Deep hard-rock mines may develop progressively softer loading environments as extraction advances, particularly near cave footprints, and highly stressed excavation boundaries. Under these conditions, brittle collapse may shift from a material-controlled failure process to a system-controlled instability driven by excess elastic strain energy released from the mine loading system. Although very stiff testing machines are essential for capturing the characteristic post-peak behaviour of brittle rock, they do not reproduce the softer loading conditions that commonly govern dynamic damage in mining environments. Numerical experiments using the discontinuum code PFC2D were conducted to investigate how loading system stiffness (LSS) influences the post-peak response of unsupported synthetic hard brittle rock. The model setup was first verified to confirm the global force–displacement response. Quasi-static and dynamic uniaxial compression simulations were then performed under different LSS conditions to quantify post-peak behaviour and damage magnification during dynamic brittle collapse. The results show that intrinsic rock properties primarily govern the response under stiff LSS conditions, such as isolated tunnel advances. Under soft loading, as may occur during late-stage extraction or near cave footprints, excess elastic strain energy from the LSS becomes increasingly influential and magnifies specimen-scale strainburst intensity (ejection velocity of breakout formations), severity (ejected mass at test termination), and bulking response (stress-induced fractured ground). The characteristic specimen-scale rupture duration for localized fragment ejection ranges from 0.10 to 0.22 ms. These sub-millisecond values reflect the small, mobilized volume and low inertial resistance of the unsupported specimen and should not be directly extrapolated to excavation-scale strainburst durations.
Talc is a common gangue mineral in copper sulfide ores; however, its high hydrophobicity hinders chalcopyrite purification and smelting. In this study, mineral separation was improved by the combined application of MnCl2 and Na3PO4, and a series of surface analyses were performed to reveal the mechanisms responsible for the reagent–mineral surface interactions. Using either the collector or frother alone, separation of the two minerals was inefficient. However, a window of separability >65% was achieved using MnCl2 and Na3PO4 as a combined depressant, indicating high selectivity. Nevertheless, the depression performance was limited depending on slurry pH and reagent molar ratio. Talc depression was attributed to the inherently low affinity of the collector toward its surface, as well as the non-uniform deposition of manganese hydrogen phosphate (MnHPO4(s))—the predominant species in the MnCl2-Na3PO4 system—on its surface via electrostatic attraction. Due to the lack of collector adsorption, partial physical shielding by MnHPO4(s) was sufficient to markedly reduce talc recovery. In contrast, although similar deposition occurred on chalcopyrite, the collector effectively adsorbed onto exposed adsorption sites adjacent to the MnHPO4(s) coating, thereby maintaining good hydrophobicity. The non-uniform deposition of MnHPO4(s) and competitive adsorption of the collector therefore play important roles in the efficient separation of chalcopyrite from talc.
To clarify the dynamic response of a local anchorage bearing unit subjected to coupled impact loading and radial confinement, the axial front-end unit of a roadway bolt support system was idealized as a bolt–resin–rock composite cylinder. Coupled static–dynamic impact tests were conducted at confining pressures of 0, 2, 4, 6, and 8 MPa using a modified Split Hopkinson Pressure Bar (SHPB) system. The incident compressive pulse was applied parallel to the bolt axis to represent compression-dominated disturbances transmitted along the reinforcement direction. At an impact gas pressure of 0.6 MPa, increasing the confining pressure from 0 to 8 MPa increased the dynamic peak stress from 61.27 to 100.79 MPa, decreased the peak strain from 0.01073 to 0.00567, increased the terminal transmitted energy from 40 to 141 J, and decreased the terminal apparent absorbed energy from 71 to 49 J. The dynamic equivalent wave impedance ratio increased from 0.085 to 0.168, indicating progressively improved impedance matching and wave transmission capacity. A confinement-dependent segmented equivalent damage constitutive model was established by combining the strain equivalence principle, Weibull statistical damage theory, and an unloading–rebound branch. The model reproduced the loading stage, peak point, and dominant post-peak response path, with an average R2 of 0.969 and an average RMSE of 3.102. The proposed model is intended to characterize equivalent responses within the calibrated laboratory range. These results provide a basis for interpreting the confinement-dependent impact response of axially anchored bearing units in deep roadway support systems.
