Micron CT Study of Pore Structure Changes and Micro-Scale Remaining Oil Distribution Characteristics During Low-Mineralization Water Flooding in Sandstone Reservoirs

Abstract

Low-salinity water flooding is a commonly used method to enhance oil recovery. At the microscopic scale, changes in pore structure and the distribution of remaining oil are critical to the effectiveness of water flooding. However, current research on the relationship between pore structure and remaining oil distribution is relatively limited. Therefore, this study employed micro-CT technology to analyze changes in pore structure and the distribution characteristics of remaining oil in sandstone cores during the water flooding process. Micron CT technology provides non-destructive, high-resolution three-dimensional imaging, clearly revealing the dynamic changes in the oil-water interface and remaining oil. The experiments included water saturation, oil saturation, and multi-stage water displacement processes in sandstone cores with different permeability values. The results show that the oil saturation in the rock core decreases during water flooding, and the morphology of remaining oil changes with increasing water flooding volume: cluster-like remaining oil decreases rapidly, while porous and membrane-like remaining oil gradually transforms, and columnar and droplet-like remaining oil increases under specific conditions. The study results indicate that at 1 PV flooding volume, the crude oil recovery rate reaches 57.56%; at 5 PV, the recovery rate increases to 64.00%; and at 100 PV, the recovery rate reaches 75.53%. This indicates that water flooding significantly improves recovery rates by enhancing wettability and capillary forces. Meanwhile, pore connectivity decreases, and particle migration becomes prominent, especially for particles smaller than 20 μm. These changes have significant impacts on remaining oil distribution and recovery rates. This study provides microscopic evidence for optimizing reservoir development strategies and holds important implications for enhancing recovery rates in mature oilfields.

1. Introduction

With the continuous growth of global energy demand and the gradual depletion of traditional oil resources, enhanced oil recovery (EOR) has become a critical issue that urgently needs to be addressed in the petroleum industry. Waterflooding, as a typical secondary recovery method for pressure maintenance and crude oil displacement, is widely applied in the middle and late stages of oilfield development due to its economic efficiency [1,2]. However, when conventional waterflooding reaches its recovery limit, enhanced oil recovery (EOR) technologies (e.g., low-salinity water flooding in this study) are required to further mobilize remaining oil. It should be noted that the waterflooding process (including conventional secondary waterflooding and EOR-related low-salinity water flooding) not only involves macro-scale fluid flow and recovery efficiency but also is accompanied by significant changes in micro-scale pore structure—these changes have a profound impact on enhancing crude oil recovery rates [3]. Therefore, conducting in-depth research into the dynamic changes in pore structure during low-salinity water flooding (an EOR technology) and their effects on remaining oil distribution and recovery rates is of great significance for optimizing EOR displacement strategies and improving the development efficiency of mature oilfields [4,5].

During waterflooding, the injection of fluids can cause complex changes in the reservoir’s pore structure, including alterations in pore connectivity, particle migration, and wettability reversal [6,7]. These changes directly impact the flow characteristics of the fluids and the distribution pattern of remaining oil [8,9]. Research indicates that the better the connectivity of the pore structure, the higher the waterflooding efficiency and recovery rate [10,11]. However, changes in pore structure may also lead to particle blockage or water lock effects, thereby restricting crude oil flow and reducing recovery rates. Therefore, elucidating the patterns of pore structure changes and their mechanisms of influence on recovery rates has attracted significant attention from numerous scholars. For example, Wang et al. [12] found that different types of pore structures correspond to markedly different waterflooding oil pathways and oil displacement efficiencies.

Changes in pore structure not only affect recovery rates but also significantly influence the transformation of remaining oil forms. Remaining oil exists in various forms within pores, such as cluster-like, porous, film-like, columnar, and droplet-like, among others. The formation and transformation of these forms are closely related to the characteristics of pore structure. For example, cluster-like remaining oil typically exists in pores with good connectivity, while membrane-like and droplet-like remaining oil are often distributed at pore edges or in micro-pores. During water flooding, changes in pore structure can lead to transformations in the form of remaining oil, thereby affecting its mobility [13,14]. Yang et al. [15] utilized CT scanning technology to measure the distribution of remaining oil in sandstone at different permeability scales, analyzing the distribution patterns of oil clusters. They found that as pore connectivity decreases, cluster-shaped remaining oil may transform into porous or membrane-shaped forms, while droplet-shaped remaining oil may be displaced or aggregated due to changes in capillary forces. Therefore, a thorough investigation of the relationship between pore structure and remaining oil morphology is of great significance for improving remaining oil recovery rates.

