Evaluation of Slope Stability and Landslide Prevention in a Closed Open-Pit Mine Used for Water Storage

Abstract

To study and quantify the impact of water storage on lake slope stability after the closure of an open-pit mine, we targeted slope control measures by large-scale parallel computing methods and strength reduction theory. This was based on a three-dimensional refined numerical model to simulate the evolution of slope stability under different water storage levels and backfilling management conditions, and to quantitatively assess the risk of slope instability through the spatial distribution of stability coefficients. This study shows that during the impoundment process, the slope stability has a nonlinear decreasing trend due to the decrease in effective stress caused by the increase in pore water pressure. When the water storage was at 0 m, the instability range is the largest, and the surface range is nearly 200 m from the edge of the pit; when the water level continued to rise to 50 m, the hydrostatic pressure of the pit lake water on the slope support effect began to appear, and the stability was improved, but there is still a wide range of unstable areas at the bottom. In view of the unstable area of the steep slope with soft rock in the north slope during the process of water storage, the management scheme of backfilling the whole bottom to −150 m was proposed, and the slope protection and pressure footing were formed by discharging the soil to −40 m in steps to improve the anti-slip ability of the slope.

1. Introduction

During open-pit mining, as excavation proceeds to greater depths, surrounding groundwater gradually converges toward the pit and discharges naturally along the steep pit walls. Prolonged dewatering operations lead to a significant decline in the groundwater table, an increase in the thickness of the vadose zone, and the eventual formation of a groundwater drawdown cone [1,2]. After mining operations cease and the open-pit mine is closed, the groundwater level gradually recovers [3]. Water storage and lake formation represent a cost-effective and practical approach for the rehabilitation of closed open-pit mines, balancing both ecological restoration and esthetic value [4,5,6].

However, the process of water impoundment may also give rise to a range of potential geological hazards. The increase in pore water pressure reduces the effective normal stress. This promotes the propagation of fractures within the rock mass and decreases the shear resistance of slope materials. Additionally, water can deteriorate the mechanical properties of slope rock and soil through various physicochemical processes. These include erosion, softening and disintegration, shrink–swell cycles, and frost heave [7,8]. Shi Zhenming et al. [9] conducted slope similarity model experiments under varying rainfall intensities and water levels, and found a consistent pattern between pore water pressure and slope surface deformation. Chen Mingliang [10], Wang Jing’e [11], and other researchers analyzed extensive datasets including rainfall, water level, surface displacement, and deep deformation, and investigated the effects of seasonal rainfall and water level fluctuations on the stability of reservoir-related landslides. Zhou Chao et al. [12] conducted coupled numerical analyses of drawdown-induced landslides based on landslide data from the Three Gorges Reservoir area and found that the delayed occurrence of sliding is often caused by the lag in pore pressure dissipation. Liu conducted slope stability analyses under different water level conditions using the numerical software GeoStudio 2018 R2, incorporating hydraulic effects. Under certain water level scenarios, multiple areas were found to have potential landslide risks. Li, H. [1] pointed out that the recovery of the groundwater level increases the pore water pressure within the slope rock mass, potentially submerging joint surfaces and thereby reducing their shear strength. During the pit lake formation process after mine closure, the coupled effects of stress and seepage can further deteriorate the slope rock mass, directly impacting slope stability [1]. The overall effect is that, over time, driven by factors such as glacial melt, rainfall, water level fluctuations, surface runoff, and groundwater activity, the pit walls may experience collapse or failure, posing risks to infrastructure or natural landscapes located above the pit [2]. Slope instability can lead to the release of additional mineral acidity and other oxidized products into the pit lake, resulting in water contamination. In extreme cases, significant slope failures may generate surge waves within the pit lake [13], which could even overflow the mine area and migrate downstream.

