Vous êtes sur la page 1sur 12

J185059 DOI: 10.

2118/185059-PA Date: 13-July-17 Stage: Page: 1 Total Pages: 12

Pore Compressibility of Shale Formations


Yuzheng Lan, Rouzbeh Ghanbarnezhad Moghanloo, and Davud Davudov, University of Oklahoma

Summary
This study introduces a novel outlook on a shale-pore system and on the potential effect of pore compressibility on the production per-
formance. We divide porosity of the system into accessible and inaccessible pores, and incorporate inaccessible pores with grains into
the part of the rock that is not accessible. In general, accessible pores contribute to flow directly, whereas inaccessible pores do not.
We present a mathematical model that uses mercury-injection capillary pressure (MICP) data to determine the accessible-pore and
inaccessible part of the rock (IRP) compressibility as a function of pressure. During MICP testing in a typical shale sample, the rock
sample experiences conformance, compression, and intrusion as effective pressure increases. We characterize the compressibility value
dependent on MICP data as a function of pressure. The calculated compressibility values for accessible pores generally appear to be
much greater (two to three orders of magnitude) than those of IRP.
Next, we evaluate how calculated accessible-pore-compressibility values affect gas recovery in several shale-gas plays. Our results
suggest that substitution of total pore compressibility with accessible-pore compressibility can significantly change the reservoir-behav-
ior prediction. The fundamental rock property used in many reservoir-engineering calculations including reserves estimates, reservoir
performance, and production forecasting is the total pore-volume (PV) compressibility, which has an approximate value typically
within the range of 1  106 to 1  104 psi1 (Mahomad 2014). By recognizing the part of the pore system that actually contributes to
production and identifying its compressibility, we can substitute total pore compressibility with accessible-pore compressibility. The
result changes the value by nearly two orders of magnitude.
The outcome of the paper changes the industrys take on prediction of reservoir performance, especially the rock-compaction mech-
anism. This study finds that production caused by rock compaction is in fact much greater than what has often been regarded, which
will change the performance evaluation on a great number of reservoirs in terms of economic feasibility.

Introduction
In North America, shale plays have become vital to the oil and gas industry, redefining the standard for evaluation and production of
reservoirs in the region. The complexity of shale-gas reservoirs can, in part, be attributed to the geological and petrophysical heteroge-
neity of the reservoir rocks themselves. Shales are composed of common minerals such as silica dioxide, but also include considerable
amounts of clays and organic matter; the latter is an essential constituent of a productive shale-gas reservoir (Potter et al. 2005). Shales
with 50% of grains smaller than 62.5 mm in diameter fall into the category of mudrocks (Kuila et al. 2012). These small grains com-
bined with the clay minerals generate multifarious pore geometry. Pores are observed at various locations inside the shale matrix; the
porosity in the Barnett, Kimmeridge, and Horn River Shales is dominantly within the organic matter (Curtis et al. 2011). However, Cur-
tis et al. (2012) and Chalmers et al. (2012) reported that the porosity in the Haynesville Shale is most prevalent in the inorganic part.
Detailed studies of scanning-electron-microscope images reveal the very-small-sized pores and hence very low connectivity in shale
matrix. By use of the 3D shale microstructure, Curtis et al. (2011) noted that only 19% of total porosity is connected. Ewing and Horton
(2002) conducted Monte Carlo simulations with random walks to mimic steady-state diffusion in porous media with sparsely connected
pore spaces, and observed a decrease in diffusivity with increasing sample size associated with both a decrease in effective porosity and
an increase in tortuosity. Hu et al. (2012) examined pore connectivity with three experimental approaches (imbibition, tracer-concentra-
tion profiles, and imaging), and they have also reported very-low connectivity in the shale matrix. Davudov et al. (2016) also studied
connectivity in shale formations by use of MICP data, and they have reported that the percentage of accessible pores in the Barnett and
Haynesville Shales is approximately 30%.
Civan (2011) used the leaky-tube model to elucidate the difference between accessible- and inaccessible-pore types. An accessible
pore is considered any part of the interconnected pores, which constitutes the hydraulic-flow tube. As shown in Fig. 1, the inaccessible
pores are of three different types: naturally isolated pores, induced isolated pores, and dead-end pores. Naturally isolated pores are sur-
rounded by grains and bonding material. Induced isolated pores were originally connected but become sealed by capillary forces. Dead-
end pores are those with one connecting pore throat to the hydraulic-flow tube but no transient flow.
This conceptual subdivision of pore types into accessible and inaccessible pores, as in the leaky-tube model, also has potential appli-
cations to improved understanding of PV compressibility. Pore compressibility is vital for reservoir engineers when estimating storage
capacity of rock because accuracy is a fundamental part of the feasibility of projects in shale plays (Mahomad 2014). Moghanloo et al.
(2015) also discussed a need for a study on connected pores that details a solid approach to model the compressibility because decom-
pression of rock and fluid is the main production mechanism in the primary production from shale plays. The pore types classified sug-
gest that pore compressibility should not be used without augmentation in material-balance systems, presenting a hindrance that needs
resolution in the material-balance system of shale plays. Situations involving primary-depletion-drive rock-pore and water compressi-
bility must be considered in estimations for lower-porosity formations (Hall 1953; Craft et al. 1991; Laurent et al. 1993). Without cor-
rected pore compressibility, the accuracy of well deliverability becomes problematic. Correcting for pore compressibility allows for
improved accuracy in both geomechanic aspects and hydrocarbon-reserves evaluation. In addition, geomechanical models of subsi-
dence and compaction can also be influenced by the relative (bulk) compressibility magnitude. Compaction and subsidence yield per-
meability decreases, fracture closure, and pore shrinkage, making an understanding of the dual-pore-compressibility system imperative
to evaluation.
Several studies have addressed the issue of pore compressibility from both theoretical and experimental considerations (Biot
1941; Dobrynin 1962; Geertsma 1966; Zimmerman et al. 1986; Andersen 1988; Laurent et al. 1993; Shafer and Neasham 2000;
Zimmerman 2000; Bailey 2009; Comisky et. al 2011). In our study, we develop a mathematical model, the dual-compressibility

Copyright V
C 2017 Society of Petroleum Engineers

This paper (SPE 185059) was accepted for presentation at the SPE Unconventional Resources Conference, Calgary, 1516 February 2017, and revised for publication. Original manuscript
received for review 5 August 2016. Revised manuscript received for review 24 April 2017. Paper peer approved 26 April 2017.

