Water-Induced Shear-Strength Degradation of Coal-Measure Rocks and Its Engineering Implications: A Case Study of the Fushun West Open-Pit Mine, China

Abstract

Slopes excavated in coal-measure strata are prone to rainfall-induced landslides because water-induced disturbances progressively degrade rock-mass shear strength. This study investigated six coal-measure lithologies from the unloading zone of the Fushun West Open-Pit Mine using triaxial compression tests after separate wetting–drying and continuous-immersion treatments. Condition-dependent cohesion, c, and internal friction angle, φ, were fitted with exponential functions and incorporated into the Mohr–Coulomb criterion. For scenario analysis, the retention ratios measured in the two separate treatment series were applied sequentially in a separable empirical parameterization. This no-interaction approximation was used to update coal-measure strength inputs in representative slope models. Cohesion and φ decreased exponentially within the tested ranges, with in-sample R² values greater than 0.85, and cohesion was generally more sensitive. Continuous immersion produced rapid early softening, whereas wetting–drying cycles produced cumulative deterioration. Limit-equilibrium and finite-element calculations showed lower safety factors and more concentrated deformation under more severe parameter-reduction scenarios. Model-specific cumulative-displacement reference values of 52.36–213.82 mm were paired with Fs levels of approximately 1.15, 1.05, and 1.00 and compared qualitatively with historical monitoring curves. The findings provide a reference for rainy-season slope-stability prediction and staged warning in open-pit coal mines.

1. Introduction

Coal-measure strata are sedimentary rock assemblages containing coal seams and associated lithologies formed during specific geological periods, and they provide the basic geological framework for coal occurrence and extraction. Compared with surrounding rocks in many other mineral deposits, coal-measure rocks generally have lower strength and are more prone to deformation and failure [1,2]. Regardless of whether coal mining is conducted in open-pit or underground settings, rock masses are inevitably affected by water-induced disturbances throughout their mining life cycle [3,4]. In practice, these disturbances are mainly related to repeated rainfall infiltration, cyclic groundwater-level fluctuations, and engineering activities such as drainage and water injection [5,6,7].

Rock masses in water-affected zones, such as shallow weathered slope rocks and surrounding rocks of underground roadways, are therefore frequently exposed to wetting–drying cycles and intermittent immersion. These processes promote rock softening and crack development, thereby weakening the mechanical properties of coal-measure rock masses [8,9,10,11]. In particular, shear strength is highly sensitive to water-induced degradation and is closely related to rock mass stability [12,13,14]. Such degradation may further contribute to mine slope instability, water inrush, and other secondary hazards, compromising the safety and stability of mining operations [15,16]. Therefore, understanding the degradation behavior of coal-measure rock masses under water–rock interactions is essential for water-induced hazard control and safe coal extraction.

Geohazards associated with sustained and cyclic water–rock interactions, including repeated rainfall and reservoir-level fluctuations, have been increasingly reported in recent years [17,18,19]. These events have motivated studies on water-induced rock-mass deterioration, mainly focusing on deformation–failure behavior, structural evolution, strength degradation, and constitutive modeling. For deformation–failure behavior, Chen et al. [20] examined limestone under cyclic hydraulic loading and showed that fluctuating high water pressure affects deformation patterns, delayed responses, strain rate, and crack propagation. Using digital image correlation (DIC), Zhang et al. [21] tracked the deformation and fracture evolution of jointed rocks and identified four failure modes: tensile, shear–tensile, buckling–shear, and buckling–shear–tensile. For structural evolution, Qin et al. [22] used scanning electron microscopy (SEM) to examine the microstructural evolution of granite under water–rock interactions, dividing the process into four stages and relating damage to the number of cycles using fractal dimensions. Mao et al. [23] combined computed tomography (CT) with 3D pore network reconstruction to investigate pore-radius and connectivity evolution in sandstone during immersion cycles. For strength characteristics, Zhang et al. [24] reported exponential P-wave velocity decay and linear compressive-strength reduction in sandstone under wetting–drying cycles and further proposed a strength prediction model incorporating wave velocity, confining pressure, and the number of cycles. Xu et al. [25] identified anisotropic shear-strength degradation in thin-bedded rock masses under cyclic water-level fluctuations through in situ shear tests. For constitutive modeling, Huang et al. [26] developed a damage constitutive model for rock masses under wetting–drying cycles by combining triaxial compression tests, the Weibull distribution, and the Drucker–Prager criterion. Fu et al. [27] proposed a temperature- and cycle-dependent damage model for siltstone under high-temperature and water-cooling cycles and validated its ability to describe strain softening and residual strength.

In mining environments, coal-measure rock masses are also susceptible to water-induced deterioration [28,29]. Although water-disturbed coal-measure rocks have been widely investigated, most studies have considered a single lithology [30,31]. Comparative responses of multiple coal-measure lithologies to distinct water-induced processes remain incompletely quantified [32,33], and links between laboratory conditioning and field mining conditions are often limited [20,21,22,23]. The Fushun West Open-Pit Mine in China is a typical landslide-prone mining area. Since 1927, 103 landslides have been recorded on its northern and southern slopes, approximately 70% of which were rainfall-induced [34]. In the Fushun Coalfield, water-sensitive strata such as mudstone, shale, coal, and tuff constitute major slope materials and commonly occur as weak interbeds. Rainfall infiltration into the discontinuity-rich unloading zone can soften these materials and reduce slope resistance [35,36]. These conditions motivate comparative quantification of strength loss under separately controlled wetting–drying and immersion treatments, together with a clear statement of the limits of any scenario combination.

This study therefore focuses on coal-measure rocks sampled from the rainfall-affected unloading zone of the Fushun West Open-Pit Mine. Triaxial compression tests were performed after wetting–drying cycles (0, 1, 3, 5, and 7 cycles) and continuous immersion (0, 1, 4, 7, and 10 days). The two treatment series were evaluated separately to quantify condition-dependent c and φ. These fitted parameters were then used in a separable, no-interaction scenario model that applies the two measured retention ratios sequentially. The resulting parameter sets supported limit-equilibrium stability calculations and finite-element deformation calculations for three representative slope profiles. The study finally examines a model-specific mapping among rainfall-related scenarios, safety-factor levels, and cumulative surface displacement. The mapping is intended for comparative scenario assessment in Fushun.

