Analysis of methane recovery through CO 2 –N 2 mixed gas injection considering gas diffusion phenomenon in coal seam

Abstract

When enhancing coalbed methane recovery using CO₂ or N₂ injection, injected gas flows into coal matrix by diffusion. Gas diffusion velocity varies, depending on gas molecular size and pore geometry which causes different sorption rates of the gas in coal seam. In this aspect, this study provides the fundamental reason for the reduction in gas permeability through cleats and methane recovery during enhanced coalbed methane (ECBM) processes. This reduction occurs not only because of the sorption affinity as reported in previous works, but also because of the characteristics of gas diffusional flow which this study attempted to examine experimentally. From the results obtained by diffusional flow experiment, diffusion coefficient is no longer increased at high pressure. Although CO₂ injection rate is very high, a large amount of CO₂ moves through cleat instead of adsorbed in matrix, which causes early CO₂ breakthrough. In ECBM, N₂ mostly acts as a displacing agent of methane, because co-diffusion of N₂ with methane is more dominant than counter-diffusion owing to its extremely low adsorption affinity. On the other hand, CO₂ is rapidly adsorbed due to its fast increasing rate of diffusion coefficient with pressure increase. Consequently, CO₂ permeability is greatly reduced at the beginning of injection and ultimately becomes the lowest value at the maximum adsorption pressure. Also, delayed methane recovery by fast diffusion and high adsorption affinity of CO₂ occurs accordingly. This study confirms that the CO₂–N₂ mixed gas injection is advisable comparing to only injecting CO₂ to pursue the prevention of CO₂ injectivity reduction and enhanced methane recovery, simultaneously.

Keywords

Diffusion displacement experiment injectivity mixed gas injection coal seam

Introduction

Previous researches on enhanced coalbed methane recovery are mainly composed of analysis on adsorption characteristics of sorbed gas to coal particles and gas diffusion in coal seam (Shi and Durucan, 2008; Sung et al., 2000; Wei et al., 2007). When injected gas flows into micropore in matrix through cleat by diffusional flow, the matrix swells depending on the magnitude of adsorption affinity onto coal particles, and this causes reduction of cleat permeability (Kim et al., 2012). Shi and Durucan (2005) proposed an improved model on Gray’s model first quantifying the permeability change caused by matrix swelling in accordance with gas adsorption, and Palmer and Mansoori model most commonly used to describe the permeability of coal seam in coalbed methane. They intended to describe the permeability change more accurately by applying the new dynamic permeability model consisted through a direct link between the amount of desorbed gas and volumetric matrix strain. CO₂ has the highest adsorption affinity among CO₂, CH₄, and N₂, which are adsorbed gas to coal, but CO₂ injectivity is decreased by the highest reduction in cleat permeability resulting from matrix swelling (Busch and Gensterblum, 2011; Chen et al., 2009; Guo et al., 2008). Thus, mixed gas injection that CO₂ to be injected is mixed with N₂ having the lowest adsorption affinity is proposed as a way to supplement the problem of efficiency reduction in CO₂ injection (Seto et al., 2006; Zhou et al., 2011; Zhu et al., 2002). Lin et al. (2008) conducted experiments of permeability alteration by injecting pure CO₂, CH₄, and N₂ gas, and various mixing ratios of CO₂–N₂ mixed gas using crushed coal sample. They ascertained that the higher content of N₂ in the gas mixture, the less permeability reduction occurred through calculating the permeability reduction by adsorption with pressure increase. In the research conducted by Kumar et al. (2012, 2015) and Wang et al. (2012), they intended to investigate permeability evolution more accurately in sorbing materials of coal and shale through observing permeability change by gas adsorption and external stress. Shale is also relevant to sorbing media with methane as adsorbed state like coal. Thus, according to the research by Kumar et al. (2010, 2016), shale is a promising area because of geological sequestration of carbon dioxide, but permeability reduction is very small than that of coal due to adsorption characteristic of matrix.

