Effects of moisture content and disturbance load on seepage characteristics of weakly consolidated rock mass

With shallow mineral resources gradually exhausted, deep mineral resource development is critical for national resource security yet faces severe challenges when tunnels cross weakly consolidated rock masses like faults and collapse columns. These rocks are prone to sudden seepage hazards under high stress and water pressure, threatening project safety. To address gaps in existing weakly consolidated rock seepage research—such as traditional theories ignoring dynamic mass loss and unclear “moisture content-disturbance load-permeability” coupling—this study explores permeability evolution mechanisms and quantifies how moisture content and disturbance load affect seepage behavior. Remolded graded gangue samples were prepared; via a THM triaxial seepage apparatus, experiments adjusted moisture content with varying immersion durations and simulated disturbance with three stress paths, monitoring axial strain, flow velocity, particle loss speed and porosity. Results show: Permeation has three stages, the response period from 150 to 180 minutes is critical for water outburst control with permeability increasing over 100 percent compared to the disturbance period. High-moisture samples with around 14 percent moisture have peak particle loss speed 2.2 times that of dry samples with 0 percent moisture. High-disturbance permeability is 1.5 to 2 times that of static loading. The non-Darcy factor rises with moisture content but falls with disturbance, approaching Darcy flow. This study fills prior gaps and supports seepage hazard prevention.

1 Introduction

Mineral resources are the core material foundation for national economic and social development (Hossain et al., 2025). Under the current shifting international landscape, enhancing the ability of mineral resource exploitation and reducing the degree of external dependence have become a major strategy for safeguarding national resource security and economic stability (Xie et al., 2017; Wu et al., 2025a). As the shallow mineral resources gradually exhaust, the resource development is progressively advancing into deeper ground strata (Kang et al., 2023). However, the deep tunnel construction often requires to cross weakly consolidated rock masses such as faults and collapse columns (Zhang et al., 2025). For instance, in China’s western deep coal mines, tunnel excavation through weakly consolidated fault zones has led to sudden water inrush events—once reaching over 500 m³/h, forcing mine shutdowns for months and causing massive economic losses; in Yunnan’s subway tunnel construction, seepage in such rock eroded surrounding rock, triggering local collapses and delaying projects. Due to the loose fragmentation and high degree of weathering within weakly consolidated rock, exhibiting complex porosity, high permeability, and sensitivity to disturbance (Nguyen et al., 2014). Under the combined effects of high stress, high water pressure, and engineering disturbance, it is prone to abrupt permeability changes (Wang Z. et al., 2025). This can lead to connectivity between aquifers and faults, triggering groundwater flow instability and water inrush hazards (Mao et al., 2025). Such phenomena pose a severe threat to the safety of deep resource development, which imply the urgent engineering significance and strategic value of studying its permeability characteristics (Chen et al., 2024).

Existing research has preliminarily revealed the micro-mechanisms of variable-mass seepage in weakly consolidated rock: on the one hand, the rock’s heterogeneous porosity makes it prone to fracture propagation and deformation under high-stress disturbances (Wang J. et al., 2025); on the other hand, the characteristic of weakly consolidated causes the consolidated material to dissolve easily under the water-bearing conditions, leading to the migration of the main particles with seepage (Cui et al., 2022). The loss of particles further accelerates the evolution of seepage pathways, forming a cyclical positive feedback process of “consolidation weakening-particle migration-pathway evolution” (Lu et al., 2024; Wu et al., 2025b). Recent studies on this topic have shown that particle migration and pore evolution in weakly consolidated media are closely coupled: seepage-driven particle migration reshapes pore structure (e.g., expanding small pores and connecting medium pores), while particle detachment from pore walls improves pore connectivity. Water content also regulates this coupling by accelerating mineral dissolution to promote particle loss. However, these studies mostly focus on single factors (e.g., seepage pressure) and ignore engineering disturbance, leaving the joint influence of moisture content and disturbance load on the coupling process unclear. Meanwhile, the academic community has accumulated considerable achievements in the field of rock mass permeability characteristics, covering aspects such as seepage theory and experimental methods (Yin et al., 2020).

However, there are still obvious shortcomings in existing research: Firstly, the difficulties in obtaining in-situ samples from weakly consolidated rock mean that experiments often rely on naturally fractured rock, resulting in highly variable mechanical and seepage properties of samples that make it difficult to establish systematic patterns (Mashinskii, 2014). Secondly, seepage in weakly consolidated rock belongs to the category of “variable-mass seepage,” and traditional seepage theories do not fully account for the dynamic impact of mass loss on pore structure, thus failing to accurately describe the evolution of its permeability characteristics (Kang et al., 2025). Thirdly, the coupling mechanism between “moisture content-disturbance load-permeability characteristics” remains poorly understood, hindering targeted disaster prevention and control in engineering practice (Sun et al., 2025; Wu et al., 2025c).

Given this, this paper aims to reveal the evolutionary patterns of permeability characteristics in weakly consolidated rock (Ma et al., 2022; 2023). Geotechnical remolding methods were employed to prepare samples, ensuring uniform mechanical properties by controlling the Talbot grading coefficient and cement content, thereby eliminating the interference of natural rock mass variability (Pan et al., 2024). A multivariate experimental system was established using THM triaxial seepage coupling equipment, introducing the “variable-mass seepage” perspective to monitor particle loss and pore evolution (Wu et al., 2021). The study aimed to clarify the quantitative effects of moisture content and disturbance loads on permeability characteristics, providing scientific basis for preventing and controlling engineering seepage hazards in weakly consolidated rock (Wu et al., 2024; Liu Z. et al., 2025; Xiao et al., 2025). It also contributes to enriching the theoretical framework of seepage behavior in low-strength consolidated rock.

2 Materials and methods

2.1 Sample preparation

The inherent characteristics of weakly consolidated rock with loose, fragmented, and low-strength properties make in-situ sampling extremely challenging. To obtain the weakly consolidated rock with mechanical properties similar to those of fractured faulted rock, geotechnical methods will be employed to prepare remolded samples. Cylindrical standard samples with a diameter of 70 mm and a height of 140 mm will be prepared to meet the compatibility requirements of subsequent mechanical and seepage testing equipment. The total mass of the rock skeleton is estimated to be 770 g.

The rock samples prepared in this paper all comprise a composite skeleton of graded-size mixed gangue (containing quartz, kaolinite, loess, etc.), where coarse-grained mixed gangue serves as the primary load-bearing structure, while fine-grained gangue particles fill voids to enhance mechanical properties. Silicate composite cement was selected as the binder, utilizing its hydration process to cement the loose gangue. Additionally, gypsum was added to appropriately delay the cement’s setting time while enhancing the strength of the hardened cement matrix. Target samples with varying strengths were achieved by adjusting the gradation ratio of the gangue skeleton and the content of the cement-gypsum binder.

Calculate the proportion of each component using the Talbot grading formula:

$P={\left(d/d_{\max }\right)}^n\times 100\%(1)$

In the formula: $d$ is the maximum particle size of the component; $d_{\max }$ is the maximum particle size; $n$ is the Talbot grading coefficient.