2. Materials and Methods

2.1. Fushun Coal Measures and Water-Induced Weakening

This study examines coal-measure rocks from the Fushun Coalfield in Liaoning Province, northeastern China. Tectonically, the Fushun Coalfield is located at the junction of the Neocathaysian Fushun–Mishan fault-depression zone and the east–west-trending Shenyang uplift zone. It is a synclinal basin formed by southward high-angle thrusting of the Hunhe Fault, a northern secondary fault of the Tan–Lu Fault Zone [37]. The coalfield syncline is underlain by Archean granitic gneiss, Jurassic clastic and effusive rocks, and Cretaceous clastic rocks, and has an asymmetric geometry with a nearly east–west-trending axis. Most of its northern limb was truncated by the Hunhe Fault, whereas the southern limb was preserved under compressional uplift and forms the main body of the Fushun coal-measure strata [38]. Stratigraphically, the coal-measure strata are designated as the Fushun Group of Paleocene–Eocene age. From bottom to top, the group consists of the Laohutai Formation (E₁¹fl) basalt, Lizigou Formation (E₁²fl) tuff, Guchengzi Formation (E₂¹fg) coal, Jijuntun Formation (E₂²fj) oil shale, Xilutian Formation (E₂³fx) mudstone and shale, and Gengjiajie Formation (E₂⁴fg) shale (Figure 1a) [39].

The Fushun Coalfield has high mining value because of the large-scale occurrence of shallow, exceptionally thick coal seams [40], and two large open-pit mines and three underground mines have been developed. Located in the western part of the coalfield, the Fushun West Open-Pit Mine is one of the largest open-pit coal mines in Asia. It has complex geological conditions, with numerous landslide hazards distributed within the slope unloading zones (Figure 1b). Time-series analysis of landslide frequency and rainfall shows that landslides increase markedly during the rainy season, indicating that rainfall is a major trigger of slope failures in the mine (Figure 1c). Historical data from June to September further show a delayed landslide response to rainfall; the offset between rainfall peaks and landslide-frequency peaks suggests that antecedent rainfall also plays an important role in triggering slope instability. For example, the repeated rainfall events preceding the July 26 north-slope landslide in 2016 represent a typical hazardous rainfall pattern in the mining area (Figure 1d) [18,35]. Compared with the external triggering effect of multiple consecutive rainfall events, the degradation of coal-measure rock masses caused by repeated water-induced disturbances is a key internal driver of slope instability [41]. Accordingly, this study focuses on coal-measure rock masses from the engineering-disturbed unloading zone of Fushun Coalfield. Mechanical tests were conducted under continuous immersion and wetting–drying cycles to simulate water-induced disturbances and quantify the resulting damage behavior.

2.2. Sample Preparation

Except for the Gengjiajie Formation (E₂⁴fg) shale, which has not been involved in coalfield development, the remaining coal-measure strata are well exposed in the Fushun West Open-Pit Mine (Figure 1a). Sampling sites were arranged within the mine to cover the principal exposed lithologies and engineering settings. Following the ISRM suggested methods [42], representative rock blocks with relatively uniform textures and typical structural planes were collected from the slope unloading zone and prepared as cylindrical specimens. The tested materials comprise Laohutai Formation (E₁¹fl) basalt, Lizigou Formation (E₁²fl) tuff, Guchengzi Formation (E₂¹fg) coal, Jijuntun Formation (E₂²fj) oil shale, and Xilutian Formation (E₂³fx) mudstone and shale (Figure 1a,b and Figure 2a).

Immediately after collection, the rock blocks were wrapped in plastic film and then in bubble wrap to limit moisture exchange and protect them from mechanical shock. They were kept at room temperature in a temperature-controlled container and transported to the laboratory together after field sampling was completed. Core orientation relative to bedding or cleavage was not systematically measured. During preparation of visibly anisotropic shale, the cutting orientation was selected, where practicable, to approximate the in situ attitude observed at the sampling location.

To ensure the reliability of the triaxial compression tests, the end-face flatness and parallelism of the specimens, as well as their perpendicularity to the specimen axis, were strictly controlled in accordance with the ISRM suggested methods [42]. During specimen preparation, cores were first drilled from large rock blocks using an RCD-250 pressure-controlled coring machine (GCTS Testing Systems, Tempe, AZ, USA). The cores were then cut and finely ground using a YMT-8 grinding machine (ISHAN Precision Ind. Co., Ltd., Taichung, China) (Figure 2b) to prepare standard cylindrical specimens with a diameter of 50 mm and a height of 100 mm. Typical prepared coal-measure rock specimens are shown in Figure 2c.

Before testing, the natural, dry, and saturated densities of coal-measure rock specimens from the study area were determined using the volumetric, oven-drying, and vacuum saturation methods [43], as listed in

Table 1. Physical properties of coal-measure rock specimens from the Fushun Coalfield.

Lithology	Natural Density, ρn (g·cm−3)	Dry Density, ρd (g·cm−3)	Saturated Density, ρsat (g·cm−3)	Saturated Water Absorption, wsat (%)	Porosity, n (%)
basalt	2.54	2.51	2.57	2.26	5.67
tuff	2.34	2.29	2.40	4.86	11.13
coal	1.45	1.42	1.48	4.23	6.01
oil shale	2.16	2.13	2.20	3.42	7.28
mudstone	2.23	2.19	2.27	3.62	7.93
shale	2.02	1.97	2.06	4.64	9.14

.

2.3. Test Equipment

Triaxial compression tests were conducted using a GCTS RTR-2000 rock triaxial testing system (GCTS Testing Systems, Tempe, AZ, USA) housed at the University of Science and Technology Beijing (Figure 3a). This system is mainly used to determine the static and dynamic mechanical properties of rocks under both conventional and high-temperature/high-pressure conditions. It can simultaneously measure the stress–strain response, acoustic parameters, and permeability properties of specimens under varying temperature and pressure conditions.

