Experimental study on variable-mass seepage model of water-rich altered granite

To deeply investigate the catastrophe mechanism of water inrush and mud outburst when tunnels traverse water-rich altered granite strata, this paper utilizes a self-developed visualized variable mass seepage test system to conduct physical model tests under varying conditions of water pressure and initial porosity. Unlike conventional geotechnical media, the study finds that the seepage failure of altered granite exhibits significant “variable mass” characteristics. Its catastrophic evolution follows a three-stage cyclic mechanism: “seepage channel connection - rock particle migration - local skeleton collapse.” The experimental results indicate that: (1) Water pressure is the dominant factor driving particle migration and mass loss; high water pressure significantly accelerates the transition from porous media seepage to pipe (conduit) flow. (2) Initial porosity has a complex non-linear impact on the catastrophic consequences. While higher initial porosity increases the volume of water inrush and the final stable porosity, it paradoxically leads to a relative decrease in the total volume of discharged mud. This is because it induces more frequent internal skeletal collapse and clogging effects. The variable mass seepage evolution model established in this study identifies particle loss as the fundamental cause of sudden surges in permeability, providing a theoretical basis for disaster prevention and control in this type of strata.

1 Introduction

The mechanisms behind water and mud inrush disasters during tunnel construction exhibit significant complexity and diversity. Particularly, such disasters occurring within granite alteration zones are progressively becoming a bottleneck constraint on the development of tunnel construction in China. Issues of granite erosion and alteration are particularly prevalent in the southwestern regions. Altered granite rock masses demonstrate remarkably poor stability, characterized by low strength, susceptibility to loosening, cracking, and disintegration, and well-developed joints and fissures. Currently, research focusing on the mechanisms of water and mud inrush in granite alteration zones, among other related aspects, remains severely inadequate.

Current research on the mechanisms of tunnel water and mud inrush focuses primarily on model testing. These studies investigate water inrush characteristics and catastrophic mechanisms by establishing models for tunnel karst and fault fracture zones (Li et al., 2018; Li et al., 2020; Li et al., 2025; Xue et al., 2022), alongside precursor early-warning systems and prevention measures (Kong et al., 2023; Xie et al., 2025; Zhong et al., 2024). Jiang et al. conducted a similar material model test study on the scenario where a confined water-dissolved cavity exists ahead of the tunnel face. They scientifically classified the failure modes of the tunnel face into two categories: central shear failure and peripheral seepage-induced failure (Jiang et al., 2017). Li et al. studied the seepage evolution law of fractured rock in collapse column under triaxial stress, by employing the triaxial seepage test equipment. Besides, they established and calculated the seepage mechanics model of broken rock in the COMSOL Multiphysics, and elucidated the water-conducting channel under mass loss condition in the collapse column furtherly (Li X. L. et al., 2023). Liu Jinquan conducted water and mud inrush evolution tests under varying water pressures and confining pressures, utilizing a self-designed large-scale laboratory test system capable of accounting for mass migration and true triaxial stress states. These tests revealed the scientific conclusion that the evolution of water and mud inrush disasters in completely weathered granite tunnels is a strongly coupled seepage-erosion process (Liu et al., 2018; Liu et al., 2016; Liu et al., 2017). Xu et al. employing laboratory model tests, successfully reproduced the catastrophic evolution process of water and mud inrush disasters in tunnels excavated through interbedded sandstone and slate of the Pre-Sinian system. They meticulously analyzed the variation patterns of surrounding rock stress-strain, seepage pressure, and flow rate with decreasing thickness of the aquitard (water-resisting layer). This work thus revealed the catastrophic triggering mechanism of water and mud inrush in interbedded sandstone-slate tunnels, initiated by hydraulic failure of the aquitard (Xu et al., 2022). Chen et al. analyzed the disaster-causing mechanism of water inrush disaster in tunnels and established the relationship between water inrush disaster and microseism. They carried out the hydraulic fracturing test of rock-like materials and verified the effectiveness of acoustic emission monitoring technology by comparing the fracture shape and AE positioning results (Chen et al., 2024).

Based on numerical simulations and model tests of water seepage in surrounding rock during tunnel excavation, Wang et al. developed a permeability evolution model capable of predicting water inrush points, incorporating variables of rock porosity and deformation-induced damage (Wang et al., 2024). Through geological surveys and mineralogical analyses, Wu et al. revealed a correlation between water and mud inrushes in tunnels and rock composition, demonstrating a higher probability of such disasters in tunnels excavated in magmatic rocks (Wu et al., 2017). Using the single-variable method, Zheng et al. examined the effects of particle gradation, initial porosity, water pressure, confining pressure, and anti-outburst layer thickness on water inrush volume, porosity, particle concentration, and water inrush velocity, revealing that particle loss serves as the dominant factor in water and mud inrushes within fully weathered water-rich granite (Zheng, 2021). Based on the Hoek-Brown nonlinear failure criterion and limit analysis theory, Yang et al. developed a method to determine the safety thickness of anti-outburst layers for water-rich high-pressure solution cavities in karst tunnels (Yang and Zhang, 2016). Through discrete element modeling, Chen et al. uncovered the process and mechanisms of clay mineral-induced mud inrush failures in karst tunnels, demonstrating that excavation-induced stress redistribution and water-induced softening of clay serve as primary contributing factors to such failures (Chen et al., 2023). By integrating strength models, catastrophe theory, and shear failure mechanisms, Li et al. established a method for determining the thickness of anti-outburst layers in tunnels (Li Z. Q. et al., 2023). Dong et al. demonstrated that faults and fractured evaporite-karst breccias exhibit high permeability. Tunnel excavation alters groundwater flow paths, prompting the proposal of sequential countermeasures: pre-drainage control at water-inrush zones, pre-grouting reinforcement, inrush sealing, and localized strengthening (Dong et al., 2023). Fan et al. identified joints as critical factors in preventing tunnel water-inrush disasters. By adopting Discontinuous Deformation Analysis (DDA) to simulate fluid-solid coupling interactions during inrush events, they revealed fluid flow patterns in jointed rock masses and the catastrophic evolution of water-inrush processes (Fan et al., 2023). Gao et al. established a near-field dynamic model of surrounding rock to investigate the process of progressive rock failure leading to water-inrush pathway development in karst tunnels, revealing the formation mechanisms and catastrophic evolution of water-inrush channels (Gao et al., 2021).