The analysis of diffusional flow as well as characteristic of CO₂ adsorption to coal is extremely important to enhance methane recovery (Kwon et al., 2001). The diffusional flow in matrix of coal seam is largely influenced by gas molecular size, pore geometry, and pore diameter. In particular, the process of methane production can be influenced by the pore structure (Sun et al., 2015). Shi and Durucan (2003) tried to interpret the flow mechanism of methane by CO₂ injection through CO₂–CH₄ displacement flow modeling applying bidisperse model composed of micropore and macropore. They demonstrated that uniform pore structure of unipore model did not accurately describe the behavior of gas diffusion in matrix of coal seam. Fathi and Akkutlu (2008) observed that the surface diffusion in the micropore is important mechanism in ECBM by injecting CO₂ through numerical modeling on counter-diffusion and competitive adsorption with methane during CO₂ injection by separating the pore within coal seam into micropore, macropore, and fracture. In addition, Cui et al. (2004) proposed that for CO₂, it is possible to diffuse more fine pore in a complex system of micropores in coal seam because kinetic diameter affecting the diffusional flow of gas molecules is 0.33 nm, it is relatively small compared to 0.36 nm for N₂ and 0.38 nm for methane.

Jessen et al. (2008) analyzed the effect of enhanced methane recovery by N₂ injection using crushed coal. They compared results in experiment of displacement flow with those of simulation using pure CO₂ and CO₂–N₂ mixture. Mazumder et al. (2008) conducted displacement experiment by using coal core sample in order to preserve the characteristics of heterogeneity and dual porosity in coal. They tried to simulate the process of methane displacement more actually, when pure CO₂ or CO₂–N₂ mixture was injected through the analysis reflecting effects of competitive adsorption, multiphase flow, and geomechanics. Schepers et al. (2010) analyzed the aspect of displacement flow caused by difference in adsorption affinity for each gas composition during CO₂–N₂ mixed gas injection in coal seam and aimed to calculate the optimum mixed ratio of injection gas for different coal rank.

As described above, CO₂ can be displaced with methane effectively due to its high adsorption affinity has been proved through several experiments and simulation studies as previous works. The injection of CO₂–N₂ mixed gas is presented to supplement permeability reduction of cleat due to high adsorption affinity of CO₂ as well. However, very few studies have been conducted to apply the effect of diffusion characteristics of each sorbed gas on displacement flow behavior experimentally.

Accordingly, in this study, the experimental apparatus was set up to analyze enhanced methane recovery as shown in Figure 1, and we performed experiments of diffusional flow using CO₂, CH₄, and N₂. The analysis of methane recovery was executed through experiments on permeability change and methane displacement for various injection condition of pure CO₂ and N₂, and CO₂–N₂ mixture. Ultimately, we investigated the effect of CO₂–N₂ injection considering gas diffusion phenomenon on enhancement of methane recovery.

Figure 1.

Experimental apparatus for analysis of enhanced methane recovery.

Coal sample preparation

In this study, coal sample obtained from East Kalimantan in Indonesia was used, and results of proximate analysis and measured maximum adsorption volume (V_L) are summarized in Table 1. The inherent moisture was measured relatively low as 0.88%, and fixed carbon and volatile matter used as classification index of coal rank showed 45.92% and 44.70%, respectively. The components shown in Table 1 converted into dry ash free (DAF) basis were applied to classification table according to the coal rank presented in Stach et al. (1982). As a result, it was estimated that the rank of coal sample corresponded to high volatile bituminous coal. In addition, a lump of coal acquired was processed into a cylindrical core with 10.83 cm in length and 3.81 cm in diameter to enable the analysis in the preserved state of pore structure. The same coal was ground to a particle size of approximately 60 meshes (mean particle size less than 0.25 mm), so that we intended to only analyze adsorption affinity of each sorbed gas except the effect of diffusion in experiment for adsorption isotherm.

Table 1.

The results of proximate analysis and measured Langmuir constant (V_L).

Parameters		Values
Inherent moisture (wt%)		0.88
Ash (wt%)		9.38
Volatile matter (wt%)		44.70
Fixed carbon (wt%)		45.92
Rank		High volatile bituminous coal
	CO₂	2,224
Maximum adsorption volume (SCF/ton)	CH₄	1,213
	N₂	485