Although the influence of pore structure on waterflooding efficiency and remaining oil distribution has been widely recognized, there remains a relative lack of microscopic research on the dynamic changes in pore structure and their relationship with remaining oil morphology [16,17]. Traditional research methods have primarily focused on the macroscopic scale, analyzing waterflooding efficiency and reservoir characteristics through numerical simulations or physical experiments. However, these methods struggle to directly observe microscopic changes at the pore scale. In recent years, with the development of micro-computed tomography (Micro-CT) technology, non-destructive, high-resolution three-dimensional imaging has become possible, providing new tools for studying the microscopic relationship between pore structure and remaining oil distribution [18,19]. However, research combining Micro-CT technology with dynamic changes in pore structure is still in its infancy, particularly in terms of systematically analyzing the quantitative changes in pore structure and their impact on remaining oil distribution [20,21].

This study aims to use digital core technology in combination with Micro-CT scanning experimental data to conduct an in-depth analysis of the changes in pore structure of sandstone reservoirs during waterflooding and their impact on the distribution characteristics of remaining oil at the microscopic level. Through in situ scanning and image processing methods, the study quantitatively describes the changes in pore structure, the morphology of remaining oil, and its distribution patterns at different stages of waterflooding. The research findings will provide microscopic evidence for optimizing waterflooding strategies and enhancing recovery rates in mature oilfields and will offer scientific guidance for future reservoir development practices and the advancement of enhanced recovery technologies.

2. Experimental Methods

2.1. Materials

A core sample (sandstone) was taken from the sandstone reservoir of a certain offshore oilfield at a depth of 1121.6 m. The gas permeability was measured to be 859.0 mD, with a porosity of 31.62%. The dimensions were 25 mm × 26 mm, and it was subsequently cut into 8 mm × 16 mm pieces. To enhance imaging quality, the experimental oil used was a mixture of crude oil produced from a specific block in the Bohai Oilfield and 10% diiodoethane (Chengdu Kelong Chemical Co., Ltd., Chengdu, China) (The X-ray attenuation coefficients of pure crude oil and experimental water exhibit minimal differences, resulting in low gray-scale contrast between oil and water phases during CT scanning, making precise differentiation challenging. In contrast, diiodoethane contains iodine with a high atomic number, significantly enhancing the X-ray attenuation coefficient of the oil phase. This creates distinct grayscale differences between the oil phase, water phase, and rock matrix within CT images. The properties of the experimental oil are shown in

Table 1. Experimental oil properties.

Acid Number (mg KOH/g)	Base Number (mg KOH/g)	Asphaltenes (wt%)	Wax (wt%)	Density (g/cm3) (Measured Temperature: 65 °C)	Viscosity (mPa·s) (Measured Temperature: 65 °C)
4.0	1.3	0.56	2.53	0.85	5.48

(the bituminous matter content is determined using the n-heptane precipitation method specified in industry standard SY/T 5119-2016 [22] “Determination of Bituminous Matter Content in Rocks and Crude Oil.”). The wax content is determined using gas chromatography-mass spectrometry (GC-MS), Thermo Fisher Scientific, Waltham, MA, USA. The experimental water was prepared based on the ion composition of formation water from a certain block in the Bohai Oilfield. The composition of the simulated formation water and low-mineralization displacement water is shown in

Table 2. Composition of experimental brines.

Brines	Na+	SO42−	HCO3−	Ca2+	Mg2+	Cl−	TDS
Formation brine (mg/L)	2396.00	87.01	95.28	591.39	254.04	6088.00	10,311.72
Low salinity water (mg/L)	599.00	21.75	23.82	147.85	63.51	1522.00	2577.93

. The mineralization of the simulated formation water was 10,311.72 mg/L, and the low-mineralization displacement water was prepared by diluting the formation water four times.