2017 SPE Journal 1

ID: jaganm Time: 19:25 I Path: S:/J###/Vol00000/170058/Comp/APPFile/SA-J###170058


J185059 DOI: 10.2118/185059-PA Date: 13-July-17 Stage: Page: 2 Total Pages: 12

model, to calculate accessible-pore-compressibility values and incorporate inaccessible pores into part of the rock matrix by use of
MICP data. Using the proposed method allows integrations of the dual-compressibility model into a mathematical system of equa-
tions. The model requires the use of a system that can apply hydrostatic stress equally without compromising the sample. To do this,
MICP is selected over similar gas systems. Mercury has a very-low compressibility, is nonwetting, and unlike brine/gas, does not
damage bonds chemically (Dastidar 2007). With an MICP apparatus selected, the use of a method developed by Bailey (2009) to cal-
culate the total pore compressibility was potentially viable. A power-law regression was used as outlined in the methodology. The
use of the power-law function to describe shale-compression behavior is performed after reviewing various geomechanical studies
performed on rock compressibility. We found hydrostatic-compression-test data from geomechanical research on different sand-
stones and shales, all showing a convincing power-law-function trend between bulk volume compressibility and confining pressure
(Andersen and Jones 1985; Jizba 1991; Niandou et al. 1997). However, in their comprehensive study on the compressibility of rocks,
Zimmerman et al. (1986) re-evaluated the concepts of compressibility by use of data gathered from different types of sandstone.
Mahomad (2014) compared the Zimmerman et al. (1986) data with shale data he had acquired, and found a convincing agreement
between MICP compressibility and geomechanically measured compressibility. Zimmerman et al. (1986) also proposed an equation,
Eq. 55 in their paper, to model pore compressibility as a function of pressure. At this time, the authors have not developed a model
dependent on the correlation found by Zimmerman et al. (1986), but future studies can be performed to evaluate the differences
between the use of the power-law-function correlation and the Zimmerman et al. (1986) correlation. Creating a protocol for accessi-
ble-pore and IRP compressibility then allows for systematic evaluation of the potential effect of discretization. A better understand-
ing of these computations leads to better reservoir estimation, with the potential for improved production prediction and fewer failed
wells caused by economic-feasibility inaccuracies.

Total pores in
media

Accessible Inaccessible
pores pores

Interconnected Induced Dead-end Naturally


pores isolated pores pore isolated pores

Fig. 1Illustration of the leaky-tube model and its subdivisions. This depicts the separate types of pores along with the
constituents.

In this study, we investigate the accessible-pore and IRP compressibility for several North American shale-gas plays by use of
MICP data. This paper is organized as follows: development of the mathematical model used for calculating compressibility values for
both accessible pores and IRP; evaluation of calculated pore-compressibility values for Barnett, Haynesville, Bakken, and Eagle Ford
shale plays; and application and evaluation of effect of accessible-pore compressibility on reservoir properties.

Methodology and Model Description


We treat the accessible-pore and IRP-compressibility problem as a dynamic problem in which the values of compressibility for each
part of the rock will change as a function of effective stress. The findings from this study are taken back into the macroscopic material-
balance equation to demonstrate their practical aspects (Moghanloo et al. 2015).
One of the major contributions of this model is that it separates out the accessible-pore compressibility and provides some support-
ive insights into the effectiveness of the compaction-production mechanism. We divide the bulk system into two parts: accessible pores
that contribute to production directly, and IRP that has no direct contribution to production. By substituting total-pore-compressibility
values with the newfound accessible-pore-compressibility values, a better understanding of reservoir potential can be reached.
We develop a Visual Basic for Applications (VBA) code to compute the compressibility function for both accessible-pore and IRP
by use of core data from four shale plays in the US. Three critical pressure points during the MICP process are considered in the build-
ing of this model, which will be explained later. This work is a continuation of our previous work (Davudov et al. 2017).

Phenomenological Model. Data from the MICP experiment were collected in the Integrated Core Characterization Center (IC3) at the
University of Oklahoma. Blank corrections were run before any data collection to eliminate the effect of mercury compressibility and
temperature during the process. Before MICP, core samples were prepared through multiple stages, including polishing, drying,
and vacuuming, which evacuates the fluids from the pore space of the rock, hence enabling the assumption that pore pressure is zero
and that confining pressure is the equivalent of effective pressure on the pores before any intrusion happens.
The dual-compressibility model is inspired by Bailey (2009), which delineates MICP data into three stages: conformance, compres-
sion, and intrusion. Fig. 2 presents a schematic of these three different stages during an MICP experiment. The conformance is simply
the amount of mercury needed to envelop the external shape of the sample before intrusion happens so when pressure reaches conform-
ance pressure (Pconf ), the volume of mercury recorded is because of the core-sample conformance and has no relation to the pore space
in that core.
In between conformance pressure and critical intrusion pressure (Pci ), the pressure is not sufficient for mercury to intrude into pores
because the characteristic pore-throat size is usually smaller than 20 nm, which can be translated into a critical intrusion pressure by use of
the Washburn (1921) equation (Tran and Sakhaee-Pour 2017). However, both accessible pores and IRP are compressed because of exter-
nal pressure. Mercury volume recorded at this stage is the sum of the volume change caused by compression in both parts of the rock.
Mercury intrusion to accessible pores starts to happen after pressure reaches Pci , because intrusion is the point at which the capillary
pressure overtakes the interfacial tension and intermolecular forces; it exceeds the critical pressure (Bailey 2009; Mahomad 2014). At
this point, mercury begins filling the rock PV and when pressure reaches final pressure (Pf ), all accessible pores are intruded. Mercury
volume measured at this pressure is the sum total of accessible PV and the volume change caused by the compression in IRP.

2 2017 SPE Journal

ID: jaganm Time: 19:25 I Path: S:/J###/Vol00000/170058/Comp/APPFile/SA-J###170058


J185059 DOI: 10.2118/185059-PA Date: 13-July-17 Stage: Page: 3 Total Pages: 12

(a) P = Pconf (b) P = Pci (c) P = Pf

Fig. 2Three schematics of the MICP process at Pconf , Pci , and Pf .

