Journal of Petroleum Science and Engineering 56 (2007) 241 – 251

www.elsevier.com/locate/petrol

A new method to acquire m exponent and tortuosity factor for

microscopically heterogeneous carbonates
M.R. Rezaee a,⁎, H. Motiei b , E. Kazemzadeh c
a
University of Tehran, University College of Science, School of Geology, Iran
b
Technology Division of R&D, NIOC, Iran
c
Research Institute of Petroleum Industry/NIOC, Iran
Received 5 April 2006; received in revised form 15 September 2006; accepted 18 September 2006

Abstract

Many factors control accurate determination of water saturation (Sw). Cementation exponent (m) and tortuosity factor (a) are
from those that have been focus of many studies. Log–log plot of porosity (ϕ) versus formation factor (F) is used to determine m
and a. The cementation exponent is determined from the negative slope of the least square fit straight line of the plotted points,
while the tortuosity factor is the intercept of the line where ϕ = 1. In heterogeneous carbonate reservoirs where pores and pore
throat networks are complex due to various diagenetic processes, F and ϕ scatter significantly on the ϕ − F plot. This will cause a
small coefficient of determination between F and ϕ and thus less reliable m and a. Although classification of data based on
petrofacies and/or permeability may improve the correlation to some extent, data still show significant scatter.
This study has established a new approach to classify ϕ and F data based on current zone indicator (CZI) and electrical flow
unit (EFU). The approach can obtain more accurate m and a and thus more realistic calculation of Sw. This study also shows
forcing a to any fixed value, will lead to both optimistic and pessimistic estimation of Sw for a reservoir.
© 2006 Elsevier B.V. All rights reserved.

Keywords: Cementation exponent; Tortuosity factor; Electrical flow unit; Formation factor; Water saturation; Carbonates

1. Introduction be measured experimentally, have been subject of many

studies. It has been shown that the use of inaccurate
Estimation of water saturation (Sw) is one of the most values for Archie's parameters have significant effects
important tasks in formation evaluation. Accurate on F and thus on Sw calculation (Hosseini-nia and
estimation of Sw and thus hydrocarbon reserve is Rezaee, 2002).
critical to reduce the uncertainty of financial forecasting In routine formation evaluation m and a are
and in developing an oil or gas field. The accuracy of Sw considered constant for a given reservoir rock. It is a
calculation depends on the accuracy of the Archie's common practice to obtain m by assuming a constant
parameters, a, m and n. These parameters, that cannot value for a and calculating m for each sample. Rocks,
mainly carbonates, display complex pore structures,
⁎ Corresponding author.
which significantly affect their electrical resistivity.
E-mail addresses: [email protected] (M.R. Rezaee),
Since physical properties of these rocks may vary sig-
[email protected] (H. Motiei), [email protected] nificantly from one sample to another, m and a values
(E. Kazemzadeh). cannot be considered constant.
0920-4105/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.petrol.2006.09.004
242 M.R. Rezaee et al. / Journal of Petroleum Science and Engineering 56 (2007) 241–251

