is the
wetting or the non-wetting fluid, and also depend on the saturation history – this phenomenon results from
successive drainage and imbibition cycles and is known as hysteresis. Relative permeability is a key input
for multiphase reservoir simulations and its impact on hydrocarbon recovery is usually very significant.
During the preparation of a dynamic reservoir simulation model, relative permeability data need to
be correctly verified, analyzed and processed for each rock-fluid region, also called saturation function
region (SATNUM), as described by Schlumberger (2016). Such rock-fluid regions are frequently defined
by combining reservoir-rock types (RRT) with reservoir compartments containing fluids of different
SPE-188260-MS 3

compositions. Therefore, different SATNUM may present different wettability behavior. RRT’s are typically
defined by classes of reservoir quality indexes, e.g. RQI, FZI and Winland, which are closely associated
to pore-throat or pore-body sizes.
Because of the uncertainty usually associated with relative permeability data, several pairs of relative
permeability curves should be available for each SATNUM. Such data are typically from different wells
and depth intervals, frequently obtained at different experimental conditions. After a proper quality check,
such relative permeability curves should be analyzed and processed to obtain a single pair of curves for
each SATNUM.
Two main different approaches may be used: fixed relative permeability curves; or normalized relative
permeability curves that incorporate distributions of saturation and relative permeability end-points, through
an end-point scaling (EPS) algorithm, as described by Schlumberger, (2016). EPS is recommended if proper
correlations are found between end-points and other reservoir properties such as absolute permeability or
RRT indexes.
Several pairs of relative permeability curves may be averaged to obtain a single pair of curves for each
SATNUM. Before the averaging, it is recommended to group the different pairs of curves and exclude
eventual outlier curves after a proper quality check. If a large variability is observed in relative permeability
data, a corresponding uncertainty envelope should also be assigned to each SATNUM, to allow reservoir
simulation sensitivity studies.
Wettability is one of the key parameters to understand relative permeability behavior and resulting
reservoir fluid flow during displacement processes. The correct wettability state, before and during the
duration of experiments, should result from experimental conditions closely matching the reservoir rock
mineralogy, in-situ fluid compositions and reservoir conditions. Therefore, samples should be submitted to
proper wettability preservation or restoration (aging after cleaning) processes. Unfortunately, such processes
are usually subjected to large uncertainties. Experimental procedures may limit the ranges of pore pressure,
temperature, fluid viscosity ratio and injection rate. As important as the core cleaning and restoration
processes, is the reservoir fluid preservation. Using synthetic or dead oil from the field, instead of preserved
live oil, may significantly alter the wettability restoration results. Oil polar components, like asphaltenes,
play an important role on wettability restoration (Amroun et al., 2001). Asphaltenes, if flocculated or
precipitated during fluid cooling or depressurization, may not go back into solution even if the fluid sample
is re-warmed and re-pressurized (Al-Qasim and Bubshait, 2017). Asphaltenes or other polar components
that came out of solution and became flocculated and dispersed in the oil sample, may artificially facilitate
the wettability change (Mullins, 2010 and Pfeifer et al., 2017). On the other hand, if recombination samples
are used, such asphaltenes may have deposited on tube walls or settled down in vessels, resulting in a final
fluid sample with low asphaltene content, and therefore, a low potential for wettability change. Façanha
et al. (2016) showed that sour gas components like CO2 and H2S may also significantly change the final
wettability of a saturated sample towards a more neutral state. Even in the case when a preserved live oil
sample is used, the duration of the wettability restoration process may not be enough to fully achieve the
final wettability state.
Although numerical reservoir simulators usually do not directly consider wettability parameters,
modelling relative permeability from multiple SCAL experiments requires a quality check that groups
results with similar wettability states. However, the relative permeability curves from several samples are
often scattered and their correlation with wettability and reservoir-rock types may not be directly apparent.
Usually, a few rules of thumb are applied to estimate wettability from water-oil relative permeability
experiments, e.g. the one based on the irreducible water saturation, the saturation at which the curves cross
and the oil and water relative permeability end-point values (Craig, 1971 and Anderson, 1987). Quantitative
estimates for wettability may be obtained by Amott-Harvey and USBM processes and their corresponding
wettability indexes, from water-oil capillary pressure curves obtained during drainage and imbibition cycles.
4 SPE-188260-MS

Such results may be obtained from different samples, at different wettability conditions. They also may not
completely represent the water-oil relative permeability behavior, especially in complex, multimodal rocks.
In such rocks, the relative permeability may be largely influenced by the pore system with larger pore sizes.
A large spread of relative permeability curves and corresponding saturation and relative permeability end-
points is usually observed, making it difficult to distinguish their behavior according to different wettability
states and RRT. Consequently, a more elaborate workflow is required to estimate the wettability state of the
rock from relative permeability data.
In the process of modeling oil-water relative permeability, other aspects besides wettability state can
also influence the results. In complex rocks, the multi-scale heterogeneities and the corresponding larger
representative element volumes (REV) may require further upscaling of relative permeability data obtained
from core-plug experiments. Relative permeability experiments may be obtained from both steady (SS)
and unsteady-state (USS) experiments. SS experiments data come from direct measurements with a whole
coverage of the saturation range, but such experiments are very time consuming. On the other hand,
the quick USS experiments only cover the saturation range after the water (or gas) breakthrough and
requires analytical or numerical methods to compensate for uneven saturation distribution, history match the
experimental raw data and obtain the resulting relative permeability curves (Johnson et al., 1959). Besides
the average pore sizes, approximately represented by reservoir rock types (RRT), relative permeability
curves are also influenced by porous-space topology, fluid interfacial tension, experimental injection rates,
fluids’ viscosity and corresponding capillary numbers.
Therefore, there are usually large uncertainties associated with relative permeability data. Such
uncertainties should be properly considered to create multiple realizations of the reservoir model and run
the corresponding sensitivity analysis. Furthermore, the relative permeability data need to be associated to
the SATNUM to populate the reservoir model with multi-phase dynamic properties. Consequently, a more
elaborate workflow is required to group, average and scale the relative permeability curves, and to assess
their corresponding uncertainty.
We propose an innovative workflow based on the water fractional flow curve (fw) concept (Dake, 1978),
which was first presented by Buckley-Leverett (1942) and Welge (1942). fw curves are one of the best and
simplest representation of the behavior of oil displacement by water injection. For this purpose, applying
the (fw) concept results in replacing a pair of relative permeability curves by a single equivalent curve. fw
curves cover the whole saturation range – from irreducible water saturation (Swirr) to residual oil saturation
to water (Sorw). They may also roughly estimate the time for water breakthrough and the linear displacement
efficiency of water injection. Agbalaka et al. (2008) had suggested that fw might be useful to assess the
wettability state of a sample during a water-oil relative permeability experiment.
The ordinary definition for water-fractional flow is:

(1)

where qw is the outlet water flow rate and qt is the total outlet flow rate at reservoir conditions (qw + q0), krw is
the relative permeability to water and krow is the relative permeability to oil at a certain saturation condition,
μw and μo are respectively water and oil viscosities.
The relationship between fw and relative permeabilities, represented by the Equation (1), considers that
the oil displacement is dominated by viscous forces and consequently disregards capillary forces, as shown
by Schmid and Alyafei (2016). Likewise, most of the methods to derive relative permeabilities curves
from experimental fractional flow data do not consider capillary forces, e.g. JBN method for unsteady-state
experiments (Deng and King, 2016).
SPE-188260-MS 5

Wettability, interfacial tension, pressure gradient during the experiment, pore-throat, pore-body
size distributions and other rock-topology parameters strongly affect water-oil relative permeability,
corresponding end-points and water fractional flow (fw). In the reservoir, wettability may significantly affect
breakthrough time, oil displacement efficiency and economic recovery factor.
The method proposed in this paper is supported by Sinha et al. (2011) results. They simulated an
oil displacement by water, with viscosity ratio equal to one, in a 2D pore-network model, considering
a randomly distributed, fractional-dalmatian wettability (Abdallah, 2007). They varied the fraction of
wettability from 100% water-wet to 100% oil-wet conditions. They also varied the capillary number Ca
from 10-2 and 10-3. Ca represents the ratio between viscous and capillary forces in a porous media. At this
capillary number range, flow behavior is dominated by viscous forces. No saturation end-points are reported,
probably due to the high capillary numbers and the 2D synthetic model used in their simulation cases. Their
results are illustrated in Figure 1.

Figure 1—Free illustration of Sinha et al. (2011) results, showing the symmetrical behavior of water-
fractional-flow curve (fw) when wettability changes from 100% water-wet (ω = 0) to 100% oil-wet state (ω =
1). It also shows the linear variation of the fw curves with the wettability fraction (ω = 0.25, 0.50 and 0.75).

They found that, considering a neutral line in which fw = Sw, fw curves show a symmetrical behavior when
wettability changes from 100% water-wet to 100% oil-wet state. Furthermore, they found that fw varies
linearly with the fraction ω of the rock surface that is strongly oil-wet (varying from 0 to 1):

(2)

Sinha et al. (2011) considered randomly distributed dalmatian wettability. However, a homogeneous
moderate wettability (corresponding to an intermediate contact angle θ) and other wettability distributions
should also be considered. Flovik et al. (2015) have extended their previous work by considering
homogeneous wettability and varying contact angles from strongly water-wet to strongly oil-wet. They
found symmetric results similar to those shown in Figure 1.
6 SPE-188260-MS

Sinha et al. (2011) also found that fw curves get closer to the neutral line as the capillary number Ca
increases as illustrated in Figure 2.

Figure 2—Free illustration of Sinha et al. (2011) results, showing the effect of the capillary number on the shape of fw curves.

The capillary number represents the balance between viscous and capillary forces and is defined as:
(3)
where µw is the water viscosity, Vw is the velocity of water flowing through the pores and γow is the water-oil
interfacial tension. Capillary number has been frequently used to successfully explain reduction in residual
oil saturation (Sorw) during desaturation experiments as water injection rate is increased. A threshold of Ca
= 10−4 has been observed above which Sorw is significantly reduced (Tiab and Donaldson, 2004). However,
not all relative permeability phenomena may be explained by capillary number alone. For instance, the
influence of oil viscosity or the existence of thin wetting films are not considered in Ca. As pore sizes
increase, capillary effects are reduced and water-oil meniscus displacement may be gradually replaced by
other multiphase flow regime. In very large pores, the different effects of capillary forces in flow dynamics
of the wetting and non-wetting phase may be significantly reduced. Furthermore, Ferreira et al. (2015) have
shown the importance of pore-throat size distribution and hydraulic tortuosity, which is closely related to
pore connectivity, to the characterization of capillary effects. Equation (4) shows the relationship between
Ca and those key rock properties, by combining Darcy and Kozeny equations with Ca definition:
(4)
where ΔP is the applied differential pressure, L is the length of the sample, τhw is the hydraulic tortuosity
to water, and rthroat is the pore-throat radius. In the laboratory, Ca is increased by rising water injection rate.
However, far from the wellbore into the reservoir, multiple rock facies are subjected to the same effective
pressure gradient, and flux rates will vary according to quality of the rock of each facies. Indeed, from
Equation (4), for constant interfacial tension γow and constant effective pressure gradient, Ca only increases
if <r2throat> increases.
SPE-188260-MS 7

