Relative Permeability Characteristics of Natural Gas and CO₂ Mixtures in Matrix and Fractured Cores: An Experimental Study

Abstract

To clarify the oil–gas multiphase flow behavior of natural gas/CO₂ composite flooding in the dual-medium system of the BZ26-6 fractured reservoir, systematic oil–gas relative permeability experiments were conducted under reservoir temperature and pressure conditions. Using the steady-state method, the effects of core type, gas composition, and reservoir pressure on relative permeability behavior were investigated. The results show that the relative permeability curves are characterized by relatively high oil-phase permeability and low gas-phase permeability. Increasing the CO₂ fraction generally enhances oil mobilization and displacement efficiency, whereas the two-phase co-flow zone may reach an optimum at an intermediate CO₂ fraction, depending on the core structure. Specifically, with increasing CO₂ fraction, displacement efficiency increased from 37.05% to 43.70% in fractured metamorphic cores and from 60.74% to 64.63% in fractured carbonate cores. In contrast, decreasing reservoir pressure may induce stress-sensitive fracture compression, narrow the co-flow zone, and reduce flow capacity. Oil–gas two-phase flow behavior is strongly controlled by reservoir structure, with fractured carbonate cores exhibiting higher displacement efficiency and a wider co-flow region than fractured metamorphic cores. Within the scope of this study, a CO₂ fraction of 40% appears to be a comparatively favorable composite-gas composition when both displacement performance and gas-source economics are considered.

1. Introduction

With the continuous growth of global energy demand and the gradual depletion of conventional oil and gas resources, fractured reservoirs have attracted increasing attention as an important target for hydrocarbon exploration and development [1,2,3,4]. Fractured reservoirs are typically developed in metamorphic or carbonate buried-hill reservoirs, characterized by storage spaces composed of both matrix pores and fracture systems, forming a typical “dual-medium” structure [1]. Although such reservoirs contain abundant oil and gas resources, their strong heterogeneity and complex flow mechanisms pose significant challenges to development [5,6,7].

Among various enhanced oil recovery (EOR) techniques, gas injection (particularly CO₂ flooding) has been proven to be an effective approach for the efficient development of fractured reservoirs [8,9,10,11,12]. After being injected into the formation, CO₂ significantly improves crude oil mobility through physicochemical mechanisms such as viscosity reduction, oil swelling, and interfacial tension reduction, thereby establishing effective displacement channels [8]. In recent years, natural gas and CO₂ composite gas flooding has gradually emerged as a research focus as a novel technical route that balances displacement efficiency with gas source economy [13]. Compared with pure CO₂ flooding, composite gas flooding can maintain satisfactory displacement performance while reducing gas source costs and corrosion risks, demonstrating promising engineering application prospects.

In recent years, extensive research has been conducted by scholars worldwide on the characteristics of relative permeability during gas injection processes. Regarding the variation patterns of relative permeability curves during gas injection, Liao et al. [14] and Liu et al. [15] systematically investigated the time-dependent behavior of relative permeability and found that a rightward shift in the curves in the early injection stage indicates enhanced crude oil mobility, whereas a leftward shift in the later stage, caused by intensified gas channeling, corresponds to poorer development performance. Their study also showed that increasing the gas injection rate causes the relative permeability curves to shift leftward. At relatively low injection rates, gas sweep efficiency and microscopic displacement efficiency can be improved, whereas excessively high injection rates reduce development performance. By contrast, increasing injection pressure shifts the relative permeability curves rightward, reduces residual oil saturation, and enhances microscopic displacement efficiency, although it may also aggravate gas channeling [16]. With respect to CO₂ flooding mechanisms and relative permeability characteristics, Pi et al. [17] used microscopic flow simulation to investigate CO₂ flooding in shale reservoirs and found that the sweep efficiency associated with the swelling and viscosity-reduction effects of CO₂ reached 37.08–41.30%, exerting the strongest influence on production and yielding significant oil recovery improvement. Seyyedi et al. [18] revealed the pore-structure alteration of carbonate reservoirs during CO₂ injection and the influence of gas-rock interactions on flow properties. Okhovat et al. [19] experimentally studied the effects of CO₂-brine-rock interactions on permeability evolution. For fractured media, Yue [20] and Yang et al. [21] carried out systematic experimental studies on relative permeability characteristics in fractured carbonate reservoirs and found that water cut and curve morphology vary significantly among different fracture-network systems; the corresponding descriptive models for relative permeability curves also differ with fracture-system type. Lian et al. [22] compared relative permeability curves before and after artificial fracturing and found that, after fracturing, both oil and water relative permeability curves declined more rapidly, residual oil saturation increased, the oil-water co-flow region narrowed, and displacement efficiency decreased. In addition, increasing confining pressure led to higher irreducible water saturation, a lower iso-permeability point, and reduced displacement efficiency. Regarding dual-medium flow mechanisms, Qi [23] established an unsteady flow model for vertical fractured gas wells in dual-porosity media and systematically analyzed the effects of matrix shape and interporosity flow pattern on pressure response. Li et al. [24] found that increasing effective stress significantly reduces fracture permeability and porosity, thereby affecting interporosity flow behavior in dual-medium systems. Zhang et al. [25] numerically investigated two-phase flow in fractured media using a discrete fracture model. In studies on injection strategy and mobility control, Xu et al. [26] found that water-alternating-gas injection with a slug ratio of 1:1 can effectively suppress in-layer gas channeling and increase recovery by 34.1–50.6 percentage points relative to CO₂ flooding alone. Pudugramam [27] evaluated the effects of hysteresis and cycle-dependent relative permeability on the performance of water-alternating-gas and foam flooding. Shafiei et al. [28] systematically reviewed direct approaches for overcoming the limitations of gas injection processes. As for the application of nuclear magnetic resonance (NMR) techniques, Chen et al. [29] revealed the coupled control effects of multiple factors, including pressure, temperature, and injection volume, on displacement efficiency. Yang et al. [30] investigated the microscopic distribution of residual oil after gas flooding under different wettability conditions. Pan et al. [31] established the relationship between pore-throat structure and flow capacity in tight sandstone reservoirs. In studies on composite gas flooding, Bai et al. [32] summarized the effects of different injected gas media on oil recovery performance. Fang et al. [33] used molecular dynamics simulations to reveal the microscopic mechanisms by which CO₂/N₂ slug injection enhances oil recovery. Liu et al. [34] employed multidimensional experiments to clarify the influence of injected gas composition on flow characteristics. Recent studies have shown that dual-medium flow behavior and miscible or near-miscible gas flooding are jointly influenced by fracture conductivity, matrix supply capacity, and oil-gas interaction. However, their coupled effects on oil-gas relative permeability in natural gas/CO₂ composite flooding systems remain insufficiently understood. However, most existing studies focus on single lithologies or single controlling factors. Systematic comparative studies on two typical buried-hill fractured reservoirs, namely metamorphic and carbonate reservoirs, remain limited. In particular, the effects of natural gas/CO₂ composite-gas ratio on oil-gas two-phase flow behavior have not been thoroughly investigated [32,35]. In addition, there is still no unified understanding of the testing methods, data processing procedures, and parameter characterization for oil-gas relative permeability under dual-medium conditions, which has constrained the optimal design of gas injection development schemes for fractured reservoirs [1,36].

