The experimental core was obtained from a gas storage layer characterized by a relatively simple broken-nose structure. The target layer consists of a braided-river delta front deposit, forming a sandstone reservoir with medium porosity and medium permeability. The average porosity of the core is 22.7%, the permeability is 167.7 × 10⁻³ μm², and the buried depth of the reservoir is −2310~−2620 m. The specific data of the core are shown in Table 1.

Distilled water, kerosene, and nitrogen were employed in the experiment due to their highly stable properties. These substances were used to simulate the alternating injection and production processes in gas storage facilities. This approach was adopted to minimize interference from extraneous physicochemical properties and thus prevent reservoir damage or blockage caused by such factors. The experimental methodology involved altering only one parameter at a time to systematically assess its impact on the reservoir’s geological structure and physical properties. Key factors examined were the core water-oil ratio and the number of alternating injection-production cycles.

Fig. 1 is an experimental device for rock damage of gas storage, which can simulate the alternating injection and production process of gas storage under multiphase conditions. The experimental device can simulate the experimental conditions of single-phase, two-phase and three-phase. It can directly perform alternating gas displacement on the core in the oil-water mixed state without replacing the displacement equipment. The sensor is connected to the computer system to record the changes of key parameters such as core gas flow, pressure difference and flow rate at both ends in real time. Real-time monitoring of reservoir damage and sand production caused by actual alternating injection and production of gas storage with different water-oil ratios is realized. The experimental device includes a gas-liquid switching device, a pressurization device, a formation simulation device, a sensor acquisition device and a sampling device.

(1)

The booster pump provides the inlet pressure required for the experiment by compressing the supply gas of the gas cylinder.

All experiments were conducted at a constant temperature of 80°C, with a confining pressure of 5 MPa, and nitrogen as the flowing phase. The outlet pore pressure was maintained at atmospheric pressure. It is important to note that the alternating pressure cycle was implemented by modulating the inlet pressure only, with no reversal of the flow direction. This approach enables the simulation of cyclic stresses on the reservoir rock framework during injection/production cycles without requiring the complexity of flow reversal.

To investigate sand production under varying water-oil ratios after alternating gas flooding, the core samples were first saturated with kerosene and then displaced with actual formation water. This process created an oil-water mixture within the core (the experimental schematic is shown in Fig. 2. By applying different displacement flow rates, various water-oil ratios were established inside the cores, enabling the observation of sand production across these different fluid saturation conditions.

The water-oil ratio is the ratio of water-bearing volume to oil-bearing volume in the core. The pore volume of the core and the mass change before and after water flooding were measured experimentally. And the water-oil ratio after different flow displacement rates is calculated using Formula (1). The formula for determining the water-bearing volume of the core is provided as follows.

V w —The volume of water in core pores, cm³

V o —The volume of oil in core pores, cm³

V p —The volume of core pore, cm³

m w —The quality of water in the core pore, g

m o —The quality of oil in the core pore, g

ρ w —density of water, g/cm³

ρ 0 —density of oil, g/cm³

By adjusting the liquid displacement time and injection flow rate, the water-oil ratio of the core can be accurately controlled. The water-oil ratio is calculated from the measured water-bearing volume of the core, which serves as the basis for critical pressure difference and alternating displacement experiments.

The purpose of this experiment is to study the influence of different injection-production flow rates on fluid distribution and sand production behavior in cores. Specifically, by changing the fluid injection rate in the three-phase experiment, different water-oil ratio environments are formed in the core, and alternating injection-production experiments are carried out under these conditions. In the experiment, the system compares the corresponding liquid flow rate and sand production under different water-oil ratio conditions. The specific experimental steps are as follows: (1)

First, the core is saturated with kerosene, and the porosity of the core is calculated by weighing the mass difference before and after.

In the alternating gas flooding experiment, since the displacing gas is vented directly after passing through the apparatus, the outlet pressure is maintained at atmospheric pressure.

Following the critical pressure difference measurements at various water-oil ratios (Section 3.2), the five core samples were subsequently used for water flooding experiments. The core was first saturated with kerosene prior to water flooding. Water samples were used to displace sandstone at a rate of 0.3, 0.4, 0.5, 0.6, 0.7 mL/min, and the water-oil ratio formed in the core was calculated after water flooding. Subsequently, gas was injected to displace the oil-water saturated core in order to determine the critical sanding pressure differential. Following the procedures outlined in Table 2, alternating displacement experiments were conducted on cores with different water-oil ratios, during which sand production was monitored in real time.

Because it is difficult to observe the sand production of the core during the actual experiment, the sand production of the core is judged by measuring the quality change of the core after displacement. Given the heterogeneity of the experimental core samples, a single representative core sample was used for each experimental condition (as detailed in Table 2) in this study. Firstly, all experiments were performed under rigorously identical and controlled conditions. Furthermore, the key trends observed were not only highly reproducible across the different experimental groups but also consistent with a clear physical mechanism. Therefore, we contend that the results provide a strong and representative characterization of the system’s behavior.