In summary, existing research achievements are concentrated on the water inflow mechanisms within soluble rock structures and fault fracture zones; however, studies targeting the unique geological medium of water-rich altered granite remain notably insufficient. Unlike karst conduits or common fracture zones, altered granite is characterized by slaking (disintegration) upon contact with water, argillization (mudding), and a high susceptibility to particle suffusion. Its water and mud inrush process is accompanied by significant loss of solid phase mass, representing a typical variable mass seepage problem. Furthermore, current experimental methods lack visualized observation of the internal microscopic “particle migration–channel formation” mechanism. Therefore, conducting visualized variable mass seepage tests that account for particle loss effects is crucial for revealing the incubation mechanisms of such disasters.

2 Experimental protocol

2.1 Engineering background

Within the research area of the Gaoligongshan Tunnel, there is a widespread distribution of granite alteration zones characterized by weak rock quality and severe weathering. The altered granite appears bluish-gray to grayish-white and mainly occurs within the contact metamorphic zone between the granitic intrusive body and the surrounding rock. This altered rock was formed by high-temperature hydrothermal metasomatism during the magma intrusion period, which induced metamorphic recrystallization reactions in the surrounding rock. The areas adjacent to tectonic fracture zones have undergone superimposed modification by structural hydrothermal fluids, forming secondary altered rock facies.

The rivers within the Gaoligongshan Tunnel site area belong to the Indian Ocean drainage system, mainly including the Nu River, the Longchuan River, and their tributaries. The groundwater types are comprehensive, primarily recharged by atmospheric precipitation with abundant rainfall, and locally recharged by surface water bodies. Granite alteration zones with severe alteration phenomena are distributed in the study area, where the rock is weak, heavily weathered, and groundwater is abundant. During tunnel excavation, the tunnel face is prone to deformation and fracturing. Simultaneously, groundwater softens the surrounding rock through seepage, triggering severe water and mud inrush problems.

A detailed analysis was conducted on the water and mud inrush accident that occurred at the section XSD1K0+352 of the Gaoligongshan Tunnel. This tunnel section is located between the Tangfang Fault and the Laodongpo Fault, with a burial depth of approximately 760 m. Cambrian metamorphic sandstone strata outcrop at the surface, while Yanshanian granite alteration zones are developed in the tunnel body and lower parts. The tunnel body is situated near the contact zone between the Cambrian strata and granite, with the surrounding rock consisting mainly of altered granite. Influenced by geological structures, joints and fissures are developed in this section; the rock mass is broken, and the integrity of the surrounding rock is poor, classified mainly as Grade V surrounding rock. A structural fracture alteration zone is developed in front of the tunnel face, with severe clay alteration of the rocks. The groundwater is bedrock fissure water occurring in granite fissures, distributed unevenly, with water-rich zones in areas of dense joints. When tunnel construction reached this section, strand-like water outflows appeared at multiple points on the tunnel crown. Subsequently, the water volume gradually increased, carrying out large amounts of mud, sand, and debris, resulting in a severe water and mud inrush accident.

2.2 Experimental procedure

Effect of water pressure on the seepage behavior of altered granite.

To simulate the phenomenon of water and mud inrush in water-rich granite alteration zones under specific water pressures, a self-developed visualization-capable variable-mass seepage model test apparatus with controllable water pressure was designed. This experimental setup primarily comprises the following core systems (Figure 1).

1. Water Pressure Loading System: Utilizes a nitrogen cylinder pressurization method to apply gas pressure and regulate water pressure, with a pressure range of 0–1 MPa.

2. Seepage System: The seepage device is primarily composed of a cover plate, an acrylic material container, and a screen mesh. The cover plate has a diameter of 15 cm and a thickness of 1.5 cm; the acrylic material container has a diameter of 15 cm, a total height of 40 cm, and a wall thickness of 1 cm. The cover plate and the acrylic container are connected by bolts and a sealing ring to ensure the airtightness during the experimental process. The screen mesh has a diameter of 15 cm and a mesh aperture of 8 mm, which is more than 1.5 times the maximum particle size of the intended filling material, to prevent interference with the loss of larger particles during the experiment (Figure 2).

3. Collection System: The ejected material collection system comprises a collection bucket, a fine muslin mesh sieve, an electronic balance, and a drying oven. Ejected material is collected at fixed time intervals using the collection bucket. A muslin mesh filter is employed to capture migrated particles. The mass of ejected material is measured with the electronic balance and collected migrated particles undergo oven-drying to determine their dry mass.