By breaking these steps into pieces, Bailey (2009) suggested that a conformance correction should be applied. Once applied, the
incremental mercury injection can be used to observe a volume change of the pores in the compression/shrinkage stage. With the incre-
mental pore compressibility established for the defined pressures, they are plotted on a log-log axial system. The result exhibits a lin-
ear-appearing trend for the shrinkage period. Any deviation from the linear line on log-log scale is because of conformance on the low-
pressure portion, and any deviation at the high-pressure region is because of intrusion. The resulting slope of the linear trend is defined
as the variable in a power-law function for calculating pore compressibility at any point.
We have made many improvements to better understand the implications for pore compressibility from MICP. In the original model,
it was assumed that volume change during MICP is the result of solely pore compression and mercury intrusion, and PV compressibility
is hence calculated by use of the total change in bulk volume as the total change in PV. We, however, looked at the bulk volume of the
rock as a whole, and plotted bulk volume compressibility against confining pressure for a sample from Barnett Shale in Fig. 3. The
authors found multiple geomechanical hydrostatic compression studies on sedimentary rocks, and extracted and plotted the published
data in Fig. 4 to fit a power-law function (Andersen and Jones 1985; Jizba 1991; Niandou et al. 1997). The region where the trend
agrees with the findings from geomechanical tests (the power-law relationship between bulk volume compressibility and confining pres-
sure) is deemed as the compression region, before and after which are considered to be the conformance and intrusion regions, respec-
tively. There exists room for further improvement to better characterize the creep behavior in the stress/strain relationship, which will
induce a secondary linear-behavior region on a log-log plot between bulk compressibility and confining pressure. Eq. 55 in Zimmerman
et al. (1986) could potentially solve part of this problem, but we expect the correlation will need further adjustment if confining pressure
is raised higher to a plastic region where the stress/strain relationship can change again.

110 3 0.035
Conformance Compression Intrusion

110 4 0.028

Cumulative Mercury (cm3/g)

110 5 0.021
Cpc (psi1)

110 6 0.014

110 7 0.007

110 8 0
6 60 600 6,000 60,000

Confining Pressure (psi)

Fig. 3MICP curve divided into three stages. Total pore compressibility calculated as a function of pressure (solid red line). Generally,
Pconf is near atmospheric pressure and Pci is in the range of 4,00010,000 psi. Pf in most cases is 60,000 psi. These data are courtesy of IC3.

Mathematical Derivation. The dual-compressibility model describes the behaviors of accessible-pore and IRP compressibility with
respect to effective stress. An analytical solution is developed to calculate compressibility values for both accessible pores and IRP, sep-
arately. Our assumptions in the making of this model include:
Pore system consists of accessible and inaccessible pores.
Pore pressure is zero before intrusion happens.
There is no mercury intrusion into pores before the critical intrusion pressure.
Compression is elastic or nearly elastic during the MICP process.

2017 SPE Journal 3

ID: jaganm Time: 19:25 I Path: S:/J###/Vol00000/170058/Comp/APPFile/SA-J###170058


J185059 DOI: 10.2118/185059-PA Date: 13-July-17 Stage: Page: 4 Total Pages: 12

1.4010 6 1.4010 6

1.2010 6
Cbc (psi1) 1.1010 6

Cbc (psi1)
1.0010 6
8.0010 7 y= 6E06x 0.279 y = 2E05x 0.442
R = 0.9904 8.0010 7 R = 0.9979

5.0010 7
6.0010 7

2.0010 7 4.0010 7
0 5,000 10,000 15,000 20,000 0 2,000 4,000 6,000 8,000 10,000

Confining Pressure (psi) Confining Pressure (psi)

Fig. 4Mechanical hydrostatic-compression-test results from different studies. Bulk compressibility related to confining pressure
is plotted against confining pressure; on the left, information from Andersen and Jones (1985), and on the right, information from
Jizba (1991).

Consider the total bulk volume as a summation of accessible pores and IRP. Therefore, total bulk volume can be written as

Vb Va VIRP ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1

where Vb represents total bulk volume, Va is the volume of accessible pores, and VIRP is the volume of IRP. Taking the derivative with
respect to confining pressure on both sides yields:
dVb dVa dVIRP
: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-2
dPc dPc dPc
Eq. A-2 shows that the change in bulk volume is simply a linear summation of the change in accessible PV and the volume of IRP.
This linearity is later found in the relations between bulk compressibility with respect to confining pressure, accessible-pore compressi-
bility, and compressibility of the IRP. The detailed derivation of this model, with the determination of accessible PV at a desired pres-
sure, is shown in Appendix A. The physical condition and balance between mercury and the pore space is the most important
foundation of this analytical model. To investigate deeper into the pore system as we defined, we must find special conditions where the
behaviors of accessible-pore and IRP compression can be identified. Our model is built with emphasis on conditions at Pconf , Pci , and
Pf . Ultimately, the system comes down to three equations with three variables:
8
> k1 ak2 1  a/k3 ;                                                            A-7
>
>
>
>
>
> a/Vbsc 1  a/Vbsc
>V
>
< Hgci Vbsc   k2 m1 m1    k3 m1 m1  ;                                        A-13
Pci  Pconf Pci  Pconf
> e m1 e m1
>
>
>
> 1  a/Vbsc
>
> VHgf Vbsc   k ;
>
>                                                     A-15
: 3
Pm1
f  Pm1
conf
e m 1

where k1 and m are the coefficient and power for the power-law function of bulk compressibility with respect to pressure; VHgci and VHgf
are volume of the mercury measured at Pci and Pf , respectively, after conformance correction; and Vbsc represents the bulk volume at the
standard condition, which is measured during MICP. The variables a, k2, and k3 are solved by use of a bracketed trial-and-error numerical
code, as explained in Appendix A. The variable a represents the accessible PV fraction; k2 is the coefficient in the power-law function
between accessible-pore compressibility and pressure; and k3 is the coefficient between IRP compressibility and pressure. Accessible-
pore and IRP compressibility can be calculated and plotted against pressure. A flow chart of the process of this model is shown in Fig. 5.