The GCTS RTR-2000 has a rated maximum axial force of 2000 kN and rated maximum confining and pore pressures of 140 MPa; these values describe the capacity of the apparatus and were not the loading conditions used in this study. The present tests used confining pressures, σ₃, of 0, 2.5, 5.0, and 7.5 MPa, and no pore-pressure loading was prescribed. At each confining pressure, specimens were loaded under axial displacement control at 0.06 mm/min until failure. Axial and radial deformation were measured with linear variable differential transformers, and the stress–strain records were acquired automatically.

2.4. Test Scheme

Triaxial compression tests were conducted to determine the shear-strength parameters of rock specimens under water–rock interactions. Two types of water-induced disturbances, namely wetting–drying cycles and continuous immersion, were considered to simulate the effects of repeated concentrated rainfall during the rainy season on the mechanical properties of coal-measure rocks from the engineering-disturbed unloading zone in the Fushun area.

Based on the local rainy-season rainfall record [44], the test matrix included 0, 1, 3, 5, and 7 wetting–drying cycles and continuous-immersion durations of 0, 1, 4, 7, and 10 days. In the present analysis, the 10-day endpoint extends the observed laboratory window from rapid early softening toward the later, slower stage of strength change. It is an experimental conditioning duration and is not assumed to reproduce a 10-day field rainfall event on a one-to-one basis. Because zero cycles and zero immersion both denote the untreated natural-state condition, the program comprised nine treatment conditions.

One wetting–drying cycle consisted of oven drying at 105 °C for 12 h, cooling to room temperature, and vacuum-assisted water treatment for 8 h. One day of immersion denotes 24 h of continuous immersion in purified water. This standardized accelerated conditioning protocol provides repeatable treatment endpoints, but it does not reproduce ambient field drying. In water-sensitive clay-bearing rocks, heating at 105 °C may contribute to thermal or mineralogical damage. The results are therefore interpreted as responses to the stated laboratory conditioning protocol. For each lithology and treatment condition, four specimens were tested, one at each of the four confining pressures, σ₃ = 0, 2.5, 5.0, and 7.5 MPa. These four specimens were pressure-level observations, not parallel replicates at a common confining pressure. The four-level range was used to fit a Mohr–Coulomb envelope; the lower pressures are more representative of the shallow unloading zone, whereas the 7.5 MPa point broadens the fitted stress range.

The tested specimens covered six lithologies and nine treatment conditions. With one specimen at each of four confining pressures for every lithology–condition combination, the program comprised 216 specimens in total. Accordingly, each lithology–condition–pressure cell contained one specimen. The following comparisons and fitted trends are therefore descriptive. The complete test scheme is summarized in

Table 2. Shear-strength test scheme for coal-measure rock specimens.

Lithology	Confining Pressure, σ3 (MPa)	Number of Wetting–Drying Cycles, t (-)	Immersion Duration, d (Days)
basalt, tuff, coal, oil shale, mudstone, shale	0.0, 2.5, 5.0, 7.5	0	0
		1	1
		3	4
		5	7
		7	10

.

3. Experimental Results and Analysis

For each lithology, peak axial strength, σ₁, was obtained under the nine treatment conditions and four prescribed confining pressures (

Table 3. Triaxial compression test results of coal-measure rock specimens.

Water-Treatment Condition	Confining Pressure, σ3 (MPa)	Peak Axial Strength, σ1 (MPa)
		Basalt	Tuff	Coal	Oil Shale	Mudstone	Shale
0 wetting–drying cycles/ 0-day immersion	0.0	14.754	1.476	3.563	1.514	1.090	0.684
	2.5	42.494	14.686	9.125	12.978	7.736	6.994
	5.0	56.107	26.486	15.104	24.913	13.968	13.480
	7.5	86.715	40.284	20.784	35.997	20.632	19.763
1 wetting–drying cycle	0.0	13.094	1.258	3.045	1.451	0.901	0.611
	2.5	40.000	13.524	8.238	12.512	7.393	6.445
	5.0	53.373	25.836	15.024	24.165	13.523	12.371
	7.5	82.725	38.030	19.420	34.803	20.053	18.315
3 wetting–drying cycles	0.0	12.124	1.103	2.787	1.381	0.834	0.592
	2.5	37.734	13.156	7.969	12.216	7.014	6.309
	5.0	50.782	24.319	13.832	23.612	12.869	12.096
	7.5	78.816	36.739	18.811	34.047	19.052	17.914
5 wetting–drying cycles	0.0	10.171	0.906	2.468	1.343	0.659	0.528
	2.5	35.951	12.859	7.425	12.094	6.673	6.054
	5.0	47.279	23.708	13.292	23.379	12.354	11.627
	7.5	75.441	36.122	17.892	33.719	18.394	17.267
7 wetting–drying cycles	0.0	9.153	0.708	2.099	1.317	0.568	0.493
	2.5	33.068	12.488	6.138	11.746	6.495	5.893
	5.0	46.362	23.018	12.229	22.915	12.066	11.386
	7.5	72.034	35.334	15.170	33.035	18.017	16.902
1-day immersion	0.0	11.596	1.101	2.747	1.457	0.811	0.536
	2.5	37.384	13.603	7.849	12.109	7.089	6.315
	5.0	51.963	24.859	14.034	23.568	13.019	12.303
	7.5	80.440	37.896	19.036	34.181	19.349	18.081
4-day immersion	0.0	10.825	0.929	2.515	1.425	0.726	0.497
	2.5	35.931	13.233	7.732	11.703	6.577	5.514
	5.0	49.875	24.268	13.861	22.902	12.119	10.763
	7.5	77.850	37.107	18.085	33.196	18.054	15.776
7-day immersion	0.0	10.406	0.785	2.334	1.377	0.692	0.469
	2.5	35.052	13.042	7.182	11.597	6.409	5.465
	5.0	48.920	24.158	13.451	22.211	11.835	10.684
	7.5	76.478	36.884	17.895	32.837	17.651	15.609
10-day immersion	0.0	10.036	0.758	2.095	1.354	0.664	0.443
	2.5	34.614	12.937	7.033	11.486	6.287	5.326
	5.0	48.237	23.851	12.903	22.154	11.621	10.428
	7.5	75.505	36.527	16.486	32.713	17.323	15.284

). Untreated specimens (zero wetting–drying cycles and zero immersion) were used as the baseline. For any strength quantity X, the deterioration degree was calculated as Dd = (X₀ − Xw)/X₀, where X₀ is the untreated value, and Xw is the value after conditioning. Positive Dd denotes an apparent reduction.