In this paper, a new method has been introduced to Blake–Kozeny model. Higher values of tortuosity have
classify F and porosity into separate electrical flow units been reported by many researchers. Wong et al. (1984)
(EFU), using current zone indicator (CZI). The method showed that tortuosity became 3.3 when ϕ was between
improves porosity and F correlation considerably. This 0.02 and 0.2 in fused-glass beads. Dubois et al. (2001)
study also shows that forcing a to be a constant value found tortuosity factor of about 9.5 for oomoldic lime-
causes m to increase. Consequently, this will lead to stones. Hirasaki (2005) reported an increase in tortuosity
both pessimistic and optimistic calculation of Sw, when sorting (standard deviation of grain size) and
depending on the reservoir tortuosity factor. porosity of sand grains decreased. He showed that a
value could reach 35 when porosity of sand grains
2. Basic concepts approach zero due to sorting reduction. Attia (2005)
suggested that tortuosity factor could not be considered
2.1. Cementation exponent constant since it depends on many factors such as the
amount of fine grains, formation resistivity factor,
Cementation exponent was first defined by Archie in cementation exponent, porosity and degree of brine
1942. Noticing that an increase in m values is associated saturation.
with sandstone consolidation, Archie named this expo- In general, the more tortuous the pore throats are, the
nent as cementation exponent. A wide range of m values harder it is for current to flow through the reservoir and
has been introduced by several authors ranging from 1 for the higher the resistivity.
fractured rocks to slightly more than 5 for highly
compacted rocks. In Archie's study (1942), m was 1.3 3. Inter-relationship of m and a, theoretical derivation
for unconsolidated sands and ranged between 1.8 and 2.0
for cemented sandstones. Timur et al. (1972) obtained Wyllie and Rose (1950) mentioned a 100-fold
values of a = 1.13 and m = 1.73 for 1800 sandstones from increase in a was accompanied by a 4-fold increase in
15 oil fields. Wong et al. (1984) worked on fused-glass m. Salem and Chilingarian (1999) suggested that a 10-
beads and showed that m values were 2.3 for fold increase in a lead to a 4-fold increase in m. They
0.02 b ϕ b 0.2, and 1 for 0.2 b ϕ b 0.4. Hamada et al. also introduced the following empirical equation:
(2002) determined values of a and m for 20 clean and
porous sandstones. They found a and m being 1.36 and m ¼ 1:5551 þ 2:1039loga ð2Þ
2.03 for one well, and 0.95 and 1.85 for the second well,
respectively. Focke and Munn (1987) demonstrated that Despite reported links between m and a by several
m depended on the petrofacies and porosity type in authors, this study suggests that there is no direct
carbonates. In their study, assuming a = 1, m ranged relationship between a and m. They have different nature
between 2 and values as high as 5.5. Dubois et al. (2001) which could not be compared. a refers to tortuosity of
introduced m = 1.36 for oomoldic limestones. pore throats whereas m defines degree of pores con-
Using Archie's equation as a base, many authors nectivity (Rezaee, in prep.). Tortuosity may vary from
(e.g., Neustaedter, 1968; Nugent et al., 1978; Sethi, one sample to another without any changes in m and vise
1979; Rasmus, 1983; Borai, 1987; Focke and Munn, versa. One cannot compare m and a with each other, but
1987) have introduced methods to determine m from log can evaluate their influence on F. In Table 1 formation
data. Log porosity and invasion corrected deep factors of rocks with different m, a, and porosity values
resistivity (Ro) were used to estimate F in wet zone. are calculated using Archie's equation (F = a / ϕm) to
In all mentioned studies, a was assumed 1. evaluate the relationship among F, m and a.
Considering a constant value for m (e.g., m = 2),
2.2. Tortuosity factor Table 2 shows tortuosity factors calculated from
Archie's equation (a = Fϕ2) using F values in Table 1.
The tortuosity (a) has been theoretically defined as It can be seen that a increases as F values increase, and
the ratio of the mean path length (La) to the straight line it is more pronounced in low porosity samples.
of porous medium length (L) (Carman, 1937): Fig. 1 is a cross-plot of porosity versus a values of
Table 2. It indicates a power law relationship between
a ¼ La=L ð1Þ the two variables (equations in Table 2). In each
equation, the exponents and intercepts are m2 − m1 and
A value of 25/12 was determined for the tortuosity of a1, respectively. m1 and a1 refer to m and a values in
uniform spherical particles by Bird et al. (1960) using Table 1 used to calculate F, m2 was considered 2. It can
 M.R. Rezaee et al. / Journal of Petroleum Science and Engineering 56 (2007) 241–251 243

Table 1 Eq. (8) can be rearranged as follow:

Formation factors for different values of porosity, m and a
Porosity (fraction) F/m1
a2 ¼ ð9Þ
0.05 0.10 0.20 0.30 0.40 /−Dm
F1 (m1 = 1.8 and 263.655 75.715 21.743 10.480 6.244
a1 = 1.2)
Eqs. (4) and (5) show the inter-relationships among
F2 (m1 = 1.9 and 444.681 119.149 31.925 14.776 8.554 porosity, F, a, and m values. As an example, in a rock
a1 = 1.5) with 15% porosity, m1 = 1.2 and a1 = 14, F will be
F3 (m1 = 2.2 and 1019.516 221.885 48.291 19.791 10.510 77.94. If m and a change to m2 = 2.2 and a2 = 1.2, with
a1 = 1.4) the same porosity, then F will be 136.4.
F4 (m1 = 2.4 and 1723.516 326.545 61.869 23.380 11.722
a1 = 1.3)
In Eqs. (6) and (8), F will vary following changes in
F5 (m1 = 2.6 and 2896.405 477.729 78.796 27.458 12.996 either a or m. When porosity and F remain unchanged,
a1 = 1.2) but m values change, a can be calculated using Eq. (9).
F6 (m1 = 3 and 8000.000 1000.000 125.000 37.037 15.625 For example, in a rock with given values of porosi-
a1 = 1) ty = 15%, a = 14, m = 1.2 and F = 136.4, if m changes
from 1.2 to 2.2, a will be 2.1.
To calculate m, Eq. (9) can be rearranged as follow:
be concluded that to remain F values unchanged, when
a2 /−Dm ¼ F/m1 ð10Þ
m changes, tortuosity factor must be adjusted according
to the following general equation: Taking logarithm from both sides of Eq. (10):
Dm
a2 ¼ a1 / ð3Þ loga2 −Dmlog/ ¼ logF þ m1 log/ ð11Þ
where Δm is m2 − m1. and rearranging Eq. (11) to:
As an example, for a rock with a porosity value of
10%, m = 2.4 and a = 1.3, F will be 326.545 (Table 1). −Dmlog/ ¼ logF þ m1 log/−loga2 ð12Þ
Again, for such a rock, with a presumed constant the following equation can be obtained :
formation factor of 326.545 and a value of 2 for m, a
must be adjusted to a value of 3.265. logF þ m1 log/−loga2
Dm ¼ − ð13Þ
Back to Archie's equation, in order to know how F log/
varies for a rock with a given porosity, if m changes
since Δm = m2 − m1, then m2 is calculated as:
from m1 to m2 and a from a1 to a2, F is:


a2 logF þ m1 log/−loga2
F¼ ð4Þ m2 ¼ − þ m1 ð14Þ
/m2 log/

or Eq. (14) is a theoretical derivation of Eq. (9), to

a1 calculate changes in m values when a is varied for a
F¼ ð5Þ given sample.
/m1 Although in these equations, m and a are related, it
if a2 in Eq. (4) is substituted with a2 in Eq. (3) then: does not necessarily mean that there is a direct
relationship between a and m, as other authors reported,
/Dm
F ¼ a1 ð6Þ
/m2 Table 2
Calculated tortuosity factor (a2) using F values in Table 1
Eq. (3) can be rearranged as:
a2 = Fϕ2 Porosity (fraction) Equations
−Dm
a1 ¼ a2 / ð7Þ 0.05 0.10 0.20 0.30 0.40
and, if a1 in Eq. (5) is substituted with a1 in Eq. (7), a2 for F1 0.659 0.757 0.870 0.943 0.999 a2 = 1.2ϕ0.2
then: a2 for F2 1.112 1.191 1.277 1.330 1.369 a2 = 1.5ϕ0.1
a2 for F3 2.549 2.219 1.932 1.781 1.682 a2 = 1.4ϕ− 0.2
a2 for F4 4.309 3.265 2.475 2.104 1.876 a2 = 1.3ϕ− 0.4
/−Dm a2 for F5 7.241 4.777 3.152 2.471 2.079 a2 = 1.2ϕ− 0.6
F ¼ a2 ð8Þ
/m1 a2 for F6 20.000 10.000 5.000 3.333 2.500 a2 = 1ϕ− 1
244 M.R. Rezaee et al. / Journal of Petroleum Science and Engineering 56 (2007) 241–251

4.1. Core analysis

Selected samples were cleaned by toluene in a

Soxhlet apparatus, and dried at 60 °C for 24 h prior to
analysis. Porosity and permeability were measured in
reservoir condition using Ultra-Porosimeter 200A,
Ultra-Permeameter and CMS-300™ (Core Measure-
ment System). Core porosity values ranged from 2.5 to
26% with a mean value of 10%. Range of permeability
was between 0.01 and 91.9 mD with a mean value of
5.4 mD.
In order to measure electrical resistivity (Ro),
samples were fully saturated by a brine with approxi-
mately the same water salinity of Asmari Formation
Fig. 1. Cross-plot of porosity versus tortuosity factor (data from Table 2). (200,000 ppm, NaCl). Using FRF Overburden Rig at a
For samples with m1 b m2, (a2 for F1 and a2 for F2), with an increase in
frequency of 1 kHz, electrical resistance (r) of the
porosity, tortuosity factor increases. For samples with m1 N m2, (a2 for F3
to a2 for F6) tortuosity factor decreases with an increase in porosity. In samples was measured along the axis of cylindrical
both cases variation of tortuosity factor is to keep the F values of Table 1 plugs in reservoir condition. Then, resistivity (R) was
unchanged. calculated from the measured resistance (r) using the
cross-sectional area (A) and the length of the core plug
since F is involved in all equations. These equations (L). Formation factor was obtained as a ratio of rock
will enable us, in the following sections, to evaluate resistivity (Ro) to brine resistivity (Rw). It ranged from
errors generated for Sw calculation, when forcing a to 24 to 1611 with a mean value of 206.
any fixed values.
4.2. Determination of m and a
4. Studied samples characteristics
The conventional determination of m and a is based
In this study, 92 clean carbonate core samples were on Archie's equation that can be rewritten as:
selected from Asmari Formation in six wells of three oil
fields, Zagros Basin, southwest Iran. Oligo-Miocene logF ¼ loga−mlog/ ð15Þ
Asmari Formation is one of the most important carbonate This equation indicates that log–log plot of F versus ϕ
oil reservoirs in Zagros Basin. Microfacies analysis of the can be used to determine a and m. The cementation
samples examined by Maghsoodi and Rezaee (2005) exponent m, is negative slope of the least square fit
showed that Asmari Limestone deposited mainly in a straight line of the plotted points, while the tortuosity
carbonate inner ramp. Petrographic studies indicated factor is the intercept of the line at ϕ = 1. Fig. 2 shows
eight petrofacies in the samples. These petrofacies in-
cluded dolomicrite, dolomitized packstone–grainstone,
dolostone, sandy dolostone, grainstone, packstone, wack-
stone and mudstone. Mechanical and chemical compac-
tion, extensive cementation, selective dissolution and
pervasive dolomitization were major diagenetic process-
es affecting original texture of the formation. Dolomiti-
zation was a common feature in most samples and in
many cases was completely overprinted the original
texture. Most of dolomite crystals varied in size from
medium to micrite resulting in very fine pores and pore
throats. Original pore spaces were mainly occluded by
different generation of calcite and anhydrite cements.
Porosity of the samples was generally fine intercrystal-
line, vugs and moldic. In most samples, no connected
visible porosity could be seen (Maghsoodi and Rezaee, Fig. 2. Log porosity versus log F. Forcing fit line to intercept y-axis at
2005). a = 1 and F = 1, has led to a higher value for m.
 M.R. Rezaee et al. / Journal of Petroleum Science and Engineering 56 (2007) 241–251 245