Energy accumulation and release in coal vary considerably between long-term high stress conditions and rapid loading conditions. As conventional macroscopic experiment methods fail to capture local energy evolution, we propose a full-field energy visualisation method via digital image correlation. We applied this method to compare coal fracture and energy dissipation under uniaxial compression and creep. Results indicate that strain energy during uniaxial compression concentrates locally, with accumulation zones strictly aligning with main macroscopic cracks, demonstrating that crack propagation is accompanied by intense energy accumulation and dissipation. Notably, individual fracture acoustic emission (AE) events yield high energy, whereas the total accumulated energy throughout the fracture process remains low. By contrast, strain energy distribution within the sample is largely uniform during creep, with energy accumulation and dissipation occurring exclusively locally before ultimate failure. Although individual AE events display low peak energy, the total accumulated AE energy is substantial. These laboratory findings were validated using a cross-scale model applied to field conditions: microseismic events in isolated pillars (creep) exhibited high frequency and low peak energy, whereas those in conventional mining faces (uniaxial loading) displayed the opposite trend. This study provides a theoretical basis for differentiated early warning of dynamic disasters in deep coal mines.
In permafrost regions including the Tibetan Plateau and polar areas, ice-filled joints are common in rock masses. The mechanical behavior of such ice-rock composite geological masses governs the dynamic stability of engineering structures in cold regions. To investigate their dynamic mechanical response and failure mechanisms, modified split Hopkinson pressure bar (SHPB) tests were conducted on rock-ice-rock (RIR) composite specimens over a temperature range of −10 °C to −40 °C. Experimental results indicate that the dynamic compressive strength of RIR specimens increases monotonically with decreasing temperature, whereas that of intact specimens exhibits a non-monotonic trend. The presence of ice layer reduces energy dissipation density but increases the energy reflection coefficient. Moreover, the stiffness of the ice layer increases and closure decreases at lower temperatures, thereby improving load transfer. Under dynamic loading, tensile failure generally dominates. As temperature decreases, the failure mode of RIR specimens transitions from multiple forms to being dominated by axial splitting. Scanning electron microscopy (SEM) observations further reveal a transition in the fracture mode from predominantly intergranular to transgranular with decreasing temperature. Finally, a theoretical model for stress wave propagation was developed, incorporating the effect of unfrozen water films, which provides theoretical interpretations for the temperature-dependent failure behavior.
In cold-region rock engineering, freeze-thaw (F-T) cycles and varying saturation levels significantly influence rock mechanical behavior and failure precursors. However, the acoustic emission (AE) characteristics and critical slowing down (CSD) precursors of red sandstone under different F-T cycles and saturation conditions remain largely unexplored. In this study, uniaxial compression tests combined with AE monitoring and CSD theory were conducted to investigate the damage evolution and failure precursors of red sandstone under varying saturation levels (0%–100%) and F-T cycles (0–30 cycles). The results show that UCS decreases by approximately 22.88% as saturation increases from 0% to 100%, and by about 40.20% after 30F-T cycles. Under high saturation and repeated F-T cycles, AE activity intensifies and crack propagation becomes more pronounced. The b-value decreases markedly prior to peak stress, while the cumulative b-value increases rapidly, indicating its potential as an instability precursor. Variance analysis based on CSD theory reveals that early-warning signals appear earlier and the warning window is extended with increasing saturation and F-T cycles. Furthermore, a conceptual dynamical framework linking initial damage (D), system recovery rate (λ), and CSD is proposed, illustrating that microcrack propagation and structural degradation under coupled F-T and saturation conditions drive critical slowing down behavior. The findings provide theoretical support and a conceptual methodology for stability assessment and precursor-based early warning in cold-region rock engineering.