The primary methods for evaluating slope stability include engineering geological analysis, limit equilibrium analysis, limit analysis based on plasticity theory, reliability analysis, and numerical simulation methods. Zhang Fei [14] integrated numerical modeling, limit equilibrium analysis, and reliability analysis to investigate the landslide mechanisms and deformation characteristics of a large-scale bedding slope failure controlled by weak layers in the southern slope of the Fushun West Open-Pit Mine. He also analyzed the probability of slope failure after backfilling remediation. Huang et al. [15] analyzed the deformation characteristics and failure modes of the Outang landslide under the influence of rainfall and reservoir water level using hydraulically coupled discrete element numerical simulations. Li Cong et al. [16] performed a statistical analysis of displacement data from numerous landslide cases exhibiting different failure modes. They summarized the correlations between various influencing factors and deformation data at different stages, and subsequently proposed early warning criteria based on the duration of deformation stages, crack characteristics, and displacement rates. Dai Keren et al. [17] proposed a landslide early warning system framework based on Earth observation technologies. Through case studies, they utilized satellite radar monitoring to detect precursory deformation of landslides, achieving early warning capabilities. Carmona et al. [18] studied the large-scale El Arrecife landslide at the Rules Reservoir in Spain and proposed a multi-sensor integrated monitoring approach based on DGPS, inclinometers, and InSAR technologies. Combined with structural kinematic analysis, this method effectively identifies landslide types, potential sliding surfaces, and estimates landslide volume.

Previous studies on the relationship between water level fluctuations and slope stability have primarily focused on reservoir shore slopes, with limited attention given to open-pit mine slopes. In fact, open-pit mine slopes differ significantly from reservoir shore slopes in terms of morphology and engineering geological conditions. The main distinctions are as follows:

(1)

Reservoir dams are generally constructed atop canyon landforms, which typically develop due to tectonic uplift and consist of valley slopes made of hard rocks. Due to long-term fluvial incision and tectonic uplift, steep cliffs are often composed of igneous or metamorphic rocks, while softer rocks are mostly sedimentary. Hydrodynamic landslides frequently occur on slopes composed of soft rocks or rock masses containing interbedded soft layers.

(2)

The morphology of high and steep open-pit mine slopes differs from that of reservoir shore slopes. After impoundment, lateral water pressure on reservoir slopes generally enhances slope stability, whereas the vertical downward water pressure acting on the flat bench faces of open-pit mines adversely affects slope stability.

It is evident that there is currently no targeted research addressing the hydrodynamic landslide mechanisms of slopes with the aforementioned hydrogeological and engineering geological characteristics. However, it is important to note that a large number of open-pit mines in China are expected to enter the phases of production cessation, closure, and comprehensive rehabilitation in the foreseeable future [19]. Given the lack of mature technical guidelines or standards for this issue, there is an urgent need to develop an efficient and feasible dynamic evaluation method for the overall slope stability at different stages of the water impoundment process after mine closure. Such a method would provide critical technical support and reference for geological hazard prevention, ecological restoration, and post-mining utilization of large open-pit mines.

This study aims to explore evaluation methods for slope stability following water impoundment in closed open-pit coal mines, with the goal of providing scientific guidance and support for mine reclamation. Based on large-scale parallel mechanical computations and the strength reduction method, the stability coefficient is defined, and its spatial distribution is simulated to visually identify potential landslide zones and quantify landslide risks. This approach enables three-dimensional dynamic stability analysis of the overall slope during the water impoundment process after mine closure, as well as effective assessment of backfilling remediation schemes.

2. Engineering Overview

Fushun West Open Pit Coal Mine is located in Fushun City, Liaoning Province, China. It is located between the two central districts of Xinfu and Wanghua in Fushun City, a narrow plain between the Hun River in the north and the foot of the Chitai Mountain in the south. The ground elevation of the northern flank of the Qiantai Mountain ranges from 102.9 m to 204.9 m, with a relative height difference of 102 m and natural slopes ranging from 20° to 30°. The north side of the mine is gently sloping and steeply sloping, with Fushun city to the north. The Fushun coalfield consists of inland sedimentary strata from the Paleogene of the Cenozoic era. The coal-bearing stratigraphic system includes the Archean Anshan Group, the Lower Cretaceous Longfengkang Formation, the Paleogene Fushun Group, and the Quaternary, as shown in Figure 1. The sedimentary strata are mainly composed of mudstone and shale, both of which are classified as soft rocks. These soft rock layers overlie granitic gneiss and volcanic rocks. The geological structure of the Fushun coalfield was formed within the Hunhe Fault Zone, which is a northern extension of the Tan-Lu Fault Zone. The structural characteristics of this area are mainly influenced by the horizontal compressive forces and strike-slip traction generated by the activity of the Hunhe Fault after the coal-forming period, resulting in a series of unique structural features.