log-porosity versus log formation factor for the samples Table 3

of this study. Two different methods were applied to Free best fit line equations, m, a, R2 and number of samples for seven
petrofacies
obtain m and a (Fig. 2). The first was to use a free best
fit line (continuous line). The slope of this free fit line is Petrofacies Equations m a R2 No. m (from
forced
− 0.94 and intercept is 14.34 with a coefficient of method)
determination of 0.57. This indicates values of 0.94 and
Dolomicrite F = 5.22ϕ− 1.11 1.11 5.22 0.97 3 2.03
14.34 for m and a, respectively. Dotted line is the best fit
Mudstone F = 0.5ϕ− 2.4 2.4 0.5 0.90 5 2.14
line that is forced to intercept y-axis at ϕ = 1 (logϕ = 0) Grainstone F = 22.56ϕ− 0.74 0.74 22.56 0.87 10 1.77
and F = 1 (logF = 0). The slope of this line is − 1.95 Packstone F = 12.53ϕ− 0.94 0.94 12.53 0.75 14 1.84
indicating a value of 1.95 for m, and 1 for a. Dolomitized F = 49.5ϕ− 0.55 0.55 49.5 0.66 16 1.87
Scattering of the data in Fig. 2, leading to small packstone–
grainstone
coefficient of determination (R2 = 0.57), represent het-
Dolostone F = 12.58ϕ− 0.1.03 1.03 12.58 0.56 36 2.08
erogeneities of the samples. Therefore, m and a derived Wackstone F = 14.8ϕ− 1.02 1.02 14.8 0.31 6 2.21
from these methods are imprecise. Many studies have
Sandy Dolostone due to very small sample size (2 samples) was
been carried out to classify rocks based on their lithol- excluded. Last column shows m values using forced fit method for
ogy, facies, porosity type and permeability to achieve a each petrofacies.
better correlation between porosity and F (e.g. Focke
and Munn, 1987; Byrnes et al., 2003).
equation, m, a, R2 and the number of samples for each
5. Results and discussions on the classification petrofacies. Free best fit line equations provided unusual
methods values of m and a for most of the petrofacies. Forced
best fit line provides more usual values for m (Table 3,
In the following sections different classifications last column).
have been applied for the samples to find more accurate
relationship between porosity and F. 5.2. Classification based on permeability and petro-
facies-permeability groups
5.1. Classification based on petrofacies
Samples were classified based on their permeability
In order to obtain more accurate correlation between classes to evaluate the effect of permeability on the ϕ − F
porosity and F, the studied samples were classified plot. The permeability classes were defined as: K b 1 mD,
based on their petrofacies. Fig. 3 shows the cross-plot of 1 b K b 5, 5 b K b 10; 10 b K b 50 and K N 50 mD. Fig. 4
porosity versus F for different petrofacies. Some petro- shows the cross-plot of porosity versus F for each class of
facies appear in narrower trends and show higher R2 permeability. Although the classes were plotted separate-
values. Table 3 presents the power law free best fit line ly to some extent, but a well defined fit line could not be

Fig. 3. Log–log plot of porosity versus formation factor for eight

petrofacies. Some of the petrofacies such as grainstone and packstone
show a narrower and relatively well defined trend. Dolostone samples Fig. 4. Log–log plot of porosity versus formation factor. Samples in
are scattered on the plot. this plot are separated based on permeability classes.
246 M.R. Rezaee et al. / Journal of Petroleum Science and Engineering 56 (2007) 241–251

obtained for each class using this method. This may be

due to a weak relationship between F and permeability.
Cross-plot of permeability versus F shows a weak cor-
relation (Fig. 5). For example samples with permeability
of 20 mD show F values of 30 to 526. This suggests that
although there are many common parameters in hydraulic
and electrical conductivity of porous media, but it seems
other factors must be taken into account for carbonates
with complex network of pores and pore throats.

