Experimental and Numerical Simulation Study on Enhancing Gas Recovery with Impure CO₂ in Gas Reservoirs

Abstract

To achieve carbon peaking and carbon neutrality goals, using CO₂ to enhance natural gas recovery has broad application prospects. However, the potential for CO₂ to increase recovery rates remains unclear, the mechanisms are not fully understood, and the cost of purifying CO₂ is high. Therefore, studying the effects of impure CO₂ on natural gas extraction is of significant importance. This study investigated the effects of injection timing and gas composition on natural gas recovery through high-temperature, high-pressure, long-core displacement experiments. Based on the experimental results, numerical simulations of CO₂-enhanced gas recovery and sequestration were conducted, examining the impact of impurity gas concentration, injection timing, injection speed, and water saturation on recovery efficiency. The results indicate that higher impurity levels in CO₂ increase gas diffusion, reducing the effectiveness of natural gas recovery and decreasing CO₂ sequestration. Earlier injection timing improves recovery efficiency but results in a lower ultimate recovery rate. Higher injection speeds and water saturation levels both effectively enhance recovery rates.

1. Introduction

With economic and social development, carbon emissions have been increasing annually, resulting in a severe greenhouse effect. Energy-related CO₂ emissions account for 88.8% of total emissions [1]. Therefore, it is imperative to transition the energy structure from high carbon to low carbon. Natural gas is the fossil fuel with the lowest carbon emissions. Increasing its proportion in energy consumption is crucial for gradually achieving this energy transition [2]. As conventional gas reservoirs undergo pressure depletion, the remaining natural gas becomes increasingly difficult to extract through standard production methods [3]. Since the 1990s, the technique of CO₂ injection to enhance gas recovery (CO₂-EGR), proposed by Burgt et al. [4], has captured significant attention. Injecting CO₂ into reservoirs can restore formation pressure and displace residual natural gas, thereby enhancing gas recovery. Moreover, the gas storage capacity of the reservoir and the sealing effectiveness of the trap are favorable for achieving geological CO₂ sequestration. This helps to reduce carbon emissions and has significant research value and application prospects [5].

Currently, several small-scale pilot projects for CO₂ sequestration and enhanced gas recovery have been conducted. Internationally, these include the Budafa Szinfelleti project in Hungary [6], the Alberta project in Canada [7], the Otway project in Australia [8], and the North Sea K12-B project in the Netherlands [9]. In China, there are pilot projects, such as the pilot project of CO₂-enhanced coalbed methane recovery in Qinshui Basin [10] and the pilot experimental project for the joint development of the Lian-4 condensate gas reservoir and the Lian-21 high-CO₂ gas reservoir in Fushan Oilfield, Hainan Province.

Studies have demonstrated the feasibility of CO₂-EGR. Previous research on CO₂-EGR mechanisms indicates that CO₂ injection can restore reservoir pressure and extend the production period [11]. The density difference between CO₂ and CH₄ under reservoir conditions causes CO₂ to settle at the bottom of the reservoir and displace CH₄ upward, forming a favorable mobility ratio during displacement [12]. CO₂ competes with CH₄ for adsorption on the rock surface and displaces the CH₄ adsorbed on the rock surface [13,14]. CO₂ dissolving in formation water can delay CO₂ breakthrough [15]. Studies on the influencing factors of CO₂-EGR have shown that CO₂ purity affects the displacement and enhancement effects [16]. The gas diffusion coefficient significantly influences the mixing of CO₂ and CH₄. Injecting CO₂ after the gas reservoir reaches abandonment pressure helps to maximize ultimate recovery [17]. Within a certain range, higher injection rates and pressures yield higher recovery rates [18]. Low injection and high production can delay CO₂ breakthrough, resulting in higher CH₄ recovery [19]. Recovery rates in the presence of bound water saturation are higher than in its absence [20].