The cumulative mass of sand produced during each experimental stage was determined using the mass difference method. For every measurement, the core was meticulously extracted from the core holder. The dry mass of the core sample was measured before the experiment and following each alternating cycle suite using a high-precision analytical balance (accuracy: ±0.1 mg). A strict cleaning and drying procedure are conducted prior to every mass measurement. This involves flushing with toluene and then vacuum-oven drying at 70°C until a constant mass is reached (change <0.01 mg per 2 h). These steps ensure that the measurement results are not influenced by residual fluids.

The most significant contribution to the measurement uncertainty stems from the balance accuracy. To mitigate this, all mass measurements were performed in triplicate, and the mean value was adopted to enhance reliability. The resulting standard uncertainty, calculated from the standard deviation of repeated measurements and balance specifications, was below ±0.5 mg. This value is considered insignificant as the measured sand production was consistently orders of magnitude larger (on the order of tens to hundreds of milligrams).

In order to explore the sand production following alternating flooding under various water-oil ratios, the core sample was initially saturated with kerosene. Subsequently, displacement was performed using actual formation water. This process established an oil-water mixture within the core, thereby simulating reservoir fluid conditions. The experimental setup is illustrated schematically in Fig. 1. The displacement cores with different flow rates are used to form different water-oil ratios inside the cores, and the sand production of the cores under different water-oil ratios is observed.

The water-oil ratio after displacement at different flow rates was determined using Eq. (1), based on measurements of the core’s pore volume and mass change before and after water flooding. The calculated results are summarized in Table 3.

The core was displaced at flow rates of 0.3, 0.4, 0.5, 0.6, and 0.7 mL/min using formation water, with each rate maintained for 5 min. The resulting water-oil ratios after flooding were 0.22, 0.25, 0.55, 0.63, and 1.21, respectively. These results indicate that, under constant displacement time, higher injection rates lead to higher water-oil ratios in the core.

Experimental investigations were conducted to evaluate the critical sand production behavior of core samples under varying water-oil ratios. Critical sanding pressure difference measurements were performed on cores with water-oil ratios of 0.22, 0.25, 0.55, 0.63, and 1.21. During the experiment, gas displacement was employed to incrementally or continuously increase the pressure differential across the core until the critical sanding threshold was attained. The mass change of the core was measured and recorded at each pressure differential level. The mass variation of each experimental core with displacement pressure differential is shown in Fig. 3.

Observation of mass variation in cores with different water-oil ratios under critical pressure differential—after accounting for mass loss during disassembly—revealed that core mass tended to stabilize once the displacement pressure differential reached a specific threshold. This indicates that the core attained a state of irreducible water saturation, at which the residual fluid could no longer be mobilized despite further increases in injection pressure. Following a period of stability, a pronounced inflection point was observed, coinciding with a marked reduction in core mass—signifying the initiation of sand production. The pressure differential at which this inflection occurred is defined as the critical sand production pressure differential.

The same experimental procedure was repeated for the remaining core samples. The results indicate that under water-oil ratios of 0.22, 0.25, 0.55, 0.63, and 1.21, the corresponding critical sanding pressure differences were 3.6, 3.7, 3.9, 4.1, and 4.4 MPa, respectively. These values demonstrate a clear trend of increasing critical sand production pressure difference with rising water-oil ratio.

The initial phase examined sand production under dry core conditions. A series of five identical alternating displacement experiments were performed, with mean sand production values serving as the basis for analyzing cyclical correlations. The primary objective of this work is to examine the correlation between sand production and displacement cycles using regression analysis. The experimental dataset is limited in size but high in measurement precision. Therefore, Newton’s interpolation method was selected for polynomial modeling. A key advantage of this method is that it ensures the fitting curve passes exactly through every data point. This property provides an accurate representation of the experimental trend. Consequently, the approach is well-aligned with the research objectives of interpolation and trend analysis. Compared to the least-squares method—which is more appropriate for extrapolation and prediction—Newton’s interpolation offers more intuitive and precise analytical outcomes for the intended scope of this work.

A polynomial model was established using Newton’s interpolation method to correlate core mass loss with the number of cycles. This model enables the prediction of sand production under various cycling conditions. Furthermore, the trend of sand production versus the number of cycles is shown in Fig. 4.

As shown in Fig. 4, the core sand production rate exhibits three distinct phases in response to the increasing number of alternating cycles: an initial increase, followed by a stabilization period, and then a final phase of significant acceleration. The trend of sand production shows a period of stability between 25 and 55 alternating cycles, after which the rate climbs markedly with further cycles. The established polynomial model was used to calculate the average sand production rates during cycles 5–55 (phase I) and cycles 55–100 (phase II), which were determined to be 0.086 g/cycle and 0.455 g/cycle, respectively. The average sand production during the phase II of alternating displacement was 5.27 times higher than that in the phase I. Concurrent visual observations confirmed that sand production from the core became progressively more severe as the number of cycles increased.