To investigate the temporal variation patterns of water inrush volume, mud inrush volume, seepage velocity, porosity, and permeability in water-rich granite alteration zones of the Gaoligongshan, this study utilizes field-collected granite alteration zone rock samples from the engineering site. By analyzing the influence of different water pressures and porosity conditions on the temporal evolution of these key parameters, the following experimental scheme was formulated.

1. Water Pressure Research Scheme: Maintain the initial porosity at 0.45. Set water pressures to 15 kPa, 30 kPa, and 45 kPa to analyze the impact of water pressure on water and mud inrush.

2. Porosity Research Scheme: Maintain water pressure at 15 kPa. Set initial porosity values to 0.4, 0.45, and 0.5 respectively to analyze the influence of porosity on water and mud inrush.

The detailed experimental procedure is as follows, it is also summarized in the form of a flowchart (Figure 3).

1. Specimen Preparation Stage: Field-collected rock samples from the granite alteration zone at the engineering site were weighed according to the required mass. To ensure homogeneous packing, specimens were loaded into the cylinder in layers up to a height of 25 cm. Prior to loading, the inner walls of the specimen cylinder were coated with a thin layer of petroleum jelly to minimize boundary effects during testing. Subsequently, water saturation was performed, with the water level maintained at 35 cm after saturation.

2. Variable-Mass Seepage Evolution Experiment: After specimen saturation, open the pressure relief valve to achieve the target control pressure. Throughout the experiment, maintain constant water pressure. As pressurized water infiltrates the specimen, soil particles migrate with the effluent flow. Migrated particles are collected via filtration, oven-dried, and weighed. Since the particle outlet is open to atmosphere and pressure losses in the pressurization line are negligible, the pressure differential across the specimen approximates the applied water pressure. Furthermore, the hydraulic gradient and seepage velocity are derived from the specimen’s height and diameter. Corresponding permeability parameters are then calculated using seepage mechanics theory.

3. Ejected Material Collection: During the experiment, data was acquired at 15 s intervals. Ejected material was collected at the collection assembly, where a muslin mesh filter captured migrated particles for subsequent oven drying. Concurrently, the corresponding water inrush volume was measured. The experiment was terminated when the effluent became clear water with no further particle migration observed.

4. Post-Experimental Processing: Following the experiment, equipment was cleaned and maintained for subsequent tests. Experimental data including water inrush volume and particle migration mass were compiled. Subsequent analysis focused on discussing the evolutionary patterns of seepage behavior.

During the experiment, continuous particle migration leads to temporal variations in the specimen’s mud inrush volume, water inrush volume, porosity, and permeability. Through in-depth analysis of the seepage parameters alongside the variation patterns of mud inrush volume and water inrush volume-combined with empirical observations from experiments-the underlying evolutionary mechanism of variable-mass seepage in altered granite can be revealed. Furthermore, by systematically altering the two key controlling factors (water pressure and porosity) and analyzing their influence on mud inrush volume (Equation 1), water inrush volume, porosity (Equation 2), and permeability (Equation 9), the governing laws dictating the evolution of water and mud inrush can be elucidated (Equation 5). This provides a scientific foundation for a deeper understanding of the intrinsic mechanisms governing water and mud inrush within granite alteration zones.

During the experiment, ejected material was collected and weighed after oven-drying at 15 s intervals. This yielded the mass of lost particles for each time segment (Δmn).

$m_n=\Delta m_1+\Delta m_2+\cdots \Delta m_n(1)$

During the experiment, particle loss leads to changes in the sample’s porosity. The relationship between the change in sample porosity and the mass of lost particles can be expressed as follows:

$\Delta \phi =\frac{\Delta m_n}{\pi {\mathrm{r}}^{2}h\rho }(2)$

where r represents the radius of the sample (m), denotes the height of the sample (m), and is the density of the sample (kg/m³).

The porosity at each moment and the rate of change of porosity are calculated using the following formulas:

${\phi }_n={\phi }_0+\frac{1}{\pi {\mathrm{r}}^{2}h\rho }\left(\Delta m_1+\Delta m_2\cdots \Delta m_n\right)(3)$

${\phi }_n^{\prime }=\frac{{\phi }_n-{\phi }_{n-1}}{\Delta t}(4)$

During the experimental process, the flow velocity was relatively low. When the flow velocity is small, the seepage process through the sample primarily conforms to Darcy’s Law, which can be expressed as:

$v=-\frac{k}{\mu }\left(\nabla p+{\rho }_{f}g\nabla z\right)(5)$

where represents the permeability of the sample (m²), denotes the hydraulic pressure (Pa).

Given the small specimen size, gravitational effects are neglected. Furthermore, seepage within the specimen can be approximated as one-dimensional flow. Thus, the following holds:

$v=-\frac{k}{\mu }\frac{\mathrm{\partial }p}{\mathrm{\partial }z}(6)$

The seepage velocity can be calculated from the water inrush flow rate Q, as follows:

$v=\frac{Q}{\pi r^2}(7)$

Assuming a uniform pressure gradient distribution within the specimen, i.e.:

$\frac{\mathrm{\partial }p}{\mathrm{\partial }z}=-\frac{p}{h}(8)$

Consequently, based on Equations 6–8, the permeability is given by:

$k=\frac{Q\mu }{\pi r^2}\frac{h}{p}(9)$

3 Experimental result

During the experiment, the specimen was saturated in layers and then observed under different water pressures (15 kPa, 30 kPa, 45 kPa) and porosities (0.40, 0.45, 0.50) (Figure 4). The onset of water and mud inrush was defined as the moment when sediment began to discharge, while the end point was marked by the effluent running clear. The experimental process exhibited the following three stages.