5.3. Classification based on flow zone indicator

Generally, there is a weak correlation between poros- Fig. 6. Semi-log cross-plot of porosity versus permeability for the
ity and permeability, especially in carbonates. To find a samples. The determination coefficient is 0.59.
better correlation between porosity and permeability and
define hydraulic flow unit (HFU), Amaefule et al.
(1993) developed an expression as: In the present study, FZI values were used for
rffiffiffiffi classification of the samples. The value for each sample
K was calculated using Eq. (16). LogFZI was applied to
0:0314
/ separate different HFUs. The samples were grouped in
FZI ¼ ð16Þ
/Z four HFUs using four FZI classes (FZI N 0.5,
0.5 N FZI N 0, 0 N FZI N − 0.5 and −0.5 N FZI N − 1). Plot-
where FZI, K and ϕ are Flow Zone Indicator (μm),
ting porosity versus permeability regardless of the sam-
permeability (mD) and porosity (fraction), respectively.
ple classification resulted in scattered plot and low
ϕZ is pore to matrix volume ratio (PMR) and can be
determination coefficient (Fig. 6). However, in the same
expressed as:
cross-plot when data were grouped in separate HFUs,
/ R2 was significantly increased (Fig. 7).
/Z ¼ ð17Þ With the same HFU groups, samples were later
1−/
plotted on the F − ϕ cross-plot (Fig. 8). Scattered data in
The equation defines a relationship between volume each HFU indicates that this approach was not also
of void space (ϕ / 1 − ϕ) and its geometric distribution successful for binning porosity and F in well-defined
(√K / ϕ). A hydraulic flow unit with identical hydraulic groups.
properties shows close FZI values. On a semi-log plot of In general, it can be stated that classification of
permeability versus porosity, samples with similar FZI samples based on permeability or FZI values is not
values normally plot together indicating close relation- successful. It is indicated that hydraulic and electrical
ship between porosity and permeability in each HFU. path are not identical and hydraulic tortuosity is much

Fig. 5. Cross-plot of permeability versus formation factor. An inverse

relationship exists between F and permeability. In general the Fig. 7. The same cross-plot as of Fig. 6, but data were grouped in four
relationship is poor and R2 is 0.25. hydraulic flow units (HFUs). Note the large R2 values for each HFU.
 M.R. Rezaee et al. / Journal of Petroleum Science and Engineering 56 (2007) 241–251 247

Fig. 8. Log–log plot of porosity versus formation factor. The samples Fig. 10. Log–log plot of porosity versus formation factor. Samples
are classified based on FZI values into four HFUs. There is no clear have been grouped in four electrical flow units (EFUs) based on their
separation for each HFU on the plot. CZI values.

larger than electrical tortuosity. This is a fact as perme- R2 and thus less reliable values for derived m and a. As
ability scales to a pore throat radii with a power of four discussed in previous sections, different approaches did
and the electric conductivity scales to a pore throat radii not improve correlation between the data.
with a power of two (David, 1993). A study by Zhang Combining Poiseuille's law for flow in cylindrical
and Knackstedt (1995) on fluid-flow and electrical con- tubes, Darcy's law for fluid flow in a porous media and
ductivity of three dimensional random porous medium Kozeny–Carman model (1937), Amaefule et al. (1993)
at a microscopic level showed that hydraulic tortuosity defined reservoir quality index (RQI) as:
is systematically larger than the electrical tortuosity, and
can differ by as much as an order of magnitude at lower sffiffiffiffi
porosities. Another study by Slater and Lesmes (2002) K
RQI ¼ 0:0314 ð18Þ
indicated that permeability does not correlate with F in /
unconsolidated sediments. Another study by Hilfer and
Manwart (2001) on three-dimensional computer tomo- where RQI is in μm, K is permeability (mD), ϕ is porosity
graphic image of Fontainebleau sandstone revealed that (fraction) and 0.0314 is the conversion factor from mD to
the permeability and F can differ significantly even in ìm. RQI is an estimation of mean hydraulic pore throat
models with identical geometrical properties. radius. In a given porosity, an increase in mean hydraulic
radius will increase permeability. In other words, higher
6. A new method for sample classification RQI values indicate better reservoir quality.
Although, a global relationship has not been found
Like porosity and permeability, in log–log plot of between permeability and F, an inverse proportionality
porosity versus F, data may scatter significantly for a has been reported by many authors (Wong et al., 1984;
mixture of heterogeneous rocks. This will lead to small Guyon et al., 1987; Kostek et al., 1992; Nettelbladt