Normalized Water-Fractional-Flow Curves

The objective of this work is to provide a new workflow for quality check, group, average, end-point
scale and assess the uncertainty of relative permeability data. We focus on water-oil relative permeability
experiments with water displacing oil, defined here as primary imbibition cycle, after water primary
drainage cycle and subsequent wettability restoration. To improve the analysis of water-oil relative
permeability curves, a new method is proposed, based on the water fractional flow concept fw, which is
an effective representation of the behavior of oil displacement by water injection, combining both relative
permeability curves to oil and water into a single, equivalent curve.
When checking, comparing and aggregating relative permeability data from different samples under
different experimental conditions, viscosity ratios will likely be different as well. From Equation (1), fw is
influenced by both the viscosity ratio and the relative permeability ratio. The symmetry regarding wettability
around the neutral line, observed by Sinha et al. (2011), was obtained at a constant viscosity ratio equal to
one. Therefore, we propose to remove viscosity ratio influence on (fw) to highlight the wettability impact
on it, by considering μw = μo to obtain the following normalized water fractional flow:

(5)

Normalizing water fractional flow curves by considering a viscosity ratio equal to one is acceptable
as relative permeability curves normally are not significantly affected by fluid viscosities (Honarpour et
al., 1986). The proposed normalization does not change the actual experimental values of viscosity ratio.
Nevertheless, increasing displacement fluid viscosity may also increase the capillary number associated to
the waterflood experiment, which may affect residual oil end-point and relative permeability curve shapes if
a certain threshold is surpassed – usually between 10-5 and 10-4 (Ramakrishnan and Wasan, 1984). Likewise,
very large oil-water viscosity ratios resulting from high viscosity oil, e.g. over 50, may also affect relative
permeability results, especially when dealing with very heterogeneous samples. Therefore, is recommended
that the relative permeability data to be compared and aggregated were obtained with viscosity ratios within
the same order of magnitude.
Saturation end-points also vary from sample to sample, and the saturation range of the waterflood
displacement process extends from irreducible water saturation (Swirr) to residual oil saturation (Sorw).
Therefore, we also normalize water saturation (Sw) according to the saturation end-points by defining:

(6)

This way, varies from 0 to 1. Equation (5) may be rearranged:

(7)

According to the chosen end-point-scaling strategy, (fw) may also be normalized by the relative
permeability end-points (krwmax = krw@ Sorw and krowmax = krow@ Swirr). After removing the influence of viscosity
ratio, mainly depends on the pore sizes, rock topology, oil-water interfacial tension and wettability state.
The non-linear aspect of relative permeability curves is a direct result of different impact of capillary
effects on wetting and non-wetting phases. Therefore, if capillary forces are reduced by reducing the
interfacial tension between phases, or because the wettability of the rock is close to neutral condition – with
contact angles close to 90°, then residual saturations are drastically reduced and both phases can present a
very symmetric, linear behavior of relative permeability with saturation (X shape curves). Capillary effects
8 SPE-188260-MS

can also be reduced with pore size increase, e.g. in large vugs and large-aperture natural fractures, likewise
resulting in symmetrical, linear behavior of relative permeability for both wetting and non-wetting phases.
It can be shown from Equation (5) that if relative permeability curves present symmetrical, linear behavior
with saturation, the corresponding curve will be coincident with the neutral line in which .
From the work of Sinha et al. (2011) and Flovik et al. (2015), in which the viscosity ratio is also considered
equal to one, we may conclude that strongly oil-wet and strongly water-wet curves should be symmetric
to each other with regards to the neutral line. Therefore, the neutral line splits the plot into two regions
of different wettability: oil-wet above and water-wet below the neutral line. It was shown that most of the
curve for strongly oil-wet samples will be above the neutral line while the curve for strongly water-
wet samples will mostly be below the neutral line (Figure 3).
Reservoir rock type indexes like Winland, RQI and FZI and corresponding RRT are somehow correlated
with rock-pore-throat sizes. Therefore, we expect that the larger is the RRT index, the closer curves get
to the neutral line (Figure 3).

Figure 3—Typical plot showing the neutral line separating the

wettability regions and the expected influence of increasing RRT

Wettability Index from Oil-Water Relative Permeability Curves

We propose a new wettability index Ikrow from the curve, based on the fraction of oil-wet surface ω
calculated by rearranging Equation 2:

(8)

which is, because of the curve symmetry, equivalent to:

(9)
SPE-188260-MS 9

In fact, Equations (8) and (9) may provide multiple values of ω, resulting in a curve of ω against .
Therefore, we propose to calculate Ikrow based on the average of the fraction of oil-wet surface 〈ω〉:
(10)
Therefore, Ikrow ranges from −1, which corresponds to a strongly oil-wet state, to +1, which corresponds
to a strongly water-wet state. Such a method requires at least two relative permeability experiments for
the same sample at different conditions: the unknown wettability condition and either strongly
water-wet or oil-wet conditions. For example, if the sample is well cleaned, relative
permeability at strongly water-wet conditions may be obtained by using a synthetic oil with no potential for
wettability change, before the sample is submitted to a wettability restoration process.
Figure 4 presents a typical plot of two curves for the same synthetic sample, the first one known to
correspond to a strongly water-wet experiment (in blue) and the second one, corresponding to a restored
wettability condition, for which 〈ω〉 is calculated to be 0.75, resulting in Ikrow= −0.5, which is indicative
of a moderate oil-wet condition.
We expect Ikrow to be more representative for viscous displacement than wettability indexes based on
water-oil capillary pressure hysteresis, e.g. Amott-Harvey and USBM.
We expect such a method to be significantly accurate. Nevertheless, it requires two relative permeability
experiments for each sample, not easily found in legacy SCAL data.