Based on the above considerations, this study focuses on the BZ26-6 oilfield block, a typical fractured buried-hill reservoir in the Bohai offshore area. This reservoir is scientifically important because it represents a dual-medium system in which oil–gas two-phase flow during natural gas/CO₂ composite flooding is jointly controlled by matrix pores, fracture networks, gas composition, and reservoir pressure. It is also of considerable engineering significance, as fractured buried-hill reservoirs constitute important hydrocarbon resources in offshore China but remain difficult to develop efficiently because of strong heterogeneity and complex matrix–fracture interactions. Nevertheless, systematic studies on oil–gas relative permeability in such reservoirs remain limited, particularly regarding the coupled effects of composite-gas ratio, reservoir pressure, and reservoir structure, as well as the respective roles of matrix and fracture systems. To address these issues, oil-gas relative permeability experiments were conducted on representative matrix and fractured cores under different CO₂ fractions and pressure conditions. The results provide experimental and theoretical support for optimizing natural gas/CO₂ composite flooding in the BZ26-6 fractured reservoir.

2. Experiment and Methodology

2.1. Materials

2.1.1. Core Samples

All natural cores collected from the study area were matrix-type samples, whereas the target reservoir is a typical fractured buried-hill reservoir in which flow behavior is jointly controlled by fracture-dominated flow and matrix-supported fluid supply. Therefore, to more realistically represent the dual-medium structural characteristics of the buried-hill metamorphic reservoir in the study area, artificial fractures with controllable geometric features were introduced into the natural matrix cores to prepare matrix–fracture cores representative of the weathered zone and inner zone. The matrix core samples obtained from the study area are shown in Figure 1.

(1)

Preparation of fractured metamorphic rock cores

Artificial fractures were generated in natural matrix-type metamorphic cores using a triaxial core testing system, as shown in Figure 2. During fracture creation, a triaxial loading procedure was applied. Cylindrical cores were placed in the triaxial apparatus, radial confining pressure was applied to simulate the in situ reservoir stress field, and axial load was gradually increased until a through-going fracture was induced under quasi-brittle failure conditions, thereby forming a conductive pathway analogous to that in the target reservoir. During loading, core permeability and porosity were measured at different stress levels to evaluate the influence of fracture development on matrix pore-permeability properties. Once the porosity and permeability of the fractured core reached the representative range of the target reservoir, loading was stopped, and the core was removed for CT scanning to obtain high-resolution images of its post-fracturing internal structure. Based on three-dimensional reconstruction and quantitative analysis, the spatial distribution, connectivity characteristics, and aperture evolution of the induced fractures were systematically characterized.

When the increase in permeability, the change in porosity, and the fracture aperture of the fractured core matched the petrophysical characteristics of the target reservoir, fracture generation was considered to have met the design requirements and the process was terminated. The resulting matrix–fracture cores had controllable apertures and good repeatability. Their representativeness was further evaluated by CT scanning and three-dimensional digital core reconstruction, through which fracture geometry, connectivity and aperture distribution were quantitatively characterized. Thus, the core design considered not only bulk porosity and permeability, but also the respective roles of matrix storage and fracture-dominated flow. However, laboratory-generated fractures cannot fully reproduce the geometric complexity and multiscale connectivity of natural fracture networks in the field.

Taking metamorphic core sample FR-1 as an example, axial stress was applied step-wise using a triaxial core testing system, while permeability and porosity were monitored in real time at different stress stages to track pore-structure evolution during fracture creation. When gas permeability increased to 8.22 mD and porosity reached 6.51%, the petro-physical properties fell within the representative ranges of the target reservoir, namely an average porosity of 4.9–6.4% and a permeability of 5.1–10.6 mD. This indicates that the fracture connectivity and pore–permeability characteristics generated in the core can reasonably represent the actual features of the target reservoir. The fractured core is shown in Figure 3.

