Jump to ContentJump to Main Navigation
Show Summary Details
More options …

Open Physics

formerly Central European Journal of Physics

Editor-in-Chief: Seidel, Sally

Managing Editor: Lesna-Szreter, Paulina

1 Issue per year


IMPACT FACTOR 2017: 0.755
5-year IMPACT FACTOR: 0.820

CiteScore 2017: 0.83

SCImago Journal Rank (SJR) 2017: 0.241
Source Normalized Impact per Paper (SNIP) 2017: 0.537

Open Access
Online
ISSN
2391-5471
See all formats and pricing
More options …
Volume 16, Issue 1

Issues

Volume 13 (2015)

Gas flow regimes judgement in nanoporous media by digital core analysis

Wenhui Song
  • State Key Laboratory of Shale Oil and Gas Enrichment Mechanisms and Effective Development; Research Centre of Multiphase Flow in Porous Media, China University of Petroleum (East China), Dongying, China
  • Email
  • Other articles by this author:
  • De Gruyter OnlineGoogle Scholar
/ Hua Liu
  • State Key Laboratory of Shale Oil and Gas Enrichment Mechanisms and Effective Development, Sinopec Petroleum Exploration and Production Research Institute, Beijing, 100083, P.R.China
  • Email
  • Other articles by this author:
  • De Gruyter OnlineGoogle Scholar
/ Weihong Wang
  • State Key Laboratory of Shale Oil and Gas Enrichment Mechanisms and Effective Development, Sinopec Petroleum Exploration and Production Research Institute, Beijing, 100083, P.R.China
  • Email
  • Other articles by this author:
  • De Gruyter OnlineGoogle Scholar
/ Jianlin Zhao / Hai Sun / Dongying Wang / Yang Li / Jun Yao
  • Corresponding author
  • Research Centre of Multiphase Flow in Porous Media, China University of Petroleum (East China), Dongying, China
  • Email
  • Other articles by this author:
  • De Gruyter OnlineGoogle Scholar
Published Online: 2018-08-13 | DOI: https://doi.org/10.1515/phys-2018-0062

Abstract

A method to judge shale gas flow regimes based on digital core analysis is proposed in this work. Firstly, three-dimensional shale digital cores in an anonymous shale formation in the Sichuan Basin are reconstructed by a Markov Chain Monte Carlo (MCMC) algorithm based on two-dimensional Scanning Electron Microscope (SEM) images. Then a voxel-based method is proposed to calculate the characteristic length of the three-dimensional shale digital core. The Knudsen number for three-dimensional shale digital cores is calculated by the ratio of the molecular mean free path to the characteristic length and is used to judge the flow regimes under different reservoir conditions. The results indicate that shale gas flow regimes are mainly located at the slip flow and transition flow region. Furthermore, adsorption has no obvious influence on the free gas flow regimes. Because adsorption only exists in organic pores, three-dimensional inorganic pores and organic pores in the Haynesville shale formation are reconstructed by a MCMC algorithm based on two-dimensional SEM images. The characteristic lengths of the three-dimensional inorganic pores and three-dimensional organic pores are both calculated and gas flow regimes in organic pores and inorganic pores are judged.

Keywords: shale gas; digital core; flow regimes; characteristic length; nanoporous media

PACS: 51.20.+d; 91.60.Np

1 Introduction

With the rapid decline in conventional reserves, unconventional resources such as tight gas and shale gas reservoirs play an increasingly important role in the North American energy industry and have gradually become a key component of the world energy supply [1, 2, 3]. In shale gas reservoirs, gas is stored in organic pores and inorganic pores with nm-scale pore size and extremely low matrix intrinsic permeability [4, 5, 6, 7]. Zou et al. [8] reported that pore sizes in shale gas reservoirs range from 1 to 200 nm and that shale matrix intrinsic permeability ranges from 10−9 to 10−3 × 10−15 m2 [9, 10]. Gas flows in the forms of free gas and adsorbed gas in an organic pore system, while gas flows in the form of free gas in an inorganic pore system [11, 12, 13]. The size of shale nanopores approach the molecular mean free path. Therefore, the continuity assumption becomes invalid [14, 15, 16]. The Knudsen number is defined as the ratio of the molecular mean free path to the pore radius [17, 18]. At different Knudsen numbers, the corresponding gas flow regime in Figure 1 differs.

Transition of flow regimes based on the Knudsen number
Figure 1

Transition of flow regimes based on the Knudsen number

Lin et al. [19] reviewed recent advances on understanding gas flow processes in unconventional porous rocks. Javadpour [20] described gas flows in the forms of Knudsen diffusion and slip flow regime and proposed a model for gas flow in a nanopore duct. Darabi and Javadpour et al. [21] incorporated Knudsen diffusion and surface roughness into the gas flow model by Maxwell theory. However, the Javadpour model [20] includes one empirical coefficient known as the tangential momentum accommodation coefficient (TMAC) and accurate prediction of TMAC for different shale samples is not available. Beskok and Karniadakis [22] developed a unified Hagen–Poiseuille-type model valid in all flow regimes and this model has been adopted by Civan et al. [18, 23, 24] to consider the intrinsic permeability, porosity, and tortuosity of porous media. Freeman et al. [25] claimed that shale gas flows in the transitional flow regime. Landry et al. [26] deemed that gas flow in nanopores at reservoir pressure – temperature conditions falls within the slip flow and early transition flow regime. Wu et al. [27] considered shale gas flows in the forms of continuum flow, slip flow and transition flow. Sun et al. [28] concluded that Knudsen diffusion can be ignored when pressure is larger than 1 MPa. Although there are different viewpoints on shale gas flow in transition flow regime, most of the current studies suggest that shale gas flows in the slip flow regime at typical reservoir conditions [3, 29, 30].