Figure 4—Typical plot of two curves for the same sample, the first one known to correspond
to a strongly water-wet experiment (in blue) and the second one, corresponding to a restored
wettability condition, for which 〈ω〉 is calculated to be 0.75, resulting in (Ikrow) = −0.5

As an alternative, we propose to calculate the new wettability index Ikrow from the curve, based on the
areas above and below the neutral line (Figure 5), according to the following equations:
10 SPE-188260-MS

(11)

(12)

(13)
where is the normalized water saturation, Aw and Ao are the areas from the curve to the neutral line, in
the water-wet and oil-wet regions, respectively. Again, Iow ranges from -1, which corresponds to a strongly
oil-wet state, to +1, which corresponds to a strongly water-wet state.

Figure 5—Typical plot showing the neutral line separating the

wettability regions, and the areas for calculate the wettability index

New Water-Oil Relative Permeability Workflow

The proposed workflow is shown in Figure 6 :
The suggested workflow steps are:
1. For each sample, obtain the routine core analysis data and classify it according to corresponding RRT
2. For each sample, obtain saturation and relative permeability end-points from the water-oil relative
permeability curves
3. For each sample, obtain the normalized water fractional flow from the water-oil relative
permeability curves, by considering viscosity ratio equal to one and normalizing water saturation
according to the saturation end-points. curves enable the comparison of relative permeability data
SPE-188260-MS 11

from different samples, wells and fields. curves also enable the comparison of results from different
experiments with the same core plug.
4. For each sample, calculate the new wettability index Ikrow from the curve according to equations
(10) or (13):
5. Define the SATNUM groups according to RRT, reservoir compartments;
6. Group the curves according to SATNUM;
7. Group the curves according to calculated wettability index Ikrow, or according to their relative
position to the neutral line separating water-wet and oil-wet behavior;
8. For each group of samples, if possible, compare the new wettability indexes Ikrow with Amott-Harvey
and USBM indexes for similar samples at similar conditions;
9. Review the SATNUM groups according to the wettability groups;
10. For each STANUM group of curves, correlate the saturation and relative permeability end-points
with RRT indexes, kabs or other rock properties;
11. If necessary, further normalize the curves by the relative permeability end-points (krow @ Swirr and
krw @ Sorw);
12. For each SATNUM group of curves, average the normalized water fractional flow curves ( ).
Alternatively, average curves may be obtained through dynamic upscaling;
13. For each SATNUM group of curves, estimate the upper and lower-boundary curves to represent
their uncertainty envelope. Such boundary curves should represent the most favorable and most
unfavorable displacement behaviors. If there are additional concerns about the wettability preservation
or restoration processes, such envelope may be further extended;
14. For each SATNUM group calculate the normalized relative permeability curves from average, lower
and upper boundaries of curves;
15. Based on the quality of end-points correlation and the uncertainty analysis objectives, choose between
de-normalized or end-point scaling (EPS) water-oil relative permeability curves
16. If de-normalized water-oil relative permeability curves are needed, combine the corresponding end-
points with the normalized relative permeability curves;
17. If end-point scaling (EPS) process is used in the reservoir simulation, keep the normalized relative
permeability curves (average, lower and upper boundaries) and apply end-point correlations to the
simulation grid;
18. Input the resulting groups of relative permeability curves and end-points to the reservoir simulation
model, creating different multiple model realizations in order to run sensitivity analysis to SCAL and
RRT data (see Figure 7).
12 SPE-188260-MS

Figure 6—Proposed workflow for estimating wettability, grouping, averaging, and scaling
water-oil relative permeability data, as well as and assessing their corresponding uncertainty

Figure 7—Typical Workflow for the Sensitivity Analysis on SCAL uncertainties, highlighting the SCAL processes

DHD
In this work, we use a Direct HydroDynamic (DHD1) simulator to obtain relative permeability curves
for digital rock models of a carbonate rock obtained from X-ray micro-CT imaging. The DHD simulator
is a numerical code that solves 3D equations of the density functional hydrodynamics, which describes
multiphase compositional mixtures, including multiphase flow at pore-scale. The equations are conservation
laws for mixture chemical components, momentum and energy with constitutive relations, which are
consistent with Helmholtz free energy or entropy being represented as functionals depending on molar
densities of chemical components (Dinariev 1995, 1998). Since 1995, the DHD has undergone extensive
SPE-188260-MS 13

validation, which is summed up in the book (Demianov et al., 2014). Practical simulation experience on
high-performance computing platforms demonstrated that the DHD can accurately predict fundamental
two-phase flow mechanisms (Armstrong et al., 2016) as well as the cooperative phenomena inherent in
microfluidic systems during two-phase flow (Armstrong et al., 2015). For the last decade, DHD simulations
were applied to oil & gas applications using Digital Rock models of rocks. DHD simulations are now
routinely used for numerical modeling of multiphase pore-scale processes of various complexity, including
evaluation of relative permeability and desaturation curves (Koroteev et al., 2013, 2015). This simulation
tool is also used commercially with practical examples including oil displacement during water alternating
gas (WAG) injection (Shandrygin et al., 2015) and polymer flooding (Shandrygin et al., 2016), evaluation
of acid stimulation efficacy in carbonates (Klemin et al., 2015), and recovery mechanisms in oil shales
(Geetan et al., 2016).

Data Setup
In this work, we apply our proposed method to digital core flooding cases simulated with DHD within
controlled wettability and capillary number, also considering the saturation end-points, and therefore
extending previous results. Subsequently, we describe such DHD simulation cases.
High resolution (2.23 micron) X-ray micro-Computed Tomography (Micro-CT) was performed on an
Indiana Limestone carbonate cylindrical mini-plug (about 8 mm in diameter) extracted from a standard 1.5"
plug. Shadow projections obtained during the imaging experiment are used to reconstruct greyscale images
of the sample's 2D cross sections (Figure 8a). These images were further processed and segmented into pore
space and rock to ultimately construct a digital rock model of the sample (Figure 8b).