The Talbot coefficient n was sequentially set to 0.4, 0.5, 0.6, and 0.7. The calculated mass percentage of mixed gangue within each particle size range is shown in Figure 1. When preparing the binder, a cement-to-gypsum mass ratio of 2:1 yields relatively uniform setting times and appropriate setting duration. Rock samples with two levels of cementation were obtained by adding 150 g and 200 g of cement, respectively. The mass composition of each component is illustrated in Table 1:

Figure 1

Figure 1. Cumulative proportion of particle size components with different Talbot Indexes.

Table 1

Table 1. Composition quality of weakly consolidated rock samples.

The sample preparation process involves six steps: crushing and grading, mixing the slurry, pouring into molds, shaking to homogenize, naturally solidifying, and demolding and curing. The detailed procedures are illustrated in Figure 2.

Figure 2

Figure 2. Remodeling process of weakly consolidated rock samples.

This paper employs a specially designed cylindrical plastic mold with a base but no cover for sample preparation, as illustrated in Figure 2. This mold enables repeated disassembly by adjusting the tightness of six symmetrically arranged screws. Compared to the metal casting molds, the specially designed plastic mold not only facilitates easier disassembly but also minimizes damage to the solidified samples during demolding. After demolding, the sample surface is smoother and flatter, requiring only light polishing before subsequent mechanical and seepage property testing.

The detailed procedures for preparing the weakly consolidated rock samples are as follows:

1. Crush the fully mixed gangue and finely screen it into various particle sizes: 0–2.5 mm, 2.5–5 mm, 5–7.5 mm, 7.5–10 mm, and 10–15 mm. Weigh the corresponding amounts of gangue, cement, and gypsum.

2. Pour materials into a mixing bucket. Add water in multiple small increments while continuously stirring until the mixture forms a distinct slurry state. Maintain stirring to ensure thorough homogeneity and delay the solidification time.

3. Lubricate the bottom and sides of the mold to reduce adhesion between the slurry and plastic walls. This prevents sample damage and extensive mold adhesion during demolding while facilitating subsequent removal. Pour the mixed slurry into the mold.

4. Place the mold immediately onto a vibrating table after molding. Initially, vibrate at low frequency and high amplitude to expel air bubbles from the slurry. Once the sample surface no longer exhibits bubbling, switch to high frequency and low amplitude for gentle vibration to ensure a smooth surface and uniform internal liquid distribution.

5. After vibration, place the sample in a cool, moderately humid location to cure naturally.

6. Once the slurry solidifies, demold and label the sample to obtain a weakly consolidated rock sample. Transfer it to a curing chamber maintained at 35 °C and 98% relative humidity for 7–10 days. After curing, remove and store in a cool, dry location.

2.2 Test equipment

To research the evolution patterns of deformation and permeability characteristics in weakly consolidated rock under different stress loading paths and to visually demonstrate the scale effects of variable-mass seepage tests, this paper constructed an experimental system centered on the THM triaxial seepage coupling apparatus for weakly consolidated rock. The experimental system primarily comprises: an axial pressure control system (a), a confining pressure control system (b), a variable-mass seepage test chamber for weakly consolidated rock (c), a seepage pressure control system (d), a data acquisition center (e), and a flow monitoring system (f). The variable-mass seepage test system for weakly consolidated rock is illustrated in Figure 3.

Figure 3

Figure 3. Experimental system of mass variation seepage of weakly consolidated rock. (a) Axial pressure control system; (b) confining pressure control system; (c) test chamber for variable-mass seepage in weakly consolidated rock mass; (d) seepage pressure control system; (e) computer data recording center; (f) flow monitoring system.

The variable-mass seepage test chamber for weakly consolidated rock is the core component of the entire apparatus. On the left side of the pressure loading frame are two hydraulic pumps: the axial compression pump and the confining pressure pump. Both pumps draw purified water from the same reservoir, applying pressure to the loading frame through preset parameters and control valves. The right side houses the injection pump, connected to the injection port via a water pipe. The drainage port at the bottom of the pressure loading frame connects to a conduit that directs seepage water into a graduated cylinder. Water volume is then determined by monitoring changes in the electronic scale reading.

2.3 Test procedures

1. Prepare the sample Take out the samples which are soaked for different durations. Lightly sand the surface with sandpaper to make it as smooth and flat as possible, ensuring the axial stress applied to the sample is balanced and stable. Select a plastic sleeve with a diameter comparable to the sample’s diameter. Cut a section slightly taller than the sample height. Place the sample into the plastic sleeve, positioning permeable plates at both the top and bottom ends. Place the prepared sample flat on the base. Secure the joint with multiple rubber bands. Finally, wrap 2-3 rubber bands around the entire plastic sleeve and base, applying 3-4 turns per band. The rubber bands and tape prevent water leakage through gaps between the sample and plastic sleeve.

2. Check and adjust equipment. Press the “Flange Up” button to raise it slowly. Avoid contact with pipelines during ascent until the cylinder approaches the flange, ensuring tight contact between the cylinder and lower flange. Connect the corresponding process pipelines, tighten all six fastening bolts, and complete the connection and debugging between the field equipment and the control computer. Subsequently, close the air valve, perform hydraulic oil pressurization and degassing, inspect the airtightness of the test apparatus, ensure the sealing integrity of the permeation pathways, and prepare for the experiment.

3. Suction of the pressure supply equipment. Activate the water pump to ensure the pressure supply equipment can deliver sufficient load.

4. Sample preloading. Activate the axial pressure and confining pressure supply systems. Set the required test parameters: first select constant pressure mode on the confining pressure pump and set the parameter values, then apply load to the sample until the target load is reached.

5. Seepage pressure loading. Open the seepage inlet valve and activate the seepage water pressure supply system to apply seepage water pressure to the sample sidewalls.

6. Adjust pump pressure at critical time points according to the predetermined stress path loading scheme.

7. Record data. While the test is running, in addition to data recorded by the computer software, also record the flow rate at the seepage outlet. Direct the outlet water into a graduated cylinder placed on an electronic scale. Manually record the scale readings every 20 s.

8. Test completion. First shut off the water pressure supply device and cease loading. Open the exhaust valve and drain valve to relieve pressure, discharging the liquid from the pressure chamber. After all water has drained from the testing apparatus, lower the sample for photography. Download data and clear the site.

2.4 Test scheme

Before conducting triaxial seepage tests, the soaking duration for each group of samples must also be determined based on prior test results. To ensure significant variations in water content among rock samples within each group, they were soaked for 0, 2, and 10 min respectively. Besides, to enhance the mass effect of the rock mass, ensure easier observation of experimental phenomena, and obtain more reliable data, the sample group with the lowest uniaxial compressive strength from the test results was selected for the seepage test.

Furthermore, to investigate the impact of mining disturbance on the variable-mass seepage characteristics of weakly consolidated rock, two additional stress paths with different disturbance intensities were designed, referencing relevant studies on stress distribution patterns in target rock masses under mining disturbance. These were compared to a control group subjected to slow loading without disturbance. DZ indicates no disturbance, DR indicates low-intensity disturbance, and GR indicates high-intensity disturbance.