Results and Discussion


Result Summary. By use of the dual-compressibility model, accessible-pore and IRP compressibility as well as bulk compressibility
are obtained as a function of pressure for the Barnett, Bakken, Eagle Ford, and Haynesville Shales. Pore pressure with respect to confin-
ing pressure is then calculated by use of an equation derived by Zimmerman et al. (1986):
Cbc  Cr
Cpc ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
/
where Cr represents the intrinsic grain compressibility with constant effective pressure. It should be noted that there is a difference
between the definition of Cr and the other types of compressibility being discussed in this paper. Whether they are defined with a con-
stant confining pressure or a constant pore pressure, they are different from this compressibility definition, which leads to the approxi-
mation of Cr with 1  a/CIRPC , where CIRPC is IRP compressibility with respect to confining pressure. This approximation stemmed
from a simple comparison between Eq. A-5 in this paper and Eq. 13 in Zimmerman et al. (1986). It has been observed that for all sam-
ples accessible-pore- and IRP-compressibility values are different, and in all cases the former has higher value than the latter. It is worth
noting that as accessible-PV fraction increases, the similarity between accessible-pore compressibility and total pore compressibility
increases. On the other hand, each field has its own signature behavior in accessible-pore and IRP compressibility and volume distribu-
tion of each type of pore. Results are shown in Figs. 6a through 6d, and summarized in Table 1. We are not able to come to rigorous
conclusions on the behaviors of each shale play because of the intrinsic heterogeneity of shale. However, the results exemplify the

4 2017 SPE Journal

ID: jaganm Time: 19:25 I Path: S:/J###/Vol00000/170058/Comp/APPFile/SA-J###170058


J185059 DOI: 10.2118/185059-PA Date: 13-July-17 Stage: Page: 5 Total Pages: 12

robustness of the dual-compressibility model and its ability to predict compressibility behavior by use of MICP data and some of the
other practical applications described later on.

Dual-Compressibility Model

1. Determine bulk compressibility 3. Calculate and plot accessible pore


2. Find volume fraction of accessible
related to confining pressure as a and inaccessible pore compressibility
pores and IRP
function of pressure as a function of pressure

2.1 Identify pressure dependency of 3.1 Plot both accessible pores and
1.1 Bailey (2009) method both accessible pores and IRP, IRP compressibility against pressure
Eq. A-7

2.2 Calculate accessible pore 3.2 Determine coefficient for


Cb = k1 P m and IRP volume at critical intrusion power-law function between
pressure and final pressure using accessible pore compressibility and
Eq. A-11 pressure, IRP compressibility
and pressure

2.3 Related mercury readings to bulk


volume, accessible pores and IRP
volume at three different pressures,
Eqs. A-12 and A-14

2.4 Determined volume fraction of


accessible pores and IRP by solving
the system of Eqs. A-7, A-12,
and A-14

Fig. 5Flow chart of the dual-compressibility model.

Accessible-pore-compressibility values calculated directly from the dual-compressibility model are compressibility values related to
confining pressure. They can be converted to compressibility values related to pore pressure by use of Eq. 2, derived from an equation
(Eq. 14) published by Zimmerman et al. (1986):
Cbc  1 a/1  a/CIRPC 
Cap ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
a/
where Cap is accessible-pore compressibility with respect to pore pressure. All discussion of accessible-pore-compressibility application
in this paper is presented after this conversion.

Discussion and Application. Pore compressibility has been neglected in most calculations because of a connotation of low signifi-
cance. On the basis of our findings, accessible-pore-compressibility values should be considered in calculations instead of total pore
compressibility as commonly recognized. In this subsection, we study how calculated accessible-pore-compressibility values will affect
some reservoir properties and how it will affect gas recovery.
PV Change. Compressibility values for accessible pores and IRP calculated in this study are higher than the industry-recognized
standards. As shown in Fig. 7, pore shrinkage occurred because accessible-pore compressibility is much higher than the pore shrinkage
that occurred because of total pore compressibility. The percentage of pore shrinkage may vary for different samples and for different
shale plays, as we found in this study. Pore shrinkage, although it may be good in assisting production through reservoir compaction,
has a tendency to cause pore closure in some shale formations. Further studies need to be conducted to understand side effects of com-
pressibility on pore systems.
Capillary Pressure Curve. Capillary pressure vs. the saturation curve should also be corrected as a result of significant pore shrink-
age and lack of intrusion before the critical intrusion pressure. Liquid saturation remains zero before the pressure reaches the conform-
ance pressure, and starts to increase as pressure surpasses said point. PV decreases as a function of effective stress because of the
compressibility of accessible pore and IRP. Fig. 8 shows results for the calculated capillary pressure curve with and without any correc-
tion. Results indicate that there is a misinterpretation in pore-size distribution from the capillary pressure curve in the industry; the
curve needs to be corrected significantly.
Effect on Gas Recovery. Furthermore, we studied how calculated accessible-pore-compressibility values affect production recovery
in several shale-gas plays. Our results suggest that substitution of accessible-pore compressibility for total compressibility can signifi-
cantly change the reservoir-behavior prediction. Reservoir compaction is generally regarded as total pore compressibility, which has
values within the magnitude range of 1  106 to 1  104 psi1 (Mahomad 2014). By recognizing the part of the pore system that
actually contributes to production and identifying its compressibility, we can substitute total pore compressibility with accessible-pore
compressibility. The result changes the value by nearly two orders of magnitude. In this study, we used macroscopic material balance
to evaluate the effect of pore compaction on gas recovery. In gas reservoirs, the main production mechanisms are fluid expansion, water

2017 SPE Journal 5

ID: jaganm Time: 19:25 I Path: S:/J###/Vol00000/170058/Comp/APPFile/SA-J###170058


J185059 DOI: 10.2118/185059-PA Date: 13-July-17 Stage: Page: 6 Total Pages: 12

expansion, rock expansion, and gas desorption in shale formations. By use of the modified macroscopic material-balance equation (Eq.
3) derived by Yuan et al. (2016), we can analyze the effect of compressibility on gas recovery.
     
Gp Bg  Bg;initial 1 Bg;initial Bg;initial bPinitial bP
Cp;acc Swi cw Pi  P qrock max  : . . . 3
Gfg i Bg 1  Swi Bg 1  Swi 1 bPinitial 1 bP

IRP Accessible pore Cbc Cpc IRP Accessible pore Cbc Cpc
1104 1104
Compressibility (psi1)

Compressibility (psi1)
1105 11055

1106 11066

1107 11077

1108 1108
1,000 3,000 5,000 1,000 3,000 5,000
Confining Pressure (psi) Confining Pressure (psi)
(a)

IRP Accessible pore Cbc Cpc IRP Accessible pore Cbc Cpc
1104 1104
Compressibility (psi1)

Compressibility (psi1) 11055


1105
11066

1107
1106
11088

1107 1109
1,000 3,000 5,000 1,000 2,000 3,000 4,000 5,000 6,000
Confining Pressure (psi) Confining Pressure (psi)
(b)
IRP Accessible pore Cbc Cpc IRP Accessible pore Cbc Cpc
1104 1104
Compressibility (psi1)

Compressibility (psi1)

1105
11055

1106
11066
1107

1107
1108
1,000 3,000 5,000 1,000 2,000 3,000 4,000 5,000 6,000
Confining Pressure (psi) (c) Confining Pressure (psi)

IRP Accessible pore Cbc Cpc IRP Accessible pore Cbc Cpc
1104 1104
Compressibility (psi1)

Compressibility (psi1)

11055
1105
11066

11077
1106
11088

1109 1107
1,000 2,000 3,000 4,000 5,000 6,000 1,000 2,000 3,000 4,000 5,000 6,000
Confining Pressure (psi) Confining Pressure (psi)
(d)

Fig. 6(a) Results from Barnett Shale. (b) Results from Haynesville Shale. (c) Results from Bakken Shale. (d) Results from Eagle
Ford Shale.