The West Open-Pit Mine is located in the western part of the Fushun coalfield, where the seam thickness ranges from 35 m to 178 m. The coal seams are embedded within the Early Tertiary rock sequence, consisting of Eocene basalts and tuffs, as well as Eocene coal, oil shale, and mudstone. The strata form a plunging syncline with an axial direction of N80° E. The north limb dips in the opposite direction, while the south limb is overturned. Compared to the north limb, which has a dip angle of approximately 48–55°, the south limb is gentler, with dip angles ranging from about 20–40°. The fold plunges slowly to the east, and the northern limb is partly cut by the F1 thrust fault.

The main faults within the coalfield are represented by the Hunhe Fault, a group of NE–ENE striking thrust faults that are almost parallel to each other, represented by the F1 and F1A faults. Fault F1 strikes NE80° with dip angles of 47–52°, while fault F1A strikes NE80–NE85° with dip angles of 70–75°. There are also associated nearly SN-trending faults and NW-trending faults, such as the F6 fault located at the eastern end of the pit slope, which dips NW85° at an angle of 35°. The groundwater in the Fushun West Open-Pit Mine area is primarily recharged by atmospheric precipitation as well as lateral inflow from the Hun River and Guchengzi River. The main discharge mechanisms include evaporation and lateral underground runoff. A hydrogeological unit is defined as a groundwater analysis region with certain boundaries and uniform conditions of recharge, flow, and discharge. It includes all aquifers involved in groundwater flow and storage, as well as aquicludes that constrain groundwater movement, together with boundary conditions such as groundwater recharge and discharge. The extent of an independent hydrogeological unit is determined by hydrogeological boundaries, which may consist of surface or subsurface watershed divides, aquicludes or impermeable rock bodies, surface water bodies, pumping wells, drainage tunnels, and other features.

The large deep open-pit coal mine measures approximately 6.6 km in length from east to west, 2.2 km in width from north to south, and has a depth of about 424 m. The overall slope angles of the southern slope range between 19° and 27°, while the northern slope angle is approximately 30°. The stratigraphic structure is shown in Figure 1. The northern slope primarily consists of green mudstone, brown shale, coal seams, and oil shale. Both the green mudstone and brown shale are soft rock formations, interbedded and exposed over more than 80% of the slope surface, and are prone to weathering, as shown in Figure 2. The underlying oil shale and coal seams are hard rock layers with thick stratification. The bedding planes dip northward at angles between 20° and 25°, mainly controlled by joints and minor faults. The bedding is indistinct, joints are unfilled, spaced over 0.5 m apart, and exhibit poor continuity. With the mine closure, water impoundment to form a lake is considered a possible reclamation strategy. Therefore, it is necessary to conduct forward-looking research on the stability and control techniques of open-pit mine slopes after water retention.

3. Theoretical Foundation and Numerical Model Construction

3.1. Theoretical Foundation

3.1.1. Percolation Control Equation

Based on the mass conservation equation and Darcy’s law, the governing equation of the seepage field in a porous elastic medium is expressed as follows [20]:

$$\begin{array}{c}{c}_1{\displaystyle \frac{\partial {\epsilon }_v}{\partial t}}+c_2{\displaystyle \frac{\partial P}{\partial t}}=\nabla [k(\nabla P+{\rho }_{1}g\nabla z)]\\ c_1=1-{\displaystyle \frac{K^{\prime }}{K_s}},c_2={\displaystyle \frac{\phi }{{\beta }_1}}+{\displaystyle \frac{(1-\phi )}{K_s}}\end{array}$$

(1)

In the equation, ${\epsilon }_v$ denotes the volumetric strain, t is time (s), p represents pore pressure, k is permeability (m²), g is gravitational acceleration (m/s²), ${\rho }_1$ is fluid density (kg/m³), z is the hydraulic head (m), and ${\beta }_1$ is the bulk modulus of the fluid (Pa). Porosity is denoted by $\phi$. $K^{\prime }$ and $K_s$ are the drained bulk modulus and the effective bulk modulus of the porous medium, respectively (Pa).

