The ability to predict the volume of methane outbursts in coal mines is critical for the prevention of methane outburst accidents and the assurance of coal-mine safety. This paper’s central argument is that existing prediction models are limited in several ways. These limitations include the complexity of the models and their poor ability to generalize. The paper proposes a methane outburst volume-prediction and early-warning method. This method is based on a secondary decomposition and improved TSMixer model. First, data smoothing is achieved through an STL decomposition–adaptive Savitzky–Golay filtering–reconstruction framework to reduce temporal complexity. Second, a CEEMDAN-Kmeans-VMD secondary decomposition strategy is adopted to integrate intrinsic mode functions (IMFs) using K-means clustering. Variational mode decomposition (VMD) parameters are optimized via a novel exponential triangular optimization (ETO) algorithm to extract multi-scale features. Additionally, a refined TSMixer model is proposed, integrating reversible instance normalization (RevIn) to bolster the model’s generalizability and employing ETO to fine-tune model hyperparameters. This approach enables multi-component joint modeling, thereby averting error accumulation. The experimental results demonstrate that the enhanced model attains RMSE, MAE, and R² values of 0.0151, 0.0117, and 0.9878 on the test set, respectively, thereby exhibiting a substantial improvement in performance when compared to the reference models. Furthermore, we propose an anomaly detection framework based on STL decomposition and dual lonely forests. This framework improves sensitivity to sudden feature changes and detection robustness through a weighted fusion strategy of global trends and residual anomalies. This method provides efficient and reliable dynamic early-warning technology support for coal-mine gas disaster prevention and control, demonstrating significant engineering application value.

It is noted that existing mine emergency-rescue algorithms have overlooked the requirement for multi-route sharing at critical nodes and have offered limited network visualisation. Consequently, a multi-team rescue-path-planning algorithm based on FA-MDPSO (Firefly Algorithm-Multiple Constraints Discrete Particle Swarm Optimisation) was proposed, and a graph-structure optimisation method combining a Force-Directed Layout with Breadth-First Search was introduced for node arrangement and visualisation. Methodologically, the superiority of the improved DPSO (Discrete Particle Swarm Optimisation) in route-planning precision was first validated on the DIMACS dataset. Subsequently, the hyperparameters of MDPSO (Multiple Constraints Discrete Particle Swarm Optimisation) were optimised by means of four intelligent algorithms—ACO (Ant Colony Optimization), FA (Firefly Algorithm), GWO (Grey Wolf Optimizer) and WOA (Whale Optimization Algorithm). Finally, simulations of one to three rescue-team deployments were conducted within a mine-fire scenario, and node-importance analysis was performed. Results indicated that FA-MDPSO achieved comprehensive superiority in route precision, search efficiency and convergence speed, with FA-based hyperparameter optimisation proving most effective in comparative experiments. The graph-structure optimisation was found to substantially reduce crossings and enhance hierarchical clarity. Moreover, the three-team deployment yielded the shortest equivalent path (56,357.02), and node-visitation frequency was observed to be highly concentrated on a small number of key nodes. This not only significantly improves the collaborative rescue efficiency but also provides intuitive and practical technical support for intelligent mine rescue operations. It lays an important foundation for optimising mine emergency rescue plans, ensuring the safety of underground personnel, and promoting the intelligent development of mines.

A methane-air premixed gas explosion is one of the most destructive disasters in the process of coal mining, and the dynamic coupling between the shock wave triggered by the explosion and the surrounding rock of the roadway can lead to the destabilization of the surrounding rock structure, the destruction of equipment, and casualties. The aim of this study is to systematically reveal the propagation characteristics of the blast wave, the spatial and temporal evolution of the wall load, and the damage mechanism of the surrounding rock by establishing a two-way fluid-solid coupling numerical model. Based on the Ansys Fluent fluid solver and Transient Structure module, a framework for the co-simulation of the fluid and solid domains has been constructed by adopting the standard $k-\epsilon$ turbulence model, finite-rate/eddy-dissipation (FR/ED) reaction model, and nonlinear finite-element theory, and by introducing a dynamic damage threshold criterion based on the Drucker–Prager and Mohr–Coulomb criteria. It is shown that methane concentration significantly affects the kinetic behavior of explosive shock wave propagation. Under chemical equivalence ratio conditions (9.5% methane), an ideal Chapman–Jouguet blast wave structure was formed, exhibiting the highest energy release efficiency. In contrast, lean ignition (7%) and rich ignition (12%) conditions resulted in lower efficiencies due to incomplete combustion or complex combustion patterns. In addition, the pressure time-history evolution of the tunnel enclosure wall after ignition triggering exhibits significant nonlinear dynamics, which can be divided into three phases: the initiation and turbulence development phase, the quasi-steady propagation phase, and the expansion and dissipation phase. Further analysis reveals that the closed end produces significant stress aggregation due to the interference of multiple reflected waves, while the open end increases the stress fluctuation due to turbulence effects. The spatial and temporal evolution of the strain field also follows a three-stage dynamic pattern: an initial strain-induced stage, a strain accumulation propagation stage, and a residual strain stabilization stage and the displacement is characterized by an initial phase of concentration followed by gradual expansion. This study not only deepens the understanding of methane-air premixed gas explosion and its interaction with the roadway’s surrounding rock, but also provides an important scientific basis and technical support for coal mine safety production.

Traditional pedestrian evacuation models struggle to balance global exit guidance with local, individual decision making under hazards. We address this by decomposing long-term objectives into Particle Swarm Optimization (PSO)-based micro-goals and proposing a hybrid Cellular Automaton (CA) and PSO model. The hybrid design reduces the decoupling between spatial discretization and individual choices and more tightly couples hazard and density fields with movement decisions. Two contributions are central. First, we develop an autonomous following pathfinding mechanism (AFPM) that linearly blends the direction toward a PSO micro-goal with a herd following direction and adds a small reward for directional consistency. This mitigates path chaos from purely autonomous moves and congestion aggregation from purely herding moves. Second, we build a multi-dimensional interpretability and robustness framework that combines the empirical Cumulative Distribution Function (CDF) and a kernel-smoothed Probability Density Function (PDF) of key evacuation times (T_clear, T_95%_alive) together with vulnerability curves, to analyze the data and assess robustness. It combines Shapley Sobol analysis to quantify parameter effects on clearance time T_clear and the 95% survival evacuation time T_95%_alive, with CDF/PDF summaries and vulnerability curves to assess anti-interference performance. Experiments use a simulated underground shopping mall. In a 60 pedestrian case, a geometry-only baseline yields T_clear $\approx 33 \mathrm{s}$; hazard- and density-aware strategies produce slightly longer T_clear but reduce peak bottleneck congestion by 20–30%. When one exit is closed, the exceedance probability at $\tau =70 \mathrm{s}$ drops from 0.44 to 0.36, reducing long tail risk. Compared with geometry-based Dijkstra, the proposed model slightly increases clearance time while lowering peak congestion by 20–30%, achieving a balance between efficiency and safety. The model and evaluation protocol provide technical support for evacuation policy, facility layout, and emergency system design in large complex buildings.