The specific experimental plan is illustrated in Table 2.

Table 2

Table 2. Triaxial mass variation seepage test scheme.

The different stress paths exhibit distinct load variation amplitudes and final stable values, as illustrated in Figure 4.

Figure 4

Figure 4. Diagram of different loading stress paths. (a) None disturbance (b) slight disturbance (c) high disturbance.

3 Test results and analysis

3.1 Soaking results for water-bearing weakly consolidated samples

To obtain the variable-mass seepage characteristics of weakly consolidated rock samples under different stress paths at varying moisture contents, the samples must be submerged in water in advance. To achieve rock samples with moisture contents of 0%, 7%, and 15%, they were fully immersed in water and left undisturbed for 0, 8, and 15 min respectively before removal. The weight changes before and after immersion were then measured. The specific calculation results are illustrated in Table 3.

Table 3

Table 3. Results of rock samples immersed in water before triaxial seepage test.

3.2 Evolution law of seepage characteristics

3.2.1 Characteristic variables

Through conducting a series of experiments, a large amount of raw data can be obtained. Therefore, it is necessary to further extract, process, and derive calculations from the raw data by integrating relevant physical definitions and fundamental mathematical theories. This process aims to obtain variables capable of characterizing the variable-mass seepage characteristics of weakly consolidated rock, followed by final analysis and interpretation. The characterization methods for the key variables involving fundamental physics, solid mechanics, and fluid mechanics that require attention during the experiments are described as follows.

3.2.1.1 Axial strain

For rock materials, being stressed causes deformation, which is typically characterized by strain ε:

$\epsilon =\underset{L\to 0}{\lim }\left(\frac{\Delta L}{L}\right)(2)$

In the formula, $L$ is the length before deformation; $\Delta L$ is the elongation of the rock material.

3.2.1.2 Flow velocity

In the triaxial seepage test, the fluid volume collected at the outlet of the collection device is denoted as $Q_i$. Based on the definition of fluid velocity, the velocity of fluid movement within the seepage chamber $q_i$ can be calculated using the following formula:

$q_i=\frac{Q_i}{\overline{A}}(3)$

In the formula, $\overline{A}$ is the average cross-sectional area through which the fluid flows.

3.2.1.3 Particle mass loss speed

$v_p\left(t_i\right)=\frac{m_p\left(t_i\right)-m_p\left(t_{i-1}\right)}{t_i-t_{i-1}}(4)$

In the formula, $m_p\left(t_i\right)$ is the cumulative mass loss at time, $m_p={\displaystyle \sum _{i=1}^n}\Delta m_i$ ; $v_p\left(t_i\right)$ is the average particle loss speed during this time interval.

3.2.1.4 Porosity

The variation in rock sample porosity is influenced by both deformation compression and mass loss. The porosity at time t_i can be calculated using the following formula:

${\varphi }_{ti}=\Delta {\varphi }_t+{\varphi }_{c0}=1-\frac{1-{\varphi }_{c0}}{1-{\epsilon }_{vi}}+\frac{m_i}{{\beta }_s{V}_0}(5)$

In the formula, ${\varphi }_{ti}$ is the rock mass porosity at time t_i; $\Delta {\varphi }_t$ is the total deformation caused by deformation and erosion, $\Delta {\varphi }_t=\Delta {\varphi }_{ci}+\Delta {\varphi }_{ei}$ ; ${\varphi }_{c0}$ is the rock mass porosity prior to deformation, ${\varphi }_{c0}=\frac{V_v}{V_s+V_v}$ , ${\epsilon }_{vi}$ denotes the volume creep generated during stabilization, ${\epsilon }_{vi=}\left(1-2{\nu }_c\right)$, ${\nu }_c$ is the Poisson’s ratio of the rock mass in the elastic deformation stage.

3.2.1.5 Permeability

Permeability can be derived from Darcy’s Law and calculated using the following formula:

$k=\frac{Q\mu L}{A\Delta P}(6)$

3.2.1.6 Non-Darcy flow control equation

In general, an increase in the seepage pressure difference across the two ends of a medium tends to make the relationship between fluid flow velocity and pressure difference exhibit nonlinearity. Many researchers have achieved remarkable results in studying the laws of fluid seepage. This manuscript adopts the Forchheimer equation, which describes nonlinear seepage, to characterize the relationship between the hydraulic gradient x and the mixed flow velocity y:

Based on previous research, empirical formulas can be used to describe the relationship for non-Darcy factor β as follows:

$\beta =\frac{{\beta }_0}{k\varphi }(7)$

In the formula, ${\beta }_0$ is the non-Darcy coefficient.

3.2.2 Evolution law of seepage characteristics

Based on the raw data and incorporating the definitions and calculation methods for each physical parameter outlined in Section 3.2.1, the evolution curves for the average flow velocity, axial strain, particle loss speed, porosity, permeability, and non-Darcy coefficient of the submerged sample during the triaxial disturbance load variable-mass seepage test can be obtained, as illustrated in Figure 5.

Figure 5

Figure 5. Evolution of fluid and solid mechanics characteristics of weakly consolidated rock. (a) Axial strain (b) Average flow velocity (c) Particle mass loss speed (d) Porosity (e) Permeability (f) Non-Darcy coefficient.

Figure 5a illustrates the variation characteristics of the average seepage velocity within the sample. It can be observed that the flow velocity exhibits an overall increasing trend over time. For samples subjected to static loading, the internal flow velocity initially increases slowly for a period before accelerating significantly and then stabilizing at a certain point. Under dynamic loading, the internal flow velocity exhibits opposite variations to the load changes: flow velocity decreases as load stress increases and recovers as stress decreases. Furthermore, after exposure to perturbed loading, the subsequent phase of flow velocity growth becomes more rapid.

Figure 5b illustrates the axial deformation patterns of each sample during seepage testing. All samples continuously deformed throughout the test, macroscopically manifesting as progressive height compression. Under static loading, initial deformation occurred rapidly before the strain rate gradually slowed. Under dynamic loading, specimen strain fluctuated inversely with load changes, exhibiting a brief surge followed by gradual deceleration after disturbance.

Figure 5c illustrates the particle loss speeds collected during each test group’s migration out of the system. The figure reveals pronounced liquefaction particle migration occurring in the latter stages of testing, while both static loading conditions and dynamic loading perturbations exerted no significant influence on particle loss speeds. At a certain point during testing, particle loss speeds exhibited a breakthrough increase and continued to rise for a period before gradually decreasing over a short timeframe and ultimately approaching zero.

Figure 5d visually illustrates the evolution of porosity across samples. During the early test phase, sample porosity exhibited a gradual decrease, inversely correlated with load magnitude. In the later phase, porosity rapidly increased to two to three times the initial value before stabilizing at a relatively constant level.

Figure 5e shows that the trend in rock permeability closely resembles the average fluid velocity curve. During the early stage of the test, rock permeability fluctuates within a small range in response to load magnitude. At a certain intermediate point, it begins to rise rapidly, stabilizing at the current level or experiencing a slight decline after reaching its peak.

