Application of White-Box Machine Learning Models for the Prediction of Blast-Induced Peak Particle Velocity

Blast-induced ground vibration is a critical environmental hazard in open-pit mining operations, capable of causing severe damage to adjacent structures and infrastructure. Accurately predicting vibration intensity is universally quantified by the peak particle velocity (PPV) index. Among the various factors affecting blast-induced ground vibrations, the distance from the blast face to the monitoring point (D) and the charge weight per delay (W) are the most influential and controllable parameters in a specific mine site. Therefore, these variables were selected as inputs for PPV estimation. The present study develops advanced white-box machine learning (ML) models, including Multi-Expression Programming (MEP), Gene Expression Programming (GEP), Multivariate Adaptive Regression Splines (MARS), and Stronger Variable Creator Machines (SVCMs), for predicting PPV. The general explicit equation was derived from the developed ML models implemented in a spreadsheet program, which can be easily used to estimate PPV. The MEP model achieves the highest accuracy, with a correlation coefficient (CC) of 0.993177 and a root mean square error (RMSE) of 0.836337, followed by the MARS, GEP, and SVCM models. The results of the present study, supported by k-fold cross-validation and parametric analysis, confirmed the potential of the proposed models for PPV estimation.