1. Initial Stage: After the experiment commenced, a mixture of particulate matter and water discharged from the specimen. Concurrently, cavity formation was observed within the specimen. This indicates that under the applied water pressure, particles within the specimen began to mobilize, migrate, and be lost. Simultaneously, the seepage flow of water transported these particles, resulting in the discharge of a water-mud mixture.

2. Intermediate Stage: As the experiment progressed, cavities progressively enlarged and interconnected, developing into pathways conducive to water inrush. This signifies that the internal structure of the specimen underwent alterations under sustained water pressure and seepage forces. The enlargement and interconnection of these cavities ultimately set the stage for subsequent water inrush.

3. Later Stage: The initial discharge was a water-mud mixture. As the experiment proceeded, the effluent gradually became clearer and ultimately ran clear. This demonstrates that with the continuous loss of particulate matter, the particle concentration within the specimen decreased, resulting in progressively clearer seepage flow.

Figure 1

Figure 1. Visualized variable-mass seepage apparatus.

Figure 2

Figure 2. Sieve mesh at the bottom of the experimental setup.

Figure 3

Figure 3. Experimental procedure.

Figure 4

Figure 4. Pictures of protrusion.

Through conducting mass-varying seepage experiments under different water pressures and initial porosities, temporal evolution curves for the following parameters were obtained: mud inrush rate, water inrush rate, porosity, permeability, and seepage velocity. This enabled the investigation of seepage evolution patterns in altered granite during the mass-varying seepage process and the analysis of the influence of water pressure and porosity on seepage.

3.1 Effect of water pressure on the seepage behavior of altered granite

As shown in Figure 5, with the initial porosity controlled at 0.45, the cumulative mud inrush mass and mud inrush velocity curves were investigated under water pressures of 15 kPa, 30 kPa, and 45 kPa, respectively. The results indicate that as seepage progressed, the mud inrush velocity exhibited a gradual decreasing trend and ultimately stabilized. Under higher water pressures, the mud inrush velocity was significantly greater than under lower pressures. Furthermore, the cumulative mass loss of the specimen increased with rising water pressure. Specifically, the final mud inrush masses at water pressures of 15 kPa, 30 kPa, and 45 kPa were 352.11 g, 618.24 g, and 883.61 g, respectively.

Figure 5

Figure 5. Cumulative mud inrush mass and inrush velocity under different water pressures.

Concurrently, the duration of mud inrush differed markedly across the different water pressures. As water pressure increased, the mud inrush duration progressively shortened, recorded as 285 s, 240 s, and 195 s, respectively. Overall, compared to the 15 kPa condition, the mud inrush mass under 30 kPa and 45 kPa pressures increased by 76% and 151%, respectively, while the inrush duration decreased by 23% and 46%, respectively.

The underlying mechanism for this phenomenon can be summarized as follows: Under the action of water pressure, fine, mobile particles within the specimen gradually detach from the granular matrix and are mobilized and transported away. When particle loss reaches a certain threshold, erosion ceases, and stable seepage channels develop. The critical flow velocity required to dislodge particles of different sizes from the matrix and transport them varies. Consequently, under lower water pressures (resulting in lower flow velocities), only a limited quantity of particles can be mobilized and lost. Conversely, under higher water pressures (producing higher flow velocities), a greater proportion of particles can detach from the matrix, leading to a substantial rise in the mud inrush mass.

Additionally, under higher water pressures, the greater seepage velocity accelerates the mass loss rate. This hastens the formation of stable seepage channels, thereby reducing the duration of the mud inrush phase. These findings collectively demonstrate the pronounced influence of water pressure on the mud inrush process.

As shown in Figure 6, with the initial porosity controlled at 0.45, the water inrush mass and water inrush velocity curves were investigated under water pressures of 15 kPa, 30 kPa, and 45 kPa, respectively. The results indicate that, generally speaking, increased water pressure exhibits a positive correlation with both the water inrush rate and the cumulative water inrush mass. That is, the greater the water pressure, the higher the water inrush rate and the larger the cumulative water inrush mass. The specific data are as follows: At a water pressure of 15 kPa, the average water inrush rate was 1.69 ml/s. At 30 kPa, the average rate was 3.55 ml/s, and at 45 kPa, it reached 5.66 ml/s. Further analysis revealed that compared to the 15 kPa condition, the water inrush rates under 30 kPa and 45 kPa pressures increased by factors of 1.10 and 2.35, respectively. The underlying mechanism for this phenomenon can be explained as follows: Under high water pressure, the rate of particle loss is significantly accelerated. As particles are lost, the resulting pore network progressively evolves into “water inrush channels.” Crucially, the higher the water pressure, the wider these water inrush channels become. This directly leads to an increase in the water inrush rate and a rise in the cumulative water inrush mass.

Figure 6

Figure 6. Cumulative water inrush mass and inrush velocity under different water pressures.

As shown in Figure 7, with the initial porosity controlled at 0.45, the porosity temporal evolution curves were investigated under water pressures of 15 kPa, 30 kPa, and 45 kPa, respectively. The results demonstrate that the porosity temporal evolution curves generally exhibited a trend of initial increase followed by stabilization. Furthermore, the greater the water pressure, the more rapidly the porosity increased, and the higher the ultimate porosity attained. The specific data are as follows: At a water pressure of 15 kPa, after 285 s of seepage evolution, the final porosity stabilized at approximately 0.478. At 30 kPa, following 240 s of seepage evolution, the final porosity stabilized at approximately 0.501. At 45 kPa, after 195 s of seepage evolution, the final porosity stabilized at approximately 0.525. These findings fully demonstrate the pronounced influence of water pressure on porosity evolution. Under higher water pressures, the increased seepage velocity within the specimen accelerates the rate of particle loss. This results in a correspondingly higher rate of porosity increase and consequently leads to a higher attained final porosity. Conversely, under lower water pressures, the erosion-driving forces acting on particles weaken. Consequently, both the amount and rate of particle loss decrease, resulting in a comparatively lower final porosity being achieved.

