Professional Documents
Culture Documents
Faculty of Science
Ain Shams University
Prepared by
Anas M. ElGendy
Abstract …………………………………….………...1
Introduction…………………………………………...3
Units…………………………….…………………….7
C H AP T E R 1
Calculation of Irreducible Water Saturation Swirr from NMR Logs
(In Tight Gas Sand)
o Preface .................…………….…………………………….…………………….………….7
o Problems of Swirr Calculation from NMR Logs in Tight Gas
Sands............……….……………………….………………….…………………….……..…8
o Novel Model of Estimating Swirr from NMR Logs ……….…………..13
o SDR Model............……….…………………………………….………………….…..…15
o Case studies ............……….…………………………………….………………….…….16
C H AP T E R 2
Calculation of Irreducible Water Saturation from Capillary Pressure
and Electrical Resistivity Measurements
Conclusion ……………………….……………….….44
Reference…….………………….……………….….46
List of Figure
Fig.1.1 Statistical graph of the T2 cutoff for 36 core samples in tight sandstones.
Fig.1.2 Comparison of the shape of the T2 spectrum for four different
conditions.
Fig.1.3 Comparison of the T2 spectra in gas-bearing formations.
Fig.1.4 Relationship between the core porosity and the permeability intight
reservoirs in four wells.
Fig.1.5 Relationship between T 2lm calculated from field NMR logs and from
NMR experimental measurement.
Fig.1.6 Relationship of T 2lm obtained from laboratory NMR experimental
measurements under two different conditions.
Fig.1.7 Calculation of irreducible water saturation from field NMR logs in
tight gas sands and watersaturated layers and the comparison with the core
analysis results.
Fig.1.8
o Fig1.8a Comparison cross plots of Swirr, and core analyzed results
o Fig1.8b Comparison cross plots of Swirr, Swi_20.75 ms, and core
analyzed results.
o Fig1.8c Comparison cross plots of Swirr, Swi_33 ms and core analyzed
results.
Fig.2.1 Typical electrical resistivity trends for non-conductive and saturated
mineral rocks as modeled with Equation 2.
Fig.2.2 Schematic of the capillary versus saturation cell (after Core Lab
Instruments, 2000).
Fig.2.3 Sketch of the electrical resistivity measurement setup.
Fig.2.4 Equivalent two-electrode circuit and the effect of blocking electrodes.
Fig.2.5 Effect of electrode polarization on impedance measurement. a)
Varying electrode Lgap for constant rock resistivity 𝜌𝑅 = 1 Ωm; b) Varying
rock resistivity for constant electrode gap Lgap = 10-6 m
Fig.2.6 _ Effect of type of electrode on the measurement of the impedance on
Berea sandstone core specimen 22. a) Al foil electrodes; b) Silver-based paint
electrodes. Electrode polarization controls the impedance measurement with
the Al foil electrodes.
Fig.2.7 Capillary pressure-brine saturation curves for all eight Berea
sandstone core specimens.
Fig.2.8 Pore size distribution for all eight Berea sandstone core specimens.
Fig.2.9 Electrical resistivity versus brine saturation for all eight Berea
sandstone core specimens.
Fig.2.10 Capillary pressure and resistivity index versus brine saturation in
Berea sandstone core specimens.
Fig.2.11 Capillary pressure and resistivity index versus brine saturation in
twelve Quartz (very fine-grained) sandstone core specimens.
Fig.2.12 Capillary pressure and resistivity index versus brine saturation in ten
Limestone core specimens.
Fig.2.13 Effect of percolation on the capillary pressure and electrical
resistivity measurements.
Fig.2.14 Effect of brine saturation the saturation exponent in Berea sandstone
specimens.
Fig.2.15 Effect of brine saturation the saturation exponent in Berea sandstone
specimens.
Fig.2.16 Effect of brine saturation the saturation exponent in Berea sandstone
specimens.
Fig.2.17 Calculated tortuosity versus degree of saturation in Berea sandstone
specimens.
Figures from 1.1 to 1.8 shows the calculation which made by:
Figures from 2.1 to 2.8 shows the calculation which made by:
1
analyzed results both in tight gas formations and water-saturated layers, the
absolute error is in the range of ±4%. The calculated results by using 20.75 ms
as the T2 cutoff are accurate in water-saturated layers but are overestimated in
gas-bearing intervals. Defining 33 ms as the T2 cutoff is unusable both in gas-
bearing and water layers.
