Predicting Blasting-Induced Ground Vibration in Mines Using Machine Learning and Empirical Models: Advancing Sustainable Mining and Minimizing Environmental Footprint

Blasting-induced ground vibrations, typically quantified by peak particle velocity (PPV), pose one of the most critical environmental challenges in surface mining and can damage nearby structures and disrupt surrounding ecosystems. Consequently, the development of reliable and accurate predictive models is essential for designing safe, environmentally responsible, and sustainable blasting operations. This study develops a robust predictive framework using a harmonized database of 506 blasting events, from which 386 high-quality records were retained after preprocessing to model PPV as a function of charge per delay (Q), monitoring distance (R), and rock mass rating (RMR). Several machine learning (ML) algorithms, including artificial neural networks trained using the Levenberg–Marquardt algorithm (ANN-LM), adaptive neuro-fuzzy inference systems (ANFIS), Gaussian process regression (GPR), and decision trees (DT), were evaluated alongside conventional empirical models such as the USBM, Ambraseys–Hendron, Langefors–Kihlstrom, and BIS. To further enhance predictive capability, two optimization strategies, Bayesian optimization (BO) and differential evolution (DE), were applied to the GPR model, producing optimized BO-GPR and DE-GPR variants. Model performance was assessed using the correlation coefficient (r), variance accounted for (VAF), mean absolute error (MAE), and relative root mean square error (RRMSE). Results indicate that the BO-GPR model achieved the best predictive performance during testing for both the two-input (Q, R) and three-input (Q, R, RMR) configurations, with r values of 0.97426 and 0.98381, respectively, and VAF values exceeding 94%. SHAP analysis revealed monitoring distance as the dominant attenuating factor controlling PPV. The optimized framework provides an accurate, interpretable tool for vibration prediction and precision blast design, supporting environmentally responsible, sustainable mining operations.