Experimental apparatus and procedures

Diffusional flow

Experimental conditions for diffusion coefficient are summarized in Table 2. Experiment was conducted at 15℃ of temperature using a water bath circulator, and measured results were recorded by data acquisition system (DAS). The transient-flow analysis method introduced in the previous paper by the authors of this study was applied to calculate the diffusion coefficient (Seomoon et al., 2015). In this experiment, after CO₂, CH₄ or N₂ gas was adsorbed and stabilized at 3.45 MPa of pressure, non-adsorbed gas of helium was injected at the same pressure. The pressure inside the coal sample was maintained at 3.45 MPa by connecting back pressure regulator (BPR) to core holder outlet. At first, helium was injected at a pressure that was about 0.03 MPa higher than the saturated pressure of 3.45 MPa to generate gas flow, but the pressure was immediately adjusted to system pressure. Thus, only diffusional flow occurred inside the coal sample through concentration gradient by gas desorption without changes of system pressure (P_s). The partial pressure of desorbed gas by helium injection at the boundary (P_b) was continuously calculated using gas mole fraction (N) to analyze diffusional flow in the coal sample. The partial pressure of desorbed gas at the boundary was calculated by multiplying the gas mole fraction and system pressure. The concentration of the discharged gas via diffusional flow was measured to obtain the gas mole fraction using gas chromatograph (GC). The diffusion coefficients were then calculated through matching the measured partial pressure of gas at each step with calculated partial pressure of gas by numerical model based on one-dimensional transient flow equation by Fick’s second law.

\frac{\partial P_{P}}{\partial t} = D (Pp) \frac{\partial 2 P_{P}}{\partial x 2}

(1)

where P_P is the gas partial pressure; D(P_p) is the gas diffusion coefficient.

Table 2.

Experimental conditions for diffusion coefficient using coal core.

Parameter	Value	Parameter	Value
Temperature (℃)	15	Core length (cm)	10.83
Confining pressure (MPa)	5.52	Core diameter (cm)	3.81
System pressure (MPa)	3.45	Injection gas	He
Effective stress (MPa)	2.07	Saturated gas	CO₂, CH₄, N₂

Additionally, initial and boundary conditions applied to the modeling are as follows:

IC : t = 0 {, P}_{P} (x, t) = P_{i}, 0 \leq x \leq L

(2)

BC : t > 0, P_{P} (x, t) |_{x = L} = P_{b} (t), P_{0}^{k + 1} = P_{1}^{k + 1}

(3)

where P_p is the gas partial pressure at the each cell, L is the coal-core length, t is the elapsed time, and k is the number of iterations.

Gas displacement

Also, an experimental system of displacement flow is the same as the diffusional flow test. The experimental conditions are summarized in Table 3. The coal sample was evacuated completely and then saturated with pure methane at experimental pressure. In this experiment, the CO₂, N₂ or CO₂–N₂ mixture was injected at a constant rate after methane was adsorbed and stabilized at 2.07 MPa of pressure. Experiments were conducted with the following gas compositions: 100% N₂, 100% CO₂, 25:75 CO₂:N₂, 50:50 CO₂:N₂, 75:25 CO₂:N₂, respectively. The binary or ternary gases are detected concurrently in displacement experiment. Thus, the composition of discharged gas is measured by GC quantitatively. By the injection of CO₂ gas, mole fraction of displaced methane by CO₂ was measured using GC with various CO₂ injection rates. For CO₂–N₂ mixed gas injection, methane recovery was obtained to pore volumes injected (PVI) for different mixing ratios, and this was compared to the result by injection of pure CO₂ or N₂ gas.

Table 3.

Experimental conditions for methane displacement by gas injection.

Parameter	Value	Parameter	Value
Temperature (℃)	15	Coal permeability (10⁻⁹ m²)	8.37
Confining pressure (MPa)	4.14	CO₂ injection rate (cm³/min)	0.2, 0.5, 0.75, 1.0, 2.0
System pressure (MPa)	2.07	Injection gas	CO₂, N₂, CO₂–N₂
Coal pore volume (cm³)	30.86

Coal permeability

Permeability alteration was observed for each sorbed gas at various pore pressures with the experimental conditions given in Table 4. The gas compositions of the test were the same as in the displacement experiment. Experimental temperature was constant at 15℃, and a BPR was connected to adjust discharged gas pressure. Experiments were conducted to analyze the permeability reduction by gas adsorption using various compositions of sorbed gas with increasing pore pressure. In addition, to exclude the external stress on the coal sample and consider only effect of the permeability reduction by gas adsorption, the effective stress was always maintained at 2.07 MPa. A wet gas meter was connected to measure the gas flow rate at the outlet of core holder. The pressure difference across the coal sample is generated by inlet pressure (p₁) from gas regulator and outlet pressure (p₂) from BPR. The permeability was calculated using sorbed gases by measuring the pressure drop across the coal sample at a series of increasing pore pressures. The permeability of the coal sample to gas was calculated through steady-state flow method according to the Darcy’s law as below:

k_{g} = \frac{2 \times 10 - 6 μ_{g} q_{a} p_{a} L}{A [(p_{1} + p_{2}) (p_{1} - p_{2})]}

(4)

where k_g is the coal permeability of gas in m²; μ_g is the gas viscosity in cp; q_g is the gas flow rate at atmosphere in cm³/s; P_a is the standard atmospheric pressure in atm; L is the length of the coal in cm; A is the cross sectional area of the coal in cm²; P₁ and P₂ are the inlet and outlet pressure of the core-holder in atm, respectively.