2
Introduction
In the upstream oil and gas industry, new ideas and solutions have been
developed as a result of computerization, technology and building on the
knowledge of prior discoveries and amassed data, especially with regard to
residual oil, irreducible water saturation and the overall process of fluid
displacement in porous media. Before the previous century (in the 1800's)
injecting water in an oil field was outlawed in the United States. However,
subsequent to the success of an accidental water-flood, the industry gradually
accepted this technology. In the mid 1900's, the preponderance of evidence
indicating the benefits of water-flooding, encouraged petroleum technologists
to study the displacement process theoretically and in the laboratory. Detailed
laboratory analyses and mathematical derivations were developed to explain
and predict the performance of a water-flood. It has been well documented that
the immiscible displacement process of oil by water behaves in a non-uniform,
channel-like fashion. However, in order to develop mathematical models such
as the "Buckley-Leverett Frontal Advance Theory", the concept of an oil bank
and flood front were defined. It had been stated by the first generation of water-
flood scientists and engineers that such mathematical models are an extreme
simplification, and should advancements in computation occur, then more
accurate estimates of water-flood performance could be performed3. In
addition to the simplifications, assumptions have been made for the purpose of
generating "conservative" results. In spite of these simplifications and
conservative bias, previous interpretations have resulted in a reasonably
accurate solution that represents a first-order or second-order approximation.
4
apprehended using seismic profiles; gamma-ray and sonic lateral logs, as well
as cores and cuttings taken in drillwells. The evaluation of residual oil
saturation, multiphase flow and oil production techniques from the Ashtart
reservoir also depend on variations and zoning of the irreducible water
saturation. Estimation of the initial water saturation and hence variations in the
capillary pressure in the reservoir, required compilations of porosity data
measured on cores, supplemented by additional but computed porosities based
on acoustic log diagrams. Furthermore, Gamma Ray, Sonic log, and well to
well correlations tied to core results and well cuttings, help recognize the
layered lithologies within the El Garia flat lying but stratified, Ypresian in
origin reservoir rocks. Abundant permeability and porosity values compiled in
the light of seismic sequence and Gamma Ray and Sonic log details, were
integrated in an empirical approach using the Leverett J function, to model the
irreducible water saturation depending on the capillary pressure distribution in
the whole reservoir. Variations of this principal hydraulic parameter in a wide
range (Swirr: 12 to 40%) compared to the preceding lithostratigraphic,
petrographical and petrophysical results help recognize four main rock pore
types in the commercial Ashtart reservoir. These vary from (a zone with a rock
pore type showing an irreducible water saturation as low as 12%, and a fairly
good reservoir character in the lower third part of the lithologic column which
is thought to channelize a multiphase fluid flow in the global oilfield, to those
zones built-up of rock pore types with higher initial water saturation amounts
which in certain cases tend to indicate zones of degraded reservoir. Our study
suggests that diagenesis prevalently controls porosity, due to operative
dissolutions of the Nummulitid tests/bioclasts, and cementation; moreover,
diagenesis exerts effects on permeability by interconnecting intergranular and
intratest pore spaces. In contrast, microfracturing enhances permeability of the
5
reservoir. This is notably the case in the fairly permeable central zone in the
Ashtart reservoir with excellent petrophysical parameters, but which were
found to degrade gradually towards its peripheries.
6
CHAPTER 1
1 Preface
Tight gas reservoirs always display characteristics of micropore body, small pore
throat radius and poor pore connectivity. The proportion of micropore space is large,
which leads to irreducible water saturation Swirr being higher than that of
conventional formations, thus resulting in low resistivity contrast between
gasbearing formations and water-saturated intervals. It is difficult to distinguish tight
gas formations from water-saturated layers. To improve the reliability of tight gas
reservoirs evaluation, it is necessary to obtain information on Swirr and nuclear
magnetic resonance (NMR) logs have a unique advantage in this aspect. Although
Swirr can be calculated from NMR T2 distribution after a T2 cutoff is defined in
conventional reservoirs, it is of great difficulty to calculate Swirr from NMR logs
even if the T2 cutoff has been obtained in low-permeability oil-bearing formations
and tight gas sands. NMR T2 spectrum is distorted due to the contribution of bulk
relaxation of light oil and diffuse relaxation of natural gas. In gas-bearing intervals,
the T2 spectrum of natural gas is overlapped with that of irreducible water. Parts of
the T2 spectrum of natural gas are considered to be that of irreducible water and the
calculated Swirr from NMR logs will be overestimated. To remove the effect of bulk
relaxation of light oil and diffuse relaxation of natural gas on the NMR T2
distribution, the best way is to estimate Swirr from NMR logs without a T2 cutoff. In
7
this study, by means of transforming Timur and Schlumberger—Doll Research
(SDR) models, a novel model of calculating Swirr from NMR logs is derived.