After fracture creation, the core was subjected to CT scanning, and a three-dimensional digital core model was reconstructed based on the scanning data to quantitatively characterize the fracture structure, connectivity pattern, and pore development features [37,38]. The CT images were acquired using a MicroXCT-400 scanner (Zeiss, Pleasanton, CA, USA) and reconstructed into a three-dimensional digital core with a voxel size of 20 μm. Digital core reconstruction and visualization were performed using Avizo software 2020.1 (Thermo Fisher Scientific, Waltham, MA, USA), and the pore space, rock skeleton, and dense mineral phases were segmented using the watershed method. As shown in Figure 4a, a high-angle fracture with strong connectivity was formed inside the core, and its geometric morphology is consistent with the typical fracture features developed in the weathered zone of the reservoir. Further calculations of pore-structure parameters based on the digital core, as shown in Figure 4b, indicate that the average fracture aperture is approximately 62.2 μm, which falls within the actual fracture-aperture range of the study area (44–179 μm). This suggests that the fracture-aperture scale of the prepared core is reasonably representative of the target reservoir at the core scale. Nevertheless, laboratory-generated fractures cannot fully reproduce the spatial variability, geometric complexity, and multiscale connectivity of reservoir-scale fracture systems. Therefore, the prepared fractured cores should be understood as representative laboratory samples for investigating fracture-controlled flow behavior and matrix-fracture interaction under controlled conditions.

Based on the above fracture-creation procedure, two natural matrix cores were artificially fractured, and their petrophysical properties are summarized in

Table 1. Basic physical parameters of the core used in oil-gas relative permeability.

Core Type	Number	Diameter/cm	Length/cm	Porosity/%	Permeability/mD	Fracture Aperture/μm
Matrix-type metamorphic rock	BZ-1	2.500	4.872	4.6	0.0185	—
	BZ-2	2.500	4.804	3.8	0.0155	—
Fractured metamorphic rock	FR-1	2.500	4.290	6.51	8.22	62.20
	FR-2	2.532	4.702	5.85	8.45	70.35
Fractured carbonate rock	FC-1	2.450	6.470	7.35	8.36	100.35

.

(2)

Preparation of fractured carbonate rock core

Because carbonate core sampling from the study area is difficult, the carbonate samples used in this study were obtained from carbonate rock specimens subjected to triaxial compression combined with real-time high-temperature treatment [39]. The cores were cleaned for oil removal in accordance with the national standard GB/T 29172-2012 Core Analysis Methods [40], and their porosity and permeability were measured after drying. Based on the petrophysical screening results, cores with appropriate porosity and permeability were selected for CT scanning. When their petrophysical characteristics matched the reservoir conditions of the study area, they were used as representative samples for subsequent oil–gas relative permeability experiments. Sample FC-1 and its CT scanning results are shown in Figure 5. The basic petrophysical properties are summarized in Table 1. Although these substitute carbonate cores were selected to match the petrophysical and fracture characteristics of the target reservoir, potential differences in mineralogical composition, pore-surface properties, and wettability from native reservoir carbonate rocks may still influence the measured relative permeability behavior. Therefore, the results obtained from these samples should be interpreted primarily as reflecting the effects of fracture structure and dual-medium flow characteristics.

2.1.2. Preparation and Properties of Fluids

According to the formation water analysis report of the target reservoir, synthetic formation water (hereinafter referred to as water) was prepared, with a measured viscosity of 0.32 mPa·s and a density of 0.80 g/cm³. To match the viscosity of the reservoir crude oil in the target reservoir, surface-degassed crude oil collected from the same oilfield was mixed with kerosene at different proportions to prepare synthetic formation oil (hereinafter referred to as oil), with a measured viscosity of 0.345 mPa·s and a density of 0.65 g/cm³. In addition, the injected composite gas used in this study was prepared by mixing field natural gas and high-purity CO₂ according to preset molar fractions for the subsequent displacement experiments. The base natural gas was taken from the first-stage separator gas of the target reservoir, with CH₄ and CO₂ as the dominant components, together with minor amounts of N₂ and C₂–C₆ hydrocarbons, as summarized in

Table 2. Composition of first-stage separator gas.

Component	CO2	N2	CH4	C2H6	C3H8	C4H10	C5H12	C6H14
Mole fraction yi (%)	15.85	0.36	70.35	8.60	2.38	1.30	1.11	0.05

.

2.2. Experimental Setup and Procedure

2.2.1. Steady-State Method for Relative Permeability Measurement

The experiments were conducted under simulated reservoir conditions. The experimental temperature was set at 145 °C and controlled using an ultra-high-temperature oven and a heating jacket installed on the core holder, while the pressure was set at the average reservoir pressure of 36 MPa. The experimental procedure followed the steady-state method for oil-gas relative permeability measurements, as illustrated in Figure 6. The experiments were carried out with reference to the national standard GB/T 28912-2012, Methods for Determining Relative Permeability of Two-Phase Fluids in Rock [36].

A total of 12 experimental runs were designed to investigate the effects of core type, natural gas and CO₂ composite-gas ratio, and pressure conditions on oil-gas two-phase flow behavior. For each experimental run, the oil and gas injection rates were not fixed as single values. Instead, five steady-state measurement points were established by stepwise adjusting the oil–gas injection ratio according to the steady-state relative permeability procedure. At each measurement point, the oil injection rate was gradually increased, and the gas injection rate was correspondingly adjusted to obtain different saturation states before the pressure drop, oil and gas flow rates, and core mass were recorded. The pressure-dependent relative permeability experiments were conducted under monotonic pressure conditions. No cyclic loading-unloading tests were performed; therefore, the reversibility or hysteresis of fracture closure was not directly evaluated in this study. In the present study, wettability was not varied as an independent experimental factor; therefore, the measured relative permeability behavior reflects the combined effects of reservoir structure, fracture characteristics, gas composition, and pressure conditions under the existing wettability state of the core samples. The detailed procedure is summarized as follows:

• The core was evacuated to 10⁻³ Pa using a vacuum pump and then saturated with formation water under reservoir pressure until the pressure in the saturation vessel stabilized. The core was then removed and weighed.
• The saturated core was placed into a high-pressure core holder, and the temperature was brought to the preset value using the ultra-high-temperature oven and heating jacket.
• The confining pressure was maintained at 4 MPa above the pore pressure throughout the experiment. Formation water was injected into the core until stable water production was achieved, and the water-phase permeability was measured. Back pressure was gradually applied to the back-pressure valve using a back-pressure pump until the experimental pressure was reached. The injection pressure was monitored continuously, and the confining pressure was maintained above the injection pressure by controlling the confining-pressure pump, ensuring that the core remained fully saturated with formation water under the target pressure.
• Formation oil was then injected to displace the water until stable oil production was obtained at the outlet and no further water was produced. The produced water volume was recorded to calculate irreducible water saturation and initial oil saturation. Oil injection was then continued for more than 10 pore volumes to ensure full saturation with the simulated oil, followed by an aging period of 24 h. The oil-phase permeability at irreducible water saturation was measured, and the initial oil saturation was determined by the gravimetric method.
• Gas was injected to displace the oil until no further oil was produced at the outlet, and the gas-phase permeability and oil saturation at this initial gas-flooding condition were determined.
• Gas and oil were then injected simultaneously according to the steady-state relative permeability procedure. For each experimental run, five steady-state measurement points were established by changing the oil–gas injection ratio step by step. During this process, the oil injection rate was gradually increased, while the gas injection rate was correspondingly reduced or adjusted to obtain different oil and gas saturation states. At each measurement point, steady state was judged when the inlet and outlet pressures and the oil and gas flow rates remained stable during the measurement period.
• After steady state was reached at each measurement point, the pressure drop across the core, inlet and outlet pressures, oil and gas flow rates, and the mass of the oil-containing core were recorded. These data were used to calculate the gas-phase effective permeability, oil-phase effective permeability, and corresponding oil and gas saturations at different saturation states. Finally, oil was injected at a constant flow rate to determine the oil-phase permeability at residual gas saturation.

2.2.2. Data Processing and Calculation

The gas-phase effective permeability (K_ge) and oil-phase effective permeability (K_oe) were calculated according to the steady-state Equations (1) and (2):

$$K_{\mathrm{ge}}={\displaystyle \frac{2p_0{Q}_{0}TL\overline{z}{\mu }_{\mathrm{g}}}{T_{0}A\left({p_1}^2-{p_2}^2\right)}}$$

(1)

$$K_{\mathrm{oe}}={\displaystyle \frac{q_{\mathrm{o}}{B}_{\mathrm{o}}{\mu }_{\mathrm{o}}L}{A(p_1-p_2)}}\times {10}^{-1}$$

(2)

where P₀ is the atmospheric pressure, 10⁻¹ MPa; Q₀ is the gas volumetric flow rate at pressure P₀, cm³·s⁻¹; T is the experimental temperature, K; L is the core length, cm; Z is the gas compressibility factor, dimensionless; μ_g is the gas viscosity at the experimental temperature, mPa·s; T₀ is room temperature, K; A is the core cross-sectional area, cm²; p₁ is the inlet pressure of the core, MPa; p₂ is the outlet pressure of the core, MPa; q_o is the oil flow rate; B_o is the oil formation volume factor, dimensionless; and μ_o is the oil viscosity at the experimental temperature, mPa·s.

The gas compressibility factor Z was calculated using NIST REFPROP software (Version 10.0) based on the temperature, pressure, and composition of the mixed gas under each experimental condition. The software employs the GERG-2008 equation of state, which provides accurate thermodynamic properties for natural gas-CO₂ mixtures over a wide range of temperatures and pressures.

The gas relative permeability (K_rg) and oil relative permeability (K_ro) were then calculated using Equations (3) and (4):

$$K_{\mathrm{rg}}={\displaystyle \frac{K_{\mathrm{ge}}}{K_{\mathrm{g}}\left(S_{\mathrm{wc}}\right)}}$$

(3)

$$K_{\mathrm{ro}}={\displaystyle \frac{K_{\mathrm{oe}}}{K_g\left(S_{\mathrm{wc}}\right)}}$$

(4)

where K_g(S_wc) is the gas effective permeability at irreducible water saturation, mD.

In this study, both gas and oil relative permeabilities were calculated by normalizing the corresponding effective phase permeabilities with a common reference permeability, namely the gas effective permeability measured at irreducible water saturation under the same experimental conditions, K_g(S_wc). Therefore, K_rg and K_ro represent the dimensionless ratios of K_g and K_o to K_g(S_wc), respectively, as expressed in Equations (3) and (4). This normalization provides a consistent basis for comparing oil-gas two-phase flow behavior under different core types, gas compositions, and pressure conditions.

The oil saturation (S_o) and gas saturation (S_g) were calculated using Equations (5) and (6), respectively:

$$S_o={\displaystyle \frac{m_i-m_w}{V_p{\rho }_o}}\times 100$$

(5)

$$S_g=100-S_o-S_{wc}$$

(6)

where m_i is the mass of the core sample at the i-th measurement point, g; m_w is the mass of the core containing irreducible water, g; and S_wc is the irreducible water saturation, %.

It should be noted that the calculated relative permeability values may still be affected by uncertainties associated with pressure, flow rate, saturation, and fluid-property measurements, as well as by the absence of replicate experiments. Therefore, the present results are mainly intended for comparative analysis of relative trends under different experimental conditions.