Based on the Split Hopkinson Pressure Bar (SHPB) and acoustic emission (AE) techniques, stress–strain curves, energy evolution and AE amplitude during coal failure were monitored to reveal the dynamic failure mechanism of coal under different impact pressures, explore the effect of impact pressure on AE characteristics during the brittle-ductile transition, and achieve accurate identification of major damage points. Crack propagation under loading and the particle size distribution of fragmented coal after failure were analyzed. The macroscopic characteristics of brittle-ductile failure were clarified, the temporal evolution of ringing counts and energy with stress–strain response was investigated, and the staged crack evolution characteristics were revealed based on the proportion of RA/AF classified cracks. Fast Fourier Transform (FFT) was adopted for the frequency-domain analysis of AE waveforms, and the differences in spectral characteristics between brittle-dominated and ductile-dominated coal failure were clarified. Based on the Stockwell Transform, time–frequency feature fusion analysis was performed to identify key damage points, and their identification accuracy was verified through b-value time-series analysis. This study clarifies the AE time–frequency response mechanism of coal brittle-ductile failure, identifies key damage points via the Stockwell Transform, and provides support for the precise early warning and mitigation of coal mine rockbursts.
Deep coal mining faces severe thermal hazards associated with high geothermal gradients. Coal-heat co-mining (CHCM) provides an effective way to recover geothermal energy while regulating the thermal environment and stress state of coal seams. However, the thermo-mechanical responses of overlying coal seams during long-term geothermal extraction remain poorly understood. In this study, a fully thermo-hydro-mechanical (THM) coupled numerical model is developed for natural fracture-based CHCM system with double horizontal wells to evaluate the influence of horizontal well opening configurations on thermal performance and coal-seam thermo-mechanical responses. Four well-opening schemes are systematically compared in terms of thermal performance, cooling and stress redistribution in the coal seam. The results indicate that shorter opening section will enhance channeling flow in fractures, leading to increased injection pressure, early cold front breakthrough, and localized thermal stress perturbations. Increasing the opened section length suppresses channeling flow, reduces injection pressure, and yields more uniform cooling. Long-term heat extraction induces cumulative coal-seam cooling and stress redistribution, with maximum temperature reductions of up to 11 °C and vertical stress decreases exceeding 1.5 MPa. These findings highlight the role of horizontal well opening configuration in regulating the thermal environment and stress state of coal seams.
To address the lack of a unified energy mapping framework for energy release and damage evolution in rocks, uniaxial compression tests were conducted on sandstone with typical pore configurations. Using acoustic emission (AE) and digital image correlation (DIC) techniques to characterize the energy release characteristics and strain evolution laws during rock fracture process. Based on the principle of energy conservation, the energy conversion efficiency α is defined to correlate AE energy with stored strain energy, and an energy type damage factor D is introduced to quantify the damage accumulation process. The results indicate that when D < 0.4, the damage evolution is relatively stable; When D ≥ 0.4, the damage rapidly intensifies, indicating the failure of the specimen. The study also found a significant time lag between AE energy release and macroscopic deformation, which is consistent with the law in engineering practice that microseismic activity precedes monitoring displacement response. Furthermore, a Lemaitre damage model incorporating α was established, which can effectively characterize the nonlinear evolution process of sandstone from elastic deformation to complete failure. Finally, an energy based rockburst warning index based on α and D was discussed, providing technical path for identifying high-risk areas and preventing rockburst disasters in mining engineering.
The dynamic evolution of in-situ stress during coalbed methane drainage governs coal seam stability, and instability-induced damage critically impacts well productivity. Triaxial deformation and methane adsorption-induced expansion experiments were conducted to investigate the stability evolution mechanisms and controlling factors in mid-deep intact and fractured seams. Results show that permeability surges at the peak stress, then declines sharply and eventually stabilizes. Adsorption pressure correlates positively with adsorption-induced strain, with the maximum volumetric strain of 1.756%. Fractured zones are more prone to instability failure than intact coal seams. Instability failure occurs under normal faulting stress regimes with moderate-to-strong desorption capacity and strike-slip stress regimes with strong desorption capacity. Seams under reverse faulting regimes remain stable. Stress difference, critical desorption pressure, elastic modulus, internal friction angle, cohesion, and fault friction coefficient are inversely related to coal seam stability. In late-stage production, failure of faulted zones in the #2 and #4 coal seams occurred at pore pressures of 1.17 and 1.70 MPa, respectively. These failures induced five abrupt bottom-hole pressure drops, each lasting 26–54 min, severely compromising well productivity. Pore pressure inversion yields fracture-wellbore distances of 15–111 m.