Figure 5f illustrates the evolution of the rock mass’s non-Darcy factor. It is observed that the non-Darcy factor exhibits an inverse relationship with permeability, declining significantly from the early test phase until it stabilizes at a relatively constant value.

Since the variation patterns of the aforementioned physical quantities are related to the loading stages of the sample, the entire test process can be divided into three stages: the disturbance period, the response period, and the post-disturbance period.

3.2.2.1 The disturbance period

The period from 0 to 150 min after the start of the test is defined as the disturbance period. During this period, the overall change in the average velocity of the internal fluid within each sample remains relatively stable, with the increase maintained below 2%. The disturbance load applied to the rock mass induces three distinct waves of flow velocity with increasing amplitude. Concurrently, the axial strain growth rate of all rock specimens accelerates significantly during this period, with strain increments exceeding 50% of the final deformation. The strain curves of non-disturbance-loaded rock masses exhibit stronger linear characteristics, indicating a greater tendency toward elastic strain during deformation. In contrast, disturbance loading induces synergistic strain-load relationships, with observable increases in strain fluctuation amplitude. This suggests that after stress reduction, subsequent stress elevation accelerates rock mass compaction. Regarding particle loss speeds, particle migration within samples during this period was extremely limited, with mass loss speeds not exceeding 0.25 mm/s. This indicates that variable-mass seepage was essentially absent within weakly consolidated rock at this stage, indirectly confirming that deformation—rather than particle migration—was the primary driver of porosity changes during this period. Therefore, during the disturbance period, rock porosity changes inversely to strain trends. As the rock mass continuously compacted during the initial test phase, internal voids were progressively compressed. Consequently, porosity decreased steadily regardless of whether subjected to disturbance or non-disturbance loads. However, this deformation-induced porosity change was limited, reducing porosity by a maximum of 0.03. The fluctuation range of rock porosity under disturbance loads also follows a pattern similar to strain fluctuations, with each disturbance causing a more pronounced fluctuation than the previous one. During this period, the permeability of the rock mass was primarily influenced by its geometric deformation. However, due to the incomplete development of permeable pathways within the weakly consolidated rock at the beginning of the test, it contained a large amount of ineffective voids. These void spaces, acting as compressed regions, did not inherently inhibit the existing stable seepage pathways. Consequently, although the porosity of the sample decreased, the permeability of the rock mass exhibited a slight increase as seepage stabilized and variable-mass seepage continuously acted upon it. Correspondingly, the non-Darcy factor also decreased rapidly during this stage. In the early test phase, due to unstable flow patterns and disordered, random channel evolution, fluid flow exhibited pronounced nonlinear characteristics.

3.2.2.2 The response period

The period between 150 and 180 min into the test constitutes the response period. By this stage, the disturbance load has fully impacted the rock mass, reaching the high stress conditions specified in the test design and stabilizing. During this period, the average flow velocity of all samples exhibits a transition from decreasing to increasing. This shift occurs because the sustained high load further restricts the fluid flow space, thereby reducing flow velocity. Simultaneously, the high stress intensifies fluid-solid interactions within the rock mass, accelerating erosion and transport of rock particles by the water flow. Particle migration promotes the development, extension, and connectivity of rock fractures, ultimately facilitating accelerated fluid flow. The strain curve during this period transitions from significant growth to gradual increase, indicating the rock mass has entered the yield stage under external high-load conditions. Particle loss speeds across samples began increasing to varying degrees, ranging from 0.25 to 2.5 g/s—representing 1 to 10 times the initial values at this stage. Rock porosity also shifted from gradual decline to rapid increase, with varying transition points, indicating that both volumetric deformation and particle loss jointly governed porosity changes during this period. During this period, the fissure networks within weakly consolidated rock, having developed in the preceding stage, now exhibit rudimentary channel networks. Due to the sharp increase in particle loss speeds, these networks expand further, leading to a rapid rise in rock permeability—typically exceeding twice the initial permeability value. Additionally, as the channels mature rapidly, fluid flow stabilizes, and the nonlinear characteristics of the fluid rapidly diminish. Given the significant and rapid increases in average fluid velocity, particle loss speed, and rock mass permeability during this stage, it represents a critical period for preventing and controlling water inrush disasters.

3.2.2.3 The post-disturbance period

The period from 180 to 330 min after the start of the test is termed the post-disturbance period. During this period, the average flow velocity of each sample rapidly increases and remains relatively stable. By the end of the test, the maximum flow velocity can exceed 6 mm/s, reaching 2 to 3 times the value at the period’s onset; the nonlinear characteristics of the strain curve intensify, with rock deformation primarily dominated by plastic deformation, exhibiting features typical of the strengthening period. Mass loss speeds exhibit varying degrees of increase. For all samples, mass loss speeds surge significantly within a short timeframe to exceed 1 g/min, with some samples approaching 3 g/min. After reaching peak values, mass loss speeds gradually decline to zero. Rock porosity continues its rapid growth trend from the previous stage, reaching a peak where the growth rate slows or even shows a slight decrease. Analysis of particle loss speeds and rock porosity changes indicates that variable-mass seepage effects significantly intensified during this period, with particle migration becoming the primary driver of porosity alteration. This resulted from long-term seepage compression weakening the cohesive bonds between rock structures, allowing fine particles to detach from the matrix, enter the fluid period, and be transported by water flow. The loss of fine particles increases the porosity of weakly consolidated rock, gradually forming internal fluid pathways. This allows more fluid to enter the rock mass, continuously increasing flow velocity. This process further enhances the erosive action of water flow and progressively weakens rock strength. It can be observed that the permeability of the rock mass continues to rise during this period, albeit at a slower rate. As particle loss and fluid velocity gradually stabilize, the permeability eventually reaches a relatively stable state. By this stage, the internal network of seepage pathways has fully developed, and the water-rock interaction tends toward equilibrium. Consequently, the fluid flow regime gradually stabilizes, transitioning from non-Darcy flow toward Darcy flow.

3.3 Influence of moisture content on rock mass seepage characteristics

Figures 3a–f, 6a–f illustrate the evolution of solid mechanics and seepage mechanics variables in weakly consolidated rock samples subjected to triaxial seepage tests under rated stress conditions after different immersion durations. Sample DZ-0 is a dry sample with zero moisture content. DZ-1 reached a moisture content of 6.85% after 8 min of immersion, while DZ-2 approached a saturated state after 15 min of immersion, achieving a moisture content of 14.35%.

Figure 6a illustrates the axial strain curves of cemented rock masses at different moisture contents. The figure reveals that in highly saturated rock, axial strain increments are larger, while the slope reduction of the curve is smaller. This indicates that the rate of axial strain increase in water-saturated rock exceeds that in dry rock. Moreover, the strain curve of highly saturated rock deviates more markedly from linear characteristics over time, exhibiting a nonlinear growth trend. Due to the damage caused by water-induced deterioration, water alters the physical properties of the rock matrix particles and cementing materials, significantly reducing the rock’s inherent strength. This makes the rock more susceptible to greater deformation under external forces. The maximum strain of the saturated sample DZ-2 is approximately 2.75%, representing an increase of about 37.5% compared to the 2% strain of the dry sample DZ-0. Moreover, higher water saturation accentuates plastic deformation characteristics in axial strain. Plastic deformation often accompanies extensive fissure formation, facilitating the expansion of seepage pathways.