The uncertainties associated with the calculated relative permeability parameters were evaluated based on the measurement accuracy of the experimental instruments and the error-propagation principle. Because replicate experiments were not conducted under each experimental condition, the present uncertainty evaluation focuses on measurement and calculation uncertainties rather than statistical repeatability. The main sources of uncertainty considered in this study are summarized in

Table 3. Main sources of uncertainty in the relative permeability experiments.

Source of Uncertainty	Variable	Instrument/Method	Accuracy
Pressure measurement	p1, p2	Pressure transducer	±0.25% FS
Gas flow rate	Q0	Gas flow meter	±1.0% of reading
Oil flow rate	q0	Displacement pump	±0.5% of reading
Temperature	T	Temperature sensor	±0.1 °C
Core length	L	Vernier caliper	±0.01 mm
Core diameter	D	Vernier caliper	±0.01 mm
Mass measurement	mi, mw	Electronic balance	±0.001 g
Oil viscosity	μo	Viscometer	±1.0%
Gas viscosity	μg	REFPROP calculation	estimated ±0.5%
Gas compressibility factor	Z	REFPROP calculation	estimated ±0.5%

, including pressure, flow rate, temperature, core dimensions, mass measurement, fluid viscosity, gas compressibility factor, and saturation calculation. These uncertainties were propagated through Equations (1)–(6) to assess their possible influence on the calculated effective permeability, relative permeability, and saturation parameters.

Although the above uncertainties do not represent statistical deviations from repeated experiments, they provide a quantitative basis for evaluating the possible measurement-induced variation in the reported characteristic parameters. Therefore, the characteristic parameters listed in

Table 4. Characteristic parameters of oil-gas relative permeability in matrix-type metamorphic rocks.

CO2 Content	Displacement Efficiency/%	Swc/%	Krg(Sor)	Residual Oil Saturation/%	Gas Saturation at Iso-Permeability Point/%	Two-Phase Co-Flow Zone Range/%
15%	60.21	53.71	0.22	18.42	28.88	40.71
40%	63.33	52.71	0.17	17.34	29.25	42.35

,

Table 5. Characteristic parameters of oil-gas relative permeability in fractured metamorphic rocks.

CO2 Content	Displacement Efficiency/%	Swc/%	Krg(Sor)	Residual Oil Saturation/%	Gas Saturation at Iso-Permeability Point/%	Two-Phase Co-Flow Zone Range/%
15%	37.05	47.63	0.56	32.97	6.798	21.68
25%	40.21	47.55	0.46	31.36	7.040	21.93
40%	41.72	47.37	0.44	30.67	7.788	22.61
100%	43.70	47.89	0.38	29.34	11.420	22.78

,

Table 6. Characteristic parameters of oil-gas relative permeability in fractured carbonate rocks.

CO2 Content	Displacement Efficiency/%	Swc/%	Krg(Sor)	Residual Oil Saturation/%	Gas Saturation at Iso-Permeability Point/%	Two-Phase Co-Flow Zone Range/%
15%	60.74	30.31	0.528	27.36	13.86	26.23
40%	61.73	33.47	0.428	25.46	14.85	28.93
100%	64.63	37.46	0.420	22.12	24.47	25.49

,

Table 7. Characteristic parameters of oil-gas relative permeability in fractured metamorphic rocks under different pressures.

Pressure/MPa	Displacement Efficiency/%	Swc/%	Krg(Sor)	Residual Oil Saturation/%	Gas Saturation at Iso-Permeability Point/%	Two-Phase Co-Flow Zone Range/%
36	41.72	47.37	0.47	30.67	13.37	25.49
31	41.55	47.08	0.29	30.93	11.89	21.98
26	40.08	46.93	0.19	31.80	8.68	21.26

and

Table 8. Characteristic parameters of oil-gas relative permeability for different core types.

Core Type	Displacement Efficiency/%	Krg(Sor)	Residual Oil Saturation/%	Gas Saturation at Iso-Permeability Point/%	Two-Phase Co-Flow Zone Range/%
Matrix-type metamorphic rock	63.33	0.17	17.34	29.25	42.35
Fractured metamorphic rock	41.72	0.44	30.67	7.788	22.61
Fractured carbonate rock	61.73	0.428	25.46	14.85	28.93

are mainly used for comparative analysis of relative trends under different experimental conditions.

3. Results and Discussion

3.1. Effect of CO₂ Content in Composite Gas on Oil-Gas Relative Permeability

During natural gas/CO₂ composite flooding, the CO₂ fraction plays a critical role in controlling oil–gas two-phase flow behavior. Variations in CO₂ proportion can significantly alter flow parameters such as relative permeability, flow resistance, and fluid distribution characteristics. In this section, oil–gas relative permeability experiments were conducted on three types of cores under different CO₂ fractions, and the effects of CO₂ proportion on oil–gas two-phase flow parameters and displacement characteristics were systematically analyzed.

3.1.1. In Matrix Metamorphic Cores

Matrix metamorphic cores BZ-1 and BZ-2 were selected for oil-gas relative permeability experiments under reservoir conditions using natural gas and CO₂ composite gases with CO₂ fractions of 15% and 40%, respectively. The results are shown in Figure 7 and Table 4. The characteristic parameters listed in Table 3, Table 4, Table 5 and Table 6 were obtained from single-run experiments under each condition. Therefore, statistical repeatability could not be evaluated from replicate measurements. Instead, the uncertainty of the calculated parameters was assessed based on the measurement accuracy of the experimental instruments and the error-propagation principle, as described in Section 2.2.2.