Overall, although some researchers have discussed the mechanisms and influencing parameters of CO₂-EGR, studies on CO₂ displacement of natural gas under high-temperature and high-pressure conditions remain limited. On the other hand, CO₂ is typically sourced from flue gas and tail gas from natural gas purification plants, which contain significant amounts of impurities such as N₂, O₂, and SO₂. The cost of purifying and separating CO₂ is high, making it important to study the effects of impure CO₂ on the recovery mechanism. This study investigates the mechanisms by which impure CO₂ enhances natural gas recovery through high-temperature and high-pressure long-core displacement experiments and numerical simulations. The aim is to optimize injection parameters to reduce costs and improve recovery efficiency.

2. High-Temperature and High-Pressure CO₂-EGR Displacement Experiments

2.1. Experimental Materials and Content

The gases used in the experiment included industrial CO₂ and industrial CH₄ (purity > 99.9%) and oxidation-absorption tail gas (28% CO₂, 67% N₂, 5% O₂) configured and supplied by Chengdu Keyuan Gas Co., Ltd., Chendu, China.

Four sets of long-core displacement experiments were designed to investigate the effects of injection timing and gas type on natural gas recovery. These experiments were conducted in a high-temperature and high-pressure long-core displacement system, utilizing six artificial fractured carbonate rock cores. The core parameters are listed in

Table 1. Carbonate rock core parameters.

SN	Core Number	Length (cm)	Permeability (mD)	Porosity (%)
1	3-22-42	4.374	2.79	2.18
2	3-19-42	4.734	1.01	1.67
3	3-19-42-b	4.504	1.16	2.24
4	3-17-42	4.342	0.981	1.85
5	4-5-127-a	3.99	0.683	2.36
6	3-19-42-a	4.852	2.76	3.02

. The permeability ranged from 0.683 to 2.79 mD, and the porosity ranged from 1.67% to 3.02%. The average permeability was 1.564 mD, average porosity was 2.222%, and total length was 26.796 cm. The cores were sorted according to the harmonic average value of permeability, and filter paper was added to the end faces of the cores to prevent end effects. The long-core displacement equipment simulated the reservoir formation temperature at 85 °C and original formation pressure at 55 MPa.

The long-core displacement experimental apparatus consisted of a displacement pump, a piston-type intermediate container, a thermostatic box, a long-core holder, a back-pressure controller, a confining pressure pump, and a flow meter, as shown in Figure 1. As the power device for the experiments, the displacement pump injected CO₂ from the intermediate container into the core by providing the corresponding pressure and flow rate. The prepared long core was placed in a corrosion-resistant rubber sleeve to protect the experimental materials and then placed in the long-core holder. The confining pressure pump provided the surrounding pressure to simulate the reservoir formation pressure. The downstream of the experimental system was equipped with a gas analysis and recording system, mainly consisting of a gas meter and a chromatograph, which were used to record the gas flow rate and gas components, respectively.

In the experiment, the effects of different injection timings and gas compositions were considered. The main steps were as follows.

First, the core was saturated with CH₄ to reach the CH₄ saturation state under original reservoir formation conditions. Multistage depletion-drive development experiments were then conducted at an original formation pressure of 55 MPa, with a pressure reduction interval of 1 MPa. Throughout the depletion process, the surrounding pressure was kept 5 MPa higher than the internal core pressure. The four experimental groups were depleted to 2, 6, and 8 MPa, followed by CO₂ injection displacement at 6 and 8 MPa and oxidation-absorption tail gas injection displacement at 2 and 8 MPa. All gases were injected at a constant speed of 0.05 mL/min. During the displacement process, a chromatograph was used to detect the components of the produced gas. The experiments ended when the CO₂ content in the produced gas reached 95%. Throughout the experiments, inlet and outlet pressures, gas production volume, outlet gas components, and CO₂ injection volume were recorded in real time. The detailed content of the experimental design is presented in

Table 2. Experimental design content.

Experimental Sequence	Influencing Factors	Experimental Content
1	Injection timing	Depletion from 55 MPa to 6 MPa, CO2 injection at 0.05 mL/min
2	Injection timing and injection components	Depletion from 55 MPa to 8 MPa, CO2 injection at 0.05 mL/min
3	Injection timing	Depletion from 55 MPa to 2 MPa, oxidation-absorption tail gas injection at 0.05 mL/min
4	Injection timing and injection components	Depletion from 55 MPa to 8 MPa, oxidation-absorption tail gas injection at 0.05 mL/min

.