2.2. Digital Core Displacement Experiment

Using a micro-CT scanning device to perform in situ scanning of the rock core (the scanning process is shown in Figure 1) to obtain high-resolution three-dimensional images of the internal structure of the rock core. The scanning voltage was set to 160 kV, the scanning power to 16 W, the pixel size to 3.23 μm, and the number of slices to 1440 frames. In situ scanning mode was employed, utilizing a specially designed core holder in conjunction with the CT scanning equipment to ensure precise, targeted scanning of the core under various conditions.

The experimental steps are as follows:

(1)

Core dry scanning: The core samples are degreased, dried, and weighed to determine dry weight, and the core diameter is recorded. The samples are then placed in a holder for the first in situ CT scan.

(2)

Saturation with water: Evacuate the core, inject simulated formation water, fully saturate the core sample with water, and perform a second in situ CT scan.

(3)

Saturation with oil: Under confined water conditions, inject simulated oil, fully saturate the core sample, and perform a third in situ CT scan.

(4)

In situ displacement scan: Conduct a water displacement experiment at a constant rate of 0.025 mL/min (The selection of this injection rate is based on two considerations: first, to match the actual flow rate of the original reservoir core (a medium-to-high permeability sandstone reservoir in the Bohai oilfield, with a permeability of 859.0 mD), which translates to a laboratory-scale rate of 0.02–0.03 mL/min according to Darcy’s law; and to ensure the observational accuracy of CT scanning for monitoring micro-scale oil-water interface evolution and remaining oil phase transformation, avoiding observation distortion or abnormal experimental cycles caused by excessively high or low rates). When the displacement volume reaches 1 PV (pore volume), 5 PV, and 100 PV, stop the pump and perform CT scans, recording the experimental parameters. The formulas for calculating the measured pore volume ${\mathrm{V}}_{\mathrm{p}}$(PV), oil saturation ${\mathrm{S}}_{\mathrm{O}}$, water saturation ${\mathrm{S}}_{\mathrm{w}}$, and crude oil recovery factor RF are as follows:

$${\mathrm{V}}_{\mathrm{p}}={\displaystyle \frac{{{\mathrm{m}}_1-\mathrm{m}}_0}{{\mathsf{\rho }}_{\mathrm{w}}}}$$

$${\mathrm{S}}_{\mathrm{O}}={\displaystyle \frac{{\mathrm{V}}_{\mathrm{O}}}{{\mathrm{V}}_{\mathrm{p}}}}\text{×}100\%$$

$${\mathrm{S}}_{\mathrm{w}}=1-{\mathrm{S}}_{\mathrm{O}}$$

$$\mathrm{RF}={\displaystyle \frac{{\mathrm{V}}_{\mathrm{Oil},\mathrm{produced}}}{{\mathrm{V}}_0}}\text{×}100\%$$

In the equation, m₁: wet weight, unit g; m₀: dry weight, unit g; ${\mathsf{\rho }}_{\mathrm{w}}$: simulated water density of the formation, unit g/cm³; ${\mathrm{V}}_{\mathrm{O}}={\mathrm{V}}_{\mathrm{P}}\times {\mathrm{S}}_{\mathrm{O}\mathrm{i}}$, where ${\mathrm{S}}_{\mathrm{Oi}}$ is the initial oil saturation, unit cm³; ${\mathrm{V}}_{\mathrm{p}}$: pore volume, unit cm³; ${\mathrm{V}}_{\mathrm{Oil},\mathrm{produced}}$: crude oil volume extracted from the rock core, unit cm³.

2.3. Displacement Image Processing

In this experiment, in situ scanning technology was used to perform micron CT scanning on sandstone cores. After scanning, the data volumes in different states were spatially aligned to ensure that the spatial positions of pores and particles remained consistent across all data volumes. During image processing, Avizo 2024.2 was used to analyze the CT scan images. First, a non-local filter was applied to the images for noise reduction and smoothing, facilitating the distinction between oil and water phases. Based on the differing X-ray attenuation coefficients of the rock matrix, oil phase, and water phase, threshold segmentation was performed using the watershed algorithm. The segmentation threshold was verified by ±2 gray values, and the oil phase recognition error was less than 5%. The voxel resolution is 10 µm per pixel, which has limitations on the characterization of micropores smaller than 5 µm. The coefficient of variation in the remaining oil saturation in the three repeated experiments was less than 3%, ensuring the reproducibility of the results [23].

