Remaining Oil Distribution Characteristics in Sandy Conglomerate Reservoirs During CO₂-WAG Flooding: Insights from Nuclear Magnetic Resonance (NMR) Technology

Abstract

Oil and gas reservoirs dominated by coarse clastic rocks, particularly conglomerates (including gravel sandstones), are commonly termed conglomerate reservoirs in both the domestic and international literature. Sandy conglomerate reservoirs generally have high thickness and high productivity per unit area, but because of their characteristics such as rapid lithology change, strong heterogeneity, low porosity, and low permeability, it is difficult to develop conventional waterflooding. There is an urgent need for an efficient development scheme for the giant sandy conglomerate reservoir. In this study, nuclear magnetic resonance (NMR) technology was employed to investigate the stratified injection-production strategy for large-scale sandy conglomerate reservoirs. Three representative cores from different strata were selected to perform CO₂ flooding and CO₂-water alternating gas (WAG) flooding experiments, respectively. The aim was to explore how different development methods affect the recovery efficiency of various core types and the distribution of remaining oil under miscible and immiscible pressure conditions. The results show that immiscible CO₂ flooding mainly displaces crude oil in large pores, and oil in micropores and mesopores is difficult to displace. After gas channeling, there is still a large area of residual oil “aggregate” in the core, and the recovery rate is low. Compared with medium-coarse sandstone, the strong heterogeneity of sandy conglomerates leads to early gas breakthrough and low recovery efficiency during gas flooding. Compared with CO₂ flooding, CO₂-WAG flooding can balance the micro-oil displacement effect between micropores and macropores, significantly improve the oil production in micropores and mesopores. Thus, CO₂-WAG flooding has a certain micropore “profile control” effect, which can delay the gas channeling and improve the core recovery efficiency of reservoirs, especially for the highly heterogeneous sandstone. Miscible CO₂ flooding can effectively extract the oil in the mesopores and micropores that immiscible CO₂ flooding is difficult to displace. The gas breakthrough is slower and the recovery is much higher in miscible CO₂-WAG flooding than that of immiscible one. Therefore, ensuring that the formation pressure is higher than the minimum miscible pressure to achieve miscible flooding is the key to reservoir stimulation.

1. Introduction

With the sustained growth in global energy demand and the increasing challenges in conventional reservoir development, enhanced oil recovery (EOR) technologies have emerged as a pivotal focus in the petroleum industry [1,2,3,4,5]. CO₂ flooding, recognized for its unique advantages in viscosity reduction, miscible effects, and environmental benefits, is a critical technique for low-permeability reservoirs [6,7,8,9]. However, in strongly heterogeneous, low-porosity/low-permeability conglomerate reservoirs characterized by complex pore structures [10,11,12,13,14], conventional CO₂ flooding exhibits limited efficacy. Although CO₂ foam flooding can effectively suppress gas channeling in heterogeneous reservoirs and significantly improve oil recovery (by 15–30%), it faces limitations such as poor foam stability under high-temperature and high-salinity conditions, insufficient adaptability to complex pore structures, and uncertain effectiveness in long-term sequestration [15,16,17]. However, while CO₂-water alternating gas (WAG) injection demonstrates improved EOR performance in heterogeneous reservoirs, significant knowledge gaps persist regarding the dynamic distribution and occurrence mechanisms of residual oil during CO₂-WAG processes in massive, heterogeneous conglomerate reservoirs.

The injection of CO₂ interacts with crude oil to alter its viscosity, reduce interfacial tension, and induce oil swelling, thereby enhancing oil mobility [18,19]. These effects of CO₂ on crude oil mainly stem from the phase changes induced by carbon dioxide in oil-bearing pores [20,21]. In conglomerate reservoirs, however, microscale pore structure heterogeneity leads to uneven CO₂–oil contact efficiency during displacement, resulting in residual oil retention within microcapillary pores or fracture blind ends [22,23]. Although CO₂-WAG has been widely implemented in field applications due to its operational advantages, the underlying EOR mechanisms across diverse reservoir types remain inadequately understood. Clarifying these mechanisms and residual oil distribution patterns is essential for optimizing field development strategies.