6 2017 SPE Journal

ID: jaganm Time: 19:25 I Path: S:/J###/Vol00000/170058/Comp/APPFile/SA-J###170058


J185059 DOI: 10.2118/185059-PA Date: 13-July-17 Stage: Page: 7 Total Pages: 12

Barnett Haynesville Bakken Eagle Ford


Sample 1 Sample 2 Sample 1 Sample 2 Sample 1 Sample 2 Sample 1 Sample 2
Accessible-PV fraction 0.561 0.886 0.265 0.570 0.287 0.610 0.741 0.592
5 5 5 5 5 5 5 5
Accessible-pore- 2.6110 2.9210 3.1510 1.7310 4.0010 3.4010 2.2610 2.0210
compressibility range to to to to to to to to
1 6 6 6 6 6 5 6 6
(psi ) 7.7210 8.9510 9.2210 3.4610 9.1010 1.0310 5.1410 6.1610
7 8 7 9 8 7 8 7
1.6710 9.7710 3.4810 7.0210 6.610 2.9210 2.6610 9.4310
IRP-compressibility to
1 to to to to to to to
range (psi ) 8 8 7 9 8 8 9 7
4.9610 3.0010 1.0210 1.4110 1.510 8.8810 6.0610 2.8810
Intrusion pressure (psi) 13,346 5,950 17,790 6,304 9,987 16,785 4,000 5,950
Coefficient for power-
0.00297 0.00292 0.0043 0.00917 0.0135 0.00398 0.0073 0.00208
law function of Cacc, k2
Exponent for power-law
0.6846 0.6617 0.6887 0.9008 0.8296 0.6675 0.8297 0.6656
function, m

Table 1Result summary for four shale plays.

Total pore compressibility Accessible pore compressibility

0.95

0.9
Vp /Vp 0

0.85

0.8

0.75
1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500 5,000 5,500 6,000
Pressure (psi)

Fig. 7Vp/Vp0 for one selected sample.

Without correction Dual-compressibility-model, corrected

100,000

10,000
Capillary Pressure (psi)

1,000

100

10

1
0.50 0.40 0.30 0.20 0.10 0.00

Liquid Saturation

Fig. 8Capillary pressure curve for one selected sample before and after correction.

Eq. 3 is used to calculate depletion efficiency for a shale-gas reservoir, which describes the potential recovery within the drainage
area. The right-hand side of the equation gives the individual terms that mathematically define different production mechanisms during
recovery: the first term expresses free-gas expansion, the second term represents rock and water expansion, and the third term denotes
gas desorption. To study and analyze each driving index, we made a synthetic reservoir model, the parameters of which are shown in
Table 2.

2017 SPE Journal 7

ID: jaganm Time: 19:25 I Path: S:/J###/Vol00000/170058/Comp/APPFile/SA-J###170058


J185059 DOI: 10.2118/185059-PA Date: 13-July-17 Stage: Page: 8 Total Pages: 12

Variable Value
Pi (psi) 6,204
P (psi) 1,450
Swi (%) 20
Cw (psi )
1
0
rock (g/cm )
3
2.68
max 6.38
b 0.1
Bgi (res cf/scf) 0.00343824
Bg (res cf/scf) 0.01201046

Table 2Values used in macroscopic material-balance equation.

The contribution from each driving index is calculated for different compressibility values. We studied two cases: when total pore
compressibility is used in calculations and when only the accessible-pore compressibility with respect to pore-pressure value is consid-
ered. With 3  106 as the total pore compressibility in the rock-expansion term, the contribution from rock expansion is less than 1%,
whereas the gas-desorption term and gas-expansion term each have 4.2 and 95.1% contribution, respectively. By increasing the com-
pressibility value from 3  106 to 5  105 psi1, the rock-expansion contribution increased dramatically to 10.2%, whereas the gas-
desorption and gas-expansion indices dropped to 3.8 and 85.9%, respectively. As a result, depletion efficiency increased by 8.0% within
the drainage area.
In addition, because the compression that occurred during the MICP experiment is under hydrostatic conditions, a correction must
be made to convert the hydrostatic compressibility value to uniaxial compressibility to mimic reservoir condition, by use of the
Zimmerman (2000) equation:
Caccuni 1 v
; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Cap 31  v

where Caccuni is the compressibility of accessible pore with respect to pore pressure in uniaxial condition, Cap represents the same vari-
able but measured in hydrostatic condition, and v is Poissons ratio, taken to be 0.23 as a typical value for shale (Jizba 1991). Note that
Eq. 4 is applied in this study under an additional assumption that the IRP is incompressible, hence a pseudo-Biot constant is assumed to
be unity. One potential problem resides in this conversion because of the difference in stress tensor of hydrostatic and uniaxial condi-
tions; for example, the increasing shear stress on the uniaxial-strain path changes the behavior of an isotropic-strain path. Future analy-
sis can be performed in regards to the conversion from hydrostatic conditions to uniaxial conditions, but it is not the main focus of
this paper.
The results plotted in Fig. 9both depletion efficiency and rock-expansion index, as explained by Moghanloo et al. (2015)show
a significant increase after we apply the calculated accessible-pore-compressibility value into the macroscopic material-balance equa-
tion. However, there is a slight decrease after the correction from hydrostatic compressibility to uniaxial compressibility is made.

Cpp Cap Caccuni Cpp Cap Caccuni

100.00% 0.50

Rock-Expansion Index
Depletion Efficiency

80.00% 0.40

60.00% 0.30

40.00% 0.20

20.00% 0.10

0.00% 0.00
6,000 5,000 4,000 3,000 2,000 1,000 6,000 5,000 4,000 3,000 2,000 1,000

Reservoir Pressure (psi) Reservoir Pressure (psi)

Fig. 9Depletion efficiency and rock-expansion index before and after correction.