Figure 8—2D slice of the Micro-CT scan of carbonate rock sample (a) and Rock model overview (b)

Rock model dimensions are 1.784x1.784x1.784 mm3. The porosity and axial permeability values
for the rock model were evaluated using the DHD simulator. Total porosity and absolute permeability
were calculated and compared to the laboratory measured permeability values on the same plug. Digital
permeability value of 634.9 mD is close to the experimental absolute permeability to gas of 661 mD. Digital
model porosity is equal to 11.3%. Physical measurements for helium porosity of the standard size core plug
resulted in porosity of 14.6%.
Digital fluid models were created for the water and oil system. The digital fluid model of oil represents
single phase oil with a viscosity of 0.85 cP and a density of 0.73 g/cc. Brine viscosity is 0.88 cP and brine
density is 1.00 g/cc. Oil-water interfacial surface tension (IFT) was set to be 40.0 mN/m.
Irreducible water saturation (Swirr) for the rock model was established using steady-state (SS) flooding
method considering the constructed digital fluid model. Wetting properties of the sample were set to
100% water-wet for the purpose of this initial steady-state (SS), drainage cycle. The contact angles for the
sample were ranging from 10 to 40 degrees. The sample was initially saturated with 100% water. During
14 SPE-188260-MS

the simulation of this first cycle two phases were injected at a set phase ratio, which is continued until
equilibrium is achieved. Consecutively the ratio is changed to a different ratio with increase in oil fraction
and injection is restarted. This cycle ended with oil injection to obtain the irreducible water saturation (Swirr).
Obtained saturations and phase velocity data were used to compute oil permeability at Swirr, at the end of
the first SS sequence. Swirr value for sample is equal to 0.105.
A second SS cycle was performed to evaluate the impact of wetting conditions of the rock on the phase
mobility. Wetting conditions for all pore surfaces were changed from the ones used in the first cycle. The
following four wetting conditions were applied to all surfaces of the rock model:
1. Water-wet: 10 - 40º;
2. Water-Wet-to-Neutral-Wet: 50 - 80 º;
3. Oil-Wet-to-Neutral-Wet: 100 - 130º;
4. Oil-wet 140 - 170 º.
The second cycle started for each sample at irreducible water saturation (Swirr), which was obtained during
the drainage cycle. During the second cycle, water/oil injection ratio was continuously increased towards
the water. The equilibrium saturation points were obtained for all second cycle floods. Obtained saturations,
phase velocities and phase pressure data were used to compute phase effective permeabilities. Relative
permeabilities were normalized to the value of absolute permeability of the digital rock model.
We also partially applied our method to a public relative permeability database published by Koederitz
and Ibrahim (2002). From this database, 274 water-oil relative permeability experiments were selected,
obtained by both steady and unsteady methods, of different lithology, and from different parts of the world.
The data includes a wettability classification based on Craig’s rules of thumb.

Results
Representative equilibrium saturation distributions for second SS cycle for each wettability range at Sorw
are presented in Figure 9.

Figure 9—Representative equilibrium saturation distributions for second SS cycle (at Sorw), for each simulation case

Water-oil steady-state (SS) relative permeability data for the investigated sample are shown in linear
scales in Figure 10.
SPE-188260-MS 15

Figure 10—Relative permeability for modeled wettability cases

Normalized water fractional flow curves are plotted against the normalized saturation in Figure 11.

Figure 11—Normalized water fractional flow plotted against the normalized saturation for modeled wettability cases

Due to the small number of actual steady-state relative permeability data points in this study, normalized
water fractional flow curves were fitted by Brooks-Corey parametric function and wettability indexes
based in area were calculated. The results are shown in Figure 12.
16 SPE-188260-MS

Figure 12—Normalized water fractional flow and corresponding wettability indexes for modeled wettability cases

Although a perfect symmetry of the cases curves was not observed, the calculated wettability indexes,
based on both average wettability fraction and areas, correspond quite well to wettability state (contact
angle ranges) for each DHD case. The lack of exact symmetry may be due to the low capillary number
in the DHD simulations (Ca = 10-7), combined with the current formulation, which does not directly
consider capillary pressure. Alternatively, it may be due to the limited amount of data points in the low
water saturation range for the oil wet cases.
Koederitz and Ibrahim (2002) database does not include some important elements, i.e. field, formation
and well identification, to properly quality check the data, group and average the results. Therefore, we
could only generate the curves and calculate the corresponding wettability index.
The wettability index based on relative permeability (Equation (13)) was calculated for the 274 oil-water
available experiments in the database, using the formulation based on areas. We have also assigned the
following values to the existing Craig’s classification: -1 to strongly oil-wet behavior, -0.5 to oil-wet, 0 to
mixed-wet, +0.5 to water-wet and +1 to strongly water-wet behavior.
Figure 13 shows the comparison plot between (sorted) wettability index Ikrow from relative permeability
and corresponding Craig’s classification for the water-oil relative permeability database (Koederitz and
Ibrahim, 2002). The qualitative results of the comparison are quite good. Indeed, the shaded areas show
samples where Ikrow and Craig’s classification significantly diverges, representing only around 14% of the
total number of samples.
SPE-188260-MS 17

Figure 13—Quite good comparison plot between (sorted) wettability index Ikrow from relative permeability and
corresponding Craig’s classification from the water-oil relative permeability database (Koederitz and Ibrahim (2002))

Figure 14 shows the curves for 51 samples, which show moderate to strongly oil-wet condition (Ikrow
<= -0.5). Figure 15 shows the curves for 176 samples, which show moderate to strongly water-wet
condition (Ikrow >= 0.5). As expected, those two plots show a remarkable segmentation of the curves between
the oil-wet and water-wet areas, according to the calculated wettability index.