The relative permeability behavior of the matrix metamorphic core is characterized by a crossover between the oil-phase and gas-phase curves and by a relatively wide two-phase co-flow region of about 40%. This indicates that the injected gas can penetrate a considerable portion of the matrix pore system and maintain two-phase flow over a relatively broad saturation interval, resulting in an oil displacement efficiency of about 60%. However, because this type of core is generally tight and dominated by fine pore-throat structures, both gas invasion and oil displacement are strongly constrained by capillary resistance and limited connectivity. As a result, the oil phase retains a relatively high flow capacity over most of the saturation range, whereas the gas phase develops conductivity only gradually, leading to relatively high residual oil saturation and low overall gas relative permeability. In the low- to intermediate-gas-saturation range, the much lower gas relative permeability indicates that gas mobility is strongly restricted in the matrix pore network. Even at residual oil saturation, gas relative permeability remains only about 0.2, further confirming that matrix flow is still controlled by fine pore-throat restriction rather than by highly conductive gas channels.

As the CO₂ content in the composite gas increases from 15% to 40%, the oil-gas two-phase flow behavior becomes more favorable for displacement. This improvement is reflected not only by the rightward shift in the relative permeability curves, but also by the expansion of the co-flow region from 40.71% to 42.35%, the reduction in residual oil saturation from 18.42% to 17.34%, and the increase in displacement efficiency from 60.21% to 63.33%. These changes indicate that a higher CO₂ fraction enhances crude-oil viscosity reduction, oil swelling, and oil-gas interaction, while lowering the interfacial tension and capillary restriction within the fine pore-throat system. Consequently, more residual oil trapped in small pores and throats can be mobilized, and the gas phase can establish effective flow pathways over a wider saturation range. Therefore, increasing CO₂ content improves not only microscopic oil mobilization but also the coordination of oil-gas two-phase flow in matrix-type metamorphic cores.

3.1.2. In Fractured Metamorphic Cores

The prepared fractured metamorphic core FR-1 was selected for oil-gas relative permeability experiments under reservoir conditions using natural gas and CO₂ composite gases with CO₂ fractions of 15%, 25%, 40%, and 100%, respectively. The results are shown in Figure 8 and Table 5.

As shown in Figure 9, the relative permeability behavior of the fractured metamorphic core is controlled by the strong conductivity contrast between the fracture system and the tight matrix. Compared with the matrix core, the two-phase co-flow region is much narrower, at about 21.68–22.78%, whereas gas relative permeability at residual oil saturation remains relatively high, ranging from 0.56 to 0.38. This indicates that once gas enters the fracture system, it preferentially migrates through highly conductive fracture pathways rather than uniformly invading the matrix pore space. As a result, gas breakthrough occurs more easily, and the matrix is swept non-uniformly, leading to insufficient oil replenishment from the matrix to the fracture channels. Therefore, although fractures enhance gas transport capacity, they also intensify gas channeling and reduce the effective oil–gas co-flow interval.

As the CO₂ fraction increases from 15% to 100%, residual oil saturation decreases from 32.97% to 29.34%, gas saturation at the iso-permeability point increases from 6.798% to 11.420%, and displacement efficiency increases from 37.05% to 43.70%. These trends indicate that higher CO₂ fractions improve crude-oil mobilization through stronger viscosity reduction, oil swelling, diffusion, and extraction effects, thereby facilitating oil transfer from the matrix toward fracture-dominated flow channels. However, the incremental improvement in displacement efficiency becomes less significant at high CO₂ fractions: when the CO₂ fraction increases from 40% to 100%, displacement efficiency rises only from 41.72% to 43.70%. This limited gain suggests a diminishing marginal benefit. Under immiscible composite-gas flooding conditions, once most movable oil has already been displaced and dominant fracture-controlled flow channels have been established, further increasing the CO₂ fraction can still enhance oil–gas interaction, but its contribution to additional recovery becomes relatively small. Meanwhile, the decrease in Krg(Sor) from 0.56 to 0.38 suggests that intrinsic gas-flow capacity is weakened by higher gas density and viscous resistance, even though overall displacement performance still improves. Therefore, the response of fractured metamorphic cores to increasing CO₂ fraction reflects the combined effect of enhanced oil mobilization and fracture-dominated channeling, rather than a simple monotonic increase in gas-flow efficiency.

3.1.3. In Fractured Carbonate Cores

To reveal the flow behavior of natural gas and CO₂ composite flooding in the inner-zone reservoir, the fractured carbonate core FC-1 was selected for oil–gas relative permeability experiments under reservoir conditions using natural gas and CO₂ composite gases with CO₂ fractions of 15%, 40%, and 100%. The results are presented in Figure 10 and Figure 11 and Table 6.

The relative permeability behavior of the fractured carbonate core reflects the combined effects of fracture-network conductivity and matrix–fracture interaction. Compared with the fractured metamorphic core, the fractured carbonate core exhibits a wider co-flow region of about 26–29% and higher displacement efficiency, indicating a more favorable balance between gas transport and oil mobilization. In the early stage of displacement, gas relative permeability increases relatively slowly, suggesting that the oil phase still occupies the main flow space and that gas flow is initially constrained by matrix storage and fracture connectivity. As gas saturation increases, gas gradually forms continuous conductive pathways within the fracture network, and gas relative permeability rises rapidly, indicating a transition toward fracture-dominated flow. The relatively wide co-flow region indicates that the fracture network in the carbonate core not only provides effective gas-flow channels but also promotes a broader sweep range and stronger matrix–fracture interaction. At the same time, gas relative permeability at residual oil saturation remains lower than that of the fractured metamorphic core, implying that although the fracture system is better connected, the combined matrix–fracture structure still imposes some restriction on pure gas flow.