2.4. Particle Transport and Aggregation Area Extraction

During low-salinity water flooding, particles in sandstone rock cores undergo migration and aggregation, which significantly affect pore structure and fluid flow characteristics. To quantitatively analyze these changes, this study uses image differential analysis combined with micron CT technology to extract particle migration and aggregation areas. The specific method is as follows:

First, micro-CT scans are performed on rock cores in different states to obtain high-resolution three-dimensional image data. Since the granular phase and fluid phase in rock cores have different gray values in CT images, with the gray values of the granular phase typically higher than those of the fluid phase, this characteristic provides the basis for image differentiation.

During image processing, core images in different states are spatially aligned to ensure that the spatial positions of pores and particles are consistent across all data volumes [24]. Subsequently, particle migration and aggregation regions are extracted using image difference methods. The specific steps are as follows: Subtract the dry-swept core image data from the displaced core image data. In the difference image, regions with positive gray values indicate that particles have moved from their pore positions during dry sweeping to their displaced positions, i.e., particle aggregation regions; regions with negative gray values indicate that particles have been eroded from their positions during dry sweeping to their displaced fluid positions, i.e., particle migration regions [25].

2.5. Study on the Distribution Pattern of Remaining Oil at the Micro Level

During the experiment, consistent with the method described in Section 2.3, a three-dimensional visualization model of remaining oil distribution in the core was constructed through consistent denoising, smoothing, and oil phase extraction. The pore space structure types of micro-scale reservoir rocks significantly influence the enrichment patterns of remaining oil, resulting in diverse distributions of remaining oil at the pore scale. By combining the three-dimensional shape factor G with the Euler number EN of the pore space, the micro-scale distribution forms of remaining oil can be quantitatively distinguished into five types: cluster-like, porous, membrane-like, columnar, and droplet-like [26,27], as shown in Figure 2.

The three-dimensional shape factor G is commonly used to characterize the structure of particles, pores, fluid flow, or oil and gas reservoirs [28]. A larger shape factor indicates a more complex object shape, with a larger surface area relative to volume, which typically affects processes such as seepage and diffusion in fluid mechanics.

$$G={\displaystyle \frac{6\sqrt{\pi }V}{S^{1.5}}}$$

In the formula, G is the shape factor of the remaining oil droplets; V is the volume of the remaining oil droplets, unit is μm³; and S is the surface area of the remaining oil droplets, unit is μm².

The Euler number is a measure of topological parameters that characterizes the connectivity of an object rather than its shape. It has the property of shape invariance, meaning that it does not change with the bending or stretching of the object’s shape. The Euler coefficient reflects the complexity of the oil droplet’s shape; the smaller the remaining oil Euler number, the more complex the shape [16,29]. The Euler number is represented by three Betti numbers:

$${\mathrm{E}}_{\mathrm{N}}={\mathrm{b}}_0-{\mathrm{b}}_1+{\mathrm{b}}_2$$

In the equation, E_N: Euler number of the remaining oil block; b₀: number of connections to the oil block; b₁: number of holes in the remaining oil block; b₂: number of cavities in the remaining oil block.

3. Results and Discussion

3.1. Sandstone Pore Structure Characteristics and Their Impact on Seepage

Since this scan uses in situ scanning, the spatial positions of pores and particles in different data sets are consistent after spatial alignment of the scanned data. Therefore, image interpolation is used to obtain pore space and particle migration. The specific principle is as follows: after CT scan reconstruction, the gray value distribution of the data volume is adjusted based on the mineral density and atomic number in the rock. The higher the density and atomic number, the higher the gray value after scanning, which appears as bright areas in the image.

The threshold segmentation method can be used to divide the pore space of rock cores into different states, thereby obtaining the three-dimensional distribution of pore space in rock cores of different states.

The CT grayscale of the rock core samples is shown in Figure 3. The rock samples are primarily composed of fine sandstone, with closely arranged skeletal particles and uniform grain size, resulting in a relatively uniform pore space structure and excellent flow conditions. The extracted CT images indicate that these rock samples have well-developed pores, primarily consisting of relatively regular intergranular pores. Although the pore sizes are small, they are almost entirely interconnected, providing excellent pathways for fluid migration. Overall, the rock samples are composed of dense sandstone, and although the pores are primarily micro- and nano-sized, they exhibit good connectivity, providing favorable flow pathways for low-salinity water flooding.