Figure 7

Figure 7. Porosity temporal evolution curves under different water pressures.

As shown in Figure 8, the temporal evolution curves of permeability under water pressures of 15 kPa, 30 kPa, and 45 kPa were investigated while controlling the initial porosity at 0.45. The results indicate that the final permeability under different water pressure conditions stabilized at approximately 1.5 × 10⁻⁹ m². This suggests that water pressure exerted a relatively minor influence on the final permeability.

Figure 8

Figure 8. Permeability temporal evolution curves under different water pressures.

As shown in Figure 9, the temporal evolution curves of seepage velocity under water pressures of 15 kPa, 30 kPa, and 45 kPa were investigated while controlling the initial porosity at 0.45. It can be observed from the figure that the water pressure is positively correlated with the seepage velocity-that is, higher water pressure leads to greater seepage velocity. The specific data are as follows: when the water pressure was 15 kPa, the final seepage velocity stabilized at approximately 1 × 10⁻⁴ m/s; at a water pressure of 30 kPa, the final seepage velocity stabilized at approximately 2 × 10⁻⁴ m/s; and at a water pressure of 45 kPa, the final seepage velocity stabilized at approximately 3.5 × 10⁻⁴ m/s.

Figure 9

Figure 9. Seepage velocity under different water pressures.

3.2 Effect of porosity on seepage in altered granite

As shown in Figure 10, the temporal evolution curves of seepage velocity under different initial porosities (0.40, 0.45, 0.50) were investigated at a water pressure of 15 kPa. The experimental results indicate that the mud inrush velocity was relatively high during the initial stage of seepage. Subsequently, as seepage progressed, the mud inrush velocity gradually decreased until the effluent became clear water. Further analysis revealed that a smaller initial porosity resulted in a greater mud inrush velocity and a longer duration of mud inrush. The specific data are as follows: at initial porosities of 0.4, 0.45, and 0.5, the final particle loss masses were 438.72 g, 352.11 g, and 240.24 g, respectively. This indicates that the mass loss of the specimen decreases with increasing initial porosity—that is, the mud inrush mass at an initial porosity of 0.4 was greater than that at an initial porosity of 0.5. Concurrently, the duration of mud inrush also varied under different initial porosity conditions. At initial porosities of 0.4, 0.45, and 0.5, the mud inrush durations were 360 s, 285 s, and 150 s, respectively. Compared to the case with an initial porosity of 0.4, the mud inrush mass increased by 47% and 82.5% at initial porosities of 0.45 and 0.5, respectively. Furthermore, the mud inrush duration increased by 0.9 times (90%) and 1.4 times (140%), respectively. The cause of this phenomenon can be attributed to the following: a smaller initial porosity results in a denser internal structure of the specimen. Consequently, a greater number of particles need to be dislodged to form seepage channels. This leads to a larger mud inrush mass. Furthermore, as more time is required to establish stable seepage channels, the duration of the mud inrush is correspondingly prolonged.

Figure 10

Figure 10. Cumulative mud inrush mass and mud inrush velocity under different initial porosities.

As shown in Figure 11, under a water pressure of 15 kPa, the water inrush mass and water inrush rate curves were investigated for specimens with different initial porosities (0.40, 0.45, 0.50). The results indicate that initial porosity exhibits a positive correlation with both the cumulative water inrush mass and the water inrush rate. That is, the higher the initial porosity, the greater the cumulative water inrush mass and the water inrush rate. Additionally, for specimens with the same initial porosity, the water inrush rate remained within a relatively stable range at different time points. The specific data are as follows: At an initial porosity of 0.40, the average water inrush rate was 2.92 ml/s. At an initial porosity of 0.45, the average rate was 1.69 ml/s, and at 0.50, it was 1.26 ml/s. Further analysis revealed that compared to the case with an initial porosity of 0.40, the average water inrush rates under initial porosities of 0.45 and 0.50 increased by 34.1% and by a factor of 1.30, respectively. This phenomenon arises because a higher initial porosity corresponds to a looser internal structure within the specimen. This facilitates the readier development of “water inrush channels” and results in wider channels, thereby increasing the cumulative water inrush mass.

Figure 11

Figure 11. Cumulative water inrush mass and velocity under different initial porosities.

As shown in Figure 12, under a water pressure of 15 kPa, the temporal evolution curves of porosity were investigated for specimens with different initial porosities (0.40, 0.45, 0.50). Overall, the higher the initial porosity, the greater the ultimate porosity attained and the shorter the porosity evolution time; however, the magnitude of the porosity increase was relatively smaller. The specific data are as follows: At an initial porosity of 0.40, after 390 s of seepage evolution, the final porosity stabilized at approximately 0.442. At an initial porosity of 0.45, following 285 s of seepage evolution, the final porosity stabilized at approximately 0.478. At an initial porosity of 0.50, after 180 s of seepage evolution, the final porosity stabilized at approximately 0.512. This phenomenon occurs because a higher initial porosity corresponds to a looser internal structure within the specimen, facilitating easier mobilization and loss of particles. Consequently, a higher ultimate porosity is attained. Simultaneously, in specimens with higher initial porosity, water inrush channels develop more readily, requiring the erosion of fewer particles to establish stable flow paths. This dual effect significantly shortens the time required to reach the final stabilized porosity.