The estimated ultimate recovery (EUR) of gas reservoirs is highly dependent on the pore structure characteristics [31, 32, 33, 34, 35, 36]. Characteristics of pore systems in Sichuan Basin shale formation in China have been studied in literature [37, 38, 39, 40]. However, the current analysis and results are mostly based on two-dimensional SEM images and the three dimensional pore structure is not fully understood. Though three dimensional organic matter and inorganic matter can be observed by Focused Ion Beam-Scanning Electron Microscopes (FIB-SEM) [41, 42, 43, 44], three dimensional organic pores and inorganic pores cannot be directly imaged and divided due to the limited resolution. On the other hand, the MCMC method is widely used for image processing research [45]. The MCMC method was applied to generate pore space of real heterogeneous porous media [46] and was also used to analyze shale pore structures [47] in our previous study.

Most of the current studies on gas flow in shale needs a predefined gas flow regime based on the Knudsen number to select the corresponding gas flow model. The Knudsen number for three-dimensional porous media is generally defined by the mean free path divided by the characteristic length of porous media. Previous work established a characteristic length analytical solution based on the two dimensional conceptual porous media [48, 49]. However, the characteristic length has not been calculated in the realistic subsurface porous media. With the recent advances in imaging techniques, detailed pore structure can be captured based on digital images, for example SEM and FIB-SEM. In this study we propose an algorithm to accurately calculate the characteristic length based on a digital image and Knudsen number for three-dimensional porous media to determine gas flow regimes and flow models. Three-dimensional shale digital cores in an anonymous shale formation in Sichuan Basin are first reconstructed by the MCMC algorithm based on two-dimensional SEM images. Then a voxel-based method is proposed to calculate the characteristic length of the shale digital core. The Knudsen number for the three-dimensional shale digital core is calculated by the ratio of the mean free path to the characteristic length and is used to judge the flow regimes under different reservoir conditions. Because adsorption only exists in organic pores, three-dimensional inorganic pores and three-dimensional organic pores are reconstructed respectively by the MCMC algorithm based on two-dimensional SEM images. Characteristic lengths of three-dimensional inorganic pores and three-dimensional organic pores are both calculated and free gas flow regimes in organic pores and inorganic pores are judged. Shale gas permeability at different reservoir conditions is also analyzed.

2 MCMC digital core reconstruction method

A Markov chain can be used for describing systems that follow a chain of linked events. Subsequent events depend only on the current state of the system. Suppose X(t) is a known random process describing the state of a process at time t0. If the state of random process X(t) in the future time t(t > t0) only depends on the state at time t0 and is uncorrelated with the state before time t0, this random process X(t) is called a Markov process. During a Markov process, the relationship between “PAST” and “FUTURE” can be described as “FUTURE” linked to the “PAST” through “NOW”, and if “NOW” is confirmed, “FUTURE” is irrelevant to the “PAST”. Suppose a system is composed of n voxels (i = 1,…,n), and X = X1,…,Xn indicates the state of the voxels. Therefore, Xi means that site i has a state xi. Specifically, for a certain voxel s, Λ_s represents all of the other points except s. The neighbor of voxel s, Ns must exist and can be described as:

p(χs|χ(Λs))p(χs|χ(Ns))(1)

Assume VLMN = {(l,m,n):0<lL,0<mM,0<nN} represents L rows, M columns, and N layers of a rectangular grid filled with cube voxels. (i,j,k) represents the intersected voxel of row i, column j and layer k, and its associated state is expressed as Xijk. Vijk represents a rectangular parallelepiped array of voxels, and its associated state is expressed as vector X(Vijk). Any (i,j,k)∈ VLMN has the joint probability function:

p(x(Vijk))=l=2im=2jn=2kp(xlmn|x(l1)mn,xl(m1)n,xlm(n1))(2)

The conditional probability of each voxel for its random Markov field is:

p(xijk|{xlmn(l,m,n)(i,j,k)})=p(xijk|{xlmn(l,m,n)N(ijk)}(3)

In the absence of three-dimensional information, three mutually perpendicular independent two-dimensional images can be used to reconstruct a three-dimensional chain. The core slice data from three mutually perpendicular planes xy, yz and xz are extracted and the corresponding binary images are obtained. The double voxel method is applied to simultaneously produce two new voxels, i.e. voxel (i, j, k) and (i, j+1, k), during the modeling process. Therefore, the reconstruction steps for the three-dimensional digital core are as follows. The porosity of the horizontal core slices is used as the conditional probability and the voxels on the first line of the first layer are simulated along the y-direction. The second voxel of the first line is simulated using the 2-neighborhood model, and after that, the 3-neighborhood model is used at the beginning of the simulation of the third voxel. Its conditional probability is calculated by the two-dimensional core slice on the xy plane (Figure 3(b)). The first layer is simulated along the x direction line by line. For double elements (i,j) and (i,j+1), 3 and 4 neighborhoods are used when simulating the edge voxels, and 5 and 6 neighborhoods are used when simulating the internal voxels. The conditional probability is derived by the two-dimensional core slice on the xy plane (Figure 3(c)). Each layer of voxels is simulated along the Z-direction and the three-dimensional model is reconstructed (Figure 3(d)).

Sketch map of a heterogeneous porous media simulation based on three perpendicular thin sections
Figure 2

Sketch map of a heterogeneous porous media simulation based on three perpendicular thin sections

The three dimensional digital core reconstruction process based on the MCMC method (white elements with lines are the “NOW” simulating voxel; white elements with dotted lines are the “FUTURE” simulating voxel; dark elements are the voxels already produced, while light elements are the neighbors of the voxel currently being simulated, and they both belong to “PAST” voxels)
Figure 3

The three dimensional digital core reconstruction process based on the MCMC method (white elements with lines are the “NOW” simulating voxel; white elements with dotted lines are the “FUTURE” simulating voxel; dark elements are the voxels already produced, while light elements are the neighbors of the voxel currently being simulated, and they both belong to “PAST” voxels)

3 Reconstruction of Sichuan Basin shale digital core

Three-dimensional shale digital cores in Figure 5 and Figure 6 are reconstructed based on binary SEM images (Figure 4) using the above mentioned MCMC algorithm. The blue colour represents pore phase and the red color represents matrix phase. In three-dimensional shale digital cores, pore phase is represented by number 0 and matrix is represented by number 1. Detailed parameters of shale digital cores are given in Table 1.