3.2. Changes in Pore Structure and Mechanisms for Improving Recovery Rates During Low-Salinity Water Flooding

A pore network model analysis was performed on a core sample taken from the same location under different conditions. The extracted pores were segmented, with connected pores marked in one color and non-connected pores marked in another color, thereby revealing changes in pore type distribution within the core under different conditions. Changes in core particle migration and aggregation are displayed in three dimensions, as shown in Figure 4. The blue areas represent particle migration regions, i.e., regions where particles move or migrate during displacement. Specifically, these regions are particle phases in the dry sweep state but transform into fluid phases after displacement. This indicates that particles are washed away by fluids and move from their original positions to other locations but are not completely displaced from the core. The red regions indicate particle aggregation zones, where particles aggregate during displacement. Specifically, these regions are pores (fluid phase) in the dry sweeping state but transform into the particle phase after displacement. This indicates that particles, after being eroded during displacement, are re-deposited and aggregate in these regions. Figure 3 clearly shows that particle migration and aggregation zones exhibit significant overlap. This phenomenon is directly related to the characteristics of the experimental core: The experimental core is relatively small (16 mm long, 8 mm diameter), featuring a short total pore network path. It is composed of medium-to-high permeability sandstone (permeability 859.0 mD, average throat diameter approximately 20 μm). After particles enter through the inlet with the displacement fluid, they only need to traverse 2–3 pore throats (short-distance migration) before encountering bottleneck effects due to partially “narrowed” throats (diameter < 10 μm). Particles then become trapped and aggregate near these narrow throats along migration paths, causing the migration endpoint to overlap with the aggregation starting point. This creates the characteristic “similar migration and aggregation zones,” indicating that in small-sized core pore networks, particles struggle to propagate over long distances and instead undergo localized aggregation after short-range migration [25,30]. The results show that by statistically analyzing particle migration and aggregation in rocks at different stages, we can observe that the greatest particle movement occurs at the onset of displacement, as this stage has the highest concentration of mobile free sand. As water displacement progresses, the concentration of mobile free sand gradually decreases. Although displacement velocity increases gradually, particle movement decreases progressively. Further observation reveals that particle migration and aggregation primarily occur among fine grains with radii less than 20 μm, while larger grains with stronger cementation resist displacement. As the displacement process continues, the overall porosity and permeability of the core decline, progressively degrading reservoir properties. The core reason for this change lies in the fact that fine particles (<20 µm), after migration under water drive, readily aggregate at pore throats and cause blockage (as shown in Figure 4, where red particle clusters are predominantly concentrated at throat locations). This phenomenon is not caused by clay swelling. Fine particle blockage directly reduces effective pore space, manifesting as decreased porosity. This phenomenon aligns closely with the fundamental properties of the experimental rock samples: the sandstone used in the experiments was primarily loosely cemented during sedimentary diagenesis, lacking strong cementing minerals to effectively bind fine particles. Consequently, particles smaller than 20 µm are more susceptible to migration under water-driven erosion.