Figure 12

Figure 12. Porosity temporal evolution under different initial porosities.

As shown in Figure 13, an experimental investigation of the temporal evolution curves of permeability under different initial porosities (0.4, 0.45, 0.5) was conducted at a constant water pressure of 15 kPa. The results demonstrate a positive correlation between initial porosity and permeability during the seepage process: the higher the initial porosity, the greater the permeability. Specific data are as follows: specimens with an initial porosity of 0.50 exhibited an average permeability of 2.78 × 10⁻⁹ m² during seepage, while those with porosities of 0.45 and 0.40 showed average permeabilities of 1.61 × 10⁻⁹ m² and 1.20 × 10⁻⁹ m², respectively. This confirms that specimens with 0.50 initial porosity have significantly higher permeability than those with 0.40 and 0.45 porosity, indicating that increased initial porosity substantially enhances specimen permeability. Higher permeability facilitates easier fluid transmission through the medium, thereby increasing the risk of water and mud inrush hazards. Based on this conclusion, practical engineering applications can employ grouting techniques to reduce formation porosity, effectively mitigating the probability of water and mud inrush disasters while ensuring project safety and structural stability.

Figure 13

Figure 13. Permeability temporal evolution under different initial porosities.

As shown in Figure 14, experimental investigation of the temporal evolution curves of seepage velocity under different initial porosities (0.4, 0.45, 0.5) was conducted at a constant water pressure of 15 kPa. The results reveal a positive correlation between initial porosity and seepage velocity: the higher the initial porosity, the greater the seepage velocity. Specifically, during the initial seepage stage, the velocity showed a brief decline before stabilizing. The final stabilized seepage velocities were approximately 5 × 10⁻⁵ m/s at 0.40 initial porosity, 1 × 10⁻⁴ m/s at 0.45, and 1.8 × 10⁻⁴ m/s at 0.50 initial porosity. The cause of this phenomenon lies in the following: a larger initial porosity results in a looser internal structure of the specimen. Consequently, the rate of particle loss increases more rapidly, seepage channels form more easily, and the water inflow volume becomes greater. This, in turn, leads to a higher seepage velocity. During the initial stage of seepage, the relatively high seepage velocity is attributed to the effluent existing as muddy water due to particle loss. The larger volume of this muddy water mixture results in a higher seepage velocity. Once stable channels are formed through particle loss and the effluent transitions to clear water, the permeability remains relatively stable. Consequently, the seepage velocity also stabilizes.

Figure 14

Figure 14. Seepage velocity under different initial porosities.

3.3 Mud inrush mechanism

Through a self-designed visual mass-varying seepage apparatus, this study investigates the characteristics and mechanisms of water and mud inrush in water-rich granite alteration zones. Analysis of the evolution patterns of mud inrush volume, water inrush volume, porosity, seepage velocity, and permeability in altered granite reveals that water pressure and porosity directly govern the permeability behavior of altered granite.

The presence of pressurized water sources within rock formations—such as water-rich faults, alteration zones, or karst cavities—constitutes a prerequisite for tunnel water and mud inrush events. Within water-rich granite alteration zones, elevated water pressure accelerates water migration velocity. This enhances the water’s capacity to impact and dissolve rock particles during flow, thereby expediting the formation of interconnected water inrush channels, as illustrated in Figure 15a. When altered granite exhibits high initial porosity, the granular matrix within the rock mass is inherently loose. Subsequent water flow and particle transport further increase porosity, reducing the load-bearing capacity of the particle skeleton. This facilitates localized collapse of loose particle clusters, blocking established water flow channels as depicted in Figure 15b. Under hydraulic pressure, locally connected channels re-form. As particle loss continues, this triggers repeated cycles of localized collapse. Under high pore pressure conditions, altered granite exhibits greater cumulative water and mud inrush volumes than under high-porosity conditions. Conversely, the mud inrush rate curve displays more pronounced oscillations in high-porosity regimes. The seepage process in altered granite follows a three-phase cycle: 1) water inrush channel interconnection, 2) rock particle migration, and 3) localized collapse.

Figure 15

Figure 15. Seepage and water inrush characteristics in altered granite. (a) High water pressure conditions; (b) High porosity Conditions.

3.4 Coupling mechanism between variable mass seepage and permeability evolution

The core of this study lies in confirming that water inrush and mud outburst in altered granite constitutes a variable mass process. The experimentally observed mass of particle loss is not an independent variable; rather, it directly governs the evolution of permeability by altering porosity. As illustrated in Figure 15, under high hydraulic pressure, the threshold velocity for fine particles is exceeded, leading to a continuous loss of skeletal mass. According to Equations 3, 4, this mass loss induces a non-linear increase in effective porosity over time. Such microstructural alterations disrupt the original seepage equilibrium, causing permeability to surge by orders of magnitude within a short period (Figure 8), thereby forming preferential flow paths. Conversely, local collapse under conditions of high porosity (Figure 15b) serves as a negative feedback mechanism, temporarily reducing permeability by re-clogging the channels. Consequently, the permeability characteristics of altered granite are the dynamic outcome of the interaction between hydrodynamic conditions and the evolution of the geomaterial mass.