Accurately determining the effective fracture toughness (Keff) of rock-concrete (R-C) bi-materials, governed by interface inclination and ambient temperature, is a prerequisite for assessing their structural stability. This study developed a hybrid NRBO-XGBoost prediction model using the Newton-Raphson-Based Optimizer (NRBO) to tune the hyperparameters of Extreme Gradient Boosting (XGBoost) model. The established model was developed based on 154 datasets obtained from laboratory tests and numerical simulations with the cracked straight-through Brazilian disc (CSTBD) specimens, including twelve input parameters. The NRBO-XGBoost model for Keff prediction was investigated and compared with seven more models. Furthermore, the Shapley Additive exPlanations (SHAP) method was employed to quantify the contributions of inputs to Keff to improve the interpretability of the developed model. Finally, new data were used to validate the model. Evaluation results demonstrate that metaheuristic optimization algorithms significantly enhance the performance of XGBoost, with NRBO-XGBoost performing the best. The models rank from highest to lowest prediction performance as follows: NRBO-XGBoost, WOA-XGBoost, PSO-XGBoost, XGBoost, RF, CatBoost, LightGBM, and AdaBoost. The interpretable analysis shows that the interface inclination angle exerts the dominant influence. The validation results demonstrate that NRBO-XGBoost achieves high predictive accuracy on a new dataset, showing promising implications for practical applications.
Understanding aggregate effects within fault fracture zones is crucial for assessing fault reactivation risks in grouted deep coal mines. This study investigates the compressive-shear damage mechanisms of grouted concrete specimens with varying aggregate contents (20%, 40%, 60%) and particle sizes (1–5 mm, 6–10 mm). Uniaxial compressive-shear tests were monitored synchronously using digital image correlation (DIC) and acoustic emission (AE). Additionally, a calibrated PFC2D discrete element model, incorporating matrix, aggregate, and interfacial transition zones, elucidated the mesoscale mechanisms. Results indicate: (1) peak shear load decreases linearly (R2>0.86) with increasing aggregate content, dropping 64.3% from 20% to 60%, driving a transition from brittle to plastic failure; (2) AE b-value evolution tracks progressive damage, while damage rate k exhibits contrasting size-dependent trends; (3) macroscopic failure modes are synergistically controlled by aggregate size and content; and (4) numerical simulations validate the mesoscale mechanical origins of this brittle-to-plastic transition. These findings reveal the micro-mechanical mechanisms of anisotropic failure and re-strengthening in grouted fault materials, offering vital geological insights into stress evolution and instability precursors during fault reactivation.
Open-pit mining in the high-altitude areas of the western Sichuan Plateau triggers fluctuations in groundwater levels, causing rocks to undergo freeze–thaw cycles at different immersion depths and posing threats to the stability of slopes and the integrity of groundwater systems. In this study, nuclear magnetic resonance (NMR) technology, scanning electron microscopy (SEM), and permeability tests were adopted to systematically investigate the influence of freeze–thaw cycles on the pore structure and permeability of slate under different immersion depths. Additionally, the distinguishing characteristics of freeze–thaw damage under various immersion states was revealed. By integrating the fractal dimension, which characterizes the complexity of pore structure, into the modeling process, a permeability model incorporating porosity, fractal dimension, number of freeze–thaw cycles, and immersion depth was established. The results showed that freeze–thaw cycles significantly influenced the pore structure and permeability of rock samples. Specifically, as the number of freeze–thaw cycles increased, both the porosity and permeability exhibited an upward trend, with these trends becoming more pronounced as immersion depth raised. Notably, under conditions of moderate to full immersion, the rock was more prone to developing internal microcracks. Furthermore, a fractal dimension-adjusted permeability model was proposed. The established model integrated four key influencing factors, namely initial porosity, fractal dimension, number of freeze–thaw cycles, and immersion depth, achieving a remarkable coefficient of determination (R2) of 0.925. The established model allowed for the estimation of permeability changes in rock subjected to varying immersion depths and different numbers of freeze–thaw cycles, serving as a reference for mitigating rock degradation and protecting groundwater resources.