Figure 6

Figure 6. Evolution of variates under constant load with different water contents. (a)Axial strain (b)Average flow velocity (c)Particle mass loss speed (d)Porosity (e)Permeability (f)Non-Darcy coefficient.

Figure 6b illustrates the average flow velocity across three water-saturated states of weakly cemented rock. The figure indicates that higher initial moisture content enables greater initial fluid velocity during the early testing phase. As the test progressed, the average flow velocity of high-water-content specimens exhibited more pronounced increases. The nearly saturated DZ-2 specimen’s average flow velocity rose from an initial 1.20 mm/s to a final 3.25 mm/s, representing a 2.8-fold increase from the initial velocity. In contrast, the flow velocity of the completely dry DZ-0 specimen increased only from 0.75 mm/s to 1.75 mm/s, representing a mere 2.3-fold increase. This suggests that when rock masses are not saturated, the formation and development of internal seepage pathways are relatively difficult. The velocity increase observed upon saturation is attributed to the rock mass’s inherently fragile strength, determined by its loose, weakly cemented structure. Upon wetting, the rock skeleton softens, and the cementing material undergoes erosion and redistribution. This accelerates the evolution and development of pre-existing spatially advantageous seepage pathways, creating a more conducive environment for rapid and stable fluid flow.

Figure 6c illustrates the variation of particle mass loss speed during the seepage in rock bodies with different water contents. The mass-dependent seepage phenomenon is initially inconspicuous in the early stages of the test, with only one minor peak occurring. This is because stresses are primarily transmitted from the rock body to internal voids, causing continuous void compression while water-rock interactions remain relatively weak. After a certain point, the particle loss speed surged dramatically. Higher water content correlated with steeper curve slopes, indicating greater mass loss per unit time. The saturated sample DZ-2 exhibited a maximum particle loss speed of 2.5 g/min^-1, which was 2.2 times the peak rate of the unsaturated sample DZ-0. Therefore, it can be inferred that water immersion causes initial damage within the rock mass. This damage significantly increases the likelihood of skeleton particle liquefaction and migration, promoting the occurrence of variable-mass seepage in the later stages of the test. Consequently, it becomes a crucial factor influencing the variable-mass permeability characteristics of the rock mass.

Figure 6d illustrates the influence of moisture content on the evolution of rock porosity. Under a continuously applied steady-state rated load, all rock specimens exhibit a porosity pattern characterized by an initial gradual decrease followed by a sudden increase. The magnitude of moisture content exerts a more pronounced influence on rock porosity during the middle and late stages compared to the initial phase, generally following the pattern that higher moisture content correlates with greater porosity variation. Notably, the saturated sample DZ-2 exhibited a final porosity of 0.3—double its initial value—while the dry sample DZ-0 reached only 1.1. The absence of particle loss during the initial test phase meant that porosity changes were primarily driven by rock strain. As particle loss speeds increased, rock porosity also rose significantly, indicating that porosity evolution in this stage became dominated by particle loss. When water-rock interactions reached equilibrium and variable-mass seepage gradually ceased, rock porosity exhibited a slight decrease due to stress effects.

Figure 6e illustrates the permeability evolution patterns for the three rock moisture contents. The rock with the highest moisture content exhibited greater permeability at the start of the test.

Figure 6f illustrates the variation of the non-Darcy factor during seepage in water-saturated rock masses. The higher the rock mass water content, the lower the non-Darcy factor observed in the experiments. Compared to water-saturated rock, seepage initiation is more difficult in dry rock, resulting in a more random and disordered mathematical expression of seepage variables.

3.4 Influence of disturbance load on rock mass seepage characteristics

Figures 7a–f illustrate the evolution patterns of solid mechanics and seepage mechanics variables in triaxial seepage tests conducted on weakly cemented rock masses under different disturbance load conditions.

Figure 7

Figure 7. Evolution of variates of absolute dry rock under different stress loading paths. (a) Axial strain (b) Average flow velocity (c) Particle mass loss speed (d) Porosity (e) Permeability (f) Non-Darcy coefficient.

Figure 7a illustrates the axial strain curve for dry rock. The figure demonstrates that stress disturbance significantly impacts the axial strain of weakly cemented rock. Higher disturbance stresses in rock masses result in greater strain levels. It can be observed that disturbance stresses transform the rock mass’s expected elastic deformation phase into a fluctuating rise, and each stress reduction fails to fully restore the rock deformation. Under the two stress loading paths, the final strains for GR-0 and DR-0 are 2.5% and 2%, which are 66.6% and 50% increases compared to the undisturbed case. Compared to the rated load, the perturbed load allows the rock mass to recover some deformation during stress release. However, during the next stress concentration event, it induces greater deformation and damage. Microscopically, this manifests as fracture development and void expansion; macroscopically, it leads to reduced rock strength, localized fatigue failure, and consequently, larger deformations.

Figure 7b illustrates the average flow velocity of weakly cemented rock under three stress path loading conditions during triaxial seepage tests. The figure reveals that rock subjected to high disturbance stress exhibits higher initial flow velocity at the onset of seepage, with noticeable fluctuations in fluid velocity corresponding to variations in load magnitude. Higher disturbance loads result in steeper slopes during the upward phase of the curve. The GR-0 sample achieved the fastest flow velocity of 3.75 mm/s, which is 3.5 times the peak velocity of the DZ-0 sample. When the rock mass was repeatedly disturbed, the internal rock skeleton was prone to minor irreversible damage. The stability of the fragile cemented structure was continuously compromised. During deformation and recovery, the cementing material became fluidized and was flushed out by water, altering the rock mass’s constitutive behavior. The greater the amplitude of the disturbance stress, the more pronounced these effects became.

Figure 7c illustrates the changes in particle mass loss speed during seepage. Similarly, particle loss speed was initially low, with only GR-0 exhibiting a slight increase during disturbance stress variations. Subsequently, all three specimens experienced a surge in particle loss speed, where higher initial disturbance stress correlated with greater loss speeds and cumulative particle mass loss. The peak particle loss speed for the highly disturbed sample GR-0 reached 1.25 g/min^-1, while the undisturbed sample DZ-0 recorded only 0.8 g/min^-1, representing a growth rate of approximately 56.2%. Therefore, it can be inferred that the stress disturbance also causes damage within the rock mass. The concentration and release of stress lead to more rock particles detaching from their original bonds and accumulating in storage fractures. When these fractures expand into water-conducting channels and fluid velocity increases, these fine particles are flushed out by water, resulting in significant particle loss during the mid-to-late stages of the test.