The common method of calculating Swirr from NMR logs is to define a T2 cutoff,
which segregates NMR T2 distribution into two volumes. Swirr is defined as the ratio
of the accumulation of T2 distribution for T2 relaxation time being lower than T2
cutoff to the sum area of T2 distribution.
2.1 Problems of Calculating Swirr from NMR Logs by Using a Defined T2 Cutoff
8
50 %
40 %
30 %
20 %
10 %
0 %
Figur.1.1 _ Statistical graph of the T2 cutoff for 36 core samples in tight sandstones
9
four saturated conditions: fully saturated with water, irreducible water saturation (the
pore volume contains irreducible water and air through centrifuge), hydrocarbon-
bearing condition (the pore space consists of irreducible water and hydrocarbon by
kerosene or transformer oil drainage) and residual oil saturation from the drainage
of the samples at hydrocarbon-bearing conditions with brine. Residual oil condition
was used to simulate the flushed zone of hydrocarbon formation in field NMR logs.
The comparison of NMR T2 distribution for two core samples is displayed in
Figure.1.2, Figure.1.2a shows the comparison of T2 distribution under four
conditions for conventional core samples, with the transformer oil used. Fig. 2b is
for low-permeability sandstone core samples with the same condition as in Fig. 2a,
with kerosene used as the oil. The black dotted lines in these two figures are the
trough of bimodal NMR T2 distribution with residual oil saturation, which is always
considered to be T2 cutoff in field NMR logs, and the black solid lines mean T2 cutoff
acquired from laboratory NMR measurements under fully water-saturated condition.
Figur.1.2 _ Comparison of the shape of the T2 spectrum for four different conditions: 1-
irreducible water saturation; 2-fully brine-saturated; 3-residual oil saturation; 4-oil-bearing
condition; 5-bulk relaxation of transformer oil; 6-bulk relaxation of kerosene.
10
As shown in Figure.1.2a two core samples have the same porosity, whereas
permeability of the core sample shown in Figure.1.2a is higher. Figure.1.2b
illustrates that for core plug with conventional porosity and permeability, when the
pore space is occupied by transformer oil, the NMR T2 spectrum under residual oil-
saturated condition will be distorted only a little, because for rocks with good pore
structure, the T2 distribution with fully water saturation is wide, and the T2 spectrum
reflecting the bulk relaxation of transformer oil overlaps with that of movable water.
The T2 cutoff acquired from above two different conditions is almost the same. It can
be concluded that the T2 cutoff acquired from laboratory NMR measurements with
fully water-saturated condition can be used in field reservoir evaluation directly for
conventional rocks or formations. However, for a low-permeability sandstone core
sample Figure.1.2b, the T2 spectrum is narrow, when the pore space is occupied by
non-wetting phase kerosene, the T2 distribution with residual oil saturation will be
wider than that with fully water saturation, and the NMR spectrum with long T2
relaxation time reflects the bulk relaxation of kerosene. T2 cutoffs obtained from
these two different saturated conditions are discrepant. If the T2 cutoff obtained from
fully water-saturated laboratory NMR measurements is used for formation
evaluation directly, the Swirr calculated from field NMR logs will deviate.
Figure.1.2b has demonstrated that calculating Swirr from field NMR logs by using a
T2 cutoff, which is acquired from NMR experimental measurement, is improper in
low-permeability sands. The T2 cutoff is also usable in tight gas sands because of the
effect of diffusion relaxation of natural gas. At present, core NMR experimental
11
Figure.1.3 _Comparison of the T2
spectra in gas-bearing formations. The
indication by and is the same as in
measurements under the gas-bearing condition cannot be carried out due to the
limitation of experimental apparatus. To illustrate the effect of natural gas on the
NMR T2 spectrum, the comparison of NMR T2 distribution obtained from
experimental measurement and from field NMR logs is displayed in Fig. 3. Figure
3 illustrates that when the pore space is occupied by natural gas, the morphology of
the T2 spectrum is distorted. The T2 spectrum moves to the left, and the signal of
natural gas overlays with that of irreducible water, which makes the amplitude of the
left peak increase and that of the right peak decrease. If the T2 cutoff acquired from
the NMR experimental data set is used for Swirr calculation directly, the Swirr will be
overestimated. To estimate Swirr in tight gas sands accurately, the best way is to
propose a novel model to calculate Swirr from NMR logs without T2 cutoff.