In-depth studies have been conducted to address these challenges. Nezhad et al. [24] investigated the effects of different development methods on secondary oil recovery using sand-packed models, and their results showed that CO₂-WAG flooding achieved the highest efficiency in secondary oil recovery. Vahid et al. [25] studied the sandstone reservoir through a core displacement experiment and evaluated the relationship between the oil displacement performance of CO₂-WAG and slug size and water–gas ratio. The results showed that a smaller slug size could obtain a higher recovery rate. In this study, the injection parameters of CO₂-WAG flooding were optimized, but the heterogeneity of the reservoir was not considered. Yu et al. [26] established a numerical simulation model for five-spot CO₂-WAG patterns to investigate production dynamics in ultra-low-permeability reservoirs with poor injectivity. Their results demonstrated that CO₂-WAG effectively replenishes formation energy, expands sweep efficiency, and significantly reduces oil saturation. Conventional numerical simulations, while capable of predicting macroscopic residual oil distribution, face limitations in resolving microscale dynamic variations and computational efficiency [27,28]. Li et al. [29] conducted CO₂-WAG experiments on carbonate rock samples with three heterogeneity levels (strong to weak), revealing that fracture-cavity carbonates with the strongest heterogeneity achieved the highest incremental recovery. CO₂-WAG altered residual oil distribution patterns, with extended injection cycles further enhancing production. Most of the above are studies on the oil recovery and injection parameters of CO₂-WAG flooding. However, there are relatively few studies on the microscopic oil displacement mechanism of CO₂-WAG flooding.

Recent advancements in nuclear magnetic resonance (NMR) technology, particularly T₂ spectral analysis, have enabled quantitative characterization of fluid distribution and migration across pore-size domains [30,31,32]. Tian et al. [33] used nuclear magnetic resonance technology to monitor the volume of crude oil in pores of different sizes of low-permeability conglomerates in real time and found that gas flooding can effectively improve the recovery efficiency of tight reservoirs, while waterflooding is only effective for conventional reservoirs with good pore structure. Wang et al. [34] conducted waterflooding and CO₂-WAG experiments on heterogeneous composite cores, finding that strong heterogeneity severely compromises waterflood efficiency but subsequent CO₂-WAG improves sweep coverage in low-permeability zones. Sun Diagenesis [35] took Bei 14 block of Hailaer Oilfield as the research object and studied the microscopic mechanism of remaining oil after WAG flooding with the help of nuclear magnetic resonance technology. The results showed that the volume proportion of large pores in the target block exceeded 85% and became a gas channel after all CO₂ was used, with a recovery rate of only 47.95%. WAG flooding after CO₂ flooding not only displaces the remaining oil in secondary macropores, but also extracts the remaining oil in micropores and mesopores to varying degrees. Qian Kun et al. [36] utilized NMR technology to analyze the microscopic residual oil distribution of the core after CO₂-WAG flooding. The results indicated that CO₂-WAG flooding made up for the drawback of waterflooding and that it was difficult to utilize crude oil in micropores, and CO₂-WAG flooding achieved better oil displacement effects in mesopores and macropores. Nevertheless, systematic NMR-based investigations into CO₂-WAG mechanisms and microscale residual oil distribution in heterogeneous conglomerate reservoirs remain scarce.

In this study, NMR T₂ spectrum analysis and imaging technology were employed to compare the microscopic flow characteristics of carbon dioxide-water alternating gas injection (CO₂-WAG) under miscible and immiscible pressures using conglomerate cores. Compared with existing NMR-based CO₂-WAG flooding studies, the innovations of this study are reflected in three aspects: (1) For three typical lithologies of conglomerate reservoirs (coarse sandstone, sandy conglomerate and onglomerate-bearing sandstone), the displacement differences under miscible/immiscible pressures were systematically compared, making up for the limitation of previous studies that focused on a single lithology. (2) The recovery factor contribution of micropores, medium pores, and large pores was quantified through T₂ spectra, clarifying the microscopic mechanism of “medium-pore-dominated enhancement” in CO₂-WAG flooding and addressing the problem of vague interpretation of pore-scale mechanisms by previous researchers. (3) A quantitative correlation between core heterogeneity and gas channeling mitigation effect was established. By systematically investigating the influence of CO₂-WAG on residual oil distribution under different pressure conditions, this work aims to clarify the microscopic displacement mechanism and provide a theoretical basis for optimizing the development strategy of conglomerate reservoirs.