Table 4.

Experimental conditions of permeability change for coal core.

Parameter	Value	Parameter	Value
Temperature (℃)	15	Core length (cm)	10.83
Effective stress (MPa)	2.07	Injection gas	CO₂, CH₄, N₂, CO₂–N₂
Core diameter (cm)	3.81

Results and discussion

Effect of CO₂ injection rate on diffusional flow

Calculation of diffusion coefficient

Helium, a non-adsorbing gas, was injected to the coal sample saturated with sorbed gas including CO₂, CH₄, or N₂ at 3.45 MPa of pressure_. Helium was injected at the constant pressure of 3.45 MPa through the inlet. At the outlet that means the boundary condition in the numerical model, the measured gas partial pressure was set to the last cell. The calculated optimum diffusion coefficient was determined by numerical matching of the estimated gas partial pressure in the last cell, which was calculated with a random diffusion coefficient, and the measured gas partial pressure (Seomoon et al., 2015).

Figure 2 shows the results of diffusion coefficient with increase of the mean partial pressure for each sorbed gas. Li et al. (2013) and Zhao (1991) showed that diffusion coefficients were increased non-linearly with increase in gas partial pressure due to change in pore size caused by gas adsorption and desorption using CH₄ and CO₂, respectively. The results of diffusional flow experiment using CO₂, CH₄, and N₂ in this study were of similar pattern to those of the references above mentioned as to change in diffusion coefficient. Because gas diffusion in coal seam has specificity which is caused by sorption-induced concentration gradient into coal particles of matrix, the increasing rate of diffusion coefficient for CO₂ was calculated as 15.53% which was about 1.7 times higher than those of CH₄ and N₂. The diffusion coefficient, however, has a certain value determined by gas molecular size and pore structure. Although CO₂ injection rate is increased, diffusion velocity is limited, thus remaining CO₂ is moved into cleat and it causes early CO₂ breakthrough (BT). Accordingly, experiments were performed judging from that change of diffusion coefficient affects CH₄ recovery which is calculated differently depending on CO₂ injection rate.

Figure 2.

The results of diffusion coefficient with mean partial pressure of CO₂, CH₄, and N₂.

Diffusional flow in CO₂-ECBM

As shown in Figure 3, for CO₂ injection at a rate 0.2 and 0.5 cm³/min, CO₂ BT occurred about 4.6 and 2.1 PVI, respectively. This is because adsorption proceeded to micropore due to CO₂ diffusion into the inside of matrix by low CO₂ injection rate. As a result of increased CO₂ injection rate at a 0.75 cm³/min, CO₂ BT was shown at approximately 1.77 PVI, and it can be identified that CO₂ BT was accelerated about 0.33 PVI than that of 0.5 cm³/min. It is judged that because CO₂ injection rate was relatively higher than that of diffusion into the matrix, early CO₂ BT occurred. Next, a displacement result from CO₂ injection rate at 1.0 cm³/min was compared to those at previous injection rates. In this case, CO₂ BT was generated about 1.18 PVI, and it represents that CO₂ BT was faster about 0.59 PVI than that of 0.75 cm³/min. In addition, for significantly increased CO₂ injection rate of 2.0 cm³/min, CO₂ BT occurred at about 0.6 PVI, and it tends to be even faster ahead with increasing CO₂ injection rate. Therefore, it is necessary to consider diffusional flow effect in estimating the optimum injection rate.

Figure 3.

The results of methane displacement behavior for various CO₂ injection rates.

Figure 4 shows the methane recovery with increase of PVI depending on CO₂ injection rate. When CO₂ was injected at 0.2 cm³/min rate, most of CO₂ were diffused to internal micropore, and displacement with CH₄ was delayed until approximately 4.5 PVI. Thus, it took a long time to recover methane initially and exhibited extremely low displacement efficiency. In addition, for 0.5 cm³/min, it shows a tendency that quite a little amount of methane was recovered until about 0.59 PVI due to the delayed displacement phenomenon as well. Comparing the results at 0.75 cm³/min and 1.0 cm³/min, it can be confirmed that displacement efficiency was reduced at 1.0 cm³/min compared to 0.75 cm³/min due to the intensified CO₂ BT. In particular, for 2.0 cm³/min, increasing rate of CH₄ recovery appears to be conspicuously reduced in the early injection period than that of 0.75 or 1.0 cm³/min. Thus, an appropriate injection rate is determined as 0.75 cm³/min considering methane recovery by injection of pure CO₂ gas.