12
3. Novel Model of Estimating Swirr from NMR Logs
Eq.1.1
Where; K is the rock permeability in units of 10-3 lm2, u is the rock porosity, and
Swirr is the irreducible water saturation, their units being fractions.
In Timur model, the rock permeability can be estimated once the values of u and
Swirr have been defined. On the contrary, Swirr can be derived with the values of K
and u.
Eq.2.2
13
Where; a, b and c are statistic model parameters. For different kinds of fields or
formations, different values will be defined to acquire permeability from porosity
and Swirr, their values should be calibrated by core samples.
Figure.1.2 displays the relationship among rock porosity, permeability and Swirr.
After the values of a, b and c have been calibrated, Swirr can be calculated once the
input parameters of porosity and permeability are acquired. Porosity can be
estimated by integrating NMR with conventional logs. The calculation of
permeability is a challenge in tight gas sands due to the generally poor correlation
between porosity and permeability Figure.1.4. To obtain accurate Swirr, Eq.1.2
should be transformed to avoid requiring permeability but to obtain information from
NMR logs.
X2_well
A
X2_well
B
X2_well
C
X2_well
D
X4_well
B
X4_well
C
X4_well D
X6_well B
X6_well D
Fig.1.4 _ Relationship between the core porosity and the permeability in tight reservoirs in four wells
14
3 SDR Model
An SDR model has been proposed by Schlumberger—Doll Research Center to
estimate permeability from NMR logs directly and is written as
Eq.1.3
Where; T2lm is the logarithmic mean of the NMR T2 spectrum in ms, C1, m1 and n1
are the statistical model parameters that can be acquired from core samples
experimental results.
Eq.1.4
Where; all the variables are the same meaning as in Eq.2, 3. Once the
parameters are defined as following:
Eq.1.5
(Eq.1.5) shows that once the values of C 2, m and n have been calibrated by NMR
experimental data set, Swirr can be estimated from NMR logs .
15
4 Case studies
With the model proposed above, the gas-bearing interval and water-saturated
layer in a well with field NMR logs in Xujiahe formation are processed. 36 core
samples were drilled for NMR experimental measurement in this region, the NMR
experiment data set being listed in Tabel.1.1.
With the data set listed in Table 1, the values of C2, m and n in Eq. (5) are calibrated
and Swirr calculation equation can be expressed as Eq.1.6.
Eq.1.6
By using this relationship, accurate Swirr can be obtained from field NMR logs once
integrating NMR with conventional logs, whereas T2lm could be decreased in gas-
this model to field applications, T 2lm should be corrected. To obtain accurate T 2lm in
16
Table. 1 36 core samples were drilled for NMR experimental measurement in Xujiahe formation
Table.1.1 _ 36 core samples were drilled for NMR experimental measurement in Xujiahe formation
17
Figure.1.5 _ Relationship between T2lm calculated from field NMR logs and from NMR experimental
measurement
Figure.1.7 shows a field example of calculating Swirr from field NMR logs by
using the model proposed in this study. Track (e) in Fig. 7 is the comparison of
T2lm calibrated from field NMR logs by using the relationship displayed in (Fig.1.5)
18
(T2lm) and analyzed from core samples (core_T 2lm). Track (f) in (Fig.1.7) displays
the comparison of the reservoir porosity calculated by integrating the interval
transit time with NMR logs (total_porosity) and acquired from core plugs
(core_porosity).
Ch.1 Fig.6 _ Relationship of T2lm obtained from laboratory NMR experimental measurements under two
different conditions.
Preferable consistency in these two tracks demonstrates the accuracy of the used
input parameters in (Eq.1.6). Track (g) in (Fig.1.7) displays the comparison of Swirr
derived by three different methods, with Swirr calculated by using the proposed
model in this study, S wi_20.75 ms calculated by using 20.75 as a T 2 cutoff and Core
Swirr analyzing irreducible water saturation from core samples. The drill stem testing
19
data displayed in the right of Fig.1.7 illustrates that Swirr matches very well with that
of core samples (core_ Swirr) in gas-bearing formation, whereas the irreducible water
saturation calculated by using 20.75 ms as a T 2 cutoff is higher. Track (h) in (Fig.1.7)
compares Swirr estimated from the proposed model with that alculated by using 33
ms as a T2 cutoff (Swi_33 ms), and the result demonstrates that the latter is not
reliable. These two comparisons mean that the T 2 cutoff obtained from laboratory
NMR measurements with full water saturation is inapplicable in gas-bearing
formations. In water-saturated
Figure.1.7 _Calculation of irreducible water saturation from field NMR logs in tight gas sands and water saturated
layers and the comparison with the core analysis results .