Fig. 9. A comparison between RQI and ERI. For most of the samples there is a good match between RQI and ERI. Note that scales are different.
248 M.R. Rezaee et al. / Journal of Petroleum Science and Engineering 56 (2007) 241–251

Table 4 1995; Hilfer and Manwart, 2001; Slater and Lesmes,

Summary of data obtained from Fig. 10 2002) which have shown that hydraulic and electrical
EFU CZI Class Equation m a R2 path are not identical.
1 CZI N 0.3 F = 5.58ϕ− 1.1 1.10 5.58 0.97 ERI is an indicator of electrical radius for each sample.
2 0.3 N CZI N 0.25 F = 7.01ϕ− 1.18 1.18 7.01 0.96 In order to separate samples with similar electrical flow
3 0.25 N CZI N 0.2 F = 9.35ϕ− 1.22 1.22 9.35 0.96 properties, ERI must be divided to volume of void space
4 CZI b 0.2 F = 14.19ϕ− 1.34 1.34 14.19 0.85
or pore to matrix volume ratio (ϕ / 1 − ϕ):
rffiffiffiffi
et al., 1995; Celzard and Marêché, 2002). Taking this /
fact into account, an identical ratio to Eq. (18) can be F
CZI ¼ ð20Þ
defined as: /Z
rffiffiffiffi
/ where ϕ, F and ϕZ are porosity (fraction), formation
ERI ¼ ð19Þ factor and pore to matrix volume ratio (PMR) respec-
F
tively. The equation defines relation between the volume
Since the ratio is an indication of electrical radius, it is of void space (ϕ / 1 − ϕ) and its electrical flow properties
called Electrical Radius Indicator (ERI). Unlike RQI (√ϕ /F). The equation has been able to separate samples
which provides mean value of hydraulic radius, ERI is with similar electrical flow properties, and for this reason
dimensionless and only quantitatively compares electri- it is named “Current Zone Indicator” (CZI). CZI is a
cal radius of samples. Eq. (19) shows that with a given factor that can be used to separate samples with relatively
ϕ, a decrease in F, will increase ERI and vice versa. A identical m and a, where the variation in F is just the
comparison of RQI and ERI represents a relatively good function of porosity. Electrical flow unit (EFU), units
match in terms of data fluctuation (Fig. 9), suggesting with the same electrical flow properties, has been pro-
that both of them may address the same properties. posed for any samples that fall within a defined range of
Lack of exact match between RQI and ERI values CZI. It is obvious that a homogenous interval of a
supports other studies findings (Zhang and Knackstedt, reservoir with similar electrical flow properties will show

Fig. 11. Comparison of measured F with calculated F from CZI (A), free (B) and forced (C) methods. Obtained m and a values from CZI resulted in
more accurate values for F.
 M.R. Rezaee et al. / Journal of Petroleum Science and Engineering 56 (2007) 241–251 249

Table 4 lists CZI classes, power law best fit line

equation, m, a and R2 between porosity and F in each
EFU. With this approach, R2 has been increased signif-
icantly which in turn enables us to obtain more reliable
m and a.

7. Comparison of F derived from different methods

In this study different methods provided different

values for m and a. A free best fit line indicated m = 0.94
and a = 14.34, and a forced fit line with a = 1, showed a
Fig. 12. Histogram showing variations of calculated tortuosity factor value of 1.95 for m (Fig. 2). Classification of the sam-
using Eq. (9) for the studied samples. For most of the samples the
tortuosity is larger or smaller than 1.
ples based on their CZI values also revealed different m
and a (Table 4). Formation factor was calculated for
close CZI values. Such an interval is a unit that has each sample using Archie's equation, with m and a
identical electrical flow properties that is named here as values derived from the three methods. Calculated F
an EFU. was then plotted against measured F (Fig. 11). A sig-
Eq. (20) was used to calculate the CZI value for each nificantly better R2 (0.95) was achieved using CZI
sample in this study. The CZI values ranged from 0.43 to method.
0.12 with an average of 0.26. Four EFUs were defined
using four CZI classes including CZI N 0.3, 8. Water saturation calculation sensitivity using
0.3 N CZI N 0.25, 0.25 N CZI N 0.20, CZI b 0.20. different methods
Fig. 10 shows the cross-plot of porosity versus F
binned in four CZI classes. The main difference between In this section, the influence of m and a on Sw
each EFU was the amount of isolated vuggy and/or calculation using different methods will be discussed.
moldic porosity. From EFU1 to EFU4 the number of With a fixed a (a = 1) the slope of the best fit line
isolated pores increases. This suggests that, increase in increases comparing to the free best fit line (Fig. 2).
tortuosity from EFU1 to EFU4 is not due to porosity Using Eq. (9), if m changes from m1 (1.1, 1.18, 1.22 and
reduction. Higher tortuosity of pore throat networks is 1.34, Table 4) to m2 = 1.95, then a must be adjusted to
due to presence of isolated and dead-end pores which in achieve the same F value for each sample. In the other
turn lead to a longer pathway. word, when m is considered 1.95, the calculated a values
must be used instead of a = 1 to obtain accurate F. Fig. 12
shows histogram of calculated a values from Eq. (9). It
shows that about 14% of a values are close to 1, 56% are
higher than 1.15 and 30% are lower than 0.9.