3.1. Uniaxial Compressive Strength

Based on the laboratory test results, the degradation effects of immersion duration and wetting–drying cycles on the uniaxial compressive strength (UCS) of six types of coal-measure rock specimens were analyzed (Figure 4). The UCS decreased nonlinearly under water-induced disturbances, with degradation patterns varying with lithology and the water-induced process.

Figure 4a shows the UCS changes under wetting–drying cycles. For UCS, Dd followed the order of tuff (52.0%) > coal (41.1%) > basalt (38.0%) > mudstone (34.4%) > shale (28.4%) > oil shale (12.6%). The specimens showed considerable early-stage loss after the first cycle. For tuff and basalt, which had the lowest first-cycle proportions, the initial loss accounted for 28.4% and 29.6% of the cumulative loss, respectively; for shale and mudstone, these proportions were 38.2% and 36.2%. UCS decreased across the measured specimens as the number of cycles increased.

Figure 4b shows the UCS degradation of the specimens under continuous immersion. For UCS, Dd followed the order of tuff (48.6%) > coal (41.2%) > mudstone (39.1%) > shale (35.2%) > basalt (32.0%) > oil shale (10.6%). Tuff showed the greatest degradation, whereas oil shale exhibited the highest resistance to immersion-induced degradation. Immersion-induced degradation showed a pronounced stage-dependent pattern, with the UCS of all specimens decreasing rapidly at the early stage and then gradually stabilizing. Within the first day of immersion, Dd exceeded 20% for all lithologies except oil shale, accounting for 52.2–66.9% of the total UCS reduction within the tested range. For oil shale, the first-day UCS reduction accounted for 35.6% of the total reduction, exceeding its average daily reduction rate.

Both separate treatment series reduced UCS within the tested ranges, but they describe different aspects of degradation. Immersion produced a larger fraction of its total measured loss during the first day, whereas wetting–drying produced progressive loss over successive cycles.

The measured UCS values retained a clear lithology-dependent hierarchy. Based on the observed UCS ranges, the specimens can be described in three strength classes: hard rock, represented by basalt; medium-hard rocks, including tuff and oil shale; and soft rocks, including coal, mudstone, and shale. Water treatment reduced UCS to different degrees, but no cross-class transition was observed in the tested specimens. This descriptive result shows that the original lithological strength ordering was retained within the experimental window; it does not establish a statistically dominant causal factor.

3.2. Peak Axial Strength

The triaxial compression results (Table 3 and Figure 5) showed three descriptive patterns consistent with the UCS results. Under the tested σ₃ conditions, (i) both continuous immersion and wetting–drying cycles reduced σ₁ in all specimens; (ii) comparison of Figure 5a–f with Figure 5g–l showed that early-stage degradation contributed more to the total σ₁ reduction under immersion than under wetting–drying cycles; and (iii) water treatment did not alter the relative σ₁ hierarchy among rock types of different hardness. The preserved hierarchy is consistent with lithology-dependent differences, while confining pressure and treatment condition modified the measured strengths.

Analysis of Dd for σ₁ under different confining pressures (

Table 4. Deterioration degree of peak axial strength, σ₁, under endpoint water-treatment conditions and different confining pressures.

Lithology	Endpoint Water-Treatment Condition	Deterioration Degree, Dd for σ1 (-)				Inhibition Ratio, σ3 IR (-)
		σ3 = 0 MPa	σ3 = 2.5 MPa	σ3 = 5 MPa	σ3 = 7.5 MPa
basalt	7 wetting–drying cycles	0.380	0.222	0.174	0.169	0.554
	10-day immersion	0.320	0.185	0.140	0.129	0.596
	ΔDd for σ1	0.060	0.036	0.033	0.040
tuff	7 wetting–drying cycles	0.520	0.150	0.131	0.123	0.764
	10-day immersion	0.486	0.119	0.099	0.093	0.808
	ΔDd for σ1	0.034	0.031	0.031	0.030
coal	7 wetting–drying cycles	0.411	0.327	0.190	0.270	0.343
	10-day immersion	0.412	0.229	0.146	0.207	0.498
	ΔDd for σ1	−0.001	0.098	0.045	0.063
oil shale	7 wetting–drying cycles	0.130	0.095	0.080	0.082	0.368
	10-day immersion	0.106	0.115	0.111	0.091	0.137
	ΔDd for σ1	0.024	−0.020	−0.031	−0.009
mudstone	7 wetting–drying cycles	0.479	0.160	0.136	0.127	0.735
	10-day immersion	0.391	0.187	0.168	0.160	0.590
	ΔDd for σ1	0.088	−0.027	−0.032	−0.034
shale	7 wetting–drying cycles	0.279	0.157	0.155	0.145	0.482
	10-day immersion	0.352	0.238	0.226	0.227	0.357
	ΔDd for σ1	−0.073	−0.081	−0.071	−0.082

Note: (1) Dd for σ₁ denotes the deterioration degree of peak axial strength, σ₁. (2) σ₃ IR represents the relative decrease in Dd for σ₁ as σ₃ increases from 0 to 7.5 MPa under the same endpoint water-treatment condition. (3) ΔDd for σ₁ is the difference in Dd for σ₁ between 7 wetting–drying cycles and 10-day immersion at the same σ₃. Positive values indicate greater degradation after wetting–drying cycles, whereas negative values indicate greater degradation after immersion.

) further reveals the following similarities and differences among the different lithologies.