20
Fig.1.8a
Fig.1.8b
Fig.1.8c
21
Figure 1.8 Comparison cross plots of Swirr, Swi_20.75 ms, Swi_33 ms and core
analyzed results (core_S wirr) for 36 core sample layers, Swirr calculated from the
proposed model and by using 20.75 ms as a T 2 cutoff almost overlap with each other.
They all match well with the core analyzed results, whereas Swirr calculated by using
33 ms as a T2 cutoff is still overestimated. This is because the formation condition is
similar to the experimental simulation condition and T2 cutoff can be used in water-
bearing formations directly. The derived method is applicative not only in gas -
bearing formations but also in water-saturated layers.
Figure 1.7 displays the comparison cross plots of Swirr, Swi_20.75 ms, Swi_33 ms
and core analyzed irreducible water saturation. Figure 8a illustrates that the majority
of calculated Swirr is close to core_ Swi, and the absolute errors between them are
lower than 4%, which meets the requirements of tight gas sands evaluation. However,
the calculated irreducible water saturations by using 20.75 and 33 ms as the T2 cutoff
are all overestimated (Fig.8b, c).
22
CHA PT ER 2
1 Petrophysical Parameter
Eq.2.1
Where; ρmineral is the resistivity of the rock mineral, ρpore liquid is the electrical
resistivity of the pore fluid, ρpore gas is the resistivity of the pore gas, φ is the porosity.
Sr is the degree of pore liquid saturation. In most oil-bearing formation rocks, the
rock mineral and pore gas are insulators.
23
One of the drawbacks of this formulation is that Equation1 fails to model the
surface conduction in a formation. The effect of surface conductance in a formation
can be estimated by adding a surface conduction term to ch.2 eq.1 (Klein and
Santamarina, 2003; Santamarina et al., 2001). Then, the resistivity of the formation
becomes:
‘
Eq.2.2
Where; Θ is the surface conduction, γmin is the mineral unit weight, is the acceleration
of gravity, and S s is the specific surface of the formation mineral. Figure 1 shows the
effect of porosity and surface conduction on different saturated rock formations. This
figure is based in (Eq.2.2), which in spite of its limitations yields boundaries to the
evaluation of the different parameters. The proper interpretation of this figure
provides a solid base for the analysis of geophysical studies. Furthermore, the
electrical resistivity of a porous medium is also related to the micro-structural
properties, including porosity, pore geometry and surface morphology of the mineral
grains lining the pores. Archie (1942) proposes the following semi-empirical
relationship:
Eq.2.3
Where; a, m, and nare experimental determined factors. The factor a varies between
0.5 and 2.5. The factor m is also known as the cementation factor and it varies
between 1.3 and 2.5 for most rocks. The cementation factor is 1.3 for unconsolidated,
clean glass beads, 2 for sandstone formation, and around 5 for carbonates (Telford
et al., 1990; Mavko et al., 1998).
24
Fig.2.1 _ Typical electrical resistivity trends for non-conductive and saturated mineral rocks as modeled
with Equation 2 (surface conduction Θ = 1.4⋅10-9 S).
The factor n is known as the water saturation exponent and it has been related to
several petrophysical parameters. Typically, its value is around 2 in water-wet
formations and it is assumed constant for clean, sandy like formations (Archie, 1942;
Mavko et al., 1998). Lewis et al. (1988), Longeron et al. (1989a, b), Donaldson and
Siddiqui (1989), and Moss et al. (2000) found that the saturation exponent n
increases to values as high as ten in oil wet formations. Furthermore, Diederix (1982),
Anderson (1986), Longeron et al. (1989a, b), and Worthington and Pallat (1992)
25
found that the saturation exponent is not always constant over the whole saturation
range in disagreement with the early work by Archie (1942). The formation
resistivity factor FR, is the ratio between the resistivity, ρo, of the rock at 100%
saturated with conducting brine to the resistivity ρpore liquid of this brine. This ratio
depends on the lithological characteristics of the rock formation and the effective
porosity. The formation resistivity factor is sometimes expressed as function of the
percolation porosity φo that is the minimum porosity that connects the pores in a rock
formation:
Eq.2.4
Eq.2.5
Swanson (1985) found that the resistivity-saturation plot shows a change in the
slope at a saturation point where the corresponding capillary pressure curve indicate
a transition from macroporosity to microporosity. Finally, tortuosity is defined as
the ratio of the ionic path Li over the distance between the Electrodes 𝐿 𝑅 .