Figure 14— plot for 51 samples, which show moderate to strongly oil-wet condition (Ikrow <= -0.5)
18 SPE-188260-MS

Figure 15— plot for 176 samples, which show moderate to strongly water-wet condition (Ikrow >= 0.5)

Discussion
Relative permeability and wettability concepts are very complex, closely related subjects. The proposed
method, although very helpful, cannot solve all quality control and processing issues commonly associated
with such experiments, i.e. the effects of sample heterogeneity and along-bedding flow. However, the
method may be very useful for data quality checking. The method, while supported by coreflood simulation
(Sinha et al. (2011) and DHD results presented in this paper), may have some limitations, e.g. very high
viscosity contrast, very high or very low injection rates and their corresponding capillary numbers. High
capillary numbers may affect saturation and relative permeability end-points, as well as the shape of relative
permeability curves. Low capillary numbers may result in experiments dominated by capillary forces, and
normalized water fractional-flow formulation (Equation (5)) should be adjusted accordingly. Nevertheless,
the DHD simulation have indicated that capillary numbers as low as 10-7 may still be acceptable. Moreover,
the proposed workflow has not been extensively applied to real field data, from different rock and fluid
environments. Such limitations should be addressed in future works.
Wettability state is one of the main drivers for water-oil relative permeability behavior. Unfortunately,
it is also one of the main sources of corresponding uncertainty. Well logs may significantly help to further
constraint the wettability uncertainty from relative permeability data. Several studies (Freedman et al.,
2003 and Flaum et al., 2005) demonstrate that nuclear magnetic resonance (NMR) measurements can
be used to estimate wettability. Wetting fluids will have enhanced relaxation due to contact with pore
surface, therefore, observed relaxation time (T2) of wetting fluids will be shorter that bulk T2. Non-
wetting fluid will not be subject to this effect, and observed T2 will be equivalent to bulk T2. However,
heavier oils T2 distribution is similar to small brine saturated pores T2 distribution and the overlap will
complicate wettability interpretation. Additionally, light oil and brine will have similar T2, also complicating
interpretation.
More recently (Abdelaal et al., 2013 and Venkataramanan et al., 2014), dielectric frequency dispersion
measurements has also been used to infer wettability. Dielectric response is sensitive to water saturated
pores. Through effective medium models, water tortuosity and water filled pores geometry can be estimated.
For example, SMD model allows estimation of "MN" water phase tortuosity exponent. If Archie "m"
exponent is known for the rock, then "n" exponent can be estimated, according to
SPE-188260-MS 19

(14)

Then "n" can be correlated to wettability state (Abdelaal et al., 2013).

Conclusions
The proposed method is simple, yet very innovative.
Using the normalized water fractional flow curves ( ) to group relative permeability curves and end-
points may be a significant improvement to the processes of analysis of water-oil relative permeability
data, wettability analysis and end-point scaling, consequently improving the resulting reservoir simulation
models and their sensitivity analysis.
The normalized water fractional flow curves may represent the wettability behavior affecting the water-
oil displacement. They enable to calculate quantitative wettability indexes from relative permeability
data. These data can be further grouped according to pore-throat sizes and corresponding RRT, reservoir
compartments and wettability states.
Wettability indication derived from well logs (e.g. NMR and Dielectric logs) may significantly help to
further constraint the wettability uncertainty from relative permeability data. The analysis of thin sections,
CT images and corresponding mineralogy may also contribute to reduce such wettability uncertainty. Using
the correct fluids composition, under the correct reservoir conditions during the laboratory experiments,
is absolutely mandatory to reduce the relative permeability uncertainties. The proposed method may
significantly improve the quality check of such data.
The obtained results so far are very encouraging and the proposed method is considered to be ready to be
tested with real field data. Nevertheless, the proposed method and corresponding workflow may be further
improved and implemented.

Abbreviations
DHD = Direct Hydro-Dynamic - digital rock simulation method
EPS = end-point scaling algorithm
FZI = flow zone indicator
JBN = method for inversion of unsteady-state experiments
Micro-CT = X-ray micro-Computed Tomography
NMR = nuclear magnetic resonance
REV = representative element volumes
RQI = reservoir quality index
RRT = reservoir-rock type
SATNUM = saturation function region
SCAL = special core analysis
SMD = dielectric effective media model after Stroud, Milton and De (1986)
SS = steady state relative permeability experiment
T2 = NMR transversal relaxation time
USBM = US Bureau of Mines wettability index
USS = unsteady state relative permeability experiment
WAG = water alternating gas injection

Symbols
Ca = capillary number
qw = water flow rate at reservoir conditions
20 SPE-188260-MS

qt = liquid flow rate at reservoir conditions

krw = relative permeability to water
krow = relative permeability to oil
μw = water dynamic viscosity
μo = oil dynamic viscosity
Swirr = irreducible water saturation
krow @ Swirr = initial oil relative permeability
krw@ Sorw = final water relative permeability
fw = water fractional flow
ω = fraction of the rock surface that is strongly oil-wet
θ = contact angle
Vw = velocity of water flowing through the pores
γow = water-oil interfacial tension
S orw = residual oil saturation
ΔP = applied differential pressure
L = length of the sample
τhw = hydraulic tortuosity to water
= averaged squared pore-throat radius
= normalized water fractional flow
= normalized water saturation
Sw = water saturation
Ikrow = wettability index from water-oil relative permeability
Aw = area from the curve to the neutral line, in the water-wet region
Ao = area from the curve to the neutral line, in the oil-wet region
MN = water phase tortuosity exponent from dielectric log
m = Archie cementation exponent
n = Archie saturation exponent
φ = porosity
Sxo = water saturation from dielectric log