Common feature: Increasing confining pressure, σ₃, consistently suppressed water-induced degradation in σ₁ across all lithologies. Under both continuous immersion and wetting–drying cycles, higher σ₃ enhanced lateral confinement and slowed crack propagation within the specimens, leading to an overall decrease in Dd for σ₁ under the tested conditions.

Distinctive feature 1: The influence of confining pressure, σ₃, on water-induced σ₁ degradation varied among lithologies. The confining-pressure inhibition ratio, σ₃ IR, was defined as the relative decrease in Dd for σ₁ as σ₃ increased from 0 to 7.5 MPa under the same endpoint water-treatment condition. A larger σ₃ IR indicates stronger suppression of water-induced σ₁ degradation by confining pressure.

(i)

σ₃ IR was higher for basalt, tuff, and mudstone (>0.5), indicating a stronger inhibition of Dd for σ₁, whereas it was lower for coal, oil shale, and shale (<0.5).

(ii)

For basalt, tuff, and coal, σ₃ IR was slightly higher after 10 days of immersion; for oil shale, mudstone, and shale, it was slightly higher after seven wetting–drying cycles. Because this comparison is made only between the two endpoint treatments, it should be interpreted as an end-stage trend rather than evidence that one water-treatment path consistently produces higher σ₃ IR throughout the process.

Distinctive feature 2: Under different σ₃ conditions, the sensitivity of σ₁ degradation to wetting–drying cycles and continuous immersion varied among lithologies. To characterize this difference, ΔDd for σ₁ was defined as the difference in Dd for σ₁ between seven wetting–drying cycles and 10-day immersion at the same σ₃. A positive value indicates greater sensitivity to wetting–drying cycles, whereas a negative value indicates greater sensitivity to continuous immersion. Although immersion duration and the number of wetting–drying cycles are not directly equivalent, this comparison is made between two fixed endpoint water-treatment conditions; for a given lithology, σ₃ is the only variable. Thus, ΔDd for σ₁ can be used to compare the relative degradation effects of the two water-treatment paths under different confining pressures.

As σ₃ increased from 0 to 7.5 MPa, ΔDd for σ₁ remained positive for basalt and tuff, meaning that the measured endpoint reduction was greater after wetting–drying cycles than after immersion, whereas it remained negative for shale, for which the immersion endpoint showed the greater reduction. For oil shale and mudstone, the sign of ΔDd changed across confining pressures, so the relative ordering of the two endpoint reductions was not stable. For coal, ΔDd changed from a near-zero negative value to positive values.

3.3. Shear Strength Parameters

Using the data in Table 3, σ₁–σ₃ relationships were fitted for each lithology and treatment condition (Figure 6). The standard Mohr–Coulomb criterion was expressed in principal-stress form to determine the strength envelope [45], as given in Equation (1).

$${\sigma }_1=m·{\sigma }_3+b={\displaystyle \frac{1+\sin \phi }{1-\sin \phi }}·{\sigma }_3+{\displaystyle \frac{2c·\cos \phi }{1-\sin \phi }}$$

(1)

where m is the slope of the σ₁–σ₃ relationship; b is the intercept on the σ₁ axis, in MPa; φ is the internal friction angle, in degrees; and c is the cohesion, in MPa.

The fitted lines, regression equations, and R² values obtained for each test condition are shown in Figure 6. The shear strength parameters derived from the Mohr–Coulomb criterion are listed in

Table 5. Shear strength parameters of coal-measure rock specimens under different water-treatment conditions.

Water-Treatment Condition	Shear-Strength Quantity, Symbol (Unit)	Basalt	Tuff	Coal	Oil Shale	Mudstone	Shale
Natural state (0 wetting–drying cycles/0-day immersion)	Internal friction angle, φ (°)	53.552	42.331	23.552	40.288	26.179	25.963
	Cohesion, c (MPa)	2.565	0.333	1.094	0.314	0.368	0.198
1 wetting–drying cycle	Internal friction angle, φ (°)	52.931	41.380	22.733	39.601	25.590	24.032
	Cohesion, c (MPa)	2.378	0.291	0.973	0.301	0.315	0.182
3 wetting–drying cycles	Internal friction angle, φ (°)	52.213	40.561	21.652	39.094	24.370	23.463
	Cohesion, c (MPa)	2.236	0.261	0.909	0.296	0.295	0.178
5 wetting–drying cycles	Internal friction angle, φ (°)	51.762	40.260	20.751	38.900	23.651	22.581
	Cohesion, c (MPa)	1.954	0.217	0.816	0.287	0.244	0.156
7 wetting–drying cycles	Internal friction angle, φ (°)	51.092	39.861	17.380	38.446	23.217	22.041
	Cohesion, c (MPa)	1.828	0.170	0.701	0.284	0.214	0.148
1-day immersion	Internal friction angle, φ (°)	52.892	41.212	22.530	39.097	24.882	23.778
	Cohesion, c (MPa)	2.047	0.254	0.809	0.302	0.275	0.159
4-day immersion	Internal friction angle, φ (°)	52.451	40.822	21.553	38.481	23.152	20.168
	Cohesion, c (MPa)	1.908	0.218	0.776	0.292	0.245	0.154
7-day immersion	Internal friction angle, φ (°)	52.230	40.790	21.451	38.252	22.563	20.027
	Cohesion, c (MPa)	1.831	0.188	0.701	0.286	0.234	0.146
10-day immersion	Internal friction angle, φ (°)	52.021	40.562	21.000	38.153	22.107	19.367
	Cohesion, c (MPa)	1.813	0.186	0.652	0.285	0.228	0.144

.

Under the present test conditions, both wetting–drying cycles and continuous immersion reduced the measured shear-strength parameters. Figure 7 summarizes the reduction coefficients of cohesion, c, and internal friction angle, φ, during the two treatment series. Here, the reduction coefficient, R, is the ratio of the current parameter value to its initial value: Rc = c_i/c₀ for cohesion and Rφ = φ_i/φ₀ for internal friction angle. With increasing immersion duration or number of wetting–drying cycles, both Rc and Rφ generally decreased across the tested specimens.