Conclusions
This study presents novel concepts of accessible-pore and IRP compressibility, and a mathematical model to calculate both pore-com-
pressibility values separately for shale formations. Results have shown that for all four shale plays, accessible-pore compressibility is
higher than the total pore compressibility; therefore, in reservoir calculations, this should be considered for accurate predictions. Some
specific conclusions in this paper are listed here:
1. We introduce a new outlook on the pore system in terms of production contribution and reservoir performance.
2. We correct for a significant overestimation in connected PV measured by MICP.
3. We present a model to obtain compressibility values of accessible pores and IRP as a function of pressure.
4. The resultant change in pore-compressibility value used in the macroscopic material-balance equation can be up to two orders of
magnitude. This finding proves that rock compaction has a more significant effect on production than traditionally perceived.

8 2017 SPE Journal

ID: jaganm Time: 19:25 I Path: S:/J###/Vol00000/170058/Comp/APPFile/SA-J###170058


J185059 DOI: 10.2118/185059-PA Date: 13-July-17 Stage: Page: 9 Total Pages: 12

Nomenclature
a accessible PV fraction, decimal
b Langmuir adsorption coefficient, psia1
Bg gas formation volume factor (FVF), res cf/scf
Bgi initial gas formation volume factor, res cf/scf, also as Bg; initial
Cac compressibility of accessible pore under hydrostatic conditions with respect to confining pressure, psia1, also as Cp; acc
Caccuni compressibility of accessible pore measured under uniaxial conditions with respect to confining pressure, psia1
Cap compressibility of accessible pore with respect to pore pressure, psia1
Cbc bulk compressibility with respect to confining pressure, psia1
CIRPC compressibility of IRP with respect to confining pressure, psia1
Cpc total pore compressibility with respect to confining pressure, psia1
Cr grain compressibility with constant effective pressure, psia1
Cw compressibility of water, psia1
Gfgi original gas in drainage area in free-gas phase, scf
Gp gas production, scf
k1 coefficient for power-law function between total pore compressibility and pressure
k2 coefficient for power-law function between accessible-pore compressibility and pressure
k3 coefficient for power-law function between IRP compressibility and pressure
m power in the power-law function between compressibility and pressure
Pc confining pressure, psi
Pci critical intrusion pressure, psi
Pconf conformance pressure, psi
Pf final pressure, psi
Pi initial pressure, psi, also Pinitial
Swi initial water saturation, decimal
Va volume of accessible pores, cm3
VaPci volume of accessible pores at critical intrusion pressure, cm3
Vb bulk volume, cm3
Vbsc bulk volume at standard condition, cm3
VHgci volume of mercury injected at critical intrusion pressure, cm3
VHgf volume of mercury injected at final pressure, cm3
VIRP volume of IRP, cm3
VIRPPci volume of IRP at critical intrusion pressure, cm3
VIRPPf volume of IRP at final pressure, cm3
Vp total PV, cm3
Vpsc total PV at standard condition, cm3
Vp0 initial total PV, cm3
v Poissons ratio, decimal
vmax maximum monolayer-volumetric capacity per unit weight of solid
qrock density of the rock, g/cm3

Acknowledgments
The authors acknowledge the support provided by IC3 and acknowledge the Unconventional Shale Gas Consortium at the University of
Oklahoma for generously sharing the laboratory data. Also, the support from Chandra Rai and the Eberly Family Chair are appreciated.
The authors appreciate the support provided through the American Chemical Society (PRF No. 56929-DN19). Our special thanks are
extended to Dalton Vice, who contributed tremendously to this work in the early stage, and we also thank Garth McLoed for editing
services.

References
Andersen, M. A. 1988. Predicting Reservoir-Condition PV Compressibility from Hydrostatics Stress Laboratory Data. SPE Res Eval & Eng 3 (3):
10781082. SPE-14213-PA. https://doi.org/10.2118/14213-PA.
Andersen, M. A. and Jones, F. O. Jr. 1985. A Comparison of Hydrostatic-Stress and Uniaxial-Strain Pore-Volume Compressibilities Using Nonlinear
Elastic Theory. Presented at the 26th US Symposium on Rock Mechanics, Rapid City, South Dakota, 2628 June. ARMA-85-0403-1.
Bailey, S. 2009. Closure and Compressibility Corrections to Capillary Pressure Data in Shales. Oral presentation given at the Denver Well Logging Soci-
ety Workshop, Golden, Colorado, 19 October.
Biot, M. A. 1941. General Theory of Three-Dimensional Consolidation. J. Appl. Phys. 12: 155164. https://doi.org/10.1063/1.1712886.
Chalmers, R. C., Bustin, R. M., and Power, I. M. 2012. Characterization of Shale Gas Pore Systems by Porosimetry, Pycnometry, Surface Area, and Field
Emission Scanning Electron Microscopy/Transmission Electron Microscopy Image Analyses: Examples from the Barnett, Woodford, Haynesville,
Marcellus, and Doig Units. AAPG Bull. 96 (6): 10991119. https://doi.org/10.1306/10171111052.
Civan, F. 2011. Transport Properties of Porous Media. Hoboken, New Jersey: John Wiley & Sons.
Comisky, J. T., Santiago, M., McCollom, B. et al. 2011. Sample Size Effects on the Application of Mercury Injection Capillary Pressure for Determining
the Storage Capacity of Tight Gas and Oil Shales. Presented at the Canadian Unconventional Resources Conference, Calgary, 1517 November.
SPE-149432-MS. https://doi.org/10.2118/149432-MS.
Craft, B. C., Hawkins, M. F., and Terry, R. E. 1991. Applied Petroleum Reservoir Engineering. Englewood Cliffs, New Jersey: Prentice Hall.
Curtis, M. E., Ambrose, R. J., Sondergeld, C. H. et al. 2011. Investigating the Microstructure of Gas Shales by FIB/SEM Tomography & STEM Imaging.
Curtis, M. E., Cardott, B. J., Sondergeld, C. H. et al. 2012. Development of Organic Porosity in the Woodford Shale with Increasing Thermal Maturity.
Int. J. Coal Geol. 103 (1 December): 2631. https://doi.org/10.1016/j.coal.2012.08.004.
Dastidar, R. 2007. Nuclear Magnetic Resonance (NMR) Study of Freezing and Thawing of Saturated Porous Media and Application To Shale and Pore
Volume Compressibility Estimation. PhD dissertation, University of Oklahoma, Norman, Oklahoma.