By separating the oil and water phases, quantitative analysis of the oil-water ratio and oil phase recovery rate within the rock core can be performed (Figure 5). The results are shown in Figure 5: at 1 PV, the recovery rate reached 57.56%, significantly higher than the conventional waterflood recovery rate for this reservoir. As the injection volume increased, more crude oil was displaced, causing the recovery rate to gradually rise, reaching 64.00% at 5 PV. This enhanced recovery effect primarily stems from the potential improvement in reservoir properties and displacement conditions achieved by low-salinity water flooding. First, cores were prepared following the procedure described in Section 2.2 (i.e., “water-saturated cores followed by crude oil saturation”), initially exhibiting oil-wetting characteristics—pre-displacement CT images showed continuous oil phase coverage on pore walls. To ensure stable establishment of oil-wetting conditions, this study employed a sequential saturation protocol: first, vacuum dehydration at ≤0.095 MPa for 2 h to remove residual moisture, followed by bound water saturation treatment, and finally, slow injection of simulated oil at 0.01 mL/min until micro-CT imaging confirmed oil saturation ≥85%. To establish oil-wetted core conditions, this study employed a sequential saturation preparation process: first, vacuum dehydration at ≤0.095 MPa for 2 h to remove residual moisture; second, bound water saturation treatment; finally, low-speed injection of simulated oil at 0.01 mL/min until oil saturation ≥ 85% (verified by micro-CT imaging). During oil injection, van der Waals forces promoted oil phase adsorption onto the core surface, forming a stable oil film. Subsequently, a 4 h static equilibrium treatment (focusing solely on oil phase homogenization without aging) was implemented to maintain the core’s oil-favoring state. Micro-CT observations further revealed that prior to water injection, the oil phase continued to diffuse within the pores without water film adhesion to the pore walls. Following the injection of 1 pore volume (PV) of low-salinity water, a continuous water film formed on the pore walls, while the oil phase dispersed into discrete oil droplets (see Figure 5b), fully consistent with the typical characteristics of oil-wetting transitioning to water-wetting. Simultaneously, under 1 PV water injection conditions, the recovery rate reached 57.56%, significantly surpassing the conventional waterflood efficiency (45–50%) for such reservoirs. The enhancement phase coincided with microscopic oil-water phase transitions, ruling out interference from chemical reactions or sudden pore structure changes. This further confirms that wettability alteration is the core mechanism driving recovery improvement [31]. During early water flooding, wettability may shift toward hydrophilic conditions by modulating interactions between rock surface charges and minerals. This potential transition may reduce oil droplet adhesion to rock surfaces, causing the oil phase to disintegrate from a continuous coating on rock walls into dispersed droplets (as shown in the post-displacement comparison in Figure 5). This subsequently lowers oil-water interfacial tension, facilitating crude displacement from pores. Second, based on observed wettability improvements (Figure 5), low-salinity water flooding may mitigate the water lock effect—where water occupies pores during displacement, obstructing crude flow [27]. This potential mitigation stems from reduced pore wall water absorption, though its exact extent requires further validation. Third, at the microscopic level, low-salinity water flooding can enhance capillary effects by increasing capillary pressure differentials within pores, thereby aiding the displacement of residual crude oil adhering to rock surfaces [32,33].

During the 5 PV to 100 PV stage, low-salinity water continues to displace remaining oil from the pores, resulting in a continued increase in recovery rate, though at a relatively slower pace. At 100 PV, the recovery rate reaches 75.53%, an increase of 11.53% compared to 5 PV. This is due to the persistent migration and aggregation of fine particles (<20 µm), which further alter the structure and permeability of the reservoir pores, thereby improving fluid flow characteristics and promoting the mobilization of remaining oil. The expanded range of red particle aggregation areas in CT images also supports this mechanism, reflecting finer particles depositing in pore throats. However, as the mobile oil in the reservoir is gradually displaced, the effectiveness of displacing remaining oil in micro-pores gradually weakens, leading to a flattening of the increase [13].

As shown in Figure 6, low-salinity water flooding can significantly improve recovery rates and mobilize more remaining oil. Although some remaining oil remains undrained during the high water cut stage, low-salinity water flooding demonstrates significant advantages in improving recovery rates through a combination of mechanisms, including changes in wettability, enhanced capillary forces, and reduced water lock effects [34].

3.3. The Effect of Changes in Sandstone Pore Structure on the Distribution of Microscopic Remaining Oil

Changes in sandstone pore structure have a significant impact on the distribution and morphology of remaining oil at the microscopic level. By analyzing the distribution of remaining oil under different PV conditions, we found that the morphology of remaining oil undergoes noticeable changes during water flooding, as shown in Figure 7 and Figure 8.

Analysis of cluster-like remaining oil changes: Clustered oil exhibits good connectivity, and its mobilization often aligns with the direction of waterflood flow lines. Due to the slow injection flow rate, the displacement effect on clustered remaining oil in the rock core is effective, resulting in a rapid decrease in the proportion of clustered oil in the rock core and a good overall coverage range. Throughout the waterflooding stage, clustered remaining oil is continuously mobilized, leading to a continuous decrease in its proportion [10,35].

Analysis of the variation patterns of porous remaining oil: Porous oil and cluster oil are interrelated, and there is a certain degree of conversion between the two. Specifically, after cluster oil is depleted, the remaining unused portion primarily forms porous remaining oil. At 1 PV, there is no porous remaining oil in the latter half, but after 5 PV, the proportion of porous remaining oil in the latter half significantly increases. Analysis suggests that the remaining oil cluster is dispersed to form a large amount of porous remaining oil.