2. Experiments

2.1. Experimental Materials

The core used in this experiment is a cylindrical rock sample from the site of the sandstone reservoir, which mainly includes three types of cores: medium coarse sandstone, gravelly sandstone, and conglomerate (Tianjin Branch, CNOOC China Limited., Tianjin, China). The cores used in the experiment are shown in Figure 1. The cores had a diameter of 25 mm and a length of 50 mm, with a permeability range of 1–150 mD and a porosity range of 0.08–0.17 (or 8–17%). The core basic parameters are shown in

Table 1. Basic property parameters of experiment cores.

Number	Core Type	Length, cm	Diameter, cm	Porosity, %	Permeability, mD
1	Sandy conglomerate	4.812	2.515	16.13	130.22
2	Sandy conglomerate	4.856	2.503	13.41	119.35
3	Sandy conglomerate	4.897	2.529	13.12	120.28
4	Sandy conglomerate	4.916	2.536	16.34	139.63
5	Sandy conglomerate	4.899	2.369	13.51	140.59
6	Conglomerate-bearing sandstone	4.939	2.489	10.54	29.66
7	Conglomerate-bearing sandstone	4.955	2.499	10.03	19.68
8	Conglomerate-bearing sandstone	5.112	2.521	10.42	15.56
9	Coarse sandstone	5.088	2.506	8.26	1.26

. The water used in this experiment was heavy water with 99.8% purity D₂O (Wuhan Isotope Technology Co., Ltd., Wuhan, China), which showed no signal in nuclear magnetic resonance detection, the experimental gas was high-purity CO₂ gas (Qingdao Deyi Gas Co., Ltd., Qingdao, China), and the experimental oil was on-site crude oil (Tianjin Branch, CNOOC China Limited., Tianjin, China). The viscosity of crude oil at the experimental temperature (70 °C) was 5.93 mPa·s.

2.2. Experimental Setup and Procedure

2.2.1. Experimental Setup

In order to explore the influence of different development methods on the recovery efficiency of sandy conglomerate reservoir, a high-temperature and high-pressure online nuclear magnetic displacement experimental system was designed in this study, as shown in Figure 2. The experimental device mainly includes injection, displacement, and collection parts. The injection part includes two high-precision plunger pumps and three intermediate vessels, which, respectively, hold crude oil, heavy water, and CO₂. Two high-precision plunger pumps are used to control the precise injection of different fluids. The displacement part includes a non-magnetic core holder, nuclear magnetic scanner, and high-temperature and high-pressure circulation system, which can satisfy the real-time monitoring of the remaining oil in the core by using nuclear magnetic resonance instrument under the experimental temperature and pressure. The gas–liquid separation device can separate gas and liquid and accurately measure the volume of oil and water. Compared with the conventional core displacement device, the experimental device can not only use nuclear magnetic instrument for online core monitoring, but also ensure the accuracy of injected fluid volume by controlling the switch of six-way valve 1 and six-way valve 2, ensuring the reliability and accuracy of the experimental results.

2.2.2. Experimental Procedure

The minimum miscibility pressure (MMP) is a core threshold parameter for achieving oil–gas miscibility in CO₂ flooding, and its value is jointly influenced by crude oil components, reservoir temperature, and pore-throat constraints. For conglomerate reservoirs, recent review studies have pointed out that the restrictive effect of their micropore throats on CO₂ diffusion makes the MMP 3–5 MPa higher than that of conventional sandstone reservoirs [37,38]. In this study, the reservoir pressure range is 20–27 MPa, and the miscible pressure between the experimental crude oil and CO₂ is 23.1 MPa as measured by the thin tube experiment. Since the minimum miscible pressure of injected gas and crude oil is close to the formation pressure, there may be two cases of miscible flooding and immiscible flooding in the gas flooding process, and this experiment will study the two cases, respectively. In addition, nuclear magnetic resonance technology was used to analyze the crude oil production in the core under different displacement methods and different injection parameters, conduct in situ quantitative analysis of placement efficiency, and clarify the enhanced oil recovery (EOR) mechanism of CO₂-WAG flooding. Preliminary experiments showed that a water–gas ratio of 1:1 yielded good result, so this ratio was adopted for all experiments in this study. The specific experimental scheme is shown in

Table 2. Experimental scheme design.