4 Conclusion

This study employs a self-designed visual mass-varying seepage apparatus to conduct physical model experiments investigating water and mud inrush mechanisms in water-rich granite alteration zones. The experiments systematically evaluate the effects of varying water pressures and initial porosities. Key parameters-including mud inrush volume, water inrush volume, porosity, permeability, and seepage velocity-were quantified throughout the inrush process, revealing their evolution patterns under different influencing factors. This work elucidates the mass-varying seepage evolution mechanisms within water-rich granite alteration zones. The main conclusions are as follows.

1. Innovative Experimental Apparatus: This study overcomes the limitations of previous research, which predominantly focused on soluble rocks (karst) or single fault fracture zones. Addressing the increasingly prominent hazards in water-rich altered granite during engineering construction in Southwest China, we successfully developed a visual mass-change seepage test apparatus. The device intuitively reproduces hidden groundwater seepage processes and confirms that water and mud inrush in altered granite is essentially a strongly coupled seepage-erosion process, where water flow drives the migration of fine particles, leading to skeleton instability.

2. Three-Stage Evolution Mechanism: The “three-stage” evolution mechanism of mass-change seepage was revealed. The study identifies that the incubation and occurrence of water and mud inrush disasters in altered granite follow a specific three-stage cyclic mechanism: 1) interconnection of water channels, 2) migration of rock particles, and 3) local collapse. This mechanistic model provides a scientific basis for understanding such non-karst geological hazards.

3. Higher initial porosity correlates with a greater risk of water inrush, yet paradoxically results in a reduced total volume of mud inrush. The underlying mechanism is that altered rock masses with high porosity possess a loose internal structure, allowing water to easily penetrate and establish preferential pathways. A stable flow path can be formed by flushing away only a small quantity of particles; consequently, water inrush occurs rapidly with high permeability, but the total sediment load transported is relatively low. Conversely, altered rock masses with low porosity exhibit a tighter internal structure. Water flow must detach and transport a significant volume of particles to create seepage channels, resulting in a substantial increase in mud discharge and a more prolonged duration of the disaster.

4. Driving Role of Water Pressure: Water pressure is the direct power source for water and mud inrush disasters. High water pressure not only directly increases the cumulative volume of water and mud discharge but also significantly accelerates the evolution rate of porosity, causing the rock and soil mass to reach a failure state more quickly. Although water pressure has a limited effect on the final permeability value, it dictates the severity of the disaster and the ultimate rate of particle loss.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Author contributions

HX: Data curation, Formal Analysis, Funding acquisition, Investigation, Writing – original draft, Writing – review and editing. YZ: Methodology, Project administration, Writing – original draft, Writing – review and editing. WS: Software, Writing – original draft, Writing – review and editing. HT: Writing – original draft, Writing – review and editing, Data curation. XB: Writing – original draft, Writing – review and editing, Data curation, Formal Analysis. QZ: Formal Analysis, Writing – original draft, Writing – review and editing. HY: Data curation, Writing – original draft, Writing – review and editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This work is supported by the Key R&D Program of Yunnan Province (Grant 202303AA080005).

Conflict of interest

Author HX was employed by Kunming Prospecting Design Institute of China Nonferrous Metals Industry Co., Ltd.

Authors YZ and HY were employed by Yungui Railway Yunnan Co., Ltd.

Author WS was employed by China Railway Eryuan Engineering Group Co., Ltd.

The remaining author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not 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

Chen, L. L., Wang, Z. F., Zhang, W., and Wang, Y. Q. (2023). Characteristics, numerical analysis and countermeasures of mud inrush geohazards of Mountain tunnel in karst region. Geomatics, Nat. Hazards Risk 14 (1), 2242691. doi:10.1080/19475705.2023.2242691

CrossRef Full Text | Google Scholar

Chen, Y. H., Cheng, S., Li, L. P., Xing, Z., Li, X., Jin, H., et al. (2024). Evolution mechanism and microseismic response characteristics of water inrush channel. KSCE J. Civ. Eng. 28, 4742–4755. doi:10.1007/s12205-024-1883-9

CrossRef Full Text | Google Scholar

Dong, L., Wang, H., Song, D., Chen, J., and Liu, C. (2023). Analysis of the catastrophe mechanism and treatment countermeasures of a sudden water inrush disaster in a long and deeply buried tunnel in the karst area. J. Perform. Constr. Facil. 37 (6), 06023002. doi:10.1061/jpcfev.cfeng-4349

CrossRef Full Text | Google Scholar

Fan, H. Y., Li, L. P., and Chen, G. Q. (2023). Analysis method of the water inrush and collapse in jointed rock mass tunnels: a case study. Eng. Analysis Bound. Elem. 146, 838–850. doi:10.1016/j.enganabound.2022.11.030

CrossRef Full Text | Google Scholar

Gao, C., Li, L., Zhou, Z., Li, Z., Cheng, S., Wang, L., et al. (2021). Peridynamics simulation of water inrush channels evolution process due to rock mass progressive failure in karst tunnels. Int. J. Geomechanics 21 (4), 04021028. doi:10.1061/(asce)gm.1943-5622.0001980

CrossRef Full Text | Google Scholar

Jiang, H. M., Li, L., Rong, X. L., Wang, M. y., Xia, Y. p., and Zhang, Z. c. (2017). Model test to investigate waterproof-resistant slab minimum safety thickness for water inrush geohazards. Tunn. Undergr. Space Technol. 62, 35–42. doi:10.1016/j.tust.2016.11.004