Binary SEM images of shale samples (white represents matrix while black represents pore phases)
Figure 4

Binary SEM images of shale samples (white represents matrix while black represents pore phases)

Three dimensional shale digital cores reconstructed by MCMC
Figure 5

Three dimensional shale digital cores reconstructed by MCMC

Pore phases of three dimensional shale digital cores
Figure 6

Pore phases of three dimensional shale digital cores

Table 1

Parameters of shale digital cores

4 Voxel based characteristic length calculation method

For single tubes, the analysis on gas flow regimes has been well studied [20, 50]. The Knudsen number for a single tube and a three-dimensional porous media can be written respectively as:

Kn=λr(4)

Kn=λlc(5)

λ=kBT2πdm2Pg(6)

lc is referred to as the characteristic length of porous media and can be calculated based on the chord length [48]. The chord lengths are those discontinuous lines distributed in pore phase when a line is used to cut porous media. The porous media characteristic length indicates the degree of gas diffusion in porous media and its physical meaning is the free path when gas transports within the porous media. The porous media characteristic length has been used to analyze gas transport ability in loose packings of spheres [51, 52] and can be calculated by [48, 49]:

lc=0zp(z)dz(7)

Chord lengths are the distributions of lengths between intersections of lines with the interface (Figure 7). A chord is then a special line segment with its end points on the interface and all other points in one of the two phases. Here chord lengths in pore phase are calculated. p(z) is the chord length distribution function in pore phases. p(z)dz is the probability of finding a chord of length between z and z+dz in pore phase.

Schematic of chord-length measurements for a cross section of porous media. The chords are defined by intersection of lines with two-phase interface, closed curve area refers to rock matrix [32]
Figure 7

Schematic of chord-length measurements for a cross section of porous media. The chords are defined by intersection of lines with two-phase interface, closed curve area refers to rock matrix [32]

Previous studies have not applied Equation (5) to real rock samples. Here we propose a method to calculate the characteristic length based on digital image. The chord length in the binary digital image can be represented as the length of continuous number 0 in the pore phase. A regional example in two dimensions is shown in Figure 8, the voxel-based chord length is represented by the number of continuous 0 on red line. In three-dimensional pore phase, the first chord length is measured in the bottom slice on the XY plane and then the measuring window moves up in the Z direction layer by layer. Finally, chord length for three-dimensional pore phase can be obtained and Equation (8) is used to calculate the characteristic length of the three-dimensional pore phase. A validation of the proposed voxel based characteristic length calculation method can be found in the next section.

lc_3D=0z_volp(z_vol)dz_vol×γ(8)

Illustration of chord length measurements in XY cross section
Figure 8

Illustration of chord length measurements in XY cross section

Gas adsorption is generally assumed to be a monolayer adsorption [53] and can be characterized with the Langmuir isotherm. Considering the influence of an adsorption layer, first adsorbed gas coverage on the pore wall is calculated by Equation (9) and the effective characteristic length of the three-dimensional pore phase is given in Equation (10).

θ=Pg/PL1+Pg/PL(9)

lc(Pg)=lc_3Ddmθ(10)

According to Equation (5), Equation (6), Equation (9) and Equation (10), the Knudsen number without adsorption and with adsorption can be given respectively as:

Kn_3D=kBT2πdm2Pg0z_volp(z_vol)dz_vol×γ(11)

Kn_3D_ads=kBT2πdm2Pg(0z_volp(z_vol)dz_voldmPg/PL1+Pg/PL)×γ(12)

5 Validation of the proposed voxel based characteristic length calculation method

The three-dimensional microscale lattice Boltzmann (LB) model with the D3Q19 discrete velocity model developed in our previous study [54] is adopted to simulate gas flow in a shale digital core. The developed LB model can be applied to simulate gas flow in slip flow and transition flow regimes. The basic evolution equation with the Bhatnagar-GrossKrook (BGK) collision approximation is shown as follows:

fα(rs+eαδt,t+δt)fα(rs,t)=1τ(fαfαeq)+δtFα(13)

Where fα is the density distribution function of α direction; α = 0, 1, 2, … , 18; rs is the spatial location of the particles; eα is the velocity of α direction; t is time; δt is time step; τ is the relaxation time; Fα is the force term; fαeq is the local equilibrium distribution function of α direction:

fαeq=wαρ[1+eαucs2+(eαu)22cs4u22cs2](14)

Where cs is the lattice sound speed; wα is the weight factor of α direction. The force term in Equation (13) can be obtained by Hermite expansion. For the D3Q19 model, it can be expressed as [55]:

Fα=wαρ[eαacs2+au:(eαeαcs2I)cs4](15)

Where a is the acceleration of the force. For gas flow in shale nanopores, the relaxation time should be determined by the Knudsen number. Considering the microscale effect and the effect of the Knudsen layer, the relaxation time can be expressed as [56]:

τe=12+6πKnψ(Kn)N(16)

ψ(Kn) is the modification function and equals 1/(1+2Kn); N is the number of lattices occupied by the characteristic length. As the solid boundaries in shale are rough, the diffuse reflection boundary condition is very appropriate for such boundaries. The discrete format of the diffuse reflection boundary condition is expressed as [57]:

fα=ξαn<0|ξαn|fαξαn>0|ξαn|fαeq(ρw,μw)fαeq(ρw,μw),(eαuw)n>0(17)

Where n is the inward unit normal vector; uw is the wall velocity; the subscript w represents the solid walls; fα is the distribution function after streaming. Gas flux Q is calculated in the outlets and the Darcy equation is applied to calculate gas permeability:

k1=QμLAΔP(18)

Detailed parameters are given in Table 2. The studied shale digital core is shown in Figure 10 and pressure drop distribution on the shale digital core is shown in Figure 11. The calculated permeability k1 is 2.7831×10−7 μm2.