3.1.2. Damage Constitutive Equation

The deterioration of materials or structures caused by microstructural defects—such as microcracks and voids—under the influence of mechanical loads and environmental factors is referred to as damage [21]. Based on Lemaitre’s strain equivalence principle [22], the damage constitutive relationship for an isotropic continuous medium can be established by introducing a damage variable D and using the nominal stress $\mathsf{\sigma }$:

$$\epsilon ={\displaystyle \frac{\tilde{\sigma }}{E}}={\displaystyle \frac{\sigma }{(1-D)E}}$$

(2)

In the equation, $\tilde{\sigma }$ and E represent the effective stress and the elastic modulus of the damaged material, respectively. The damage variable D ranges from 0 to 1, where D = 0 indicates an undamaged state and D = 1, corresponds to complete damage. Furthermore, the degree of damage can be characterized by relating the elastic modulus of the damaged element, which is expressed as follows:

$$E=E_0(1-D)$$

(3)

This study refers to the elastic-softening damage model implemented in the RFPA 3D v5.0 software, as illustrated in Figure 3. Tensile stress (or strain) is considered positive, while compressive stress (or strain) is negative. Before the damage threshold is reached, each mesoscopic element is assumed to behave in a purely elastic manner, exhibiting a linear stress–strain relationship. At the macroscopic level, this corresponds to linear elastic behavior. Once the damage threshold is exceeded, the material enters a nonlinear deformation stage, characterized by strain softening—where the stress level decreases with increasing strain until it reaches a constant residual strength [23].

The Mohr–Coulomb strength criterion can be expressed as follows:

$${-\sigma }_3+{\sigma }_1{\displaystyle \frac{1+\mathit{sin}\varphi }{1-\mathit{sin}\varphi }}-{\sigma }_c=0$$

(4)

In the equation, ${\sigma }_c$ is the uniaxial compressive strength of the rock, $\varphi$ is the internal friction angle, ${\sigma }_1$ and ${\sigma }_3$ represent the maximum and minimum principal stresses, respectively.

3.2. Numerical Model

In this study, the previously established damage constitutive equation is embedded into the COMSOL Multiphysics^® 6.1 Multiphysics finite element analysis platform in the form of an external material model, and the nonlinear relationship between stress and strain is established by calling a dynamic link library (DLL) written in C language to realize the numerical simulation of slope stability. The model takes the geomechanical response of the slope as the main research object, and takes into account the influence of different water storage conditions on the slope stability. By applying different head boundary conditions, four different water levels, namely, status quo, −50 m, 0 m, and +50 m, are simulated to systematically analyze their effects on the overall stability of the slope. On the basis of 3D fine geological modeling and high-quality mesh discretization, a massively parallel finite element simulation is carried out with the help of an ASUS high-performance server configured with an Intel Xeon W-3245M CPU (16 cores and 32 threads) and 512 GB of RAM, which can support up to one billion degrees of freedom computational tasks. The numerical solution is based on the Strength Reduction Method (SRM), which iteratively calculates the slope damage evolution process under different strength reduction factors (SRFs), and defines intermediate variables to record the SRF values corresponding to the damage of the units, so as to realize the quantitative characterization of the safety reserve of the units. The final results can not only give the overall slope stability evaluation, but also visualize the spatial distribution and damage pattern of potential sliding zones, which can provide high-precision mechanical information support and risk identification basis for open-pit mine slope engineering.