Figure 7d illustrates the effect of perturbed stress on rock mass porosity evolution. Regardless of whether a steady rated load or a perturbed load was applied externally, the rock mass porosity exhibited a pattern of initially slow decline followed by a sudden increase. Similarly, the impact of external disturbance stress on rock porosity becomes more pronounced during the mid-to-late stages compared to the initial phase. Overall, a higher magnitude of stress variation correlates with a greater change in porosity. For instance, the porosity of the highly disturbed specimen GR-2 ultimately increased to 0.2, representing an 1.8-fold increase from its initial porosity, whereas the undisturbed specimen DZ-0 exhibited only a 1.3-fold increase. The absence of particle loss during the initial test phase meant that porosity changes were primarily driven by rock strain. As particle loss speeds increased, rock porosity also rose significantly, indicating that porosity changes in this stage were dominated by particle loss. When water-rock interactions reached equilibrium and variable-mass seepage gradually ceased, rock porosity exhibited a slight decrease due to stress effects.

Figure 7e illustrates the permeability evolution patterns for the three rock moisture contents. The rock with the highest moisture content exhibited greater permeability at the start of the test.

Figure 7f illustrates the variation of the non-Darcy factor during seepage in water-saturated rock masses. The higher the rock mass water content, the lower the non-Darcy factor observed in the experiments. Compared to water-saturated rock, dry rock exhibits greater difficulty in initiating seepage, with its seepage variables exhibiting more random and disordered mathematical relationships.

3.5 Future research directions

This study clarifies the effects of moisture content and disturbance load on weakly consolidated rock seepage, while follow-up work can be expanded in three aspects:

1. Integrate temperature factors with current findings, using a THM triaxial seepage apparatus to explore synergistic effects of temperature, moisture content, and disturbance load on permeability, establishing a full THM coupling seepage model.

2. Combine microscopic techniques to verify the link between microscale pore evolution and macroscopic permeability, supplementing multiscale mechanism support.

3. Conduct long-term seepage tests to monitor long-term permeability changes, providing a basis for engineering long-term seepage hazard prediction.

4 Discussion

This study identifies three distinct stages in the seepage evolution of weakly consolidated rock masses, namely, the disturbance period, response period, and post-disturbance period, each governed by distinct dominant mechanisms. Specifically, consolidation deformation governs the disturbance period, particle migration acts as the core driver during the response period, and channel stabilization characterizes the post-disturbance period. Experimental results confirm that particle migration serves as the direct driver of pore structure reconstruction and permeability fluctuations. Concurrently, the non-Darcy factor exhibits a significant negative correlation with permeability: as seepage processes stabilize, the non-Darcy factor decreases, inducing a transition in the flow regime from strong nonlinearity to quasi-linearity. These findings elucidate the intrinsic coupling relationship between mass loss, pore evolution, and permeability variation in variable-mass seepage, addressing the inadequacies of traditional seepage theories in characterizing this complex interrelationship.

1. In the context of existing research on weakly consolidated rock seepage, this study aligns with the core consensus that moisture content and stress disturbance are pivotal external factors regulating seepage characteristics (Sellers and Klerck, 2000; Wang and Park, 2003). Both this work and prior studies recognize the cyclic mechanism of “consolidation weakening–particle migration–flow pathway evolution,” wherein the attenuation of consolidation strength accelerates particle migration, thereby facilitating the formation and expansion of flow pathways (Zhao et al., 2022). Additionally, relevant investigations have consistently observed non-Darcy flow phenomena that deviate from Darcy’s law (Hansbo, 1997). However, this study advances beyond previous research in three critical aspects: first, it shifts the analytical focus from single-factor assessments to the exploration of the coupled effects of moisture content, disturbance load, and permeability, quantitatively delineating their synergistic impacts on seepage evolution; second, unlike existing studies that merely describe macroscopic permeability trends, this research refines the division of seepage phases by explicitly defining the dominant mechanism of each stage; third, it replaces qualitative recommendations for seepage hazard prevention with quantitative critical control thresholds for moisture content and disturbance load, thereby enhancing the practical applicability of the research outcomes.

2. The theoretical contribution of this study resides in refining the analytical framework for variable-mass seepage in weakly consolidated rock masses. By establishing the three-stage seepage model and clarifying the regulatory roles of particle migration and non-Darcy factor evolution (Abbas et al., 2014; Li et al., 2019), this work fills the gaps in traditional theories that overlook the dynamic effects of mass loss. Methodologically, geotechnical remolding techniques are adopted; within this approach, grading coefficients and binder content are precisely controlled to fabricate specimens with homogeneous mechanical properties, and this eliminates the interference arising from the high variability of natural rock samples. Furthermore, the multi-parameter synchronous monitoring system, which captures axial strain, seepage velocity, particle loss rate, and porosity, avoids the limitations of single-parameter monitoring, thereby improving the reliability of experimental data and the comprehensiveness of mechanism analysis. For engineering applications, the identification of the response period as a critical window for water inrush prevention, coupled with the quantification of critical control values, provides actionable indicators for seepage hazard mitigation in deep tunnel engineering, facilitating a paradigm shift from post-event remediation to pre-event prevention.

3. This study also has inherent limitations. First, it does not incorporate temperature as a variable; in actual deep engineering environments, temperature variations can alter the pore structure and seepage characteristics of rock masses, resulting in conclusions that cannot fully encompass complex high-temperature scenarios (Liu X. et al., 2025; Tomás et al., 2025; Wang B. et al., 2025). Second, the verification of microscale mechanisms is insufficient: inferences regarding pore evolution and particle migration rely solely on macroscopic parameters, lacking corroboration from micro-observational data, which constrains the depth of explanatory analysis. Third, the experimental duration is relatively short, failing to capture the long-term seepage evolution patterns of weakly consolidated rock masses during extended engineering service cycles, thus precluding the provision of comprehensive evidence for long-term engineering safety assessments.

5 Conclusion

1. The evolution of permeability characteristics in weakly consolidated rock masses exhibits distinct periods, which can be divided into three stages: disturbance period, response period, and post-disturbance period. During the disturbance period, consolidation deformation predominates, with porosity gradually decreasing while permeability remains relatively stable. The response period sees significantly enhanced particle migration and pore expansion, marked by a simultaneous sharp increase in flow velocity, permeability, and particle loss speed—representing the critical stage for flow regime transition. In the post-disturbance period, flow pathways become fully connected, fluid flow patterns stabilize, and permeability reaches its peak before maintaining a quasi-steady state.

2. Water content is the primary factor governing permeability characteristics in weakly consolidated rock masses. As water content increases, cement hydration and dissolution intensify, significantly reducing inter-particle cohesion and transforming the structure from dense to porous. In high-water-content specimens, both porosity and permeability markedly increase, flow velocity grows more rapidly, and particle loss becomes more severe. Water-induced softening plays a dominant role in permeation instability and significantly alters the development rate of flow pathways within the rock mass.

3. Stress perturbations enhance the permeation process in weakly consolidated rock masses. Under high-perturbation loading conditions, alternating opening and closing of fractures repeatedly damages the internal skeletal structure, leading to greater fluctuations in porosity. The larger the perturbation amplitude, the more pronounced the increase in permeability and the more active the particle migration. Experiments demonstrate that disturbance stresses can increase permeability by 1.5–2 times compared to static loading conditions, making them a crucial external factor inducing flow discontinuities.