Based on the variations in Rc and Rφ under continuous immersion and wetting–drying cycles (Figure 7), three main features can be identified.

First, the continuous-immersion series showed a more pronounced early-stage decrease in Rc and Rφ than the wetting–drying series, consistent with the descriptive UCS and peak axial-strength patterns. The rapid decrease in the immersion-series shear-strength retention coefficients extended over approximately 4 days, compared with the 1-day early stage observed for the compressive-strength indices.

Second, Rc decreased more markedly and rapidly than Rφ under both water-induced processes, indicating that cohesion was more sensitive to water-induced degradation. The reduction pattern of c was generally consistent among lithologies, following the order of tuff > mudstone ≈ coal > basalt > shale > oil shale.

Third, the reduction pattern of φ depended on the water-induced process. Under wetting–drying cycles, the early-stage reduction sequence (cycles ≤ 5) was shale > coal > mudstone > tuff > oil shale > basalt; after five cycles, the reduction in φ for coal exceeded that for shale, suggesting structural damage induced by cyclic wetting–drying. Under continuous immersion, the reduction sequence of φ was shale > mudstone > coal > oil shale > tuff > basalt, indicating different φ responses to the two water-induced processes.

3.4. Degradation Model of Shear Strength Parameters

For each lithology and each separate treatment series, the standard Mohr–Coulomb principal-stress relationship was evaluated using condition-dependent c and φ. Equation (2) gives the two distinct forms for wetting–drying cycles and continuous immersion. Reference [46] supports the use of treatment-dependent strength deterioration functions.

σ₁,_i(t, 0) = σ₃[1 + sin φ_i(t, 0)]/[1 − sin φ_i(t, 0)] + 2c_i(t, 0)cos φ_i(t, 0)/[1 − sin φ_i(t, 0)]
σ₁,_i(0, d) = σ₃[1 + sin φ_i(0, d)]/[1 − sin φ_i(0, d)] + 2c_i(0, d)cos φ_i(0, d)/[1 − sin φ_i(0, d)]

(2)

where i denotes lithology; σ_1,i is the peak axial strength; σ₃ is the confining pressure; t is the number of wetting–drying cycles; d is the immersion duration; and c_i and φ_i are the condition-dependent cohesion and internal friction angle. The first line applies only to the wetting–drying series, and the second applies only to the continuous-immersion series.

The condition-dependent c and φ functions in Equation (2) were fitted separately against t or d by least squares. An exponential decay form was used to describe the five treatment-level estimates in each separate series. Its general form is given in Equation (3).

$$y=A·{\mathrm{e}}^{(-x/B)}+C$$

(3)

where y is the dependent variable of the fitting function, representing c or φ; x is the independent variable, representing t or d; and A, B, and C are fitting coefficients.

Figure 8 summarizes the conceptual link between water-related microstructural change and the measured condition-dependent Mohr–Coulomb parameters. The present tests quantify macroscopic strength changes; they do not directly identify a coupled damage mechanism between the two treatment paths.

Based on Equation (3), the treatment-level c and φ estimates were fitted separately for each lithology and treatment series (Figure 9). Both parameters decreased with increasing t or d within the tested ranges. All curves had in-sample R² values greater than 0.85. These R² values describe agreement with the five fitted treatment levels.

The fitted functions describe the observed c and φ trends within the experimental ranges. They provide empirical inputs for the scenario calculations below.

4. Discussion and Application

4.1. Separable Scenario Parameterization of Strength Reduction

Previous studies by Liu et al. and Fu et al. showed that rock masses soften and lose strength when exposed to water, with slight strength recovery during subsequent drying [36,47,48]. Rainfall-induced dry–wet cycling has also been reported to affect slope deformation and landslide evolution in open-pit mines [49]. In the Fushun area, rainy and dry seasons are distinct. Because the rock masses generally remain dry from the end of one rainy season to the onset of the next, one year can be treated as a complete cycle: the rainy season is the strength-degradation stage, whereas the dry season is a relatively stable stage.

The wetting–drying and continuous-immersion series were performed independently; no factorial tests combined both treatments on the same specimens. To explore rainfall-related parameter scenarios without claiming an unmeasured interaction law, the two separate retention ratios were applied sequentially in a separable empirical parameterization. This is a no-interaction baseline for scenario analysis.

For a strength parameter p (c or φ), the single-process retention ratios are Rp(t, 0) = p(t, 0)/p(0, 0) and Rp(0, d) = p(0, d)/p(0, 0). Sequential application to the same baseline gives p(t, d) = p(0, 0) Rp(t, 0) Rp(0, d), and therefore the separable scenario factor in Equation (4).

Rφ,_i(t, d) = Rφ,_i(t, 0) Rφ,_i(0, d)
  Rc,_i(t, d) = Rc,_i(t, 0) Rc,_i(0, d)

(4)

where Rφ,_i and Rc,_i are the retained fractions of φ and c for lithology i. The equation satisfies the measured single-process limits when either t = 0 or d = 0.

Figure 10 shows the scenario factors obtained by multiplying the two independently fitted retention ratios. The surfaces illustrate the numerical consequence of the separability assumption; they are not measurements of specimens exposed to both treatments.

The separable parameterization supplies internally consistent strength inputs for comparative slope scenarios in Fushun. Synergistic or antagonistic effects associated with evolving pores, cracks, mineral hydration, or seepage cannot be resolved from the present dataset. The calculated surfaces should therefore be read as a transparent no-interaction baseline that requires factorial testing before mechanistic or transferable use.

4.2. Rainy-Season Slope Stability and Warning

The Fushun West Open-Pit Mine was used to examine how the laboratory-derived relative reductions affect comparative slope-stability scenarios. Limit-equilibrium calculations were used to evaluate changes in Fs, and finite-element calculations were used to examine the corresponding deformation patterns. These calculations do not simulate a complete rainfall-infiltration process and are not presented as a hydromechanical forecast.