CrossRef Full Text | Google Scholar

Kong, F., Xue, Y., Gong, H., Jiang, X., Song, Q., Fu, Y., et al. (2023). The formation mechanism of dynamic water and mud inrush geohazard triggered by deep-buried tunnel crossing active fault: insights from the geomechanical model test. Tunn. Undergr. Space Technol. doi:10.1016/j.tust.2023.105437

CrossRef Full Text | Google Scholar

Li, S., Gao, C., Zhou, Z., Li, L., Wang, M., Yuan, Y., et al. (2018). Analysis on the precursor information of water inrush in karst tunnels: a true triaxial model test study. Rock Mech. Rock Eng. doi:10.1007/s00603-018-1582-2

CrossRef Full Text | Google Scholar

Li, L., Sun, S., Wang, J., Song, S., Fang, Z., and Zhang, M. (2020). Development of compound epb shield model test system for studying the water inrushes in karst regions. Tunn. Undergr. Space Technol. doi:10.1016/j.tust.2020.103404

CrossRef Full Text | Google Scholar

Li, Z. Q., Xue, Y., Li, G., Qiu, D., Xu, L., Liu, Q., et al. (2023). Probabilistic determination and application of rock thickness resisting water inrush from karst cave. Process Saf. Environ. Prot. 172, 462–472. doi:10.1016/j.psep.2023.02.041

CrossRef Full Text | Google Scholar

Li, Z., Nie, L., Xue, Y., Li, W., and Fan, K. (2025). Model testing on the processes, characteristics, and mechanism of water inrush induced by karst caves ahead and alongside a tunnel. Rock Mech. Rock Eng. doi:10.1007/s00603-025-04414-x

CrossRef Full Text | Google Scholar

Li, X. L., Li, X. L., and Wang, Y. (2023). The seepage evolution mechanism of variable mass of broken rock in karst collapse column under the influence of mining stress. Geofluids 2023, 7256937. doi:10.1155/2023/7256937

CrossRef Full Text | Google Scholar

Liu, J. Q., Chen, W. Z., and Yuan, J. Q. (2016). Test on anti-scouring property of grouting reinforced bodyin completely weathered granite. Chin. J. Rock Mech. Eng. 35 (09), 1767–1775. doi:10.13722/j.cnki.jrme.2015.1537

CrossRef Full Text | Google Scholar

Liu, J. Q., Yang, D. S., and Chen, W. Z. (2017). Research on particle starting velocity in the expansion of water inrush channel in completely weathered granite. Rock Soil Mech. 38 (04), 1179–1187. doi:10.16285/j.rsm.2017.04.032

CrossRef Full Text | Google Scholar

Liu, J. Q., Chen, W. Z., and Zhen, W. H. (2018). Test of water and mud inrush in completely weathered granite tunnels considering variable mass properties. China J. Highw. Transp. 31 (10), 190–196. Available online at: https://zgglxb.chd.edu.cn/CN/Y2018/V31/I10/190.

Google Scholar

Wang, C., Zhang, D., Tan, D., Ye, J., Wang, X., Ren, F., et al. (2024). Investigation on seepage evolution property of surrounding rock under the tunneling and water surge. Tunn. Undergr. Space Technol. 153, 105970. doi:10.1016/j.tust.2024.105970

CrossRef Full Text | Google Scholar

Wu, G. J., Chen, W. Z., Yuan, J. Q., Yang, D. s., and Bian, H. b. (2017). Formation mechanisms of water inrush and mud burst in a migmatite tunnel: a case study in China. J. Mt. Sci. 14, 188–195. doi:10.1007/s11629-016-4070-8

CrossRef Full Text | Google Scholar

Xie, J., Yin, P., Yang, X., Yan, C., Li, H., Yan, D., et al. (2025). Study on the damage mechanism of water and mud inrush in a tunnel with water-rich fault zones based on experiment and numerical modeling. Tunn. Undergr. Space Technol. doi:10.1016/j.tust.2025.106575

CrossRef Full Text | Google Scholar

Xu, P., Peng, P., Wei, R. H., and Zhang, Z. q. (2022). Model test of the mechanism underpinning water-and-mud inrush disasters during tunnel excavation in sandstone and slate interbedded Presinian strata. J. Zhejiang University-Science A 23 (11), 882–899. doi:10.1631/jzus.a2200134

CrossRef Full Text | Google Scholar

Xue, T., Zhang, Q., Duan, K., Wang, P., Lin, H., Fan, Q., et al. (2022). Geo-mechanical model test on the water inrush induced by zonal disintegration of deep tunnel under hydro-mechanical coupling. Int. J. Rock Mech. Min. Sci. doi:10.1016/j.ijrmms.2022.105278

CrossRef Full Text | Google Scholar

Yang, Z. H., and Zhang, J. H. (2016). Minimum safe thickness of rock plug in karst tunnel according to upper bound theorem. J. Central South Univ. 23 (9), 2346–2353. doi:10.1007/s11771-016-3293-8

CrossRef Full Text | Google Scholar

Zheng, S. (2021). Experimental investigation of water-rich fully weathered granite on water bursting and mud bursting. Front. Earth Sci. 9, 658855. doi:10.3389/feart.2021.658855

CrossRef Full Text | Google Scholar

Zhong, Z., Shen, Z., Qiao, H., Li, Y., and Zhu, K. (2024). Study on mechanism of water and mud inrush in deep-buried large-section tunnel crossing water-rich fault fracture zone. Rock Mech. Rock Eng. doi:10.1007/s00603-024-04176-y

CrossRef Full Text | Google Scholar