2017 SPE Journal 9

ID: jaganm Time: 19:25 I Path: S:/J###/Vol00000/170058/Comp/APPFile/SA-J###170058


J185059 DOI: 10.2118/185059-PA Date: 13-July-17 Stage: Page: 10 Total Pages: 12

Davudov, D., Moghanloo, R. G., and Yuan, B. 2016. Impact of Pore Connectivity and Topology on Gas Productivity in Barnett and Haynesville Shale
Plays. Presented at the Unconventional Resources Technology Conference, San Antonio, Texas, 13 August. URTEC-2461331-MS. https://doi.org/
10.15530/URTEC-2016-2461331.
Davudov, D., Moghanloo, R. G., Lan, Y. et al. 2017. Investigation of Shale Pore Compressibility Impact on Production with Reservoir Simulation. Pre-
sented at the SPE Unconventional Resources Conference, Calgary, 1516 February. SPE-185059-MS. https://doi.org/10.2118/185059-MS.
Dobrynin, V. M. 1962. Effect of Overburden Pressure on Some Properties of Sandstones. SPE J. 2 (4): 360366. SPE-461-PA. https://doi.org/10.2118/461-PA.
Ewing, R. P. and Horton, R. 2002. Diffusion in Sparsely Connected Pore Spaces: Temporal and Spatial Scaling. Water Resour. Res. 38 (12): 21-121-13.
https://doi.org/10.1029/2002WR001412.
Geertsma, J. 1966. Problems of Rock Mechanics in Petroleum Production Engineering. Presented at the 1st ISRM Congress, Lisbon, Portugal, 25
September1 October. ISRM-1CONGRESS-1966-099.
Hall, H. N. 1953. Compressibility of Reservoir Rocks. J Pet Technol 5 (1): 1719. SPE-953309-G. https://doi.org/10.2118/953309-G.
Hu, Q., Ewing, R. P., and Dultz, S. 2012. Low Pore Connectivity in Natural Rock. J. Contam. Hydrol. 133 (15 May): 7683. https://doi.org/10.1016/
j.jconhyd.2012.03.006.
Jizba, L. 1991. Mechanical and Acoustical Properties of Sandstones and Shales. PhD dissertation, Stanford University, Stanford, California.
Kuila, U., Prasad, M., Derkowski, A. et al. 2012. Compositional Controls on Mudrock Pore-Size Distribution: An Example from Niobrara Formation.
Presented at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas, 810 October. SPE-160141-MS. https://doi.org/10.2118/
160141-MS.
Laurent, J., Bouteca, M. J., Sarda, J.-P. et al. 1993. Pore-Pressure Influence in the Poroelastic Behavior of Rocks: Experimental Studies and Results. SPE
Form Eval 8 (2): 117122. SPE-20922-PA. https://doi.org/10.2118/20922-PA.
Mahomad, B. 2014. Evaluation of Shale Compressibility From NMR and MICP Measurements. Masters thesis, University of Oklahoma, Norman,
Oklahoma.
Moghanloo, R. G., Yuan, B., Ingrahama, N. et al. 2015. Applying Macroscopic Material Balance to Evaluate Interplay Between Dynamic Drainage Vol-
ume and Well Performance in Tight Formations. J. Nat. Gas Sci. Eng. 27 (2): 466478. https://doi.org/10.1016/j.jngse.2015.07.047.
Niandou, H., Shoa, J. F., Henry, J. P. et al. 1997. Laboratory Investigation of the Mechanical Behaviour of Tournemire Shale. Int. J. Rock Mech. Min. 34
(1): 316. https://doi.org/10.1016/S1365-1609(97)80029-9.
Potter, P. E., Maynard, J. B., and Depetris, P. J. 2005. Mud and Mudstones: Introduction and Overview. Berlin: Springer-Verlag.
Shafer, J., and Neasham, J. 2000. Mercury Porosimetery Protocol for Rapid Determination of Petrophysical and Reservoir Quality Properties. Oral pre-
sentation of paper SCA2000-21 given at the 2000 International Symposium of Society of Core Analysts, Abu Dhabi, 1822 October.
Tran, H. and Sakhaee-Pour, A. 2017. Viscosity of Shale Gas. Fuel 191 (1 March): 8796. https://doi.org/10.1016/j.fuel.2016.11.062.
Washburn, E. W. 1921. Note on a Method of Determining the Distribution of Pore Sizes in a Porous Material. Proc. Natl. Acad. Sci. USA 7 (4):
115116. https://doi.org/10.1073/pnas.7.4.115.
Yuan, B., Moghanloo, R. G., and Shariff, E. 2016. Integrated Investigation of Dynamic Drainage Volume (DDV) and Inflow Performance Relationship
(Transient IPR) to Optimize Multi-Stage Fractured Horizontal Wells in Shale Oil. J. Energy Resour. Technol. 138 (5): 052901052901-9. https://
doi.org/10.1115/1.4032237.
Zimmerman, R. 2000. Implications of Static Poroelasticity for Reservoir Compaction. Presented at the 4th North American Rock Mechanics Symposium,
Seattle, Washington, 31 July3 August. ARMA-2000-0169.
Zimmerman, R., Somerton, W., King, M. 1986. Compressibility of Porous Rocks. J. Geophys. Res.-Sol. Ea. 91 (12): 1276512777. https://doi.org/
10.1029/JB091iB12p12765.

Appendix AMathematical Derivations for the Dual-Compressibility Model


In this study, the bulk volume is divided into accessible pores and IRP. Hence, the bulk volume is the sum of accessible PV and volume
of the IRP:
Vb Va VIRP ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1

where Vb represents bulk volume, Va is the volume of accessible pores, and VIRP is the volume of IRP. Taking the derivative with
respect to confining pressure on both sides of Eq. A-1 yields
dVb dVa dVIRP
; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-2
dPc dPc dPc
where Pc represents confining pressure. Eq. A-2 can be rewritten as
1 dVb Va Vp 1 dVa VIRP Vp 1 dVIRP
; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-3
Vb dPc Vp Vb Va dPc Vp Vb VIRP dPc

hence,
Va VIRP
Cbc /Cac /CIRPC ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-4
Vp Vp
where Cbc ; Cac , and CIRPC
represent compressibility of the bulk, accessible pores, and IRP, respectively, all with respect to confining pressure, and / is the porosity of
the rock measured from low-pressure porosimetry testing, which is considered as absolute porosity of the rock at standard condition. Setting
Va VIRP 1  a/
the accessible PV fraction as variable a, should then be equal to . Eq. A-4 then becomes
Vp Vp /
Cbc a/Cac 1  a/CIRPC : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-5