As the CO₂ fraction increases from 15% to 100%, the displacement efficiency of the fractured carbonate core increases from 60.74% to 64.63%, while the residual oil saturation decreases from 27.36% to 22.12%. This indicates that increasing the CO₂ fraction enhances gas–oil interaction, including crude-oil swelling, viscosity reduction, diffusion, and extraction effects, thereby improving oil mobilization from both the matrix and fracture-associated pore spaces. However, the two-phase co-flow zone does not increase monotonically with CO₂ fraction. It expands from 26.23% at 15% CO₂ to 28.93% at 40% CO₂, but decreases to 25.49% at 100% CO₂. This suggests that a moderate CO₂ fraction is more favorable for maintaining coordinated gas–oil two-phase flow, whereas pure CO₂ flooding may enhance microscopic oil mobilization but also alter gas mobility and flow resistance, resulting in a narrower effective co-flow interval. Therefore, for fractured carbonate cores, 40% CO₂ provides the widest two-phase co-flow zone, while 100% CO₂ gives the highest displacement efficiency but with reduced co-flow stability.

Overall, fractured carbonate cores respond strongly to CO₂ composite flooding. Increasing the CO₂ fraction is beneficial for reducing residual oil saturation and improving displacement efficiency, whereas the sweep range reaches a maximum at an intermediate CO₂ fraction. However, under high-CO₂ conditions, gas-phase flow resistance also increases, indicating that the effectiveness of composite-gas flooding in this type of reservoir depends on the overall balance between enhanced crude-oil mobilization and restricted gas-phase flow.

3.2. Effect of Displacement Pressure on Oil-Gas Relative Permeability

To clarify the effect of reservoir pressure variation on oil-gas two-phase flow characteristics, the fractured metamorphic core FR-2 was selected for oil-gas relative permeability experiments at three pressure levels, 26, 31, and 36 MPa, under reservoir temperature conditions. The injected gas was a natural gas/CO₂ composite gas with a CO₂ fraction of 40%. The results are shown in Figure 12 and Figure 13 and Table 7.

As shown in Figure 12, when reservoir pressure decreased from 36 MPa to 26 MPa, both the oil and gas relative permeability curves shifted downward as a whole, and the two-phase co-flow region gradually narrowed. This indicates that pressure depletion severely impedes oil–gas two-phase transmissibility in the fractured metamorphic core. This effect is mainly because pore-pressure decline increases effective stress, resulting in mechanical compaction of fracture and pore structures, weakening of dominant flow channels, and increased oil–gas flow resistance. As a result, the flow system gradually shifts from fracture-dominated flow toward matrix-pore-dominated flow, making crude oil more likely to be trapped in fine pore throats and thereby increasing residual oil saturation.

The variations in characteristic parameters further confirm this trend. As shown in Figure 13a,b, with decreasing pressure, residual oil saturation increased from 30.67% to 31.80%, while displacement efficiency decreased from 41.72% to 40.08%, indicating that pressure reduction is unfavorable for crude-oil mobilization. This is because overall displacement efficiency represents the cumulative recovery over the entire flooding process, whereas some movable oil had already been produced during the earlier stages of displacement. Therefore, pressure depletion mainly suppresses subsequent transport and mobilization capacity rather than causing an equally large decline in total recovery.

As shown in Figure 13c, gas saturation at the iso-permeability point decreased from 13.37% to 8.68%, suggesting that with increasing effective stress, deterioration in gas-phase flow capacity becomes more pronounced, and the oil phase can maintain flow capacity comparable to that of the gas phase at a lower gas saturation. Figure 13d shows that gas relative permeability at residual oil saturation decreased from 0.47 to 0.19, corresponding to a reduction of 59.57%, indicating that the suppressive effect of pressure decline on gas-phase flow capacity is particularly significant. It should be noted that the large reduction in Krg(Sor) mainly reflects deterioration of gas-phase endpoint flow capacity under reduced pressure, whereas displacement efficiency represents cumulative recovery over the entire flooding process; therefore, the two parameters describe different aspects of displacement behavior and are not expected to vary proportionally. This difference is associated not only with the reduction in flow space caused by fracture compression but also with increased two-phase flow resistance resulting from weaker crude-oil swelling, higher oil viscosity, and increased oil–gas interfacial tension.

Although this study was not designed to determine the minimum miscibility pressure (MMP), miscibility pressure is important for interpreting the pressure-dependent displacement behavior of natural gas/CO₂ composite flooding. The experimental pressure relative to the MMP controls the extent of oil–gas interaction. When the displacement pressure approaches near-miscible conditions, CO₂-rich composite gas can more effectively promote crude-oil swelling, viscosity reduction, interfacial-tension reduction, and light-component extraction, thereby improving oil mobilization. Conversely, pressure depletion may move the system farther away from near-miscible conditions, weakening these favorable oil–gas interactions and reducing displacement efficiency. Therefore, the observed pressure effect in this study should be interpreted as the combined result of pressure-dependent oil–gas interaction and stress-sensitive fracture/pore compression. Because the MMP was not directly measured in this work, its quantitative influence on the relative permeability behavior requires further investigation through slim-tube experiments, interfacial-tension measurements, or equation-of-state-based phase-behavior calculations.

Overall, a decrease in reservoir pressure simultaneously weakens fracture-controlled flow and the displacement capacity of the composite gas, resulting in lower flow capacity, a narrower co-flow region, higher residual oil saturation, and reduced displacement efficiency. It should be noted that fracture closure and permeability evolution were not monitored directly during the relative permeability experiments. Instead, their effects were inferred from the observed changes in relative permeability characteristics and the known stress sensitivity of fractured cores. For this type of fractured reservoir, maintaining a relatively high reservoir pressure is therefore important for preserving gas-phase flow capacity, sustaining the effective sweep range, and improving the performance of natural gas/CO₂ composite flooding.

3.3. Effect of Core Type on Oil-Gas Two-Phase Flow Behavior

To compare oil-gas two-phase flow characteristics under different reservoir structural conditions, the relative permeability results of three core types obtained under identical experimental conditions were analyzed comparatively. The injected gas in all cases was a natural gas and CO₂ composite gas with a CO₂ fraction of 40%. The experimental results are shown in Figure 14 and Table 8.