Alternatively, according to Tiab and Donaldson (1996), tortuosity is defined as:
Eq.2.6
26
Where 𝜑𝑡𝑟𝑎𝑝𝑝𝑒𝑑 and 𝜑𝑐ℎ𝑎𝑛𝑛𝑒𝑙 are porosities that correspond to the porosity of
the trapped fluid that does not contribute to electrical conduction and to the
porosity of the fluid that contributes to the conduction. In addition, Attia (2005)
found that the tortuosity factor is affected by porosity, cementation factor and
degree of brine saturation.
2 Experimental Tests
The porous plate method was used to measure the capillary pressure-saturation
curves for different types of core specimens (Berea sandstone, Quartz and
27
Limestone) with 5% NaCl brine solution. The procedure has several steps, including:
– Porous plate saturation: the procedure indicated in the manual “Soil Moisture
Capillary Pressure Cell” manual (Core Lab Instruments, 2000) is used to fully
saturate the porous stone saturation.
– Core specimen preparation: the core specimens are measured, cleaned, dried,
weighed, and saturated with the NaCl brine solution. After the saturation process is
completed the core specimens are re-weighed to determine the porosity-φ (Table1).
– Testing procedure: with the core specimen placed on top of the porous plate, the
air-brine capillary pressure procedure is ready to begin (Fig.2.2). A certain air
pressure pa is applied to the top of the core specimen and the water within the core
specimen is allowed to drain through the porous stone. The capillary pressure pc is
equal to the difference between the air pressure and the water pressure. When
equilibrium is achieved, the air pressure and the top of the cell are removed. Then
the core specimen is weighed and the degree saturation is calculated at the applied
capillary pressure Pc. The specimen is returned to the cell and the procedure is
repeated for higher values of the capillary pressure Pc. The capillary pressure versus
saturation curve is obtained and then used to estimate the pore size distribution.
28
Fig.2.2 _ Schematic of the
under varying degrees of brine saturation. The LCZ meter is capable of measuring
inductance, capacitance, dissipation, quality factor, impedance magnitude, and
phase angle in equivalent parallel and series circuits (Hewlett Packard, 1984). These
measurements cover a frequency range from 100 Hz to 20 kHz. The HP 4297A
model is controlled by a personal computer via an HP-IB card and the Agilent VEE
Pro software. The card and the software allow the digitalization of the data for the
analysis and interpretation of the measurement results. Fig.2.3 presents a sketch of
the testing setup, measurement system, and data collection system. The
measurements include impedance magnitude and phase angle from 100 Hz to 20
kHz in steps of 100 Hz. The impedance analyzer is equipped with four coaxial BNC
connectors and with an HP 16047A Test Feature. This feature combines the high
and low voltage and current terminals and permits two terminal measurements of the
electrical properties of the specimens. The two-terminal measurement method is a
simple procedure to implement, however it has some drawbacks over competitive
methods (i.e., four-electrode method – see for example Santamarina et al., 2001).
One of the biggest drawbacks is the problem of electrode polarization. This
29
phenomenon is explained next. The measurement of the electrical resistivity of rock
specimens involves the analysis of an equivalent electrical circuit where the rock
core specimen is in series with the electrodes and the required peripheral electronics.
The rock electrode interface feels the movement of ions on the rock specimen, and
the movement of electrons on the electronic circuit. At the interface, where the use
blocking electrodes is common (i.e., electrodes that do not experience
oxidationreduction), ions accumulate at the interface causing electrode polarization.
The phenomenon of electrode polarization may be modeled by adding a thin air-gap
Lgap at the interface between the electrode and the rock (Klein and Santamarina,
1997). Consider a parallel plate capacitor with a specimen of thickness LR and cross
sectional area A (Fig.2.4) The rock can be represented as a resistor and a capacitor
in parallel, with effective electric resistivity ρR (Ωm) and dielectric permittivity εR
(F/m). Then, the electrical impedance for the equivalent circuit Zeq is:
Eq.2.7
30
Fig.2.3 _ Sketch of the electrical resistivity measurement setup.
31
the effects of electrode polarization become negligible with smaller separation gap
Lgap (Fig.2.5a) and that the transition frequency is smaller with the greater rock
resistivity 𝜌𝑅 (Fig.2.5a – Santamarina et al., 2001). Therefore, electrode
polarization effects are not expected in high-frequency laboratory measurements.