In this study, the determination of geomechanical parameters takes into account the laboratory test results, the structural characteristics of the rock body, and the actual engineering practice in the field. Firstly, the initial rock strength parameters were obtained based on the indoor mechanical tests of the rock samples; then, by combining the structural features and GSI obtained from the field investigation, the Hoek–Brown Rock Strength Criterion was used to complete the parameter conversion from the experimental scale to the engineering scale, and then to obtain the representative rock strength parameters reflecting the geometric dimensions and structural features of the actual slopes. On this basis, we further refer to a large number of published research results on Fushun West Open-Pit Mine, which is one of the most important large-scale open-pit coal mines in China. Hundreds of related studies have been published to systematically study the geologic conditions and slope stability of this mine, and the mechanical parameters in this paper are also based on the widely accepted research data. In the process of numerical modeling, the simulation results are compared and adjusted with the actual slope stability condition many times, and the parameter combinations with engineering representativeness and field adaptability are finally determined to ensure the accuracy and reliability of the simulation results.

3.2.1. Construction of Numerical Model

Based on the three-dimensional geological model of the mine, high-precision mesh discretization was performed using HyperMesh 2023 software. Considering the lithological diversity and multi-scale characteristics of the mining area, unstructured tetrahedral meshes were employed to accommodate the complex and irregular geological structures, such as multiple faults and folds.

The modeling process was as follows: first, stratigraphic boundaries of lithologically similar layers, faults and secondary structures in non-critical study areas, as well as cleanup benches and safety platforms, were appropriately simplified and merged. This allowed for the construction of a geometric model suitable for 3D numerical simulation while preserving the essential geological features of the open-pit mine. After solid geometry repair (it refers to the necessary checking and repair operations on the imported 3D solid model during the numerical simulation modeling process to ensure the integrity and correctness of the model geometry. Specifically, this includes repairing cracks in the geometry, eliminating overlapping surfaces, closing openings, and correcting unreasonable boundaries to ensure the quality of finite element or finite difference meshing and the stability of subsequent calculations), coplanarity adjustment of stratigraphic layers, and edge smoothing, a surface mesh was generated. The minimum mesh size was controlled at one-fourth of the thickness of the corresponding stratum. In areas with intense stratigraphic undulations or fault intrusions, control lines were manually added, and sub-blocks were divided to optimize the mesh. Mesh quality was checked, and issues such as distorted elements, degenerated cells, and free edges were corrected.

The location of artificial boundaries is a factor to be considered in numerical modeling. Slopes are located in native rock within the site, and the slope model should include as much native rock as possible. However, this usually results in models that are too large and computationally time consuming to be feasible. In practical modeling, the location of artificial boundaries should follow the principle of not affecting the slope damage pattern [22]. The open pit quarry has an east–west length of 6.6 km, a north–south width of 2.2 km, and a height of 420 m. To minimize the boundary effect, the final model size is 15 km × 10 km × 1.5 km, as shown in Figure 4.

After completing the aforementioned preprocessing steps, tetrahedral mesh discretization was carried out. The resulting mesh model consists of approximately 41.18 million nodes and 230 million tetrahedral elements. It incorporates major geological structures and topographic features, including granitic gneiss, Cretaceous sandstone, interbedded mudstone and shale, oil shale, coal seams, basalt, tuff, primary faults (F1, F1A, F5), the unloading-relaxation zone, and 19 major mining benches. The mesh model is shown in Figure 5.

3.2.2. Pore Water Pressure

Under water impoundment conditions, hydrostatic pressure varying with elevation is applied on the slope surface below the water table. The groundwater level within the rock mass is determined from seepage field simulation results under different water level scenarios. Yang Tianhong et al. [25] investigated the distribution of water pressure on open-pit mine slopes and summarized the variation in pore pressure with depth measured by borehole piezometers. They found that the ratio of deep slope pore water pressure to hydrostatic pressure is approximately 0.7. Figure 6 illustrates the variation in pore water pressure in the slope rock mass with piezometer burial depth. The slope of the linearly fitted curve, representing the water pressure coefficient, is approximately 0.61. Therefore, in the simulation, neglecting capillary effects in the unsaturated zone, the pore water pressure in the slope rock mass is assigned according to the following equation:

$$P\left(x,y\right)=k_w·\mathrm{M}\mathrm{a}\mathrm{x}(f\left(x\right)-y,0)·{\gamma }_{w}g$$

(5)

In the equation, $\left(x,y\right)$ denotes the coordinates of an arbitrary point in the numerical model; $k_w$ is the water pressure coefficient; $P\left(x,y\right)$ represents the pore water pressure at that point, with units of Pascal (Pa); $f\left(x\right)$ is the interpolation function of the phreatic surface; the Max () function returns the maximum value of its two arguments; ${\gamma }_w\mathrm{a}\mathrm{n}\mathrm{d}g$ are the unit weight of water and gravitational acceleration, respectively, with units of kg/m³ and m/s².