Fig. 13. A comparison of formation factors calculated from forced and

CZI methods. As calculated tortuosity factor (bars at the top of the
figure) departure from 1, F values calculated from two methods depart
from each other. When a N 1 (left side of figure) F from forced method Fig. 14. A comparison between formation factors calculated from free,
generally shows smaller values. When a b 1 (right side of figure) F forced and CZI methods. F calculated from free method varies
from forced method show larger values. considerably and do not match with F derived from CZI method.
250 M.R. Rezaee et al. / Journal of Petroleum Science and Engineering 56 (2007) 241–251

Fig. 13 compares F values obtained from forced and from EFU1 to EFU4 is mostly due to an increase in the
CZI methods. According to Archie's water saturation isolated and dead-end pores. This study suggests, unlike
equation, F has direct relation with Sw in a given Rw rocks with intergranular and well-connected pores, for
and Rt. Where a values depart from 1, calculated for- rocks with complex pore networks where most of the
mation factors from two methods depart form each other intergranular pores are occluded by cements and
progressively (Fig. 13). With a values of larger than 1, irregularly-distributed secondary pores are either isolated
56% of calculated F from forced method shows lower or connected by tortoise path, tortuosity plays an
values comparing to CZI method (left part of Fig. 13) important role controlling electrical conductivity.
resulting in underestimation of Sw. When a values are This suggests that, increase in tortuosity from EFU1
less than 1, 30% of F from forced method show higher to EFU4 is not due to porosity reduction. Higher
values (right part of Fig. 13) leading to Sw overestima- tortuosity of pore throat networks is due to presence of
tion. Only 14% of calculated F values from forced isolated and dead-end pores which in turn lead to a
method are valid and gain correct Sw for the studied longer pathway.
samples.
In the meanwhile, F values calculated from the free Acknowledgements
method does not match the F values from CZI method
(Fig. 14). The small correlation between porosity and F The authors are grateful for sponsoring and financial
(Fig. 2) has led to small fit between these values. support by the NIOC – Research & Development
Directorate, the University of Tehran and Research
9. Conclusion Institute of Petroleum Industry (RIPI). Mr. M. Rahimi
and K. Saadat are acknowledged for conducting lab
This study shows that for heterogeneous carbonates tests. We thank Dr. S. Etemadi for the critical review of
with microscopically complex pore networks the the article. The authors acknowledge NIOC – Research
relation between F and porosity is not straightforward. & Development Directorate for permission to publish
For such complex reservoirs, using a unique value for m this paper.
and a will lead to an inaccurate estimation of
hydrocarbon reserves. References
The most basic and widely used form of Archie's
equation for carbonates is: Amaefule, J.O., Altunbay, M., Tiab, D., Kersey, D.G., Keelan, D.K.,
1993. Enhanced reservoir description: using core and log data to
identify hydraulic (flow) units and predict permeability in uncored
1
F¼ ð21Þ intervals/wells. SPE Pap. 26436, 1–16.
/2 Archie, G.E., 1942. The electrical resistivity log as an aid in
determining some reservoir characteristics. Trans. Am. Inst. Min.
Although, an assumed m = 2 and a = 1 is relatively Metall. Pet. Eng. Inc. 146, 54–62.
Attia, M.A., 2005. Effects of petrophysical rock properties on
fair choice for carbonates with dominantly interparticle tortuosity factor. J. Pet. Sci. Eng. 48, 185–198.
and intercrystalline porosity, for carbonates with Bird, R.B., Stewart, W.E., Lightfoot, E.N., 1960. Transport Phenom-
secondary porosity and complex network of pores and ena. John Wiley & Sons, New York.
pore throat however it may cause a significant error in Borai, A.M., 1987. A new correlation for the cementation factor in
Sw estimation. low-porosity carbonates. SPE Form. Eval. 2, 495–499.
Byrnes, A.P., Franseen, E.K., Wateney, W.L., Dubois, M.K., 2003. The role
This study shows that classification of rocks based on of moldic porosity in paleozoic kansas reservoirs and the association of
petrofacies, permeability and FZI is inadequate to obtain original depositional facies and early diagenesis with reservoir
accurate values for m and a. In additions, using free best properties. AAPG Annual Convention in Salt Lake City, Utah.
fit line without sample classification and or fixing a to a Carman, P.C., 1937. Fluid flow through granular beds. Trans. Inst.
constant value, which causes m to increase, may lead to Chem. Eng. 15, 150–167.
Celzard, A., Marêché, J.F., 2002. Fluid flow in highly porous
both over and underestimation of Sw. anisotropic graphites. J. Phys., Condens. Matter 14, 1119–1129.
Application of CZI method to classify the samples in David, C., 1993. Geometry of flow path for fluid transport in rocks.
well defined groups (EFUs) provided a suitable method to J. Geophys. Res. 98, 12267–12278.
obtain accurate a and m and thus better estimation of Sw. Dubois, M.K., Byrnes, A.P., Watney, W.L., 2001. Field development
CZI method has shown that the samples fall into multiple and renewed reservoir characterization for Co2 flooding of the
Hall-Gurney Field, Central Kansas. AAPG Annual Convention in
groups where the variation is mostly due to tortuosity Denver, Colorado.
factor. For each group, unlike m that varies slightly from Focke, J.W., Munn, D., 1987. Cementation exponents in middle
1.1 to 1.3, a changes from 5 to 19. The wide variation of a eastern carbonate reservoirs. SPE Form. Eval. 2, 155–167.
 M.R. Rezaee et al. / Journal of Petroleum Science and Engineering 56 (2007) 241–251 251