Illustration of chord length measurements in three orthogonal X,Y,Z directions
Figure 9

Illustration of chord length measurements in three orthogonal X,Y,Z directions

Shale digital core (blue represents pore phase, red represents matrix phase)
Figure 10

Shale digital core (blue represents pore phase, red represents matrix phase)

Pressure drop distribution on the shale digital core
Figure 11

Pressure drop distribution on the shale digital core

Table 2

Parameters of shale digital core for model validation

The Civan et al. model [18] in Equation (19) is applied to calculate gas permeability in slip flow regime and transition flow regime. The characteristic length is calculated based on the proposed voxel based method in Equation (8) and the Knudsen number is calculated using Equation (11).

k2=ϕrc28τf(Kn)=ϕlc232τf(Kn)(19)

The flow condition function f(Kn) is given by [22]:

f(Kn)=(1+αKn)(1+4Kn1βKn)(20)

The parameter α in Equation (20) is a dimensionless rarefaction coefficient and can be written as:

α=12815π2tan1[4.0Kn0.4](21)

Though the slip coefficient β = −1 is initially considered only to be applicable to slip flow condition, evidence from the DSMC simulations and Boltzmann solutions [58] showed that β = −1 is valid within the full range of flow regimes. The calculated characteristic length in Equation (8) is 11.72nm and the calculated permeability k2 in Equation (19) is 2.9784×10−7 μm2. k2 is very close to k1 which suggests the proposed characteristic length calculation algorithm is correct.

6 Gas transport mechanism judgement and permeability analysis

6.1 Gas flow regimes judgement of Sichuan Basin shale

Firstly characteristic lengths are calculated by Equation (8) based on three-dimensional pore phase of shale digital cores shown in Figure 6. From the calculated results listed in Table 3, the average characteristic length for Sichuan Basin shale is 14.87 nm. Subsequently, the average characteristic length is used in Equation (11) and Equation (12) to study the change in the Knudsen number under different formation pressure and temperature without adsorption effect and with adsorption effect. The relative degree of deviation is defined to study the free gas transport differences in the above two cases:

dev=|Kn_3DKn_3D_ads|Kn_3D×100%(22)

Table 3

Characteristic lengths of anonymous shale formation, Sichuan Basin

The Knudsen number values in Figure 12, Figure 13 are compared with the Knudsen number values in Figure 1. It is can be seen that shale gas flow regimes mainly locate at the slip flow and transition flow region. The value of the relative degree of deviation is larger in high pore pressure than that in low pore pressure. However, the maximum value of the relative degree of deviation is 5%. Therefore, the adsorption effect has no obvious influence on the free gas transport mechanism in the Sichuan Basin shale gas reservoir.

Knudsen number versus temperature and pressure with adsorption effects
Figure 12

Knudsen number versus temperature and pressure with adsorption effects

Knudsen number versus temperature and pressure without adsorption effects
Figure 13

Knudsen number versus temperature and pressure without adsorption effects

6.2 Gas flow regimes judgement based on three dimensional organic pores system and inorganic pores system

Based on the aforementioned MCMC method, SEM images (Figure 15) in Haynesville shale [59] are used to reconstruct three-dimensional organic pores and inorganic pores. First we identify organic pores and inorganic pores by the pixel value of red and green colour. Then, the corresponding binary images can be obtained using the maximum class separation distance method proposed by Otsu [60]. Taking a pixel as the basic unit, the pixel value is set to 1 when the pixel is located in the matrix phase and the pixel value is set to 0 when the pixel is located in the pore phase (Figure 16, Figure 17). Finally, the MCMC method is applied to reconstruct three-dimensional organic pores and inorganic pores based on binary images (Figure 18). The voxel sizes of the three-dimensional organic pores and inorganic pores are 800 × 800 × 800.

Relative degree of deviation comparing the above two results
Figure 14

Relative degree of deviation comparing the above two results

American Haynesville Shale SEM Image
Figure 15

American Haynesville Shale SEM Image

Binary image of inorganic pore in Haynesville shale rocks (white is the rock matrix, and black is the pore space)
Figure 16

Binary image of inorganic pore in Haynesville shale rocks (white is the rock matrix, and black is the pore space)

Binary image of organic pore in Haynesville shale rocks (white is the rock matrix, and black is the pore space)
Figure 17

Binary image of organic pore in Haynesville shale rocks (white is the rock matrix, and black is the pore space)

(a) Three dimensional inorganic pores (b) Three dimensional organic pores
Figure 18

(a) Three dimensional inorganic pores (b) Three dimensional organic pores

Based on the statement that gas transports in the forms of free gas and adsorbed gas in organic pores and gas transports only in the form of free gas in inorganic pores, Equation (8) is applied to calculate the characteristic length without adsorption effect in three-dimensional inorganic pores and organic pores. Our calculation results indicate characteristic lengths without adsorption effect in three-dimensional inorganic pores and organic pores are 49.5 nm and 35 nm respectively. Then, the value of the characteristic length without adsorption effect in three-dimensional organic pores of 35 nm is revised by Equation (10) to consider the influence of the adsorption layer. Finally, Equation (11) and Equation (12) are used respectively to calculate the Knudsen number in three-dimensional inorganic pores and organic pores. In Figure 19 and Figure 20, the Knudsen number differs a little in low pore pressure (< 5 MPa). But for shale gas reservoirs, pore pressure usually ranges from 10 MPa to 40 MPa [61] and the Knudsen number is almost the same in three-dimensional organic pores and inorganic pores under this condition. Therefore, free gas flow regimes can be deemed as the same in inorganic pores and organic pores. This may indicate that in macro scale gas flow numerical simulation, inorganic pores and organic pores can be viewed as a continuous system and a dual medium (matrix-micro fracture) mathematical model is suitable to simulate the shale gas production process.