3.2.3. Boundary Conditions and Physico-Mechanical Parameters

The model boundaries are constrained using roller supports: the lateral boundaries restrict normal displacement while remaining free in the z-direction; the bottom boundary is fixed in the z-direction but unrestrained horizontally. Under water impoundment conditions, hydrostatic pressure varying with elevation is applied to the slope surface below the water level, and corresponding pore water pressures are assigned to the rock mass below the phreatic surface. The physical and mechanical parameters used in the numerical model are listed in

Table 1. The mechanical parameters for numerical calculations.

Lithology	Density [g/cm3]	Elastic Modulus [GPa]	Poisson’s Ratio	Cohesion [MPa]	Tensile Strength [MPa]	Friction Angle [°]	Permeability Coefficient [m/s]
Backfill Material	1.8	0.8	0.3	0.3	0.1	30	3 × 10−2
Green Mudstone	2.25	2	0.24	0.7	0.25	28	2.1 × 10−7
Brown Shale	2.2	1.2	0.3	1	0.3	25	9 × 10−7
Coal	1.5	1.2	0.24	1.4	0.6	40	1.04 × 10−6
Oil Shale	2.1	3.4	0.26	1	0.4	40	3.47 × 10−6
Sandstone	2.3	4	0.25	2	1	45	8.68 × 10−7
Tuff	2.0	0.4	0.2	9.5	3	25	9 × 10−8
Basalt	2.8	7	0.14	10.3	5	45	5 × 10−7
Granite	2.6	8	0.2	10.3	5	45	7.6 × 10−7
Fault Zones (F1, F1A)	2.0	2	0.3	0.7	0.3	20	1.18 × 10−6

.

4. Slope Stability Evaluation During the Water Impoundment Process

The slope stability simulation results under the current condition and water levels of −50 m, 0 m, and +50 m are shown in Figure 7. Due to the northern slope being mainly composed of green mudstone and brown shale, which have relatively low mechanical strength parameters, its stability remains lower than that of the southern slope throughout the entire water impoundment process. As shown in Figure 7a, the minimum stability coefficient is 1.15, located in the central-eastern section of the northern slope, below −150 m elevation. In regions above −50 m, the slope stability coefficients are all greater than 1.3. This indicates that under current conditions, the slope is generally in a relatively stable state. It also suggests that prior to water impoundment, the existing slope has not been significantly affected by pore water pressure, resulting in higher effective stress and stronger resistance to sliding.

At the water level of −50 m (Figure 7b), the stability coefficients of the northern slope from the middle-western to the eastern end of the open-pit mine, below the pit bottom and the pit lake water level, are all below 1, indicating a high likelihood of landslides. The western part of the northern slope above the water level, as well as the southern slope, remains generally stable. On the eastern northern slope, the stability coefficient gradually increases with elevation; however, some areas still have stability coefficients below 1.1, highlighting potential landslide risks. These unstable zones are confined to the shallow surface layers of the slope and have not yet affected the ground surface. This indicates that during the impoundment process, increasing pore water pressure reduces the effective stress, thereby weakening the slope’s shear resistance.

At the water level of 0 m (Figure 7c), the stability coefficients of the northern slope below the water table remain below 1, while the unstable zones above the water level have significantly expanded. A marginally stable zone with a stability coefficient below 1.1 appears in the eastern part of the southern slope. A large-scale instability area emerges from the eastern part of the northern slope to the eastern end wall, where the stability coefficients above the water table are generally less than 1. A wide surface area with stability coefficients between 1.1 and 1.2 is observed, extending up to approximately 200 m from the pit edge. These results indicate that as the water level continues to rise, pore water pressure increases further, leading to a significant reduction in effective stress and a marked decline in the slope’s shear resistance. Moreover, the instability in the submerged zones exerts a traction effect, triggering a decrease in the stability of the slope areas above the water level.