Number	Core Type	Flooding Mode	Production Pressure/MPa	Injection Rate/(mL·min−1)	Gas–Water Ratio
1	Coarse sandstone	CO2 flooding	20	2	/
2	Conglomerate-bearing sandstone	CO2 flooding	20	2	/
3	Sandy conglomerate	CO2 flooding	20	2	/
4	Coarse sandstone	CO2-WAG flooding	20	2	1:1
5	Conglomerate-bearing sandstone	CO2-WAG flooding	20	2	1:1
6	Sandy conglomerate	CO2-WAG flooding	20	2	1:1
7	Coarse sandstone	CO2 flooding	27	2	/
8	Conglomerate-bearing sandstone	CO2 flooding	27	2	/
9	Sandy conglomerate	CO2 flooding	27	2	/
10	Coarse sandstone	CO2-WAG flooding	27	2	1:1
11	Conglomerate-bearing sandstone	CO2-WAG flooding	27	2	1:1
12	Sandy conglomerate	CO2-WAG flooding	27	2	1:1

.

In this experiment, the online nuclear magnetic resonance instrument is used to monitor the core displacement process. The specific experimental steps are as follows:

(1)

Wash and dry the experimental core, measure the length and diameter of the core, and determine the porosity and permeability of the experimental core.

(2)

Each core was vacuumed and then saturated with water at a displacement rate of 0.05 mL/min. The volume of saturated water was measured, after which the core was saturated with oil until no water was produced at the outlet. The volume of saturated oil was then measured to determine the bound water saturation.

(3)

Before the experiment, anhydrous ethanol was used to clean the pipeline of the test system, and experimental water was used to continue washing after cleaning.

(4)

Leak detection of the experimental system was performed to ensure the airtightness of the test system, to ensure the accuracy of the experiment and facilitate the smooth conduct of subsequent experiments.

(5)

Before displacement, the core was scanned with a NMR scanner to see the distribution of oil and water and calculate the porosity and permeability. The NMR-related parameters were set as follows: echo time of 0.1 ms; number of echoes of 25,000; and number of accumulations of 64.

(6)

The core was placed into the core holder, the instrument installed, and the switching of each valve checked, to prepare for the experiment.

(7)

Confining pressure and back pressure was added to the core according to the experimental scheme, the constant pressure and constant speed pump was opened, and the pump parameters adjusted according to the formulated experimental scheme.

(8)

Under the condition of keeping the pressure unchanged, the displacement was started. During the displacement process, the nuclear magnetic scanner was used to scan the core and observe the distribution of oil and water.

(9)

After displacement, the pressure in the experimental system was relieved, the desktop cleaned, and the experimental equipment organized.

3. Results and Discussion

3.1. Immiscible Displacement Experiment

In order to explore the EOR enhancement mechanism of CO₂-WAG flooding and CO₂-WAG flooding, three kinds of cores (medium coarse sandstone, pebbled sandstone, and sandy conglomerate) in different layers in the sandy conglomerate reservoir were studied under miscible pressure (20 MPa). Firstly, the high-temperature and high-pressure online nuclear magnetic resonance system was used to monitor the CO₂ displacement process of the three kinds of cores, and core imaging was performed at key nodes in the displacement process (initial stage, breakthrough, 1 PV, 3 PV, 5 PV, 10 PV), as shown in Figure 3. The red in the figure represents crude oil with a nuclear magnetic signal. Blue represents the unsignaled water and rock skeleton, and the gradual process from blue to red represents the change in oil saturation from small to large. With the progress of displacement, the oil saturation in the core gradually decreases. When the displacement reaches 10 PV, the residual oil with more oil in the core is still not produced, and the overall oil recovery efficiency is relatively low. Among the three core types, the displacement of crude oil in medium-coarse sandstone was the most uniform, resulting in the highest recovery efficiency. For conglomerate-bearing sandstone, recovery efficiency was relatively high in most regions, with only a small area containing large-sized residual oil aggregates. In contrast, sandy conglomerate exhibited the lowest recovery efficiency, and the core still contained a greater amount of residual oil aggregates after displacement. This is because medium-coarse sandstones have high porosity, high permeability, and low heterogeneity, allowing CO₂ to sweep crude oil out of the core relatively uniformly and thus achieving higher recovery efficiency. In contrast, conglomerate-bearing sandstones and sandy conglomerates have low permeability and strong heterogeneity, with sandy conglomerates exhibiting the most significant heterogeneity. In the process of CO₂ displacement, gas breakthrough is easy in the high-permeability area. As a result, the recovery effect of crude oil in the low-permeability area is poor, and more residual oil aggregates are formed.