Knudsen number versus temperature and pressure in three dimensional organic pores
Figure 19

Knudsen number versus temperature and pressure in three dimensional organic pores

Knudsen number versus temperature and pressure in three dimensional inorganic pores
Figure 20

Knudsen number versus temperature and pressure in three dimensional inorganic pores

6.3 Shale gas permeability at different reservoir conditions

According to Equation (10), Equation (12) and Equation (19), free gas permeability in consideration of adsorption can be given as:

kfree=ϕlc2(Pg)32τf(Kn_3D_ads)(23)

Surface diffusion takes place within the adsorbed gas to enhance the transport of gas molecules along molecular concentration gradients [62]. Based on Hwang and Kammermeyer’s model [63], combined with methane adsorption experimental data, the surface diffusion coefficient for methane Ds0 when gas coverage is zero can be expressed as:

Ds0=8.29×107T0.5exp(ΔH0.8RT)(24)

The isosteric adsorption heat ΔH is a function of gas coverage. According to Equation (25), the isosteric adsorption heat and gas coverage have a linear relationship and can be given as [64]:

ΔH=ηθ+ΔH(0)(25)

The surface diffusion coefficient in Equation (24) is obtained under a low pressure condition by theory and experiments, and is a function of gas molecular weight, temperature, and gas activation energy, isosteric adsorption heat and is independent of pressure [63]. In order to describe the gas surface diffusion in nanopores of shale gas reservoirs under a high pressure condition, the influence of gas coverage on surface diffusion is considered. Chen et al. [65] used the kinetic method to calculate the surface diffusion coefficient:

Ds=Ds0(1θ)+κ2θ(2θ)+{H(1κ)}(1κ)κ2θ2(1θ+κ2θ)2(26)

H(1κ)=0,κ1;1,0κ1(27)

Where H(1-κ) is Heaviside function. The surface diffusion coefficient at different pressure and temperature conditions is calculated according to Equation (26). Surface diffusion of adsorbed gas molecules can be modelled as the general diffusion process, using the molar flow rate per unit area of the concentration gradient within the adsorbed monolayer as developed in [62]:

Ja=DsdCadx(28)

Ca is calculated assuming Langmuir adsorption and is given by:

Ca=Camaxθ(29)

Camax can be expressed as [66]:

Camax=Cmaxεks(30)

Combining Equation (28) and Equation (29), molar flow rate in the adsorbed layer is then expressed below:

JA=DsCamaxdθdpπ(lc2lc2(Pg))dpdx(31)

From Equation (31), volumetric flow rate VA is:

VA=MρDsCamaxdθdpπ(lc2lc2(Pg))dpdx(32)

According to Equation (26) and the Darcy law, adsorbed gas permeability can be written as:

ksurface=ϕμMτρDsCamaxdθdp(1(lc(Pg)lc)2)(33)

Shale gas permeability in Equation (34) can be derived by combining free gas permeability in consideration of adsorption by Equation (23) with surface diffusion of adsorbed gas by Equation (33). Typical shale gas reservoir parameters are given in Table 4. Figure 21 shows that gas permeability increases with the decrease of pressure and the increase of temperature. The gas permeability value ranges from 4.3685 × 10−7 μm2 to 2.2451 × 10−6 μm2 for different reservoir conditions.

kt=ϕlc2(Pg)32τf(Kn_3D_ads)+ϕμMτρDsCamaxdθdp(1(lc(Pg)lc)2)(34)

Shale gas permeability versus temperature and pressure
Figure 21

Shale gas permeability versus temperature and pressure

Table 4

Typical shale gas reservoir parameters

7 Conclusions

A method to judge shale gas flow regimes based on digital core analysis is studied. A Voxel-based method is proposed to the calculate characteristic length of a three-dimensional shale digital core. The Knudsen number for MCMC reconstructed three-dimensional shale digital core is calculated by the ratio of the molecular mean free path to the characteristic length and is used to judge the gas flow regime under different reservoir conditions. The results indicate that shale gas flow regimes mainly locate at the slip flow and transition flow region. Furthermore, adsorption has no obvious influence on the free gas flow regimes. Moreover, free gas flow regimes in organic pores and inorganic pores are judged. Our analysis results show that free gas flow regimes can be deemed as the same in two different pore systems and inorganic-organic pores can be viewed as the continuous system in a macro scale gas flow simulation. Gas permeability increases with the decrease of pressure and the increase of temperature.

Acknowledgement

This project was supported by the National Natural Science Foundation of China (No. 51504276, No.51490654), the Fundamental Research Funds for the Central Universities (No.16CX05018A, No. 18CX06008A), the Major Projects of the National Science and Technology (2016ZX05061), Graduate School Innovation Program of China University of Petroleum (YCX2017019).

Nomenclature

ϵks

total organic grain volume per total grain volume, dimensionless

κ

ratio of the rate constant for blockage to the rate constant for forward migration

ΔH(0)

isosteric adsorption heat at zero gas coverage (J/mol)

η

fitting coefficients of isosteric adsorption heat (J/mol)

γ

resolution(m)

λ

mean free path of molecules(m)

μ

viscosity(Pa⋅s)

Φ

porosity, dimensionless

ρ

gas density (kg/m3)

τ

tortuosity, dimensionless

θ

gas coverage on the pore wall, dimensionless

A

cross section area of shale digital core (m2)

Camax

maximum adsorbed gas concentration inside the total core sample (mol/m3)

Ca

adsorbed gas concentration (mol/m3)

Cmax

maximum adsorbed gas concentration inside the organic matter (mol/m3)

dm

gas molecular diameter (m) 4 × 10−10

Ds

surface diffusion coefficient (m2/s)