At a water level of +50 m (Figure 7d), the pit lake level approaches the groundwater level within the slope, and the stabilizing effect provided by hydrostatic pressure is significantly greater than that at the 0 m water level. As a result, the overall stability coefficients exhibit a slight increase. However, due to the weakening effect of water saturation, a large area below the water surface still shows stability coefficients less than 1. The marginally stable zones above the water level have not been significantly reduced, although the surface stability coefficients are all greater than 1.2.

5. Research on Slope Treatment Plan

5.1. Management Approach

The bottom slopes of the open-pit mine are relatively steep. The slope stability evaluation results during the water impoundment process (Figure 7) indicate that, under different water levels, the bottom section of the northern slope—from the roof of the oil shale upward—remains in an unstable state. To ensure the stability of the bottom slopes of both the southern and northern walls, as well as the interbedded green mudstone and brown shale zone, and to prevent environmental contamination caused by coal–oil mixed pollutants originating from the coal seams and oil shale at the pit bottom, a comprehensive backfilling of the pit bottom is proposed, along with toe reinforcement of the slope in the green mudstone and brown shale interbedded zone [26].

The preliminary design of the backfilling scheme is as follows: high-density fill materials (such as waste rock or concrete) will be used to backfill the pit bottom up to an elevation of −150 m, forming a toe structure to increase the self-weight of the bottom slope and thereby enhance its sliding resistance. Below the elevation of −40 m on the northern slope, stepped toe-reinforcement will be applied to mitigate the adverse effects of water on the stability of the interbedded green mudstone and brown shale slopes. Between −150 m and −40 m, three stepped internal benches will be constructed. The overall internal slope angle is designed as 16°, with a single bench face angle of 30°. Starting from −150 m, the first and second benches are each 40 m high, while the third is 30 m high. The berm width is 90 m. The upper slope section requires complete drainage, and the surface needs to be sealed with concrete or impermeabilized using an HDPE geomembrane. The detailed layout of the scheme is shown in Figure 8.

5.2. Evaluation of Stability of the Slope After Treatment

The slope stability under two water level scenarios, −50 m and +50 m, after implementation of the backfilling scheme, is discussed herein. The simulation results of the slope stability after treatment are shown in the corresponding figures. Although the backfilled body was included in the numerical simulations, to better evaluate the effect of the internal benching measures on slope stability, the backfill is rendered semi-transparent in the visualization. This allows for clearer observation of changes in the stability of the original slope beneath the backfilled area.

The simulation results after backfilling are shown in Figure 9. The comparison results before and after backfilling show that the backfilling scheme significantly improves the stability of the water side slopes. Before backfilling, the minimum safety factors (MSFs) of the north slopes were 0.85 and 0.78 at water depths of −50 m and 50 m, respectively, which showed high landslide risks. After backfilling, the minimum safety coefficients increased to 1.05 and 1.02, respectively, indicating that backfilling effectively improved the slope stability. In addition, the application of backfill material significantly reduced the erosion of the original rock mass by water and the accumulation of pore water pressure, only the slope rock mass near the water level was affected by water erosion, and the local instability of the shallow surface layer appeared, but the range of its influence was small. Compared with the area before backfilling, the unstable area in the eastern part of the north slope was reduced by about 66%.

Further analysis of water storage to −50 m and 50 m shows that the unstable range of the slope expands with the increase in water level. When the water storage reaches −50 m, the eastern part of the north slope becomes unstable; while when the water reaches 50 m, not only is the eastern part unstable, but also the western part starts to appear unstable area. Combined with the significant increase in the safety factor after backfilling and the simulation results, it is recommended that the maximum storage level be limited to less than 50 m. This water level limit can effectively prevent the sidewalls from becoming unstable. This water level limit can effectively avoid the further expansion of the unstable area of the slope and ensure the long-term stability and safety of the slope.