4. Pore volume, permeability, and particle loss processes exhibit consistent evolutionary characteristics. Pore volume slightly decreases at the experiment’s onset, rapidly increases during the middle phase, and stabilizes in the late phase. Permeability grows synchronously with pore volume, peaking when particle loss reaches its maximum. Particle migration serves as the direct driving force for pore structure reconstruction and is the fundamental cause of permeability fluctuations. Variable-mass flow effects significantly intensify during the middle and late phases, promoting the transformation of rock mass from closed pores to interconnected channels.

5. Variations in the non-Darcy factor reveal the nonlinear evolution mechanism of seepage. During early seepage stages, flow turbulence and inertial effects are prominent, resulting in a high β value. As conduits form and flow stabilizes, the β value rapidly decreases, transitioning fluid motion from strongly nonlinear to quasi-linear. The non-Darcy factor exhibits a significant inverse relationship with permeability, serving as an indicator for characterizing the nonlinearity intensity of seepage in weakly consolidated rock masses.

6. Engineering significance and field advancements: For seepage hazard prevention in weakly consolidated rock engineering such as deep tunnels, this study offers three practical references. First, strengthen monitoring in the response period which is critical for water inrush. Second, control surrounding rock moisture content below 7% via pre-excavation dewatering. Third, limit excavation disturbance amplitude to axial pressure no more than 6 MPa and confining pressure no more than 4 MPa. Compared with prior studies that focus on passive grouting or qualitative control, this research quantifies the effects of moisture content and disturbance load, provides operable thresholds, fills the “quantitative guidance gap” in existing strategies, and promotes seepage control from “remedial” to “preventive”.

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author contributions

PX: Writing – original draft, Writing – review and editing. JJ: Writing – review and editing. HS: Conceptualization, Data curation, Writing – review and editing. GW: Formal Analysis, Funding acquisition, Writing – review and editing. XiL: Investigation, Methodology, Writing – review and editing. WF: Project administration, Resources, Writing – review and editing. XuL: Software, Supervision, Validation, Writing – review and editing.

Funding

The authors declare that financial support was received for the research and/or publication of this article. This work was supported by the National Natural Science Foundation of China (Nos. 52074259, 52174092) and the Yunlong Lake Laboratory of Deep Underground Science and Engineering Project (No. 104023002). The authors gratefully appreciate this support. This work is supported by the Major Science and Technology Projects of China Coal, China (20231BY001).

Conflict of interest

Authors PX, JJ, HS, GW, XiL, WF, and XuL were employed by China Coal No.5 Construction Co., Ltd. and China Coal Construction Group Limited Corporation.

Generative AI statement

The authors declare that no Generative AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Abbas, M., Magaud, P., Gao, Y., and Geoffroy, S. (2014). Migration of finite sized particles in a laminar square channel flow from low to high reynolds numbers. Phys. Fluids 26, 123301. doi:10.1063/1.4902952

CrossRef Full Text | Google Scholar

Chen, Z., Huang, L., Li, X., Liu, Z., and Chen, J. (2024). Numerical investigations of mechanical behavior and permeability evolution in mine geothermal reservoir with blasting fractures using fully coupled thermal–hydraulic-mechanical modeling. Appl. Therm. Eng. 252, 123596. doi:10.1016/j.applthermaleng.2024.123596

CrossRef Full Text | Google Scholar

Cui, G., Ning, F., Dou, B., Li, T., and Zhou, Q. (2022). Particle migration and formation damage during geothermal exploitation from weakly consolidated sandstone reservoirs via water and CO2 recycling. Energy 240, 122507. doi:10.1016/j.energy.2021.122507

CrossRef Full Text | Google Scholar

Hansbo, S. (1997). Aspects of vertical drain design: darcian or Non-Darcian flow. Géotechnique 47, 983–992. doi:10.1680/geot.1997.47.5.983

CrossRef Full Text | Google Scholar

Hossain, Md. E., Yağiş, O., Haseeb, M., Khudoykulov, K., Wibowo, W., and Esquivias, M. A. (2025). Striking the balance in resource management: exploring the impact of natural and mineral resources on financial development in BRICS-T nations. Environ. Sustain. Indic. 28, 100905. doi:10.1016/j.indic.2025.100905

CrossRef Full Text | Google Scholar

Kang, H., Gao, F., Xu, G., and Ren, H. (2023). Mechanical behaviors of coal measures and ground control technologies for China’s deep coal mines – a review. J. Rock Mech. Geotech. Eng. 15, 37–65. doi:10.1016/j.jrmge.2022.11.004

CrossRef Full Text | Google Scholar

Kang, X., Xie, X., and Zeng, K. (2025). Analysis of semi-analytical solution for two-stage failure of surrounding rock and field test of staged control in water-rich weakly cemented sandstone formation. Tunn. Undergr. Space Technol. 166, 106951. doi:10.1016/j.tust.2025.106951

CrossRef Full Text | Google Scholar

Li, S., Wang, J., Li, L., Shi, S., and Zhou, Z. (2019). The theoretical and numerical analysis of water inrush through filling structures. Math. Comput. Simul. 162, 115–134. doi:10.1016/j.matcom.2018.12.014

CrossRef Full Text | Google Scholar

Liu, X., Tan, J., Wang, L., Song, X., Luo, R., Yu, J., et al. (2025). Impact of supercritical CO2 transport mode in fractured rock on seepage capacity and heat transfer efficiency. Appl. Therm. Eng. 280, 128447. doi:10.1016/j.applthermaleng.2025.128447

CrossRef Full Text | Google Scholar

Liu, Z., Ma, C., Long, S., and Wei, X. (2025). Permeability characteristics of rock subjected to cyclic disturbance loading. Eng. Geol. 356, 108270. doi:10.1016/j.enggeo.2025.108270

CrossRef Full Text | Google Scholar

Lu, C., Cao, X., Liu, H., Guo, S., and Lu, Z. (2024). A unified model for predicting the diffusion distance of cement slurry in weakly consolidated formation considering different diffusion patterns. Constr. Build. Mater. 441, 137484. doi:10.1016/j.conbuildmat.2024.137484

CrossRef Full Text | Google Scholar

Ma, D., Duan, H., Zhang, J., Liu, X., and Li, Z. (2022). Numerical simulation of water–silt inrush hazard of fault rock: a three-phase flow model. Rock Mech. Rock Eng. 55, 5163–5182. doi:10.1007/s00603-022-02878-9

CrossRef Full Text | Google Scholar

Ma, D., Duan, H., Li, Q., Wu, J., Zhong, W., and Huang, Z. (2023). Water–rock two-phase flow model for water inrush and instability of fault rocks during mine tunnelling. Int. J. Coal Sci. Technol. 10, 77. doi:10.1007/s40789-023-00612-6