2.2. Experimental Results and Analysis

2.2.1. Depletion Experiment

The recovery rate and the pressure difference between the inlet and outlet during the depletion experiment are shown in Figure 2. The horizontal axis represents the inlet pressure during the depletion process. Depletion-drive development was achieved by adjusting the outlet pressure using a control valve, with a pressure reduction gradient of 1 MPa. The pressure was reduced by 1 MPa every half hour, with the entire process taking 24.5 h. Generally, the recovery curve was nearly linear, indicating a stable depletion process. The recovery rate increased with the degree of depletion, reaching a maximum of 96.27% at 2 MPa. Because the outlet pressure was reduced by the control valve, the inlet pressure could not immediately align with the outlet pressure, creating a pressure difference within a certain range. Figure 2 (pressure variations) shows the pressure differences between the inlet and outlet during depletion, with a maximum pressure difference of 0.5 MPa. Overall, the pressure difference was relatively small, suggesting that gas can flow smoothly in the artificial fractured core.

2.2.2. CO₂ Injection for Gas Displacement—Ultimate Recovery

Figure 3 and

Table 3. CH₄ recovery rates at different injection volumes.

Experiment Name	Depletion Recovery	Ultimate Recovery	Increment
Depletion from 55 MPa to 6 MPa, CO2 injection at 0.05 mL/min	91.11%	96.07%	4.96%
Depletion from 55 MPa to 8 MPa, CO2 injection at 0.05 mL/min	82.66%	91.07%	8.41%
Depletion from 55 MPa to 2 MPa, oxidation-absorption tail gas injection at 0.05 mL/min	94.91%	96.27%	1.36%
Depletion from 55 MPa to 8 MPa, oxidation-absorption tail gas injection at 0.05 mL/min	82.10%	86.69%	4.59%

show the variation in CH₄ recovery with the volume of injected gas occupying the core pores. Generally, as the gas injection displacement progressed, the CH₄ recovery rate gradually increased. The gas injection displacement can be divided into three stages.

(1) Initial injection stage: The gas had just been injected, and the displacement effect was not yet evident. There was relatively little produced gas at the outlet, and CH₄ recovery increased slowly.

(2) Middle injection stage: As injection continued, the displacement effect became pronounced. There was stable gas production at the outlet, and CH₄ recovery increased rapidly.

(3) Late injection stage: Compared with the middle stage, although the gas production at the outlet remained on the same level, the increasing proportion of non-CH₄ gases (CO₂ and oxidation-absorption tail gas) in the core pores slowed the increase in CH₄ recovery. When the CO₂ and tail gas contents reached 95%, the CH₄ recovery rate stabilized.

Figure 3 shows that the ultimate recovery rate ranged from 86.69% to 96.27%—that is, gas injection enhanced the recovery by 1.36% to 8.41%. CO₂ injection at 6 MPa had a 3.45% lower enhancement effect than at 8 MPa but achieved a 5% higher ultimate recovery. Similarly, oxidation-absorption tail gas injection at 2 MPa had a 3.23% lower enhancement effect than at 8 MPa but achieved a 9.58% higher ultimate recovery. Furthermore, CO₂ injection at 8 MPa had a 3.82% higher enhancement effect and a 4.38% higher ultimate recovery than tail gas injection at 8 MPa. These results indicate that gas injection significantly improves CH₄ recovery. The earlier the injection timing, the more pronounced the enhancement effect. For better ultimate recovery, the highest possible depletion recovery should be achieved first during depletion-drive development. Moreover, under identical experimental conditions, pure CO₂ injection is more effective in recovery enhancement than impurity gas (oxidation-absorption tail gas) injection.

2.2.3. CO₂ Injection for Gas Displacement—Gas Components

Figure 4 shows the variation of the gas components at the outlet as a function of the injected gas volume occupying the core pores. The gas component variation can be divided into two stages.

(1) No breakthrough stage: In the initial injection phase, the injected gas had not yet flowed through the core pores and fractures to the outlet. At that time, the gas at the outlet was 100% CH₄.