6. Discussion

This study aims to investigate the mechanisms by which pit lake formation after open-pit mine closure affects slope stability, and to propose engineering-feasible slope stabilization strategies. The central hypothesis is that as the water level gradually rises, pore water pressure within the slope rock mass increases significantly, reducing effective stress and consequently diminishing slope stability. However, appropriately designed backfilling and toe-reinforcement measures can effectively enhance slope resistance to sliding. To test this hypothesis, the study employs high-resolution 3D geological modeling and unstructured mesh discretization, integrated with large-scale parallel mechanical computation and strength reduction theory, to simulate the spatial evolution of slope stability factors under different water impoundment scenarios and to assess the effectiveness of backfilling treatments.

The numerical simulation results reveal a nonlinear decline in slope stability with rising water levels. Notably, when the impoundment reaches 0 m, the slope exhibits the largest extent of instability, with surface deformations extending up to 200 m from the pit edge. After the water level reaches 50 m, a slight recovery in slope stability is observed, attributed to the increasing hydrostatic pressure exerted by the water body, which provides passive support to the slope. These findings confirm the fundamental mechanical principle that elevated pore water pressure reduces shear resistance, consistent with previous studies on reservoir bank landslide evolution [1,27,28].

Moreover, the proposed “integrated backfilling combined with stepped toe reinforcement” strategy effectively mitigates slope failure in the simulation, demonstrating promising engineering applicability. The novelty of this study lies in two key aspects: (1) the development of a high-fidelity 3D numerical model with 230 million elements, which, supported by a large-scale parallel computing platform, enables the first full-domain simulation of slope stability evolution in an open-pit mine; and (2) the design of a stepped toe backfilling approach for the soft-rock, high-angle northern slope, which simultaneously addresses environmental containment and mechanical stability, offering strong practical value and potential for broader application.

Nevertheless, this study has certain limitations. First, the simulation focuses solely on the impact of rising water levels during the impoundment phase, without addressing the drawdown stage that may occur during lake operation due to regulation, pumping, or natural evaporation. Numerous engineering cases have shown that rapid water level decline can generate steep hydraulic gradients, leading to delayed pore pressure dissipation and increasing the risk of landslides and other geohazards [1,27,29]. In addition, some rock mass parameters in the current model were estimated based on literature or laboratory tests, which may differ from in situ conditions. Future work should incorporate field monitoring data and apply inverse modeling techniques to improve parameter accuracy and model reliability.

Moreover, the slope’s response to external disturbances such as seismic loading, long-term wet–dry cycling, and freeze–thaw effects should also be considered in future studies. In summary, this study provides theoretical insights and technical strategies for the dynamic evaluation and stabilization of pit slopes during post-closure water impoundment. However, further advancements are needed in multi-condition, multi-stage, and multi-source coupled modeling to more comprehensively support real-world engineering applications.

7. Conclusions

In this study, large-scale parallel mechanical computation was integrated with the strength reduction method to evaluate slope stability and assess backfilling treatment schemes during the water impoundment process in a decommissioned open-pit mine. The following conclusions were drawn:

(1)

The current slope of the open-pit mine is basically stable, the stability coefficient of the slope from the bottom of the pit to the water level of the pit lake below the north slope is less than 1, which is very prone to landslides, and with the rise in the water level of the water storage, the range of unstable area gradually increases, and the stability coefficient of the slope above the water level when the water is stored to more than 0 m is less than 1, and the wave of the ground surface from the edge of the pit at nearly 200 m. The current slope of the open-pit mine is basically stable.

(2)

When the water is stored to 50 m, the pit lake water is located close to the groundwater level in the slope, and the stabilizing support provided to the slope by hydrostatic pressure is significantly higher than that in the case of water storage to 0 m, so the stability coefficient is slightly improved.

(3)

A management plan of a step berm pressure footing was proposed for the unstable area to reduce the adverse effect of water on the stability of the north slope.

(4)

The overall backfilling of the bottom can not only prevent the coal and oil mixed pollutants from the coal seam and oil mother shale at the bottom of the pit from seeping into the lake, but also effectively improve the stability of the north and south slopes.

(5)

For the north slope soft rock high steep slope multi-step soil discharge, not only the north slope play a slope protection pressure foot role, but also prevents the pit lake erosion north original rock slope disintegration to form larger landslides.