Representative slope selection considered strata attitude, structural framework, and post-mining terrain modification. Post-mining changes included high-fill treatment west of E350 and low-fill treatment between E350 and E1000. On this basis, the existing slopes were classified into three structural types: the overturned synclinal anti-dip slope on the north wall (E350–E1300), the monoclinal anti-dip slope on the north wall (E1300–E3000), and the monoclinal dip slope on the south wall (E350–E3000).

Computational model construction: The slope geometries of the three profiles were extracted from UAV photogrammetry data, and the stratigraphic structures were determined from borehole data. Considering the effects of excavation unloading and weathering on coal-measure strata, the slope rock masses were divided into unloading and non-unloading zones. Based on the comprehensive identification of slope unloading characteristics by Zhou et al. [50,51] and Gao et al. [52] using borehole televiewers, acoustic logging, ground-penetrating radar, and core RQD indices, the unloading-zone depth was idealized as approximately 100 m for the soft-rock slopes on the north wall and 90 m for the relatively hard-rock slopes on the south wall. Accordingly, computational models of the representative pit-slope profiles were established and discretized with a uniform 5 m element size. The E1300(N), E1800(N), and E2000(S) models contained 22,987, 22,743, and 23,252 elements, respectively. The bottom boundary was fixed in both the X and Y directions, the lateral boundaries were fixed in the X direction, and the slope surface was free, as shown in Figure 11b–d.

Calculation cases: Wetting–drying cycles and immersion duration were used as scenario coordinates for antecedent repeated rainfall disturbance and subsequent continuous rainfall or infiltration. Values inside the experimental domain were obtained from the fitted exponential functions. Any value outside that domain is identified as extrapolated. The coordinates do not represent a unique conversion from laboratory treatment time to field rainfall duration.

Parameter determination: Historical landslides in the mining area indicate that slope instability mainly occurred within the unloading zone, where the rock-mass quality is relatively poor. For example, the 26 July 2016 large north-wall landslide was approximately 25–65 m thick [18], and the large south-wall landslide that continued to deform after August 2010 was approximately 60–120 m thick [53]. These cases indicate that the main slip zones in the mining area are generally located above the unloading boundary. Therefore, in the stability calculations, rainfall-related parameter updating was mainly applied to geomaterials within the unloading zone. The non-unloading zone was assigned natural-state parameters only, without water-induced reduction coefficients, and was idealized as relatively stable bedrock.

For coal-measure lithological units within the unloading zone, the natural and saturated densities were taken from Table 1. The initial shear-strength parameters were adopted from the zero-cycle and zero-day condition in Table 5 and were then updated for each calculation case using the separable scenario factors shown in Figure 10. For non-coal geomaterials within the unloading zone, the physical and mechanical parameters were simplified into natural and saturated states: natural-state parameters were used before the rainfall-related scenario and saturated-state parameters afterward. These values were obtained from laboratory tests (

Table 6. Physical and mechanical parameters of non-coal-measure geomaterials.

Quantity, Symbol (Unit)	Gravelly Soil		Fault Zone		Glutenite
	Natural	Saturated	Natural	Saturated	Natural	Saturated
Density, ρ (g·cm−3)	1.80	1.86	2.38	2.44	2.38	2.45
Internal friction angle, φ (°)	19.50	17.80	21.60	17.28	38.20	36.60
Cohesion, c (MPa)	0.12	0.07	0.22	0.17	2.50	1.90

). This binary assignment is a scenario simplification rather than a transient infiltration calculation.

The numerical inputs were laboratory-derived baseline strengths rather than in situ rock-mass parameters. For each scenario, one c and φ pair was assigned uniformly to all elements of the same coal-measure lithology inside the predefined unloading zone. The update was static and simultaneous within that zone; propagation of degradation from exposed surfaces was not simulated. No transient seepage, pore-pressure evolution, or hydromechanical coupling was included. Non-unloading materials retained natural-state parameters as a simplifying boundary assumption, not because deeper rock is physically immune to infiltration. These choices make the results suitable for relative scenario comparison but may produce different Fs-displacement relations from a spatially progressive or coupled analysis.

The numerical analyses were performed in GeoStudio 2018. Slope stability was evaluated with the Morgenstern–Price limit-equilibrium method. Deformation was calculated separately by the finite-element method using quadrilateral elements and a Mohr–Coulomb constitutive model.

Figure 12 presents the stability analysis results obtained using the Morgenstern–Price method. With increasing wetting–drying cycles and prolonged immersion duration, the safety factors, Fs, of the three representative slopes decrease from the lower-left to the upper-right of the contour plots.

The three profile models showed different Fs responses across the scenario grid. E1300(N) had the largest decrease in Fs and entered the lower reference states at less severe model coordinates than the other profiles. E1800(N) showed a similar pattern but retained slightly higher Fs values, whereas E2000(S) remained higher across the evaluated grid.

Within the adopted parameterization, increasing the immersion coordinate produced a slightly larger reduction in Fs than increasing the wetting–drying coordinate for the north-wall profiles, whereas the two coordinates produced more balanced changes for the south-wall profile. The early steep and later gradual decline along the immersion coordinate mirrors the fitted immersion-retention functions used as model inputs.

Across the evaluated scenario space, Fs decreased as the imposed parameter reductions increased. Following the adopted stability-state classification [54], Fs = 1.15, 1.05, and 1.00 were used as reference levels for advisory, watch, and warning scenarios.

To examine deformation at the three reference Fs levels, the closest parameter-reduction scenarios were selected for finite-element calculations. The stiffness values are listed in

Table 7. Stiffness parameters of unloading-zone geomaterials.

Quantity, Symbol (Unit)	Coal-Measure Strata						Non-Coal-Measure Strata
	Basalt	Tuff	Coal	Oil Shale	Mudstone	Shale	Gravelly Soil	Fault Zone	Glutenite
Elastic modulus, E (GPa)	17.3	11.8	0.4	3.4	1.2	1.8	0.1	0.6	5.5
Poisson’s ratio, ν (-)	0.22	0.24	0.26	0.26	0.28	0.25	0.2	0.3	0.25

[55], and the strength parameters matched those used in the corresponding stability scenario.