(2) Continuous breakthrough stage: As the injected gas moved through the core pores and fractures to the outlet, the CO₂ and oxidation-absorption tail gas contents at the outlet gradually increased, while the CH₄ content in the core pores decreased. When the CO₂ and tail gas contents at the outlet reached 95%, the gas injection displacement experiment concluded.

Comparing the CO₂ injections at 6 and 8 MPa reveals that the breakthrough occurred 0.13 hydrocarbon pore volume (HCPV) later at 8 MPa than at 6 MPa. Similarly, oxidation-absorption tail gas injection at 8 MPa delayed the breakthrough by 0.443 HCPV from that of tail gas injection at 2 MPa. The experimental results indicate that the earlier the injection timing, the later the breakthrough. During the gas injection displacement process, more CO₂ and oxidation-absorption tail gas were sequestered as the breakthrough time was delayed.

2.2.4. CO₂ Injection for Gas Displacement—Pressure Variations

Figure 5 shows the inlet and outlet pressure variations with the volume of injected CO₂. Because of the good gas flow characteristics of the artificial fractured core, the pressure difference between the inlet and outlet fluctuated within the range of 0–0.4 MPa during the displacement process. The overall pressure changes were relatively small. When the CO₂ injection volume reached 0.5 HCPV, the pressure difference between the inlet and outlet gradually increased during gas injection displacement at 6 and 8 MPa.

3. CO₂ Injection in Long-Core Models for Enhanced Gas Recovery and Sequestration Simulation

3.1. Core Model Parameters

Based on the long-core displacement experiments, a numerical model of a long core was developed using the GEM compositional module of the numerical simulation software CMG 2021.10 (Computer Modelling Group Ltd. Alberta, Canada), considering the porosity and permeability parameters and the bound water saturation of the actual core. The model consists of nine core sections, each 54 cm long, and a total of 18,000 grids. An injection well was placed at the first grid and a production well at the last grid. Pure CH₄ was used to simulate natural gas within the model.

The core model parameter distribution is shown in Figure 6, and the core porosity parameters are listed in

Table 4. Core porosity parameters.

Core Number	Porosity	Permeability
1	2.1	0.25
2	2.8	0.31
3	3.2	0.32
4	4.3	0.42
5	3.7	0.36
6	4.0	0.39
7	7.5	0.51
8	6.0	0.47
9	3.6	0.38

. The porosity ranged from 2.1% to 7.5%, the permeability ranged from 0.25 to 0.51 mD, and the bound water saturation was set at 15%.

The diffusion behavior of the multicomponent CO₂ mixture within natural gas can lead to an early breakthrough, causing mixing, which reduces the economic value of natural gas and ultimately affects gas recovery. In addition, CO₂ dissolution in formation water delays CO₂ breakthrough, which also impacts ultimate recovery. For better accuracy, the proposed model considers both the diffusion of the CO₂ mixture in natural gas and CO₂ dissolution in water.

3.2. Mass-Conservation Equation

$${\displaystyle \frac{\partial }{\partial t}}\left({{\displaystyle \sum }}_{\beta =A,G}\varphi S_{\beta }{\rho }_{\beta }{X}_{\beta }^{\kappa }\right)+\nabla \cdot \left({{\displaystyle \sum }}_{\beta =A,G}{X}_{\beta }^{\kappa }{v}_{\beta }\right)+\nabla \cdot J_{\beta }^{\kappa }={{\displaystyle \sum }}_{\beta =A,G}{X}_{\beta }^{\kappa }{q}_{\beta }$$

where t represents time in seconds (s); β denotes the phase, with β = A, G representing the aqueous and gaseous phases, respectively; $\kappa$ denotes the mass component, with = m, c, $\omega$ representing CH₄, CO₂, and water components, respectively; $\varphi$ is porosity; $S_{\beta }$, is the saturation of phase; ${\rho }_{\beta }$ is the density of phase $\beta$, in kilograms per cubic meter (kg/m³); $x_{k\beta }$ is the mass fraction of component x in phase $\beta$; $v_{\beta }$ is the mass flow velocity of phase $\beta$, in kilograms per square meter per second (kg/m²/s); $J_{\beta }^{\kappa }$ is the mass diffusion rate of component $\kappa$ in phase $\beta$, in kilograms per square meter per second (kg/m²/s); and $q_{\beta }$ is the mass flow rate per unit volume of phase $\beta$ in the production or injection well, in kilograms per cubic meter per second (kg/m³/s).