However, due to specifications of the HP Impedance Analyzer, the range of
measurement extend the low frequency range, where electrode polarization
dominates the impedance measurement unless the deportation gap Lgap is very small
or the resistivity of the measured material is high. Two different types of electrodes
are used in these measurements: Al foil and silver-based paint. The Al foil may act
as a blocking electrode if low rate of chemical reaction occurs. However the problem
with the Al foil electrodes is that separation gap Lgap is about the same order of
magnitude as the sandstone grain size (D50 ≈0.1 mm) and this separation gap may
change for different specimen and the applied normal pressure. The silver paint
electrodes have been used in the past (see for example Longeron et al., 1989a). The
advantage of this type of electrodes is that they reduce the separation gap.
Furthermore, the gap remains constant throughout the testing series and it does not
depend on the normal pressure. Figure 6 shows the measured impedance on Berea
sandstone core specimen 22 using the Al foil electrodes and silver-based paint for
different saturation Sr. It is clear that the Al foil electrode measurements are
controlled by the effect of electrode polarization (Fig.2.6a) while the silver-based
paint electrode measurement removes most of the effect of electrode polarization in
the impedance response (Fig.2.6b). Furthermore, the transition frequency is smaller
when the resistivity of the specimen is greater, as observed by comparing side-by-
side the two set of results. In light of these observations the silver-based paint
electrodes are selected for the electrical resistivity measurements.
32
2.3 Description of Electrical Resistivity Measurement Methodology
The core specimens are cleaned, dried in the oven, painted on each side, evacuated,
and then saturated with 5% NaCl brine solution (ρpore liquid = 0.161 Ωm) by
applying vacuum for 12 hours to the submerged specimens. The specimens are then
removed and their surfaces are dried with tissue paper. This action renders the
specimen saturated-surface dry. Then the specimens are weighed to determine the
degree of saturation and placed in the cell to measure the impedance. After each
measurement, the specimen is placed in a covered container to stabilize its moisture
content till the next set of measurement. This technique reduces the instability of the
capillary pressure when the measurements are taken.
3 Tests Results
33
Fig.2.6 _ Effect of type of electrode on the measurement of the impedance on Berea sandstone core
specimen 22. a) Al foil electrodes; b) Silver-based paint electrodes. Electrode polarization controls the
impedance measurement with the Al foil electrodes.
Fig.2.7 _ Capillary pressure-brine saturation curves for all eight Berea sandstone core specimens
34
Table.2.2 _ Summary of porosity, tortuosity, ionic length, formation resistivity factor, cementation factor,
and water saturation exponent for all tested Berea sandstone core specimens (Attia, 2005)
Table.2.3 _ Summary of porosity, quartz flour content, formation resistivity factor, and water saturation
exponents for all tested Quartz sandstone core specimens (after Attia, 2005)
Table.2.3 Summary of porosity, tortuosity factor, formation resistivity factor, and water saturation
exponents for all tested Limestone core specimens
Ch.2 Table.4 _ Summary of porosity, tortuosity factor, formation resistivity factor, and water saturation
exponents for all tested Limestone core specimens
35
4 Discussion
The electrical resistivity and capillary pressure curves for the Berea sandstone,
Quartz, and Limestone cores show similar behavior: as the brine saturation decreases,
the values of the petrophysical parameters increase. However, the point of
irreducible brine saturation in the capillary pressure curve differs from the point
where there is a break in the electrical conductivity path. This difference is due to
the different nature of the two processes. The capillary pressure versus saturation
curve is obtained by applying gas pressure to the brine phase until the air percolates
through the porous matrix. This percolation point is the irreducible brine saturation.
It is “irreducible” only in the sense that the air pressure cannot “push” the brine out
beyond this point. Nevertheless, at the irreducible brine saturation point, the
electrolyte paths are still active and there are no sudden changes in the electrical
conduction behavior. The change in the electrical conduction only occurs at a lower
degree of saturation, when the conduction paths are broken and the resistivity rapidly
increases. The observations presented in Figures 10 through 12 can be clearly
explained with the help of Figure 13. This figure shows the process of evacuation
(drying) using the porous stone method. As the capillary pressure (air pressure)
increases, it pushes the brine out of the porous medium decreasing the saturation and
increasing the electrical resistivity of the medium (Figs.2.13a-c). This process
continues until the air paths between the top and bottom of the specimen are
continuous (Fig. 13d). At this point, the irreducible water saturation in the “capillary
pressure sense” is reached. A further increase in the capillary pressure cannot
remove anymore brine and the electrical resistivity should remain constant (Fig. 13e).