Guyon, E., Oger, L., Plona, T.J., 1987. Transport properties in sintered Nugent, W.H., Coates, G.R., Peebler, R.P., 1978. A new approach to
porous media composed of two particle sizes. J. Phys., D. Appl. carbonate analysis. 19th SPWLA Symposium, Paper O.
Phys. 20, 1637–1644. Rasmus, J.C., 1983. A variable cementation exponent, m, for fractured
Hamada, G.M., Al-Awad, M.N.J., Alsughayer, A.A., 2002. Water carbonates. Log Anal. 24, 13–23.
saturation computation from laboratory, 3D regression. Oil Gas Salem, H.S., Chilingarian, G.V., 1999. The cementation factor of
Sci. Technol. – Rev. IFP 57, 637–651. Archie's Equation for shaly sandstone reservoirs. J. Pet. Sci. Eng.
Hilfer, R., Manwart, C., 2001. Permeability and conductivity for 23, 83–93.
reconstruction models of porous media. Phys. Rev., E 64, 021304. Sethi, D.K., 1979. Some considerations about the formation resistivity
Hirasaki, G.J., 2005. Flow and Transport Through Porous Media, factor-porosity relations. 20th SPWLA Symposium, paper L.
Course note. Slater, L., Lesmes, D.P., 2002. Electrical-hydraulic relationships
Hosseini-nia, T., Rezaee, M.R., 2002. Error sensitivity of petrophy- observed for unconsolidated sediments. Water Resour. Res. 38,
sical parameters on water saturation calculation for hydrocarbon 31-1.
reservoirs. J. Sci. Univ. Tehran 28, 69–91. Timur, A., Hempkins, W.B., Worthington, A.E., 1972. Porosity and
Kostek, S., Schwartz, L.M., Johnson, D.L., 1992. Fluid permeability in pressure dependence of formation resistivity factor for sandstones.
porous media: comparison of electrical estimates with hydrody- Presented at Form. Eval. Symp. Can. Well Log. Soc., 4th, Calgary
namical calculations. Phys. Rev., B 45, 186–195. (Paper, D).
Maghsoodi, F., Rezaee, M.R., 2005. Effects of microfacies, sedimen- Wong, P.-Z., Koplik, J., Tomanic, J.P., 1984. Conductivity and
tary environment and digenesis on the reservoir quality of the permeability of rocks. Phys. Rev., B 30, 6606–6614.
Asmari Formation, Gachsaran Field. Geol. Soc. Iran 683–698. Wyllie, M.R.J., Rose, W.D., 1950. Some theoretical considerations
Nettelbladt, B., Ahlent, B., Niklassons, G.A., Holtll, R.M., 1995. related to the quantitative evaluation of the physical characteristics
Approximate determination of surface conductivity in porous of reservoir rock from electrical log data. Trans. AIME 189,
media. J. Phys., D. Appl. Phys. 28, 2037–2045. 105–118.
Neustaedter, R.H., 1968. Log evaluation of deep Ellenburger gas Zhang, X., Knackstedt, M.A., 1995. Direct simulation of electrical and
zones. SPE Paper 2071, presented at the Deep Drilling and hydraulic tortuosity in porous solids. Geophys. Res. Lett. 22,
Development Symposium-Delaware Basin of the SPE of AIME, 2333–2336.
Monahans, Texas.