Figure 4.

The results of methane recovery for various CO₂ injection rates.

Methane displacement behavior considering diffusion characteristics

Delayed displacement by pure CO₂ injection

Jessen et al. (2008) performed the displacement experiments using crushed coal sample, and the results were compared to those of the simulation. In this result, a very fast diffusion occurred due to the destroyed pore structure by using crushed coal. Also, Mazumder et al. (2008) conducted experiments for displacement analysis by the differences of adsorption affinity for each gas and moisture content using coal core, but they did not consider diffusional flow in coal core. In this study, for CO₂, delayed displacement with CH₄ was confirmed due to fast diffusion into the micropore and high adsorption affinity considering the diffusion phenomenon caused by concentration gradient in the preserved pore structure of the core sample. As a way to compensate the defect, through experiment of mixed gas injection including N₂ which is slow diffusion than CO₂ and has also low adsorption affinity, it was intended to compare with the result of pure CO₂ gas injection. Therefore, CO₂ BT time and methane recovery were compared on the basis of the injection rate of 0.5 cm³/min that generated delayed methane recovery by injection of pure CO₂ gas.

Diffusional flow in CO₂–N₂ ECBM

Figure 5 shows displacement behavior of pure CO₂ and N₂, and various mixing ratios of CO₂–N₂ mixture at a rate of 0.5 cm³/min. For pure N₂ gas injection, N₂ BT occurred at about 1.48 PVI, and it can be ascertained that N₂ increased sharply after that point. This is because co-diffusion with methane of N₂ was dominant rather than displacement with methane, and it acted as a displacing agent of methane. In the analysis result of methane displacement behavior using pure CO₂ with the same injection rate, CO₂ BT occurred at 2.07 PVI, and it is verified that effective displacement of methane by counter-diffusion between CO₂ and methane was possible. This means that CH₄ recovery by pure N₂ injection is a slow process, and thus, it indicates that a large amount of N₂ is to be injected as compared to CO₂. In addition, in the case of pure N₂ injection, additional cost is needed to separate CH₄ and N₂ due to extremely fast BT of N₂. Next, it is examined for the result of displacement experiment by using 75:25 CO₂–N₂ of high purity mixture. In this case, CO₂ BT occurred at 2.2 PVI of mixed gas, and it is confirmed that CO₂ BT time was delayed about 0.1 PVI than that of pure CO₂ injection. Therefore, the CO₂ BT time was somewhat delayed by enhanced methane production through N₂ component included in mixture. For 50:50 CO₂–N₂, CO₂ BT time occurred at about 2.95 PVI, and it is found that delayed effect of CO₂ BT happened by reduction of CO₂ ratio in mixed gas. Meanwhile, for 25:75 CO₂–N₂ as a rate of 0.5 cm³/min, it is identified that the increasing amount of N₂ generated N₂ BT rapidly, and it caused further ahead the time as displacing agent one of the important roles of N₂. Furthermore, CO₂ BT occurred at about 2.80 PVI of mixed gas, and it was delayed about 0.7 PVI than pure CO₂ injection under the same condition. From this result, it is elucidated that co-diffusion with methane caused by low adsorption affinity of N₂ in ECBM acts as a factor of enhanced displacement efficiency.

Figure 5.

The results of CH₄ displacement behavior by injecting mixed gas. (a) N₂ 100%, (b) CO₂ 100%, (c) CO₂:N₂ = 75:25, (d) CO₂:N₂ = 50:50, (e) CO₂:N₂ = 25:75.

Improvement of CO₂ injectivity and methane recovery

Permeability reduction considering gas diffusion

This study was based on a permeability change in the Palmer and Mansoori model represented by the sum of pore and sorption strain components (Palmer and Mansoori, 1998).

\frac{φ}{φ_{0}} = 1 + C_{f} (P - P_{o}) + \frac{1}{φ_{0}} (1 - \frac{K}{M}) (ɛ - ɛ_{0})

(5)

where Φ is the natural fracture porosity; Φ₀ is the porosity at initial reservoir pressure; c_f is cleat compressibility in MPa⁻¹; P is the reservoir pressure in MPa; P₀ is the initial reservoir pressure in MPa; K is the bulk modulus in MPa; M is the constrained axial modulus in MPa; ɛ is the volume strain; ɛ₀ is the volume strain at initial reservoir pressure.