Ds0

surface diffusion coefficient when gas coverage is zero (m2/s)

dev

relative degree of deviation, dimensionless

Ja

molar flow rate per unit area (mol/(m2⋅s))

kB

Boltzmann constant (J/K),1.3805 × 10−23

Kn

Knudsen number, dimensionless

Kn_3D

Knudsen number without adsorption effect

Kn_3D_ads

Knudsen number with adsorption effect

k

permeability (μm2)

lc

characteristic length (m)

lc_3D

characteristic length of three dimensional pore phases (m)

lc(Pg)

characteristic length of three dimensional pore phases with adsorption effect(m)

Mg

gas molecular weight (g/mol)

N(ijk)

neighbor of (i,j,k)

n

voxel number

Pg

pore pressure(MPa)

pL

Langmuir pressure(MPa)

p(z)

chord length distribution function in pore phase, dimensionless

Q

gas flux (m3/s)

r

single tube radius (m)

s

certain voxel, dimensionless

t

time(s)

vg

gas velocity (m/s)

Vijk

rectangular parallelepiped array of voxels

VLMN

rectangular grid filled with cube voxels, dimensionless

X(t)

Markov process

z_vol

chord length based on voxel unit in three dimensional digital image

z

chord length (m)

References

  • [1]

    Zhang K., Ma X., Li Y, et al., Parameter prediction of hydraulic fracture for tight reservoir based on micro-seismic and history 65 matching, Fractals, 2018: 1840009. Google Scholar

  • [2]

    Zhang K., Zhang X., Zhang L, et al., Assisted history matching for the inversion of fractures based on discrete fracture-matrix model with different combinations of inversion parameters, Computational Geosciences, 2017, 21(5-6): 1365-1383. CrossrefGoogle Scholar

  • [3]

    Wang L., Wang S., Zhang R, et al., Review of multi-scale and multi-physical simulation technologies for shale and tight gas reservoirs, Journal of Natural Gas Science and Engineering, 2017, 37: 560-578. CrossrefGoogle Scholar

  • [4]

    Singh H., Javadpour F. Langmuir slip-Langmuir sorption permeability model of shale, Fuel, 2016, 164: 28-37. CrossrefGoogle Scholar

  • [5]

    Ma J., Couples G. D., Jiang Z, et al., Predicting the Effective Gas Flow Properties of Gas Shale, from Nano-Pores to Shale Parasequences, (12-14 August, Denver 2013, Colorado, USA), Proceedings of URTEC: Unconventional Resources Technology Conference, Society of Petroleum Engineers, 2013. Google Scholar

  • [6]

    Song W., Yao J., Ma J, et al., Pore-scale numerical investigation into the impacts of the spatial and pore-size distributions of organic matter on shale gas flow and their implications on multiscale characterisation, Fuel, 2018, 216: 707-721. CrossrefGoogle Scholar

  • [7]

    Zhu H., Ju Y., Lu W, et al., The Characteristics and Evolution of Micro-Nano Scale Pores in Shales and Coals, Journal of Nanoscience and Nanotechnology, 2017, 17(9): 6124-6138. CrossrefGoogle Scholar

  • [8]

    Zou C., Zhu R., Wu S, et al., Types, characteristics, genesis and prospects of conventional and unconventional hydrocarbon accumulations: Taking tight oil and tight gas in China as an instance, Acta Petrolei Sinica, 2012, 33(2): 173-187. Google Scholar

  • [9]

    Loucks R. G., Reed R. M., Ruppel S. C, et al., Morphology, genesis, and distribution of nanometer-scale pores in siliceous mudstones of the Mississippian Barnett Shale, Journal of sedimentary research, 2009, 79(12): 848-861. CrossrefGoogle Scholar

  • [10]

    Wang F. P., Reed R. M., Pore networks and fluid flowin gas shales, (4-7 October, 2009, New Orleans, Louisiana, USA), Proceedings of the SPE Annual Technical Conference and Exhibition, Society 14 Ë W. Song et al. of Petroleum Engineers, 2009. Google Scholar

  • [11]

    Ma J., Sanchez J. P., Wu K, et al., A pore network model for simulating non-ideal gas flow in micro- and nano-porous materials, Fuel, 2014, 116(Supplement C): 498-508. CrossrefGoogle Scholar

  • [12]

    Ma J., Couples G. D., Zhao X, et al., Assessing Impact of Shale Gas Adsorption on Free-Gas Permeability via a Pore Network Flow Mode, (20-22 July, 2015, San Antonio, Texas, USA), Proceedings of, URTEC: Unconventional Resources Technology Conference, Society of Petroleum Engineers, 2015. Google Scholar

  • [13]

    Riewchotisakul S., Akkutlu I. Y. Adsorption-Enhanced Transport of Hydrocarbons in Organic Nanopores, SPE Journal, 2016, 21(07): 1960-1969. CrossrefGoogle Scholar

  • [14]

    Ho C.-M., Tai Y.-C. Micro-electro-mechanical-systems (MEMS) and fluid flows, Annual Review of Fluid Mechanics, 1998, 30(1): 579-612. CrossrefGoogle Scholar

  • [15]

    Song W., Yao J., Li Y, et al., Apparent gas permeability in an organic-rich shale reservoir, Fuel, 2016, 181: 973-984. CrossrefGoogle Scholar

  • [16]

    Song W., Yao J., Ma J, et al., Assessing relative contributions of transport mechanisms and real gas properties to gas flow in nanoscale organic pores in shales by pore network modelling, International Journal of Heat and Mass Transfer, 2017, 113: 524-537. CrossrefGoogle Scholar

  • [17]

    Ziarani A. S., Aguilera R. Knudsen’s permeability correction for tight porous media, Transport in Porous Media, 2012, 91(1): 239-260. CrossrefGoogle Scholar

  • [18]

    Civan F. Effective correlation of apparent gas permeability in tight porous media, Transport in Porous Media, 2010, 82(2): 375-384. CrossrefGoogle Scholar

  • [19]

    Lin D., Wang J., Yuan B, et al., Review on gas flow and recovery in unconventional porous rocks, Adv Geo-Energy Res, 2017, 1: 39-53. CrossrefGoogle Scholar