Analysis of the variation in membrane-like remaining oil: Membrane-like oil is primarily distributed at the edges of the initial confined water pore zones, and most membrane-like oil exists in a remaining oil state [36,37]. At 1 PV, the distribution of membrane-like remaining oil is relatively concentrated at both ends. At 5 PV, as other types of remaining oil are continuously depleted, the proportion of membrane-like remaining oil in the central region increases. At 100 PV, with the continuous injection and flushing of water, membrane-like remaining oil is continuously depleted, resulting in a significant decrease in the proportion of membrane-like remaining oil in the middle and rear sections.

Analysis of the variation in droplet-like remaining oil: Droplet-like remaining oil consists of small oil droplets distributed in microscopic pores, constrained by the Jamin effect [38]. Under initial water-flooded conditions, the content of droplet-type remaining oil is relatively low. At 5 PV, there is a significant increase in droplet-type remaining oil in some areas. Analysis suggests that other types of remaining oil are being mobilized, forming new droplet-type remaining oil. However, when the water injection volume reaches 100 PV, the droplet-type remaining oil is quickly carried away by the continuously injected fluid, causing the proportion of droplet-type remaining oil to begin decreasing again.

Analysis of the variation patterns of columnar remaining oil: Columnar remaining oil is primarily distributed in the central regions of pores, surrounded by water and constrained by capillary forces. The initial content of this type of remaining oil is very low. At 1 PV, the proportion of columnar remaining oil at the rear end is relatively high. Analysis suggests that as various types of remaining oil are passively displaced, a large amount of columnar remaining oil easily forms at the rear end. However, after 5 PV, the proportion of columnar remaining oil at the rear end decreases significantly. Analysis indicates that as the injection water volume increases, remaining oil becomes difficult to passively mobilize, and the columnar remaining oil in the back end is continuously displaced and cannot form new columnar remaining oil, resulting in a continuous decrease in the proportion of columnar remaining oil.

In general, marine sandstone reservoirs exhibit excellent porosity connectivity. Low-salinity water flooding can effectively alter rock wettability and dissolve clay minerals, particularly demonstrating significant effectiveness in mobilizing droplet-shaped, film-shaped, and columnar remaining oil. During production, a good injection-production pressure differential can be established, and low-salinity water flooding is economically viable. During development, production efficiency can be further improved through injection-production control techniques. Although a significant amount of droplet-shaped and porous remaining oil remains during the medium-to-high water cut period, its connectivity is good, providing favorable conditions for further enhancing recovery rates. However, the stable pressure field formed by long-term water injection keeps such remaining oil in a balanced state, unable to generate sufficient driving force. It is necessary to disrupt the stable pressure field of oil in the pores to mobilize it. For droplet-shaped, film-shaped, and columnar remaining oil dispersed in small pores, surfactant flooding or polymer flooding methods can be considered in later stages to further enhance recovery rates [39].

4. Conclusions

This study utilized micro-CT technology to analyze the impact of pore structure changes in sandstone cores during water flooding on the distribution of residual oil at the microscopic level. Results indicate that as water injection volume increases, core oil saturation decreases while water saturation rises, leading to significant evolution in residual oil morphology: clustered residual oil rapidly diminishes due to enhanced connectivity, while porous and film-like residual oil gradually transform; columnar and droplet-like residual oil increase under specific conditions. The recovery rate reached 57.56% at 1 PV, increased to 64.00% at 5 PV, and reached 75.53% at 100 PV, confirming the advantages of water flooding in improving rock wettability, mitigating water lock effects, and enhancing capillary forces.

This study offers direct guidance for mature field development: Based on changes in pore structure (e.g., connectivity decline due to fine-powder migration), field injection parameters can be optimized (e.g., initial flow rate control at 0.025 mL/min to delay blockage), providing microscopic evidence for adjusting development strategies in old oilfields. Furthermore, by integrating insights into the evolution of the residual oil phase, it offers practical support for subsequent optimization of waterflooding schemes and enhanced recovery in mature fields, thereby boosting the economic viability and sustainability of old oilfield development.