The oil-gas relative permeability curves of all three core types exhibit a typical X-shaped pattern with a pronounced concave-downward form. However, their curve morphologies and characteristic parameters differ significantly, indicating that core type strongly controls oil-gas two-phase flow behavior. Among the three core types, the matrix metamorphic core shows the widest two-phase co-flow region (42.35%), the lowest residual oil saturation (17.34%), and the highest displacement efficiency (63.33%). These results indicate that in the matrix core, gas can enter the pore system more effectively, resulting in better coordinated oil–gas two-phase flow and a larger sweep range. However, because of the fine pore-throat structure and complex connectivity of this type of core, the displacement process requires relatively high flow pressure, making development more difficult.

The fractured metamorphic core exhibits distinctly different flow behavior. Its two-phase co-flow region narrows to 22.61%, residual oil saturation increases to 30.67%, and displacement efficiency decreases to 41.72%, whereas gas relative permeability at residual oil saturation increases markedly. This indicates that once fractures are formed, gas preferentially migrates along the fracture pathways. Although fracture conductivity enhances gas-phase flow capacity, it also intensifies gas channeling, weakens effective mobilization of matrix oil, and consequently reduces overall displacement efficiency.

The fractured carbonate core, similar to the fractured metamorphic core, belongs to a fracture-dominated flow system. However, owing to its larger fracture aperture, stronger connectivity, and better-developed fracture-network geometry, the effective oil–gas co-flow interval is enlarged. As shown in Table 8, the two-phase co-flow zone of the fractured carbonate core reaches 28.93%, and the displacement efficiency increases to 61.73%, which is significantly higher than that of the fractured metamorphic core and close to that of the matrix metamorphic core. These results indicate that a better-connected fracture network is favorable for enlarging the gas sweep range and enhancing matrix–fracture interaction during displacement. In contrast, the fracture system in metamorphic cores is relatively simpler and more localized, making gas flow more prone to preferential channeling rather than broad matrix–fracture sweep. Nevertheless, the residual oil saturation of the fractured carbonate core is 25.46%, which is lower than that of the fractured metamorphic core but higher than that of the matrix metamorphic core. This indicates that the better-connected fracture network improves sweep efficiency, whereas matrix oil supply remains partly restricted by pore-structure heterogeneity. Therefore, the displacement process in fractured carbonate cores is controlled by the combined effect of strong fracture conductivity and restricted matrix supply, and its improved displacement efficiency mainly reflects a wider sweep range rather than uniformly efficient mobilization of all matrix oil.

Overall, core type significantly affects oil–gas two-phase flow behavior by controlling pore structure, fracture development, and medium heterogeneity. The matrix metamorphic core shows the strongest oil–gas co-flow coordination, although its absolute flow capacity remains limited by the tight pore-throat structure. The fractured metamorphic core is dominated by fracture-controlled flow and is more prone to gas channeling, resulting in a much narrower co-flow region and lower displacement efficiency than the matrix metamorphic core. In contrast, the fractured carbonate core, owing to its larger fracture apertures, stronger connectivity, and more complex fracture-network geometry, exhibits a wider co-flow region and higher displacement efficiency than the fractured metamorphic core. However, its flow behavior is also more strongly controlled by fractures, and the influence of heterogeneity is more pronounced.

4. Conclusions

This study focused on the BZ26-6 fractured reservoir and investigated the flow behavior and displacement characteristics of natural gas/CO₂ composite flooding using laboratory physical experiments and digital-core characterization. It should be noted that the fractured core samples used in this study were artificially fractured cores prepared from natural matrix cores rather than naturally fractured reservoir cores. On this basis, the findings provide guidance for optimizing gas-injection development in the BZ oilfield and similar buried-hill reservoirs.

(1)

The CO₂ fraction in the composite gas has a significant effect on relative permeability behavior and displacement efficiency. Increasing the CO₂ fraction generally improves oil mobilization and displacement efficiency, whereas the two-phase co-flow zone may reach an optimum at an intermediate CO₂ fraction depending on core structure. Within the scope of this study, 40% CO₂ appears to be a comparatively favorable option when both flooding performance and gas-source economics are considered.

(2)

Reservoir pressure strongly controls oil-gas two-phase flow capacity in fractured reservoirs. Pressure depletion weakens the oil–gas interaction and potential near-miscible displacement capacity of the composite gas and may induce stress-sensitive compression of pore-fracture structures, resulting in reduced gas relative permeability, a narrower co-flow region, higher residual oil saturation, and lower displacement efficiency. Maintaining relatively high reservoir pressure is therefore beneficial for composite-gas flooding.

(3)

Reservoir structure is a key factor controlling flooding effectiveness. Matrix metamorphic cores show relatively strong oil–gas coordination but are difficult to develop because of their fine pore-throat structure. Fractured metamorphic cores are prone to preferential flow and gas channeling, which limit matrix oil mobilization. For fractured carbonate cores, increasing the CO₂ fraction improved displacement efficiency and reduced residual oil saturation. However, the two-phase co-flow zone reached its maximum at 40% CO₂ rather than increasing monotonically. This indicates that a moderate CO₂ fraction is more favorable for maintaining coordinated oil–gas flow, whereas pure CO₂ flooding provides stronger oil mobilization but may narrow the effective co-flow interval.

Overall, the effectiveness of natural gas/CO₂ composite flooding is jointly controlled by gas composition, reservoir pressure, and reservoir structure. It should also be noted that the present conclusions are mainly based on core-scale experiments, and further multiscale upscaling and field-scale validation are still needed.