  • [20]

    Javadpour F., Nanopores and apparent permeability of gas flow in mudrocks (shales and siltstone), Journal of Canadian Petroleum Technology, 2009, 48(08): 16-21. CrossrefGoogle Scholar

  • [21]

    Darabi H., Ettehad A., Javadpour F, et al., Gas flow in ultra-tight shale strata, Journal of Fluid Mechanics, 2012, 710: 641-658. CrossrefGoogle Scholar

  • [22]

    Beskok A., Karniadakis G. E., Report: a model for flows in channels, pipes, and ducts at micro and nano scales, Microscale Thermophysical Engineering, 1999, 3(1): 43-77. CrossrefGoogle Scholar

  • [23]

    Civan F., Devegowda D., Sigal R. F., Critical evaluation and improvement of methods for determination of matrix permeability of shale,(1-2 September October, 2013, New Orleans, Louisiana, USA), Proceedings of the SPE Annual Technical Conference and Exhibition, Society of Petroleum Engineers, 2013. Google Scholar

  • [24]

    Civan F., Rai C. S., Sondergeld C. H., Shale-gas permeability and diffusivity inferred by improved formulation of relevant retention and transport mechanisms, Transport in Porous Media, 2011, 86(3): 925-944 CrossrefGoogle Scholar

  • [25]

    Freeman C., Moridis G., Blasingame T., A numerical study of mi50 croscale flow behavior in tight gas and shale gas reservoir systems, Transport in Porous Media, 2011, 90(1): 253. CrossrefGoogle Scholar

  • [26]

    Landry C. J., Prodanović M., Eichhubl P., Direct simulation of supercritical gas flow in complex nanoporous media and prediction of apparent permeability, International Journal of Coal Ge55 ology, 2016, 159: 120-134. CrossrefGoogle Scholar

  • [27]

    Wu K., Chen Z., Li X, et al., A model for multiple transport mechanisms through nanopores of shale gas reservoirs with real gas effect–adsorption-mechanic coupling, International Journal of Heat and Mass Transfer, 2016, 93: 408-426. CrossrefGoogle Scholar

  • [28]

    Sun H., Yao J., Cao Y.-c, et al., Characterization of gas transport behaviors in shale gas and tight gas reservoirs by digital rock analysis, International Journal of Heat and Mass Transfer, 2017, 104: 227-239. CrossrefGoogle Scholar

  • [29]

    Ilgen A. G., Heath J. E., Akkutlu I. Y, et al., Shales at all scales: Exploring coupled processes in mudrocks, Earth-Science Reviews, 2017, 166: 132-152. CrossrefGoogle Scholar

  • [30]

    Song W., Yao J., Li Y, et al., Fractal models for gas slippage factor in porous media considering second-order slip and surface adsorption, International Journal of Heat and Mass Transfer, 2018, 118: 948-960. CrossrefGoogle Scholar

  • [31]

    Wang C., Yao J., Wu K, et al., Pore-scale characterization of carbonate rock heterogeneity and prediction of petrophysical properties, (27-30 August, 2012, Aberdeen, Scotland, UK), Proceedings of the International Symposium of the Society of Core Analysts, 2012. Google Scholar

  • [32]

    Jian-Chao C., A fractal approach to low velocity non-Darcy flow in a low permeability porous medium, Chinese Physics B, 2014, 23(4): 044701. CrossrefGoogle Scholar

  • [33]

    Yao J., Wang C., Yang Y, et al., The construction of carbonate digital rock with hybrid superposition method, Journal of Petroleum Science and Engineering, 2013, 110: 263-267. CrossrefGoogle Scholar

  • [34]

    Cai J., Luo L., Ye R, et al., Recent advances on fractal modeling of permeability for fibrous porous media, Fractals, 2015, 23(01): 1540006. CrossrefGoogle Scholar

  • [35]

    Chen D., Yao Y., Fu G, et al., A new model for predicting liquid loading in deviated gas wells, Journal of Natural Gas Science and Engineering, 2016, 34: 178-184. CrossrefGoogle Scholar

  • [36]

    Xia Y., Cai J., Wei W, et al., A new method for calculating fractal dimensions of porous media based on pore size distribution, Fractals, 2018, 26(01): 1850006. CrossrefGoogle Scholar

  • [37]

    Guo X., Li Y., Liu R, et al., Characteristics and controlling factors of micropore structures of the Longmaxi Shale in the Jiaoshiba area, Sichuan Basin, Natural Gas Industry B, 2014, 1(2): 165-171. CrossrefGoogle Scholar

  • [38]

    Jiao K., Yao S., Liu C, et al., The characterization and quantitative analysis of nanopores in unconventional gas reservoirs utilizing FESEM–FIB and image processing: An example from the lower Silurian Longmaxi Shale, upper Yangtze region, China, International Journal of Coal Geology, 2014, 128: 1-11. Google Scholar

  • [39]

    Liang C., Jiang Z., Zhang C, et al., The shale characteristics and shale gas exploration prospects of the Lower Silurian Longmaxi shale, Sichuan Basin, South China, Journal of Natural Gas Science and Engineering, 2014, 21: 636-648. CrossrefGoogle Scholar

  • [40]

    Shi M., Yu B., Xue Z, et al., Pore characteristics of organic-rich shales with high thermal maturity: A case study of the Longmaxi gas shale reservoirs from well Yuye-1 in southeastern Chongqing, China, Journal of Natural Gas Science and Engineering, 2015, 26: 948-959. CrossrefGoogle Scholar

  • [41]

    Song W., Yao J., Li Y, et al., New pore size distribution calculation model based on chord length and digital image, Journal of Natural Gas Science and Engineering, 2017, 48, 111-118. CrossrefGoogle Scholar

  • [42]

    Milliken K. L., Rudnicki M., Awwiller D. N, et al., Organic matter–hosted pore system, Marcellus formation (Devonian), Pennsylvania, AAPG bulletin, 2013, 97(2): 177-200. CrossrefGoogle Scholar