References
Abdallah, W., Buckley, J. S., Carnegie, A., Edwards, J., Herold, B., Fordham, E., Graue A., et al 2007. Fundamentals of
Wettability. Schlumberger Oilfield Review, 44–61.
Agbalaka, C. C., Dandekar, A. Y., Patil, S. L., Khataniar, S., and Hemsath, J. 2008. The Effect of Wettability on
Oil Recovery: A Review. SPE-114496-MS, Asia Pacific Oil and Gas Conference and Exhibition, 1–13. Society of
Petroleum Engineers. 10.2118/114496-MS.
Al-Qasim, A., Bubshait, A. and Saudi Aramco. 2017. Asphaltenes:What Do We Know So Far, Definition of Asphaltene,
Structure of Asphaltene, SPE-185690-MS.
Amroun, H. and Tiab, D. 2001. Alteration of Reservoir Wettability Due to Asphaltene Deposition in Rhourd-Nouss Sud
Est Field, Algeria. SPE-71060-MS, Proceedings - SPE Rocky Mountain Petroleum Technology Conference, Petroleum
Technology - Leads the Way, 261–71. 10.2523/71060-MS
Anderson, W. G. 1987. Wettability Literature Survey–Part 5: The Effects of Wettability on Relative Permeability. J Pet
Technology 39 (11): 1453–1468. SPE-16323-PA. http://dx.doi.org/10.2118/16323-PA.
Abdelaal, A. F., Khadija, A., Daghar A., Bin M., Khalfan, A., and Faivre O. 2013. Integration of Dielectric Dispersion
and 3D NMR Characterizes the Texture and Wettability of a Cretaceous Carbonate Reservoir. SPE-164150-MS
Armstrong, R. T., Evseev, N., Koroteev, D., Berg, S. 2015. Modeling the velocity field during Haines jumps in porous
media, Adv. Water Resour., 77, 57–68
Armstrong, R. T., Berg, S., Dinariev, O., Evseev, N., Klemin, D., Koroteev, D., and Safonov, S. 2016. Modeling of pore-
scale two-phase phenomena using density functional hydrodynamics. Transp. Porous Med., 112(3), 577.
SPE-188260-MS 21

Baker, P. E. 1960. Discussion of effect of viscosity ratio on relative permeability, J. Pet. Technol., 219, 65
Buckley, S. E. & Leverett, M. C. 1942. Mechanism of Fluid Displacement in Sands, Trans., AIME, Vol. 146, 187–196.
Carvalho, M. S. 2015. Flow of Complex Fluids through Porous Media: Application in Oil, OTC-26079-MS
Constantinides, G. N. and Payatakes, A. C. 1996. Network Simulation of Steady-State Two-Phase Flow in Consolidated
Porous Media, AICHE Journal, 42 (2), 369.
Craig, F. F. 1971. The Reservoir Engineering Aspect of Waterflooding, Monograph Series, SPE, Richardson, TX.
Dake, L. P. 1978. Fundamentals of Reservoir Engineering, Elsevier, Amsterdam.
Demianov, A. Y., Dinariev, O. Y., and Evseev, N. V. 2014. Introduction to the density functional method in hydrodynamics.
Fizmatlit, Moscow.
Deng, L., King, M. J., and Texas A. 2016. Estimation of Relative Permeability from Laboratory Displacement Experiments
Application of the Analytic Solution with Capillary Corrections. SPE-183139-MS.
Dinariev, O. Y. 1995. A hydrodynamic description of a multicomponent multiphase mixture in narrow pores and thin
layers. J. Appl. Math. Mech. 59(5), 745–752.
Dinariev, O. Y. 1998. Thermal effects in the description of a multicomponent mixture using the density functional method,
J. Appl. Math. Mech. 62(3), 397–405.
Facanha, J., Sohrabi, M., Drexler, S., Couto, P. 2016. Experimental Investigation of Wettability for Brazilian Pre-Salt
Carbonates, Rio Oil&Gas, IBP1455_16
Ferer, M., Anna S. L., Tortora, P., Kadambi, J. R., Oliver, M., Bromhal, G. S., and Smith, D. H. 2011. Two-Phase Flow
in Porous Media: Predicting Its Dependence on Capillary Number and Viscosity Ratio, Transp. Porous Media, vol.
86, no. 1, pp. 243–259.
Ferreira, F. C., Booth, R., Oliveira, R., SPE, Schlumberger, Carneiro, G., Bize-Forest, N., and Wahanik, H., Schlumberger,
2015. New Rock-Typing Index Based on Hydraulic and Electric Tortuosity Data for Multi-Scale Dynamic
Characterization of Complex Carbonate Reservoirs, presented at ATCE 2015, Houston, SPE-175014-MS
Flaum, M., Chen, J., and Hirasaki, G. J. 2005. NMR diffusion editing for D-T2 maps: Application to recognition of
wettability change. Petrophysics, 46 (2), pp. 113–123.
Flovik, V., Sinha, S. and Hansen, A. 2015. Dynamic Wettability Alteration in Immiscible Two-Phase Flow in Porous
Media: Effect on Transport Properties and Critical Slowing Down. Frontiers in Physics 3 (November): 1–10. 10.3389/
fphy.2015.00086
Freedman, R., Heaton, N. J., Flaum, M., Hirasaki, G. J., Flaum, C., and Hürlimann, M. D. 2003. Wettability, Saturation,
and Viscosity from NMR Measurements. SPE Journal 8 (4): 317–27. 10.2118/87340-PA.
Fulcher, R. A., Turgay Ertekin, and C. D. Stahl. 1985. Effect of Capillary Number and Its Constituents on Two-Phase
Relative Permeability Curves. Journal of Petroleum Technology 37 (2): 249–60. 10.2118/12170-PA
Geetan, S. I., MacDonald, R. M., and Klemin, D. 2016. Insights Into Recovery Mechanisms in Shales Through Digital
Rock Technology, URTeC: #2461115
Hansen, Alex, Sinha Santanu, Bedeaux Dick, Kjelstrup Signe, Savani Isha, and Vassvik Morten. 2016. A New Set
of Equations Describing Immiscible Two-Phase Flow in Isotropic Porous Media, no. 7491 (May). http://arxiv.org/
abs/1605.02874.
Honarpour, Mehdi, Koederitz Leonard, and a Herbert Harvey. 1986. Relative Permeability of Reservoirs Petroleum, 1–
1. CRC Press, Inc. Boca Raton, Florida
Honarpour, M. M.; Cullick, A. S. 1994. Influence of Small-Scale Rock Laminations on Core Plug Oil/Water Relative
Permeability and Capillary Pressure SPE 27968, Tulsa.
Honarpour, M. M.; Cullick, A. S. 1995. Effect of Rock Heterogeneity on Relative Permeability: Implications for Scale-
Up, SPE 29311.
Johnson, E. F., Bossler, D. P. and Naumann Bossler, V. O. 1959. Calculation of Relative Permeability from Displacement
Experiments. Petroleum Transactions, AIME 216: 370–372, SPE-1023-G. http://dx.doi.org/10.2118/1023-G
Klemin, D., Nadeev, A., and Ziauddin, M. 2015. Digital Rock Technology for Quantitative Prediction of Acid Stimulation
in Carbonates, SPE-174807-MS
Koederitz, L. F., & Mohamad M. N. 2002. Developing a Proficient Relative Permeability Resource From Historical Data.
Petroleum Society of Canada, PETSOC-2002-213. 10.2118/2002-213
Koroteev, D., Dinariev, O., Evseev, N., Klemin, D., Nadeev, A., Safonov, S., Gurpinar, O., Berg, S., van Kruijsdijk, C.,
Armstrong, R., Myers, M. T., Hathon, L., de Jong H. 2014. Direct Hydrodynamic Simulation of Multiphase Flow in
Porous Rock, Petrophysics 55(4)
Koroteev, D., Dinariev, O., Evseev, N., Klemin, D., Safonov S., Gurpinar, O., Berg, S., van Kruijsdijk, C., Myers, M.,
Hathon, L., de Jong, H. 2013. Application of Digital Rock Technology for Chemical EOR Screening, SPE enhanced
oil recovery conference, SPE-165258-MS
Mullins Oliver C. 2010. The Modified Yen Model. Energy and Fuels 24 (4): 2179–2207. 10.1021/ef900975e
22 SPE-188260-MS