Similarly, CO₂-WAG flooding was carried out on three kinds of cores, respectively, and nuclear magnetic resonance instrument was used to monitor the core displacement process, and nuclear magnetic imaging of cores during the displacement process was conducted, as shown in Figure 4. It can be seen from the figure that, compared with CO₂ flooding, CO₂-WAG flooding has better recovery effect on the three kinds of cores, especially in the conglomerate-bearing sandstone and sandy conglomerate, where the crude oil saturation has decreased significantly and there are almost no large residual oil aggregates, and the oil recovery effect is better.

Secondly, the breakthrough time and recovery rate of CO₂-WAG flooding core are significantly different from those of CO₂ flooding. As shown in Figure 5, from the perspective of breakthrough time, the gas breakthrough point of medium and coarse sandstone under CO₂ flooding is 0.26 PV, and that under CO₂-WAG flooding is 0.36 PV, which delays 0.1 PV compared with CO₂ flooding. For pebbled sandstone cores, the breakthrough point of CO₂ flooding is 0.14 PV, which is earlier than that of medium coarse sandstone. Under CO₂-WAG flooding, the breakthrough point of pebbled sandstone is 0.32 PV. In the case of CO₂ flooding, the breakthrough time of sandy conglomerate core is 0.12 PV, while in the case of alternating water and gas flooding, the breakthrough time is delayed to 0.31 PV. Therefore, compared with CO₂ flooding, CO₂-WAG flooding can significantly delay gas breakthrough, among which the breakthrough delay effect is the best for sandy conglomerate, and the breakthrough time is delayed by 0.19 PV. Its microscopic mechanism is directly related to the heterogeneous pore structure of coarse sandstone: coarse sandstone a dual-pore system consisting of “large pores and micro-throats”. During continuous CO₂ flooding, CO₂ due to its low viscosity, tends to channel rapidly along the large pores. In contrast, in CO₂-WAG flooding, the aqueous phase can first block the large pores for microscopic profile control, forcing the subsequent CO₂ to divert to the micro-throats. This enables CO₂ to more easily displace the crude oil in the micropores and further extend the gas channeling path.

Although NMR technology can provide intuitive characteristics of oil and water changes in the core, it cannot quantify the efficiency of oil displacement in different types of pores. Therefore, in this study, the nuclear magnetic resonance CPMG sequence is used to detect and obtain the nuclear magnetic T₂ spectrum of the core, which can more accurately quantify the oil displacement efficiency during the displacement process. In this experiment, the dynamic T₂ spectrum of core is measured during displacement, and the measured nuclear magnetic signal is converted into saturated oil quantity, and the quantitative analysis of oil displacement efficiency is realized.

During the experiment, when the core is saturated with different volumes of crude oil, its nuclear magnetic resonance T₂ signal will show certain regular changes, as shown in Figure 6. By analyzing the changes in these signals, the linear relationship between saturated oil volume and nuclear magnetic semaphore can be established, and the conversion relationship can be established as shown in Equation (1):

$$V=0.0001Q-0.1392$$

(1)

where V is the saturated oil content, mL; Q is the total amount of saturated oil NMR signals.

Therefore, we can obtain the calculation formula of oil recovery efficiency as shown in Equation (2):

$$R=1-{\displaystyle \frac{v_{\mathrm{r}}}{v_{\mathrm{s}}}}=1-{\displaystyle \frac{0.0001Q_{\mathrm{r}}-0.1392}{0.0001Q_{\mathrm{i}}-0.1392}}$$

(2)

where v_r is the volume of remaining oil; v_s is the volume of the initial saturated oil; Q_r is the semaphore of the remaining oil; Q_i is the semaphore of the initial saturated oil.