  • [43]

    Bai B., Elgmati M., Zhang H, et al., Rock characterization of Fayetteville shale gas plays, Fuel, 2013, 105: 645-652. CrossrefGoogle Scholar

  • [44]

    Song W., Yao J., Ma J, et al., Numerical simulation of multiphase flow in nanoporous organic matter with application to coal and gas shale systems, Water Resources Research, 2018, 54(2): 1077-1092. CrossrefGoogle Scholar

  • [45]

    Geman S., Geman D., Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images, Pattern Analysis and Machine Intelligence, IEEE Transactions on, 1984, (6): 721-741. Google Scholar

  • [46]

    Wu K., Van Dijke M. I., Couples G. D, et al., 3D stochastic modelling of heterogeneous porous media–applications to reservoir rocks, Transport in Porous Media, 2006, 65(3): 443-467. CrossrefGoogle Scholar

  • [47]

    Yang Y., Yao J., Wang C, et al., New pore space characterization method of shale matrix formation by considering organic and inorganic pores, Journal of Natural Gas Science and Engineering, 2015, 27: 496-503. CrossrefGoogle Scholar

  • [48]

    Lu B., Torquato S., Chord-length and free-path distribution functions for many-body systems, The Journal of chemical physics, 1993, 98(8): 6472-6482. CrossrefGoogle Scholar

  • [49]

    Torquato S., Random heterogeneous materials: microstructure and macroscopic properties[M]. Springer Science & Business Media, 2013 Google Scholar

  • [50]

    Javadpour F., Fisher D., Unsworth M., Nanoscale gas flow in shale gas sediments, Journal of Canadian Petroleum Technology, 2007, 46(10):55-61. Google Scholar

  • [51]

    Zalc J. M., Reyes S. C., Iglesia E., The effects of diffusion mechanism and void structure on transport rates and tortuosity factors in complex porous structures, Chemical Engineering Science, 2004, 59(14): 2947-2960. CrossrefGoogle Scholar

  • [52]

    Levitz P., Knudsen diffusion and excitation transfer in random porous media, The Journal of Physical Chemistry, 1993, 97(15): 3813-3818. CrossrefGoogle Scholar

  • [53]

    Heller R., Zoback M. Adsorption of methane and carbon dioxide on gas shale and pure mineral samples, Journal of Unconventional Oil and Gas Resources, 2014, 8: 14-24. CrossrefGoogle Scholar

  • [54]

    Zhao J., Yao J., Zhang M, et al., Study of gas flow characteristics in tight porous media with a microscale lattice boltzmann model, Scientific Reports, 2016, 6: 32393. CrossrefGoogle Scholar

  • [55]

    Suga K., Takenaka S., Ito T, et al., Evaluation of a lattice Boltzmann method in a complex nanoflow, Physical Review E, 2010, 82(1): 016701. CrossrefGoogle Scholar

  • [56]

    Li Q., He Y., Tang G, et al., Lattice Boltzmann modeling of microchannel flows in the transition flow regime, Microfluidics and nanofluidics, 2011, 10(3): 607-618. CrossrefGoogle Scholar

  • [57]

    Tang G., TaoW., He Y., Lattice Boltzmann method for gaseous microflows using kinetic theory boundary conditions, Physics of 40 Fluids, 2005, 17(5): 058101. CrossrefGoogle Scholar

  • [58]

    Karniadakis G., Beskok A., Aluru N., Microflows and nanoflows: fundamentals and simulation[M]. Springer Science & Business Media, 2006. Google Scholar

  • [59]

    Rine J., Dorsey W., Floyd M, et al., A Comparative SEM Study of Pore Types and Porosity Distribution in High to Low Porosity Samples from Selected Gas-Shale Formations, Gulf Coast Association of Geological Societies Transactions, 2010, 60:825. Google Scholar

  • [60]

    Otsu N., A threshold selection method from gray-level histograms, Automatica, 1975, 11(285-296): 23-27. Google Scholar

  • [61]

    Freeman C., Moridis G., Blasingame T., A numerical study of microscale flow behavior in tight gas and shale gas reservoir systems, Transport in Porous Media, 2011, 90(1): 253-268. CrossrefGoogle Scholar

  • [62]

    Cunningham R. E., Williams R., Diffusion in gases and porous media[M]. Springer, 1980. Google Scholar

  • [63]

    Hwang S. T., Kammermeyer K., Surface diffusion in microporous media, The Canadian Journal of Chemical Engineering, 1966, 44(2): 82-89. CrossrefGoogle Scholar

  • [64]

    Nodzeński A., Sorption and desorption of gases (CH4, CO2) on hard coal and active carbon at elevated pressures, Fuel, 1998, 77(11): 1243-1246 CrossrefGoogle Scholar

  • [65]

    Chen Y. D., Yang R. T., Concentration dependence of surface diffusion and zeolitic diffusion, AIChE Journal, 1991, 37(10): 1579-1582. CrossrefGoogle Scholar

  • [66]

    Wasaki A., Akkutlu I. Y., Permeability of organic-rich shale, SPE Journal, 2015, 20(06): 1,384-381,396. Google Scholar

  • [67]

    Wu K., Li X., Wang C, et al., Model for surface diffusion of adsorbed gas in nanopores of shale gas reservoirs, Industrial & Engineering Chemistry Research, 2015, 54(12): 3225-3236. CrossrefGoogle Scholar

About the article

Received: 2017-12-19

Accepted: 2018-06-12

Published Online: 2018-08-13


Conflict of InterestConflict of Interests: The authors declare no conflict of interest.


Citation Information: Open Physics, Volume 16, Issue 1, Pages 448–462, ISSN (Online) 2391-5471, DOI: https://doi.org/10.1515/phys-2018-0062.

Export Citation

© 2018 W. Song et al., published by De Gruyter. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License. BY-NC-ND 4.0

Comments (0)

Please log in or register to comment.
Log in