Machine Learning and Cyber Physical Systems for Rock Mass Classification: A Critical Review

Critical assessment on challenges and advances in coal pillar strength estimation techniques: A focus on geological discontinuities

Accurate estimation of coal pillar strength is essential for ensuring safety and operational efficiency in underground mining. Current assessment methods often face limitations in addressing time-dependent failure mechanisms, geological discontinuities, and dynamic loading conditions. This paper identifies future research directions aimed at enhancing the reliability of pillar strength evaluations. One important focus is the development of advanced numerical models that account for time-dependent behaviours such as creep and fatigue. Incorporating multi-scale modelling techniques, which connect micro-scale material responses to macro-scale structural performance, may lead to more precise predictions. The integration of real-time monitoring systems measuring stress, deformation, and environmental factors into predictive models can enable continuous assessment and proactive management. When combined with machine learning algorithms analysing large datasets and recognizing patterns, these approaches can optimize predictive accuracy and maintenance strategies. Improved geological characterization using advanced mapping technologies, such as geophysical surveys, is critical to account for weak planes, fractures, and faults. Additionally, long-term field monitoring and laboratory experiments are necessary to validate and refine models. Establishing standardized regulatory guidelines will help ensure consistency, particularly in challenging mining environments. Collaboration between academia and industry is essential to drive innovation and develop robust, reliable methods for coal pillar strength estimation.

From Imperfect Geotechnical Data to Accurate Prediction: Rockburst Classification Using the Pretrained Transformer TabPFN Model

Predicting underground mining rockbursts remains challenging because geotechnical data sets are often small, sparse, incomplete, and noisy, resulting from inconsistent testing standards, high acquisition costs, measurement errors, and geological variability. Consequently, traditional machine learning (ML) models frequently exhibit poor predictive accuracy and reliability. To address these issues, this study introduces the Tabular Prior-Data Fitted Network (TabPFN), a novel pretrained transformer that leverages approximate Bayesian inference to generalize effectively from small, noisy data sets. Moreover, this model can natively handle missing values, mitigate overfitting, reduce computational time, and simplify practical deployment, all without requiring fine-tuning. The model was evaluated on 150 rockburst cases using five key parameters to predict four rockburst intensity classes. To quantify its advantages over traditional ML methods, TabPFN was systematically benchmarked against support vector machine, multilayer perceptron, decision tree, random forest, and extreme gradient boosting. The results demonstrated TabPFN’s superior predictive accuracy on noisy and imbalanced data sets, and this advantage persisted in limited-data scenarios. It also performed robustly on incomplete data sets, with further gains when paired with imputation methods like k -nearest neighbors. Furthermore, the model provided strong interpretability, identifying the elastic energy index as the most critical predictive factor, followed by tangential stress and buried depth. TabPFN’s practical utility was further confirmed through successful application to real-world engineering projects (93% accuracy). Ultimately, this study establishes TabPFN as a powerful and reliable tool for rockburst prediction and highlights the broader potential of this pretrained model to effectively address the persistent challenge of scarce, poor-quality data in engineering geology.

Machine Learning for Bridge Fire Performance Assessment: Methods, Challenges, and Future Directions

The Evolution and Future of Machine Learning in Rock Slope Stability: A Systematic Review and Meta-Analysis

A Review of Machine Learning Applications in Monitoring Data Processing for Underground Engineering

With the acceleration of global urbanization and the large-scale development of underground spaces, underground engineering faces extremely complex and variable geological environments and high-risk construction disturbances. The widespread application of the Internet of Things and novel sensing technologies has given rise to structural health monitoring data increasingly characterized by massive volume, high dimensionality, multi-source heterogeneity, and strong spatiotemporal coupling. Traditional data processing methods based on mechanical analysis, empirical formulas, or numerical simulation have increasingly exposed limitations of insufficient accuracy, lengthy computation times, and weak generalization capability when confronted with such engineering big data. Machine learning and deep learning technologies, by virtue of their superior nonlinear mapping capability, advantages in feature extraction from massive data, and flexible architectural design, provide solutions for efficient knowledge extraction and intelligent assessment of underground engineering monitoring data. This paper reviews the current application status and frontier advances of machine learning technologies in the field of underground engineering monitoring data processing in recent years. First, the development trajectory of analytical algorithms evolving from classical shallow machine learning, through temporal and spatial deep learning, to physics-data dual-driven approaches is delineated. Second, targeting the critical challenges of missing field data and sparse sensor deployment, spatiotemporal fusion imputation techniques and spatial reconstruction methods incorporating mechanical prior knowledge are thoroughly evaluated, elucidating the paradigm shift in monitoring philosophy from discrete point-based alarming to inference-augmented sparse sensing that approximates full-field state awareness through model-dependent estimation rather than direct measurement. Third, the applications of machine learning in underground structural deformation mechanism interpretation, key influencing factor identification based on explainable artificial intelligence (AI), and rapid back-analysis of geomechanical parameters are summarized. Finally, composite network architectures and physics-constrained guidance strategies for non-stationary deformation time series prediction under complex and variable working conditions are discussed. A methodological audit of the 73 included studies—of which 33 enter the quantitative comparison tables—reveals that 26 of the 33 audited studies (78.8%) validate exclusively on single-project data, only 1 study conducts rigorous out-of-distribution generalization testing, and none of the 33 studies (0%) provides uncertainty quantification. These findings highlight cross-project generalization and probabilistic prediction as important methodological challenges. This paper aims to provide theoretical references and methodological guidance for safety early warning, intelligent construction, and full life-cycle health management of underground engineering.

Geological Hazard Prediction and Prevention: A Review of Mechanisms, Monitoring, and Mitigation

As industrialization expands, modern infrastructure increasingly collides with volatile geological environments, driving a sharp escalation of complex, cascading geohazards. These destructive hazard chains represent a critical global threat to structural and economic resilience. Despite this escalating risk, current early warning systems predominantly rely on fragmented, static methodologies that fail to capture the dynamic reality of temporal hazard evolution. Furthermore, contemporary predictive models are bifurcated between deterministic physical models demanding exhaustive geotechnical parameters and data-driven artificial intelligence algorithms functioning as opaque black boxes devoid of physical interpretability. To resolve these limitations, this review systematically evaluates the modern geohazard landscape through a coupled active-passive conceptual model. This framework systematically evaluates how primary high-energy failures trigger subsequent instability in surrounding geomaterials. The synthesis reveals that integrating space-air-ground multi-scale monitoring including orbital InSAR, UAV photogrammetry, and distributed ground sensors establishes a vital surveillance continuum for early hazard identification. Because these heterogeneous data streams possess significant environmental noise, rigorous multi-source data fusion remains essential to minimize false alarms and resolve spatial discontinuities. Analytically, physically based models provide indispensable mechanical transparency by explicitly simulating material deformation, whereas data-driven ensemble algorithms excel at processing high-dimensional, nonlinear geospatial inputs. Ultimately, transitioning toward proactive, real-time risk reduction demands the integration of physics-informed neural networks (PINNs) that embed geomechanical constraints into computational pipelines. Coupling these hybrid architectures with interactive digital twin technologies will transform static hazard mapping into dynamic virtual environments, empowering engineers to successfully mitigate evolving disaster chains globally.