Figure 13 presents the finite-element results for the three representative slope profiles at the near-critical state, where Fs ≈ 1.00. The high-displacement zones broadly coincide with the potential sliding bodies used in the stability interpretation, and shear strain concentrates along the corresponding potential sliding paths.

The E1300(N) and E1800(N) profiles on the north wall showed similar modeled deformation–strain patterns: displacement concentrated in the middle-to-lower parts of the potential sliding bodies, tensile zones occurred near the rear margins, and shear-strain bands extended along fault-fracture zones toward the slope toe. The maximum-displacement zones occurred near the toe, with peak values of 213.82 mm and 154.82 mm, respectively. E1300(N) showed larger modeled displacement and more continuous shear-strain concentration than E1800(N). The E2000(S) model showed a lower maximum displacement of 72.50 mm, with deformation and shear strain concentrated near soft coal interbeds and fault-fracture zones.

The selected near-critical cases for E1300(N), E1800(N), and E2000(S), where Fs ≈ 1.00, corresponded to (t, d) = (3, 4 days), (5, 4 days), and (10, 3 days), respectively. These values are coordinates in the scenario model, not field rainfall conditions required for failure. Because 10 wetting–drying cycles exceed the experimental range, the E2000(S) parameters were extrapolated from the functions in Section 3.4 and Section 4.1. E1300(N) reached the selected Fs level at a lower t coordinate than E1800(N).

For each profile and reference level, the calculation case with Fs closest to 1.15, 1.05, or 1.00 was identified. The maximum displacement from the paired finite-element calculation was then recorded as a model-specific reference value (

Table 8. Model-specific cumulative-displacement reference values and paired rainfall-related scenarios.

Representative Slope Profile	Slope-Section Type	Advisory, u (mm); Fs ≈ 1.15	Watch, u (mm); Fs ≈ 1.05	Warning, u (mm); Fs ≈ 1.00
E1300(N)	Overturned synclinal anti-dip slope	86.76 [(1, 1)]	170.92 [(1, 4)]	213.82 [(3, 4)]
E1800(N)	Monoclinal anti-dip slope	62.20 [(3, 1)]	124.81 [(1, 10)]	154.82 [(5, 4)]
E2000(S)	Monoclinal dip slope	52.36 [(7, 1)]	62.20 [(7, 10)]	72.50 [(10, 3)]

Note: (1) Values are model-specific cumulative-displacement reference values in millimeters; they are not independently calibrated field thresholds. (2) Brackets indicate representative rainfall-related cases, expressed as (t, d), where t is the number of wetting–drying cycles and d is the immersion duration in days. (3) The cases were selected according to their proximity to the target Fs levels.

).

The resulting table is a scenario-based reference map linking parameter-reduction cases, Fs levels, and paired displacement outputs. Operational thresholds should not be transferred directly from this map; they require calibration against site monitoring and should be updated when the numerical model or input parameters change.

Historical displacement curves from three landslide-prone sections were compared with the model-specific displacement reference levels [56,57]. In all three panels of Figure 14, the horizontal lines reproduce the values in Table 8, which were obtained from paired stability and finite-element calculations. The deformation-stage annotations follow the observed sequence of each monitoring curve. The two therefore show an approximate correspondence in stage progression.

For field use, cumulative displacement should be interpreted together with displacement velocity and acceleration, rainfall, and changes in subsurface conditions. Velocity and acceleration can identify transitions toward accelerated movement and are widely used in slope-warning practice [49,58,59]. The present figure-based records are insufficient for reliable numerical differentiation; quantitative derivative thresholds require the original time-displacement series and smoothing choices. A conservative operational rule would adopt the highest warning level triggered by any validated indicator rather than relying on cumulative displacement alone.

Scope and limitations: The laboratory protocols, separable parameterization, and numerical reference values were developed for the Fushun coal-measure unloading zone. Application to other mines requires local lithological testing, rock-mass scaling, structural and hydraulic characterization, and monitoring-based calibration. Snowmelt-dominated regions also require explicit treatment of freeze–thaw damage, snowmelt infiltration, and rain-on-snow sequences, none of which were examined here.

5. Conclusions

This study investigated water-induced shear-strength degradation of coal-measure rocks from the unloading zone of the Fushun West Open-Pit Mine and examined its slope-engineering implications. Triaxial testing, empirical strength-reduction functions, and numerical scenario analyses were used to compare degradation under wetting–drying cycles and continuous immersion and to examine a staged warning-reference framework for rainy-season slopes. The main conclusions are as follows.

(1)

Within the tested ranges, the measured shear-strength parameters of Fushun coal-measure rock specimens were described by exponential functions with in-sample R² values greater than 0.85. Cohesion, c, was generally more sensitive to water-induced degradation than internal friction angle, φ. Lithology-dependent degradation was evident: tuff, mudstone, and coal showed stronger c degradation, shale showed greater φ sensitivity, and basalt and oil shale showed stronger water resistance. Process-dependent patterns were also observed: immersion caused rapid softening mainly within 0–4 days, whereas wetting–drying cycles produced cumulative degradation, with accelerated φ degradation in coal after five cycles.

(2)

Condition-dependent c and φ functions were incorporated into the standard Mohr–Coulomb principal-stress relation. For exploratory slope scenarios, the independently measured wetting–drying and immersion retention ratios were applied sequentially in a separable no-interaction parameterization. This empirical baseline is not a coupled damage law and requires combined-treatment experiments before interaction effects can be evaluated.

(3)

Model-specific pairs of Fs levels and cumulative displacement were obtained for three representative slope profiles. The 52.36–213.82 mm values are outputs of the stated numerical models and were compared qualitatively with historical deformation stages. They should be used as local scenario references; field warning should combine displacement, velocity or acceleration, rainfall, and site-specific calibration.

(4)

The main limitations include the lack of within-condition replication, the accelerated 105 °C conditioning and unverified saturation degree, specimen-to-rock-mass scale effects, and the untested separability assumption with static uncoupled numerical modeling. Future work should include replicated experiments, more field-representative conditioning, site-specific calibration, and coupled seepage–deformation analysis.