Back‐break (BB), a significant environmental concern in mining operations, arises from the improper execution of blasting techniques. The primary cause of BB is the inadequate design of blasting pattern parameters, which leads to structural issues such as fractures, overhangs, and underhangs on the construction face. These issues escalate the costs associated with creating a smooth high wall and a clear face for subsequent blasts. Given the critical need to address these challenges, developing a reliable predictive model for BB is essential. Such a model can effectively represent the inherent characteristics of the construction face and guide more precise blasting operations. This study introduces an advanced predictive model leveraging a tree‐based approach, light gradient boosting machine (LightGBM), optimized through the whale optimization algorithm (WOA). The unique integration of WOA in this model ensures the identification of optimal hyperparameters, enhancing the LightGBM model’s predictive accuracy. The model was trained using a comprehensive dataset of BB measurements obtained from the Sungun mine in Iran, a prominent mining site known for its complex geological features. The findings from this study underscore the efficacy of the proposed hybrid LightGBM–WOA model in accurately forecasting BB. On the training dataset, the model achieved a coefficient of determination ( R 2 ) of 0.997, a value account for (VAF) of 99.651%, and a root mean square error (RMSE) of 0.152. The model’s performance on the test dataset was equally robust, with an ( R 2 ) of 0.990, a VAF of 99.030%, and an RMSE of 0.230. These results demonstrate the model’s potential to significantly improve the precision of blasting operations, thereby reducing environmental impact and operational costs.
Yari et al. (Thu,) studied this question.