Pfeiffer, T., Di Primio, R., Achourov, V., Mullins, O. C. Scanning electron micrographs of tar-mat intervals formed by
asphaltene phase transition (2017) Petrophysics, 58 (2), pp. 141–152.
Prisco, Giuseppe De, Toelke Jonas, and Dernaika Moustafa R. 2012. Computation of Relative Permeability Functions in
3D Digital Rocks by a Fractional Flow Approach Using the Lattice Boltzmann Mehod. Int. Symp. Soc. Core Analysts,
1–12.
Ramakrishnan, T. S., and D. T. Wasan. 1984. The Relative Permeability Function for Two-Phase Flow in Porous Media:
Effect of Capillary Number. SPE. https://www.onepetro.org/conference-paper/SPE-12693-MS
Schlumberger. 2016. ECLIPSE 2016.1 User Guide and Technical Description, http://www.slb.com
Schmid, K. S., The Boston, Consulting Group, N. Alyafei, and A. Texas. 2016. Analytical Solutions for Spontaneous
Imbibition : Fractional-Flow Theory and Experimental Analysis, August: 2308–16.
Sinha, S., Grøva, M., Ødegården, T. B., Skjetne, E., Hansen, A. 2011. Local wettability reversal during steady-state two-
phase flow in porous media Physical Review E - Statistical, Nonlinear, and Soft Matter Physics, 84 (3), art. no. 037303,
10.1103/PhysRevE.84.037303
Shandrygin, A., Shelepov, V., Ramazanov, R., Andrianov, N., Klemin, D., Nadeev, A., Safonov, S., Yakimchuk, I. 2015.
Mechanism of Oil Displacement During WAG in Porous Media with Micro-Inhomogeneities, SPE-176629-MS
Shandrygin, A., Shelepov, V., Ramazanov, R., Andrianov, N., Klemin, D., Nadeev, A., Safonov, S., Yakimchuk, I.
2016. Mechanism of Oil Displacement During Polymer Flooding in Porous Media with Micro-Inhomogeneities,
SPE-182037-MS
Stroud, D., Milton, G. W., and De, B. R.. 1986. Analytical model for the dielectric response of brine-saturated rocks.
Phys. Rev. B 34: 5145.
Tiab, D., & Donaldson, E. C. 2004. Petrophysics: Theory and Practice of Measuring Reservoir Rock and Fluid Transport
Properties(2nd). Elsevier.
Tørå, Glenn, and Alex Pål-eric Øren. 2012. "A Dynamic Network Model for Two-Phase Flow in Porous Media," 145–64.
10.1007/s11242-011-9895-6
Venkataramanan, Lalitha, Martin D. Hürlimann, Jeffrey A. Tarvin, Fellah Kamilla, Acero-allard Diana, and Nikita
V. Seleznev. 2014. "Experimental Study of the Effects of Wettability and Fluid Saturation on Nuclear Magnetic
Resonance and Dielectric Measurements in Limestone." Petrophysics 55 (6): 572–86.
Welge, H. J. 1952. "A Simplified Method for Computing Oil Recovery by Gas or Water Drive" Trans. AIME 195 91–98.
Zaitoun, A., Kalaydjian, F., and Jacquin, C. 1990. Two-Phase Flow Through Porous Media: Influence of Viscosity Ratio
on Relative Permeabilities; 5th IFP Research Conference on Expl./Prod. ARLES