CrossRef Full Text | Google Scholar

Mao, Y., Zhang, X., Zhang, Z., Tao, Z., He, M., and Zhang, C. (2025). Research on genesis, evolution and prevention of water and mud inrush in the Non-karst tunnel: a case study in Yunnan, southwest China. Tunn. Undergr. Space Technol. 166, 106967. doi:10.1016/j.tust.2025.106967

CrossRef Full Text | Google Scholar

Mashinskii, E. I. (2014). Elastic–microplastic nature of wave propagation in the weakly consolidated rock. J. Appl. Geophys. 101, 11–19. doi:10.1016/j.jappgeo.2013.11.007

CrossRef Full Text | Google Scholar

Nguyen, V. H., Gland, N., Dautriat, J., David, C., Wassermann, J., and Guélard, J. (2014). Compaction, permeability evolution and stress path effects in unconsolidated sand and weakly consolidated sandstone. Int. J. Rock Mech. Min. Sci. 67, 226–239. doi:10.1016/j.ijrmms.2013.07.001

CrossRef Full Text | Google Scholar

Pan, Z., Wang, Y., Zhou, K., Jiang, Z., Saleem, F., and Wang, P. (2024). Evaluation of MSA mortar pore structure characteristics in water-saturated curing environment based on talbot gradation theory. Constr. Build. Mater. 447, 138109. doi:10.1016/j.conbuildmat.2024.138109

CrossRef Full Text | Google Scholar

Sellers, E. J., and Klerck, P. (2000). Modelling of the effect of discontinuities on the extent of the fracture zone surrounding deep tunnels. Tunn. Undergr. Space Technol. 15, 463–469. doi:10.1016/S0886-7798(01)00015-3

CrossRef Full Text | Google Scholar

Sun, W., Shi, G., Liu, Q., Fu, K., Ma, Y., and Liang, B. (2025). Crack growth behavior and stress intensity factor (SIF) prediction in water-saturated sandstone under stress-seepage coupling. Results Eng. 28, 107254. doi:10.1016/j.rineng.2025.107254

CrossRef Full Text | Google Scholar

Tomás, R., Benavente, D., Martínez-Ibáñez, V., and Garrido, M. E. (2025). How do high temperatures affect rock properties? A comprehensive review of experimental thermal effects and underlying mechanisms. Eng. Geol. 357, 108323. doi:10.1016/j.enggeo.2025.108323

CrossRef Full Text | Google Scholar

Wang, J.-A., and Park, H. D. (2003). Coal mining above a confined aquifer. Int. J. Rock Mech. Min. Sci. 40, 537–551. doi:10.1016/S1365-1609(03)00029-7

CrossRef Full Text | Google Scholar

Wang, B., Yang, W., Zhang, X., Yang, Y., Zhang, L., and Yao, J. (2025). Flow behavior of a rough single rock fracture under high-temperature, high-stress, and high-seepage pressure coupling conditions. Int. J. Min. Sci. Technol., S2095268625001508. doi:10.1016/j.ijmst.2025.09.001

CrossRef Full Text | Google Scholar

Wang, J., Xie, X., Li, W., Chen, X., Zhang, X., Fan, N., et al. (2025). Damage-fracture evolution mechanism of rock bridge structures in flawed sandstone under cyclic disturbance: insights from DIC-AE. Theor. Appl. Fract. Mech. 140, 105180. doi:10.1016/j.tafmec.2025.105180

CrossRef Full Text | Google Scholar

Wang, Z., Zhao, G., Xie, Y., Wu, B., and Xia, K. (2025). Permeability enhancement of reservoir rocks loaded by repeated low-amplitude stress waves. Int. J. Rock Mech. Min. Sci. 194, 106243. doi:10.1016/j.ijrmms.2025.106243

CrossRef Full Text | Google Scholar

Wu, F., Zhang, D., Ma, L., Meng, T., Zhao, G., Liu, P., et al. (2021). Thermo-hydro-mechanical (THM) evolution law and development of permeability and pore structure of enhanced geothermal systems at ultra-high temperatures. Geothermics 97, 102253. doi:10.1016/j.geothermics.2021.102253

CrossRef Full Text | Google Scholar

Wu, J., Wong, H. S., Zhang, H., Yin, Q., Jing, H., and Ma, D. (2024). Improvement of cemented rockfill by premixing low-alkalinity activator and fly ash for recycling gangue and partially replacing cement. Cem. Concr. Compos. 145, 105345. doi:10.1016/j.cemconcomp.2023.105345

CrossRef Full Text | Google Scholar

Wu, J., Yang, S., Williamson, M., Wong, H. S., Bhudia, T., Pu, H., et al. (2025a). Microscopic mechanism of cellulose nanofibers modified cemented gangue backfill materials. Adv. Compos. Hybrid. Mater. 8, 177. doi:10.1007/s42114-025-01270-9

CrossRef Full Text | Google Scholar

Wu, J., Zhang, W., Wang, Y., Ju, F., Pu, H., Riabokon, E., et al. (2025b). Effect of composite alkali activator proportion on macroscopic and microscopic properties of gangue cemented rockfill: experiments and molecular dynamic modelling. Int. J. Min. Metall. Mater. 32, 1813–1825. doi:10.1007/s12613-025-3140-8

CrossRef Full Text | Google Scholar

Wu, J., Zhang, W., Yang, S., Jivkov, A. P., Kearsley, E., Pu, H., et al. (2025c). Polycarboxylate superplasticizer instead of ultrasonic treatment for dispersing cellulose nanofibers to strengthen cemented rockfill.

Google Scholar

Xiao, Z., Wang, G., Liu, J., Deng, H., Jiang, Y., Jiang, F., et al. (2025). Experimental study and model improvement on coal permeability: the influence of effective stress, slip effect, and water content. Int. J. Rock Mech. Min. Sci. 185, 105972. doi:10.1016/j.ijrmms.2024.105972

CrossRef Full Text | Google Scholar

Xie, H., Ju, Y., Gao, F., Gao, M., and Zhang, R. (2017). Groundbreaking theoretical and technical conceptualization of fluidized mining of deep underground solid mineral resources. Tunn. Undergr. Space Technol. 67, 68–70. doi:10.1016/j.tust.2017.04.021

CrossRef Full Text | Google Scholar

Yin, W., Zhao, Y., and Feng, Z. (2020). Experimental research on the permeability of fractured-subsequently-filled granite under high temperature-high pressure and the application to HDR geothermal mining. Renew. Energy 153, 499–508. doi:10.1016/j.renene.2020.02.009

CrossRef Full Text | Google Scholar

Zhang, W., Liu, X., and Zhang, J. (2025). Intelligent decision-making of shield tunneling machine using deep reinforcement learning in complex geological environment. Eng. Appl. Artif. Intell. 156, 111237. doi:10.1016/j.engappai.2025.111237

CrossRef Full Text | Google Scholar

Zhao, J., Tan, Z., Wang, X., Zhou, Z., and Li, G. (2022). Engineering characteristics of water-bearing weakly cemented sandstone and dewatering technology in tunnel excavation. Tunn. Undergr. Space Technol. 121, 104316. doi:10.1016/j.tust.2021.104316

CrossRef Full Text | Google Scholar