Low-permeability coalbed methane recovery is restricted by dense pore-fracture structures and the limited cavitation efficiency of conventional water media. This study introduces MIL-100(Fe) nanofluid as an active ultrasonic cavitation medium to regulate cavitation nucleation and strengthen coal pore-network reconstruction. Two coal samples were treated by water- and nanofluid-mediated ultrasonic cavitation and characterized using in-situ X-ray nano-CT, SEM, FTIR, XRD, BET/FHH analysis, nanoindentation, and high-pressure methane adsorption. Compared with water treatment, nanofluid-mediated ultrasonic cavitation more effectively weakened aromatic C–H adsorption sites, increased hydroxyl-related absorption, and optimized microcrystalline parameters, with d002/fa reaching 0.0819 nm/0.146 for BG coal and 0.0805 nm/0.142 for SJZ coal. It increased porosity to 29.10% and 30.10%, raised connected-pore volume fractions to 95.23% and 95.64%, and formed mesopore–macropore structures containing 47.5–59.8% mesopores and 40.2–52.5% macropores. The Langmuir adsorption capacity decreased from 16.77 to 10.48 mg g−1 for BG coal and from 24.52 to 17.70 mg g−1 for SJZ coal, while Young’s modulus decreased from 7.96 to 7.01 GPa and from 6.95 to 6.53 GPa, respectively. These results reveal that MIL-100(Fe) nanofluid enhances ultrasonic cavitation through coupled porous multicenter nucleation, thermal regulation, and particle scouring, thereby converting isolated pores into connected methane desorption pathways.
To investigate the evolution of pore-fracture structure (PFS) in coal under stress constraints, real-time nuclear magnetic resonance tests were conducted on coal samples subjected to four stress paths (SP-I to IV). The PFS evolution, fractal characteristics, failure morphology, and permeability contribution were analyzed using T2 spectra, nuclear magnetic resonance images (NMRIs), and fractal theory. The results show that stress paths significantly influence pore compaction, dilation, and fracture development. SP-III exhibits the most significant promotion of pore dilation before coal failure, followed by SP-IV and SP-II, whereas SP-I suppresses pore dilation. Confining pressure unloading promotes pore dilation, whereas the effect of axial loading depends on the confining pressure reduction. Coal failure significantly reduces pore heterogeneity, accompanied by a decrease in the pore fractal dimension (Db). Multifractal parameters Δα and Ac were associated with pore compression, dilation and microfracture nucleation, while Hc served as a robust index of pore connectivity. The failure mode of SP-I was shear-dominant, while the remaining stress paths manifested a combined tensile-shear failure. The failure zone of SP-III exhibited the highest morphological complexity, followed by SP-IV, SP-II, and SP-I. The stress path significantly influenced the permeability contribution of PFS within the 100–1000 ms, highlighting distinct permeability responses under different stress paths. These findings provide valuable insights into the mechanisms governing gas migration in coal seams.
To effectively control the large deformation of the surrounding rock under complex conditions, it is often necessary to apply prestress to anchor cables. However, due to the influence of surrounding rock deformation, mining disturbance, and strong impact, anchor cables are often in a dynamic and static coupling stress state. Therefore, it is crucial to study the dynamic and static coupling mechanical characteristics of anchor cables. Based on this, the self-developed dynamic and static coupling test equipment is developed. The dynamic and static coupling mechanical test of anchor cables is conducted. Test results indicate that the energy absorption for anchor cables under the initial load of 350 kN decreased by 69.8% compared to the condition without initial load, and the energy absorption efficiency increased by 6.6 times. The increase of initial load can improve its energy absorption efficiency, but it can also lead to a decrease in its energy absorption. The energy absorption and energy absorption efficiency shows a bilinear variation law with the increase of initial load. On this basis, the energy absorption calculation formula and the support design model of the anchor cable are established. It provides new ideas for the safety control of dynamic disasters in deep engineering.