In the process of CO₂-flooding and CO₂-WAG displacement, the T₂ spectrum of core dynamics was measured by nuclear magnetic resonance instrument, as shown in Figure 7. In the T₂ spectrum, the horizontal coordinate is relaxation time, and the vertical coordinate is signal amplitude. In the T₂ spectrum, the longer the relaxation time is, the larger the pore diameter of the corresponding crude oil is. In this study, 0.001~1 ms is defined as the crude oil in micropores, 1~100 ms is defined as the crude oil in medium pores, and 100~10,000 ms is defined as the crude oil in large pores. As shown in Figure 7, compared with CO₂ flooding, the NMR T₂ signal in the short relaxation time range after CO₂-WAG flooding decreases considerably. This indicates that CO₂-WAG flooding improves overall recovery efficiency primarily by enhancing the displacement efficiency of crude oil in mesopores. Among the three core types, sandy conglomerate cores exhibit the most significant improvement in mesopore oil recovery efficiency, followed by conglomerate-bearing sandstones, with coarse sandstones showing the least improvement.

The oil recovery rate in different types of pores (micropores, mesopores and macropores) can be obtained by quantifying the T₂ spectral curve using Equations (1) and (2), as shown in Figure 8. It can be seen from the figure that the microscopic oil displacement law of reservoir pores is that the oil displacement efficiency of macropores is the highest, followed by mesopores, and micropores have the lowest. Compared with CO₂ flooding, CO₂-WAG flooding can significantly improve the oil recovery of mesopores and micropores, and CO₂-WAG flooding can effectively improve the microscopic oil displacement efficiency of mesopores and micropores. Among them, CO₂-WAG flooding has the best effect on crude oil extraction from mesopores and micropores of sandy conglomerate, and can evenly displace crude oil from various pores. This is because when CO₂-WAG flooding is carried out in heterogeneous cores, due to the large viscosity of water, it has a certain sealing effect on the hyperpermeability area, delaying the breakthrough of CO₂ and increasing the spread range of CO₂, resulting in more crude oil being extracted from micropores and improving recovery.

In general, CO₂-WAG flooding under immiscible pressure can improve oil recovery for sandy conglomerate reservoirs, especially for conglomerate-bearing sandstone and sandy conglomerate layers with strong heterogeneity. However, considering that the large-scale CO₂-WAG flooding development method is prone to problems such as cumbersome operation methods, high development cost, and equipment corrosion, the layered injection and production method can be considered, the CO₂-WAG flooding development method can be used for sandy conglomerate layer, and the water and CO₂ flooding development methods can be used for other layers.

3.2. Miscible Displacement Experiment

Similarly to Section 3.1, CO₂ flooding and CO₂-WAG flooding experiments were carried out on the cores of medium coarse sandstone, conglomerate-bearing sandstone and sandy conglomerate under miscible pressure, respectively, and the displacement process of cores was monitored using a high-temperature and high-pressure online NMR system. Core imaging is performed at key points in the displacement process (initial state, breakthrough, 1 PV, 3 PV, 5 PV), as shown in Figure 9 and Figure 10. As can be seen from Figure 9, CO₂ flooding under miscible pressure for all kinds of cores in the conglomerate reservoir has a good recovery effect. When the displacement reaches 10 PV, there is almost no residual oil in the core of the medium coarse sandstone, and the recovery effect is the best. After CO₂ breakthrough, with the progress of displacement, the residual oil saturation in the core changes little, and the oil displacement efficiency decreases greatly. As can be seen from Figure 10, after CO₂-WAG flooding is used to displace three types of core, there is almost no residual oil in the core after displacement, and the recovery effect is excellent. Different from CO₂ flooding, after core breakthrough in CO₂-WAG flooding, the residual oil saturation in the core still gradually decreases with the progress of displacement, and CO₂-WAG flooding still has better oil displacement efficiency after gas breakthrough.