However, if the medium is allowed to dry due to evaporation the electrical resistivity
would still slowly increase because the ionic paths will be strangled as the brine
saturation decreases. This process continues until all the ionic paths are broken and
36
electrical resistivity reaches its final plateau. From Equation 2.1, the theoretical
electrical resistivity of the plateau is approximately equal to:
Eq.2.8
The irreducible brine saturation obtained with either method cannot be directly
used to evaluate the production potential of a reservoir. They can only be considered
as upper and lower limits of the irreducible brine saturation. These two measured
levels of brine saturation yield an indication of the percolation phenomena. The true
value of the irreducible water saturation will depend on the production method and
the affinity of the mineral to the fluid. One common parameter used in the evaluation
of the production potential of a formation is the saturation exponent. The saturation
exponent in the measured core specimens is not constant for different degree of
saturations as indicated by Archie (1942) even for clean Berea sandstone (Anderson,
1986). Figures 14a, 15a and 16a show that average saturation coefficient changes
from 1.60 to 2.70, 2.00 to 3.50 and 2.09 to 3.84 for Berea sandstone, Quartz, and
Limestone core specimens respectively, when the degree of saturation drops below
a certain value. The break in the slope of log(IR) versus log(Sr) occurs at the point
that coincides with the capillary pressure-based irreducible brine saturation (see Figs.
14b, 15b and 16b) and it does help in the evaluation of the electrical resistivity-based
irreducible saturation that provides the higher bound in the production potential of
the formation.
37
Fig.2.8 _ Pore size
specimens.
Fig.2.9 _ Electrical
specimens.
38
The change in slope seems to indicate the beginning of the process that produces
the rupture of electrolyte paths in the porous media. At degrees of saturation higher
than the point of sudden change in the saturation exponent, the increase of electrical
resistivity seems to occurs only due to the decrease in the cross sectional area of
continuous electrolyte paths. Then a change in the behavior seems to occur at the
point of capillary pressure-based irreducible brine saturation where the brine phase
still percolates at a minimum through the surface of the porous. At this point the
isolating non-wetting phase zone starts to become dominant over the conductive
wetting phase.
Eq.2.9
where 𝑆𝑟𝑖𝐸 is the irreducible water in the electrical resistivity sense. This new
equation assumes that the trapped porosity does not contribute to the electrical
conductivity and the tortuosity reaches infinite value when all the brine in the
channel is removed (Fig.2.17a). The normal-normal plot (Fig.2.17a) clearly shows
the value of lower value of the irreducible water saturation, while it does not present
a clear break in the log-log plot (Fig.2.17b). The tortuosity provides an indication of
the twisting of ionic path and it appears as it only changes at a slowly rate and it does
not suddenly alter the conduction paths until the last brine percolation path is broken.
39
Fig.2.10 _ Capillary pressure and resistivity index versus brine saturation in Berea sandstone core
specimens. Two irreducible water saturations are identified: one for capillary pressure and another one for
the electrical properties.
Fig.2.11 _ Capillary pressure and resistivity index versus brine saturation in twelve Quartz (very fine-
grained) sandstone core specimens. Two irreducible water saturations are identified: one for capillary
pressure and another one for the electrical properties.
Fig.2.12 _ Capillary pressure and resistivity index versus brine saturation in ten Limestone core specimens.
Two irreducible water saturations are identified: one for capillary pressure and another one for the electrical
properties.
40
References
Fig.2.13 _ Effect of percolation on the capillary pressure and electrical resistivity measurements.
41
Fig.2.14 _ Effect of brine saturation the saturation exponent in Berea sandstone
specimens. a) Log-log plot of the resistivity index versus brine saturation. b)
Saturation exponent versus porosity.
42
‘
43
Conclusion
Capillary pressure and electrical resistivity versus saturation curves show that
the irreducible brine saturation is a parameter that depends on the type of conduction
and its corresponding value of saturation at percolation. The two values of
irreducible brine saturation yield an upper and lower bound limits that can be used
to estimate limits to the production capacity of the porous medium. The detailed
analysis of the data shows that the change in the value of the saturation exponent.
This change occurs at the irreducible brine saturation as obtained with capillary
pressure data. The change in the slope seems to indicate the presence of
discontinuous ionic path in porous media, however this observation must be
confirmed with complementary testing techniques (e.g., micro-tomography). Finally,
an equation to evaluate the tortuosity versus degree of saturation is proposed.
Tortuosity results provide an indication of the distortion of the ionic path. Tortuosity
44
appears to change at slow rates and the conduction paths show no sudden changes
until the last brine percolation path is broken.
45
REFERENCES
Textbooks
Papers
46