According to the previous research, it is known that the reduction in permeability halts at a critical pressure corresponding to the point at which maximum adsorption is achieved and then increase as a consequence of diminishing effective stress (Kumar et al., 2010, 2016; Wang et al., 2012). In this study, analysis of permeability change was based on the Palmer and Mansoori model which consists of the sum of pore strain and sorption strain, but experiments were conducted under same effective stress to analyze the permeability change by only sorption strain considering diffusion phenomenon in coal. Therefore, because the effect of external stress is excluded, permeability is reduced continuously with increasing pore pressure until it reaches the maximum amount of adsorption, and after that, the permeability change did not occur due to constant effective stress. From the experimental results shown in Figure 6, the drastic permeability reduction at the early time of injection is induced by diffusion characteristic shown in Figure 2. In other words, because fast increase in diffusion coefficient increases adsorption rate, it results in a drastic permeability reduction. In addition, the permeability at the maximum adsorption pressure is ultimately determined by the difference in adsorption affinity of each gas. For the case of injecting N₂, movement of gas molecules caused by diffusion is slower than that of CO₂, and early BT occurs due to low adsorption affinity. It leads to the result that permeability decreases gradually. For CO₂–N₂ mixed gas injection, adsorbed N₂ is desorbed easily due to the high sorption capacity of CO₂. Accordingly, permeability change appears somewhat irregularly, especially for 25% CO₂. For injecting mixed gas including N₂, permeability reduction in the experimental pressure range were shown as 73.2% for 75% CO₂, 70.7% for 25% CO₂. It was improved as compared to permeability reduction as 76.5% for injecting pure CO₂ case. It is confirmed that N₂ delays CO₂ BT and acts as a factor for inhibition in permeability reduction simultaneously. As a result, since the increase in diffusion coefficient with gas partial pressure is different depending on diffusion properties, gas flow also occurs differently, and then it affects the aspect of permeability change and displacement flow. It can be seen that diffusional flow must be considered in permeability change and displacement flow in coal seam.

Figure 6.

The results of permeability change by sorbed gas injection.

Enhanced methane recovery by CO₂–N₂ injection

Figure 7 shows the result of methane recovery for pure CO₂ and N₂, and CO₂–N₂ mixture with various mixing ratios. For pure CO₂ injection, although CH₄ recovery increases slowly, displacement efficiency is enhanced by only CH₄ production due to complete displacement between CO₂ and CH₄. However, injectivity problem occurs near injection well owing to the high adsorption affinity of CO₂. As known from previous studies, coal is more vulnerable to the injectivity problem and reduction of displacement efficiency with CH₄ due to high matrix swelling of coal caused by sorption-induced strain in CO₂-ECBM. Also, diffusion phenomenon depending on pore structure in coal seam affects aspects of permeability change and displacement flow conclusively, diffusion properties of each sorbed gas are very important in ECBM accordingly. Therefore, this study was conducted with a focus on CO₂–N₂ mixed gas injection using the low adsorption property of N₂ considering diffusion phenomenon. As shown in Figure 7, for mixed gas injection consisted of 25%, 50%, and 75% in N₂, it tends to increase methane recovery than that of pure CO₂ injection, and N₂ serves as a displacing agent heavily for 75% N₂ than 25% N₂. Because CO₂ content is low among the compositions of injected gas, however, it reveals rather low increasing behavior of methane recovery compared to 25% N₂. Furthermore, for 50% N₂, it shows a similar increasing behavior of methane recovery for 75% N₂ during initial injection period, but difference in methane recovery occurred gradually on account of delayed displacement depending on counter-diffusion with methane which was generated by CO₂ content as increase of injection volume. Nevertheless, the recovery difference appears to be decreased progressively toward the latter. Thus, it is confirmed that the co-diffusion with methane by N₂ at the initial period of injection and the counter-diffusion by CO₂ in the latter affected methane production significantly. For pure N₂ gas injection, methane was produced through role of N₂ as a displacing agent for desorbed methane with reduction of partial pressure so that methane recovery tends to increase rapidly. Nitrogen, however, has a disadvantage that a large amount of N₂ must be injected compared to the other gases due to the increased role of co-diffusion with methane as well as the most rapid BT time of N₂. Accordingly, it is ascertained that N₂ is effective for enhanced methane recovery when it is injected as CO₂–N₂ mixture than pure gas resulting in generating a greater concentration gradient between CO₂ and methane, and facilitating counter-diffusion further.