Darcy’s law

$$v_{\beta }=-K{\displaystyle \frac{K_{r\beta }{\rho }_{\beta }}{{\mu }_{\beta }}}(\nabla p_{\beta }-{\rho }_{\beta }g)$$

where $K$ is the absolute permeability, in square meters (m²); $K_{r\beta }$ is the relative permeability of the $\beta$ phase; ${\mu }_{\beta }$ is the viscosity of the $\beta$ phase, in pascal-seconds (Pa. s); $p_{\beta }$ is the pressure of the $\beta$ phase, in pascals (Pa); and $g$ is the gravity acceleration vector, in meters per second squared (m/s²).

3.3. Impact of Injection Purity on CO₂ Displacement Efficiency and Sequestration

Given the high cost of CO₂ separation and purification, CO₂ is often captured from such sources as oxidizer boiler flue gas, oxidation-absorption tail gas, and hydrogen desulfurization tail gas. However, the gas thus acquired typically has a low purity. Thus, in this study, the effects of varying CO₂ purity on displacement efficiency and sequestration were investigated.

After dewatering, the main impurities in the three types of tail gas (oxidizer boiler flue gas, oxidation-absorption tail gas, and hydrogen desulfurization tail gas) were N₂ and O₂. There were very few harmful gases, such as H₂S and SO₂. Ignoring these harmful gases, the average concentrations of the main components were normalized, resulting in the following compositions: oxidizer boiler flue gas (13% CO₂ + 82% N₂ + 5% O₂), oxidation-absorption tail gas (28% CO₂ + 67% N₂ + 5% O₂), and hydrogen desulfurization tail gas (32% CO₂ + 68% N₂).

Following depletion to 6 MPa, CO₂ was injected at 0.2 mL/min to 1.2 times the HCPV using the three types of tail gas and pure CO₂. The variation in injection gas content within the produced gas with injection volume is shown in Figure 7, while the CH₄ content distribution for different gases is depicted in Figure 8.

As Figure 7 shows, higher impurity content led to earlier breakthrough and slower post-breakthrough increase in injection gas content in the produced gas. The CH₄ content distribution indicates that higher impurity content resulted in greater mixing due to diffusion, extending the CO₂–CH₄ mixing zone and causing earlier injection gas breakthrough.

The simulation results in Figure 8 reveal that impurities in injected CO₂ enhanced CO₂–CH₄ mixing, reducing the improvement in recovery. They also lowered the CO₂ content in the injected gas, which in turn suppressed the final CO₂ storage rate.

3.4. Impact of Injection Timing on CO₂ Displacement Efficiency and Sequestration

After the reservoir pressure depleted to 2, 4, 6, and 8 MPa, CO₂ was injected at 0.2 mL/min to 1.2 times the HCPV to study the impact on gas recovery. The variation in CO₂ content in the produced gas with injection volume, the CO₂/CH₄ density and viscosity ratios with pressure, and CO₂ content distribution are shown in Figure 9. The variations in residual gas recovery, its cumulative recovery, and CO₂ storage rate with injection volume are illustrated in Figure 10.

As shown in Figure 9, increasing the injection pressure delayed CO₂ breakthrough but accelerated CO₂ content growth after breakthrough. Under higher pressure, CO₂ displayed its supercritical characteristics, which increased the CO₂/CH₄ density and viscosity ratios. The growing physical property difference shortened the CO₂–CH₄ mixing zone, delayed CO₂ breakthrough, and slowed down the breakthrough speed thereafter.

Figure 10 shows that when gas was injected at reservoir pressures of 2, 4, 6, and 8 MPa, the higher the injection pressure, the greater the residual gas content. Furthermore, as the physical property differences between CO₂ and CH₄ increased, CO₂ acted like a piston in the displacement of CH₄. The recovery rate of residual gas increased after CO₂ injection, enhancing the recovery degree, but the ultimate recovery decreased. With higher injection pressures, CO₂ was sequestered in a supercritical state, increasing the final storage rate.