Eq. A-5 shows a linear relationship between bulk compressibility, accessible-pore compressibility, and IRP compressibility. We
found among geomechanical research that the bulk compressibility of shale can be represented with a power-law function. Because of
the linearity of the relationship, the equation can be rewritten as a function of pressure:

10 2017 SPE Journal

ID: jaganm Time: 19:25 I Path: S:/J###/Vol00000/170058/Comp/APPFile/SA-J###170058


J185059 DOI: 10.2118/185059-PA Date: 13-July-17 Stage: Page: 11 Total Pages: 12

k1 Pm ak2 Pm 1  a/k3 Pm ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-6


where k1 , k2 , and k3 are the coefficients in the power-law function with respect to confining pressure for bulk, accessible pores, and
IRP, respectively. The value m is the exponent of the function, which remains the same for all three compressibility values because of
the linearity of the relationship.
Simplifying Eq. A-6 yields

k1 ak2 1  a/k3 : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-7

The pressure region between Pconf and Pci is considered as the compression region, where all the accessible pores and IRP are com-
pressed. After Pci , a portion of the accessible pores start to get intruded by mercury, depending on pore-throat sizes, which control the
capillary pressure or entry pressure for the pores connected to them. Pressure will equalize for pores that have been intruded, causing
them to rebound to the original volume, but pore pressure still remains at zero for pores that have not been intruded. Thus, IRP and the
not-yet-intruded accessible pores are still compressed because of effective pressure. MICP reading at the Pci consists of the mercury
conforming to the surface and the volume change in both accessible pores and IRP because of the compression under Pci . We review
the compressibility definition:
1 dVa
Cac  : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-8
Va dPc
We rewrite compressibility as a function of pressure:
1 dVa
k2 P m  : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-9
Va dPc
We solve this ordinary-differential equation:
k2
Pm1  Pm1 lnV1  lnV2 : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-10
m1 2 1

Therefore, volume of accessible pores at Pci can be written as


a/Vbsc
VaPci 
k2
 ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-11
Pm1  Pm1
conf
e m 1 ci
where VaPci represents volume of accessible pores at Pci , and Vbsc represents the bulk volume under standard conditions, which is deter-
mined from MICP readings at standard conditions. Similarly, the volume of IRP at Pci can be written as
1  a/Vbsc
VIRPPci 
k2
 ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-12
Pm1  Pm1
conf
e m 1 ci

where VIRPPci is the volume of IRP at Pci . Hence, mercury reading at Pci can be written as
a/Vbsc 1  a/Vbsc
VHgci Vbsc  
k2
  
k3
 ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-13
Pm1
ci  Pm1
conf Pm1  Pm1
conf
e m1 e m 1 ci

where VHgci is the mercury reading at Pci corrected by subtracting MICP reading at Pconf . At the final pressure in MICP testing, Pf is
reached in the cell as confining pressure, whereas all the accessible pores have been intruded by mercury, and IRP remains under com-
pression because of effective pressure. Volume of IRP, by use of the same method described previously, can be written as
1  a/Vbsc
VIRPPf 
k3
 ; .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-14
Pm1  Pm1
conf
e m1 f

where VIRPPf represents the volume of IRP at Pf . Therefore, mercury reading at Pf can be written as
1  a/Vbsc
VHgf Vbsc  
k3
 ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-15
Pm1
f  Pm1
conf
e m1

where VHgf is the mercury reading at Pf corrected by subtracting the MICP reading at Pconf . Bulk volume compressibility can be found
from the first part of this model, which is inspired by and similar to the Bailey (2009) model, which means the values of k1 and m can
be acquired by plotting Cb against pressure under the critical intrusion pressure. Porosity at standard conditions, /, is determined by use
of the low-pressure porosimetry testing. VHgci and VHgf are, respectively, the reading of cumulative MICP at Pci and Pf , with the cumu-
lative MICP reading at Pconf being subtracted from both. Combining Eqs. A-7, A-13, and A-15, a system of three equations can be
solved for the three variables a, k2 , and k3 . In this study, a VBA macro is developed to solve this system of equations by use of a numer-
ical bracketed trial-and-error method. A flow chart of logic and calculation details of the VBA macro is shown in Fig. A-1.

2017 SPE Journal 11

ID: jaganm Time: 19:25 I Path: S:/J###/Vol00000/170058/Comp/APPFile/SA-J###170058


J185059 DOI: 10.2118/185059-PA Date: 13-July-17 Stage: Page: 12 Total Pages: 12

Set a = 0.001

Calculate k3 (Eq. 15)

Calculate k2 (Eq. 13) Calculate k2 (Eq. 7)

Calculate error between the k2 values


from two different equations
Output
recorded a, k2,
k3, and error
value

a = a + 0.001
If error is smaller than recorded
(k2 and k3 are both positive),
then record this error and all
values calculated.

No Yes
Is a = 1?

Fig. A-1Flow chart of the macro used to solve the system of equations.

Yuzheng Lan holds a bachelors degree in petroleum engineering from the University of Oklahoma. His research interests include
petrophysics of unconventional formations, reservoir engineering, and fluid transport. Lan is a member of SPE.
Rouzbeh Ghanbarnezhad Moghanloo has been an assistant professor in the Melbourne School of Petroleum and Geological Engi-
neering at the University of Oklahoma since 2013. Previously, he worked as a petroleum engineer for 6 years in the areas of reservoir
engineering, enhanced oil recovery, and field development. Moghanloo is an active member of numerous technical societies,
serving as associate editor for multiple technical journals. He holds bachelors and masters degrees in chemical engineering from
Amirkabir University of Technology, Iran, and a PhD degree in petroleum engineering from the University of Texas at Austin.
Davud Davudov is a PhD degree candidate in petroleum engineering at the University of Oklahoma. His main research interests
are pore connectivity, fluid flow, and transport in tight formations. Davudov serves as a technical editor for Journal of Petroleum
and Science Engineering. He holds a bachelors degree from the Middle East Technical University, Turkey, and a masters
degree from Texas A&M University, Kingsville.

12 2017 SPE Journal

ID: jaganm Time: 19:25 I Path: S:/J###/Vol00000/170058/Comp/APPFile/SA-J###170058

Vous aimerez peut-être aussi