Figure 7.

The results of methane recovery by sorbed gas injection.

Conclusions

This study provides the fundamental reason for the reduction of gas permeability and methane recovery during ECBM processes. From the results obtained from the experiments of diffusional flow, displacement flow, and gas permeability, we conclude the followings:

Because the diffusion coefficient is in general a function of gas molecular size and pore geometry of coal matrix, even if CO₂ injection rate is very high, it is no longer increased at high pressure. Therefore, a large amount of CO₂ moves through cleat causing an early CO₂ breakthrough, and correspondingly, it is necessary to consider diffusional flow effect in estimating the optimum injection rate.

When injecting a CO₂–N₂ mixed gas into coal seam, N₂ acts as a displacing agent of methane, because co-diffusion of N₂ with methane is more dominant than counter-diffusion owing to its extremely low adsorption affinity. It can be seen that such diffusional flow behavior of N₂ is important in enhanced methane recovery.

CO₂ is rapidly adsorbed due to its fast increasing rate of diffusion coefficient with pressure increase. Thus, CO₂ permeability is reduced rapidly and ultimately becomes the lowest value at the maximum adsorption pressure. Also, delayed methane recovery by fast diffusion and high adsorption affinity of CO₂ occurs accordingly. This study confirms that the CO₂–N₂ mixed gas injection is advisable compared to CO₂ injection to pursue the prevention of CO₂ injectivity reduction and enhanced methane recovery, simultaneously.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Korea Evaluation Institute of Industrial Technology grant from the Korean Government’s Ministry of Knowledge Economy (No. 10039231).

References

Busch

Gensterblum

(2011) CBM and CO₂-ECBM related sorption processes in coal: A review. International Journal of Coal Geology 87(2): 49–71.

Chen Z, Liu J, Elsworth D, et al. (2009) Investigation of CO₂ injection induced coal–gas interactions. In: 43rd U.S. rock mechanics symposium & 4th U.S.–Canada rock mechanics symposium, Asheville, NC, 28 June–1 July, pp. 1–14.

Cui

Bustin

Dipple

(2004) Selective transport of CO₂, CH₄, and N₂ in coals: Insights from modeling of experimental gas adsorption data. Fuel 83(3): 293–303.

Fathi E and Akkutlu IY (2008) Counter diffusion and competitive adsorption effects during CO₂ injection and coalbed methane production. In: SPE annual technical conference and exhibition, Denver, CO, 21–24 September, pp. 1–15.

Guo

Mannhardt

Kantzas

(2008) Laboratory investigation on the permeability of coal during primary and enhanced coalbed methane production. Journal of Canadian Petroleum Technology 47(10): 27–32.

Jessen

Tang

Kovscek

(2008) Laboratory and simulation investigation of enhanced coalbed methane recovery by gas injection. Transport in Porous Media 73(2): 141–159.

Kim

Sung

Han

(2012) The effect of aquifer condition on methane recovery for CO₂ injection in coalbed. Geosystem Engineering 15(4): 247–260.

Kumar

Elsworth

Liu

(2012) Optimizing enhanced coalbed methane recovery for unhindered production and CO₂ injectivity. International Journal of Greenhouse Gas Control 11: 86–97.

Kumar

Elsworth

Liu

(2015) Permeability evolution of propped artificial fractures in coal on injection of CO₂. Journal of Petroleum Science and Engineering 133: 695–704.

10.

Kumar H, Elsworth D, Marone C, et al. (2010) Permeability evolution of shale and coal under differential sorption of He, CH₄ and CO₂. In: American Geophysical Union Fall meeting, San Francisco, CA, 13–17 December, p. 1042.

11.

Kumar

Elsworth

Mathews

(2016) Permeability evolution in sorbing media: Analogies between organic-rich shale and coal. Geofluids 16(1): 43–55.

12.

Kwon

Ryou

Sung

(2001) Numerical modeling study for the analysis of transient flow. Korean Journal of Chemical Engineering 18(1): 88–93.

13.

Li W, Cheng Y, Wang L, et al. (2013) A study of coal swelling-controlled CO₂ diffusion processes. In: Proceedings of 13th coal operators' conference, University of Wollongong, The Australasian Institute of Mining and Metallurgy & Mine Managers Association of Australia, 14–15 February, pp. 324–330.

14.