CO₂ flooding and CO₂-WAG flooding experiments were carried out under miscible pressure to compare the breakthrough timing and recovery efficiency of cores of different reservoir types, as shown in Figure 11. As can be seen from the figure, the gas breakthrough point of medium and coarse sandstone is 0.61 PV under CO₂ flooding, and it is delayed to 0.68 PV under CO₂-WAG flooding. The gas breakthrough point of conglomerate-bearing sandstone is 0.46 PV under CO₂ flooding, and it is delayed to 0.5 PV under CO₂-WAG flooding. In the case of CO₂ flooding, the gas breakthrough point is 0.18 PV, and in the case of CO₂-WAG flooding, the delay point is 0.33 PV, among which CO₂-WAG flooding has the best delay effect on the sandy conglomerate. This is due to the relative homogeneity of coarse sandstones and gravel sandstones, while the sandy conglomerate core has strong heterogeneity, so in the process of CO₂ flooding, it is easy to form a dominant channel in the hyperpermeability area of the sandy conglomerate core, resulting in premature CO₂ breakthrough. CO₂-WAG flooding has a certain profile control effect, which can delay the CO₂ breakthrough in the sandy conglomerate and improve the early oil recovery efficiency.

Under miscible pressure, the dynamic T₂ spectral curves of various core CO₂-flooding and CO₂-WAG flooding processes were measured by the NMR instrument, as shown in Figure 12. As can be seen from the figure, for coarse sandstone, T₂ spectrum curve decreases significantly in the whole range with displacement, and both development methods have good recovery effects. For conglomerate-bearing sandstone and conglomerates, the T₂ curve of short relaxation range after CO₂-WAG flooding is significantly reduced compared with CO₂ flooding, which indicates that CO₂-WAG flooding can increase the oil recovery in micropores.

Equations (1) and (2) were used to quantitatively calculate the above T₂ spectrum, and the total recovery efficiency under different development modes and the oil recovery efficiency in various pores could be obtained, as shown in Figure 13. It can be seen from the figure that under miscible pressure, the recovery effect of the two displacement methods is relatively good, up to about 75%. Compared with CO₂ flooding, CO₂-WAG flooding can significantly improve the oil recovery rate in medium and micropores, especially in the sandstone reservoir with strong heterogeneity. Miscible gas flooding can effectively extract the oil in the medium and micropores that is difficult to displace by waterflooding and immiscible gas flooding. The gas breakthrough is slow, and the recovery rate is much higher than that of immiscible gas flooding. Therefore, ensuring that the formation pressure is greater than the minimum pressure to achieve miscible gas flooding is the key to reservoir stimulation.

4. Conclusions

In this study, high-temperature and high-pressure online nuclear magnetic resonance technology was used to study the distribution characteristics of CO₂-WAG flooding remaining oil in sandy conglomerate reservoirs from macro and micro perspectives, and the following key conclusions are drawn:

(1)

Pore heterogeneity affects the NMR displacement trend through the dual effects of preferential channel development and fluid production uniformity. Immiscible CO₂ flooding mainly recovers crude oil from macropores, while crude oil in micropores and mesopores is difficult to recover. After gas channeling, large-scale residual oil aggregates still remain in the core, resulting in low recovery efficiency. Compared with coarse sandstone, the strong heterogeneity of sandy conglomerate leads to faster gas breakthrough in gas flooding and low recovery rate.

(2)

Compared with CO₂ flooding, CO₂-water alternating gas (WAG) flooding can balance the microscopic oil displacement effect between micropores and macropores, and significantly increase oil production from micropores and medium pores. The core mechanism by which CO₂-WAG flooding improves crude oil recovery in micropores lies in the aqueous phase’s ‘profile control’ effect in micropores, which blocks preferential channeling paths. This not only delays gas channeling during immiscible CO₂ flooding and enhances core oil recovery but also exhibits a more pronounced improvement effect for sandy conglomerates with strong heterogeneity.

(3)

Miscible CO₂ flooding can effectively extract the oil in the mesopores and micropores that is difficult to displace using immiscible CO₂ flooding. Also, the gas breakthrough is slower, and the recovery rate is much higher in miscible CO₂-WAG flooding than that of immiscible one. Therefore, ensuring that the formation pressure is higher than the minimum pressure to achieve miscible flooding is the key to reservoir stimulation.