Following the investigation into the relationship between alternating displacement cycles and sand production, the influence of the core water-oil ratio on sand production was examined. According to the results of Experiment 4.2, the critical pressure difference of the core sample was measured to be 3.6, 3.7, 3.9, 4.1, and 4.4 MPa corresponding to water-oil ratios of 0.22, 0.25, 0.55, 0.63, and 1.21, respectively. A gas pressure alternating displacement experiment was conducted on cores with different water-oil ratios using a pressure alternation interval of 0.5 MPa and 3.5 MPa. The changes in core mass during the alternating process and the variation in gas flow through the core under constant pressure are presented in Fig. 5.

Fig. 5 shows that the lower the water-oil ratio of the core (i.e., the higher the oil content), the more efficient the gas displacement is in draining liquid from the core. The early stage of displacement is characterized by a rapid decline in core mass and high liquid production, which later stabilizes. Additionally, a higher water-oil ratio correlates with a lower initial gas flow rate and a more gradual increase in gas flow. Therefore, it can be concluded that when the core exhibits a high water-oil ratio, indicating high water content in the reservoir, liquids become more difficult to displace from the core. This will further complicate the injection and production operations in gas storage facilities. The core was dried and characterized for mass and permeability after the alternating displacement experiment to evaluate its internal structural changes. The corresponding data are summarized in Table 4.

Changes in core mass and permeability were observed before and after the experiment under various water-oil ratios. The mass difference before and after the experiment was considered as the sand production. A polynomial model relating sand production, permeability change, and water-oil ratio was established using Newton’s interpolation method. The polynomial models established in this study are based on the current experimental dataset and are primarily intended to reveal variation trends among the parameters. The resulting trends are presented in Fig. 6.

As shown in Fig. 6, the sand production of the core exhibits a non-monotonic relationship with the water-oil ratio, characterized by an initial increase, followed by a decrease, and a subsequent rebound. Within the water-oil ratio range of 0.3 to 1, a gradual reduction in sand production is observed. The observed reduction in sand production at this stage is attributed to the increased water-oil ratio. This increase enhances water erosion and, more importantly, strengthens the stabilizing effect of the water film. Under cyclic stress loading, the core framework undergoes progressive fatigue damage. This process begins with the detachment of weakly cemented particles. Once mobilized, these particles are transported by the flowing gas phase. They can then form bridge blockages at pore throats, which intensifies localized damage and alters flow pathways. Throughout this process, the residual water films coating sand and gravel surfaces contribute critical cohesive strength via capillary forces, significantly influencing particle mobilization and retention. To isolate these physical mechanisms from complex chemical interactions, inert fluids (kerosene, distilled water, and dry nitrogen) were deliberately employed as displacement media. This experimental design effectively excludes confounding factors such as mineral dissolution/precipitation and formation water evaporation. Consequently, observed particle adhesion behaviors and mobilization resistance can be attributed more conclusively to the physical role of water films rather than active chemical processes. Simultaneously, particles are immobilized through physicochemical adsorption, collectively forming a retention mechanism that suppresses sand production.

The relationship curve in Fig. 6a indicates an increasing trend in sand production as the the water-oil ratio exceeded 1. However, due to the limited sample size under high water-oil ratio conditions. The universality of this trend and its quantitative relationship require further validation with more samples in the future. As shown in Fig. 6, the model exhibits excellent goodness-of-fit, with R² values of 0.9389 and 0.9928 for the two relationships, both indicating outstanding performance within the observed data range.

The degree of permeability changes in the core before and after the experiment was observed to increase continuously with higher water-oil ratios. A positive correlation was identified, wherein a higher initial water-oil ratio resulted in a more significant alteration in permeability. The post-experimental rise in core permeability is interpreted as being principally caused by grain migration induced by internal rock damage. The proposed mechanism is twofold. Firstly, alternating stress initiates the detachment and migration of weakly cemented skeleton particles. These mobilized particles subsequently form bridge blockages at pore throats, representing a form of macroscopic internal damage. Although such damage does not immediately lead to significant sand production, it considerably influences the core’s seepage capacity. This is evidenced by the observed increase in core permeability post-experiment. Concurrently, particle migration scours and widens the original pore channels, potentially creating new micro-fractures or pore-throat networks. The resulting reorganization of the pore microstructure effectively expands fluid flow pathways and inherently enhances porosity. While wettability changes could influence permeability, their role is considered minor in this inert fluid (kerosene, nitrogen) displacement system. Consequently, the marked permeability improvement is predominantly ascribed to the augmented porosity and enhanced connectivity arising from grain migration.