Lin

Tang

Kovscek

(2008) Sorption-induced permeability change of coal during gas-injection processes. SPE Reservoir Evaluation & Engineering 11(4): 792–802.

15.

Mazumder

Wolf

KHAA

Hemert

(2008) Laboratory experiments on environmental friendly means to improve coalbed methane production by carbon dioxide/flue gas injection. Transport in Porous Media 75(1): 63–92.

16.

Palmer

Mansoori

(1998) How permeability depends on stress and pore pressure in coalbeds: A new model. SPE Reservoir Evaluation & Engineering 1(6): 539–544.

17.

Schepers K, Oudinot A and Ripepi N (2010) Enhanced gas recovery and CO₂ storage in coalbed-methane reservoirs: optimized injected-gas composition for mature basins of various coal rank. In: SPE international conference on CO₂ capture, storage, and utilization, New Orleans, LA, 10–12 November, pp. 1–21.

18.

Seomoon

Lee

Sung

(2015) Analysis of sorption-induced permeability reduction considering gas diffusion phenomenon in coal seam reservoir. Transport in Porous Media 108(3): 713–729.

19.

Seto CJ, Jessen K and Orr FM (2006) A four-component, two-phase flow model for CO₂ storage and enhanced coalbed methane recovery. In: SPE annual technical conference and exhibition, San Antonio, TX, 24–27 September, pp. 1–12.

20.

Shi

Durucan

(2003) A bidisperse pore diffusion model for methane displacement desorption in coal by CO₂ injection. Fuel 82(10): 1219–1229.

21.

Shi

Durucan

(2005) A model for changes in coalbed permeability during primary and enhanced methane recovery. SPE Reservoir Evaluation & Engineering 8(4): 291–300.

22.

Shi JQ and Durucan S (2008) Modelling of mixed-gas adsorption and diffusion in coalbed reservoirs. In: SPE unconventional reservoirs conference, Keystone, CO, 10–12 February, pp. 1–14.

23.

Stach

Mackowsky

MTh

Teichmûller

(1982) Stach’s Textbook of Coal Petrology, Berlin: Gebrûder Borntraeger.

24.

Sun

Zhang

Guo

(2015) A study of the formation morphology and phase equilibrium of fractured methane hydrates. Energy Exploration & Exploitation 33(5): 745–754.

25.

Sung

Huh

Ryu

(2000) Development and application of gas hydrate reservoir simulator based on depressurizing mechanism. Korean Journal of Chemical Engineering 17(3): 344–350.

26.

Wang

Elsworth

Liu

(2012) A mechanistic model for permeability evolution in fractured sorbing media. Journal of Geophysical Research 117: 1–17.

27.

Wei

Wang

Massarotto

(2007) A review on recent advances in the numerical simulation for coalbed-methane-recovery process. SPE Reservoir Evaluation & Engineering 10(6): 657–666.

28.

Zhao X (1991) An experimental study of methane diffusion in coal using transient approach. PhD Thesis, The University of Arizona.

29.

Zhou

Yao

Tang

(2011) Influence and sensitivity study of matrix shrinkage and swelling on enhanced coalbed methane production and CO₂ sequestration with mixed gas injection. Energy Exploration and Exploitation 29(6): 759–776.

30.

Zhu J, Jessen K, Kovscek AR, et al. (2002) Recovery of coalbed methane by gas injection. In: SPE/DOE improved oil recovery symposium, Tulsa, OK, April 13-17, pp. 1–15.

Analysis of methane recovery through CO 2 –N 2 mixed gas injection considering gas diffusion phenomenon in coal seam

Abstract

Keywords

Introduction

Coal sample preparation

Experimental apparatus and procedures

Diffusional flow

Gas displacement

Coal permeability

Results and discussion

Effect of CO2 injection rate on diffusional flow

Calculation of diffusion coefficient

Diffusional flow in CO2-ECBM

Methane displacement behavior considering diffusion characteristics

Delayed displacement by pure CO2 injection

Diffusional flow in CO2–N2 ECBM

Improvement of CO2 injectivity and methane recovery

Permeability reduction considering gas diffusion

Enhanced methane recovery by CO2–N2 injection

Conclusions

Footnotes

Declaration of conflicting interests

Funding

References

Effect of CO₂ injection rate on diffusional flow

Diffusional flow in CO₂-ECBM

Delayed displacement by pure CO₂ injection

Diffusional flow in CO₂–N₂ ECBM

Improvement of CO₂ injectivity and methane recovery

Enhanced methane recovery by CO₂–N₂ injection