3.5. Impact of Injection Speed on CO₂ Displacement Efficiency and Sequestration

When the reservoir pressure was depleted to 6 MPa, CO₂ was injected at speeds of 0.1, 0.2, 0.4, and 0.8 mL/min to 1.2 times the HCPV to simulate the enhancement effect on gas recovery. The variation in CO₂ content in the produced gas with injection volume and CO₂ content distribution is shown in Figure 11, while the variations in residual gas recovery, its cumulative recovery, and CO₂ storage rate with injection volume are illustrated in Figure 12.

Figure 11 indicates that higher injection speeds corresponded to larger injection volumes at the time of CO₂ breakthrough. With the same injection volume, slower injection speeds caused earlier CO₂ breakthrough. The CO₂ distribution shows stronger diffusion effects at lower injection speeds, extending the CO₂–CH₄ transition zone.

As shown in Figure 12, residual gas recovery and the degree of enhanced recovery increased with injection speed, but this increase slowed significantly after the injection speed surpassed 0.2 mL/min. Higher injection speeds caused earlier CO₂ breakthrough at the production well. Therefore, the CO₂ storage rate decreased with increasing injection speed when the injection volume remained stable. The impact of injection speed on storage rate decreased after it moved beyond 0.2 mL/min.

3.6. Impact of Reservoir Bound Water Saturation on CO₂ Displacement Efficiency and Sequestration

Bound water saturation levels were set at 0, 0.15, 0.25, and 0.35. When the reservoir pressure depleted to 6 MPa, CO₂ was injected at 0.2 mL/min to 1.2 times the HCPV to simulate the enhancement of gas recovery and carbon sequestration. The variation in CO₂ content in the produced gas with injection volume and CH₄ content distribution is shown in Figure 13, while the variations in residual gas recovery, its cumulative recovery, and CO₂ storage rate with injection volume at different bound water saturation levels are illustrated in Figure 14.

Figure 13 shows that higher bound water saturation caused more CO₂ to dissolve in water. The growth of dissolved CO₂ reduced the amount of free CO₂ participating in displacement. This shortened the CO₂–CH₄ mixing zone and delayed CO₂ breakthrough.

As shown in Figure 14, because of the presence of bound water, the dissolution loss of CO₂ in water reduced its sweep efficiency during CH₄ displacement, suppressing the effectiveness of recovery enhancement. As bound water saturation increased, more CO₂ dissolved in water. This delayed CO₂ breakthrough and boosted the CO₂ storage rate.

4. Conclusions

(1) The long-core displacement experiments indicated that, as gas reservoirs undergo depletion-drive development to lower formation pressures, the degree of CH₄ recovery increases. The later the injection timing, the lower the formation pressure, the less CO₂ injection volume required to produce flow at the outlet end, and the sooner the CO₂ breakthrough, leading to a shorter stable production period. The degree of enhanced recovery with CO₂ injection decreases, but the ultimate recovery increases. Under identical experimental conditions, pure CO₂ shows better enhancement and ultimate recovery effects than impurity gases.

(2) Numerical simulations revealed that the impurity gases in the injected CO₂ increase CO₂–CH₄ mixing, reducing the degree of enhanced recovery and the CO₂ content in the injection gas. This ultimately lowers the CO₂ storage rate.

(3) With higher formation pressure during CO₂ injection, the CO₂–CH₄ density and viscosity ratios increase, which in turn delays CO₂ breakthrough. After the breakthrough, the breakthrough speed slows, enhancing the degree of recovery and ultimate recovery.

(4) Higher CO₂ injection speeds result in greater CO₂ injection volumes at breakthrough, leading to higher degrees of CH₄ recovery and CO₂ sequestration. At lower injection speeds, CO₂ diffusion is more pronounced, extending the CO₂–CH₄ mixing zone.

(5) Higher bound water content causes more CO₂ to dissolve in water, which shortens the CO₂–CH₄ mixing zone and delays CO₂ breakthrough. In this case, more CO₂ is sequestered in water, reducing the degree of CH₄ production enhancement.