Accurate in-situ identification of coal and gangue is critical for intelligent mining, particularly in longwall top coal caving (LTCC) mining, where it enables precise control of the gangue mixed ratio and enhances resource recovery. This study introduces a Secondary Intervention strategy to augment the conventional "liquid intervention + infrared detection" approach. Results demonstrate that Secondary Intervention can consistently enhance the thermal contrast between coal and gangue, and the average accuracy of infrared image recognition for coal and gangue increased from 79.14% after the First Intervention to 93.75% after the Secondary Intervention, representing an improvement of 14.61%. Furthermore, the average contact angle difference between coal and gangue expanded from 16.64 degrees after the First Intervention to 33.80 degrees after the Secondary Intervention, an increase of 17.16 degrees. Meanwhile, the area difference between coal and gangue increased by 4.94 times. Moreover, based on comprehensive analysis of the temperature difference, accuracy of morphological identification, as well as the contact angle and area of droplets, the eco-friendly compound surfactant (EFCS) of Soapnut Saponin (SS) + CTAB with a concentration of 0.06 wt% demonstrated optimal performance. These findings advance liquid intervention techniques for infrared-based recognition and support the development of greener, more intelligent coal production systems. (c) 2026 China University of Mining & Technology. Publishing services by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Rock fragment size distribution (FSD) plays an important role in various engineering applications, such as mining, tunnelling, and other underground construction scenarios. While vision-based deep learning approaches have been increasingly applied to FSD analysis, they are often case-specific, showing limited cross-site generalization despite their accuracy. To address these challenges, FragSAM, an end-to-end, fully automated framework is proposed for near real-time rock fragment segmentation and FSD analysis across diverse engineering environments. FragSAM integrates the generalization power of Segment Anything Model (SAM) with a context-aware prompting mechanism and lightweight architecture for efficient dense fragment segmentation. In Stage 1, an enhanced SAM automatically generates high-quality annotations, which are used to train a modified CenterNet for precise centroid prediction. In Stage 2, these centroids serve as prompts for EdgeSAM, a lightweight SAM variant optimized for real-time inference. This two-stage design eliminates dense grid prompting and reduces reliance on heavy post-processing, enabling efficient and scalable segmentation. Experimental results show that FragSAM achieves competitive segmentation performance with significantly lower latency and model complexity compared to existing SAM-based methods. In comparison with supervised learning approaches, it also demonstrates superior generalization and performs better in low-quality or unseen scenarios. Furthermore, case studies on blasting fragmentation, TBM muck, and coastal rock surfaces confirm its robustness and seamless cross-site adaptability, requiring no tuning or retraining, making it highly practical for on-site applications.
Although hydraulic fracturing of water-bearing tight sandstone gas reservoirs has been extensively investigated, little attention has been paid to the influence of water saturation (ws) on pore structure and fluid infiltration behavior during the fracturing process. To address this gap, hydraulic fracturing experiments were conducted on sandstone specimens with different water saturations (ws = 0, 25%, 50%, 75%, and 100%) using a real-time nuclear magnetic resonance (NMR) system. Results show that increasing ws reduces both breakdown pressure and breakdown time. Fluid injection promotes progressive micropores dilation and their transformation into mesopores and macropores at ws = 0-50%, while higher saturation enhances macropores modification by coalescence of pre-existing smaller pores. Capillary tension at dry-wet interfaces and clay mineral dissolution are suggested as dominant mechanisms governing pore-scale damage at low and high ws, respectively. Fluid infiltration is enhanced at low ws due to improved pore connectivity and strong water absorption effect, but is suppressed at high ws owing to water-locking effects. An apparent transition in mesopore evolution and preferential infiltration direction is observed between ws = 50% and 75%. These findings provide mechanistic insights into the role of ws in hydraulic fracturing of water-bearing tight sandstone gas reservoirs. (c) 2026 China University of Mining & Technology. Publishing services by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
