Feasibility of particle swarm optimization and multiple regression for the prediction of an environmental issue of mine blasting

Purpose Blasting is an economical method for rock breakage in open-pit mines. Backbreak is an undesirable phenomenon induced by blasting operations and has several unsuitable effects such as equipment instability and decreased performance of the blasting. Therefore, accurate estimation of backbreak is required for minimizing the environmental problems. The primary purpose of this paper is to propose a novel predictive model for estimating the backbreak at Shur River Dam region, Iran, using particle swarm optimization (PSO). Design/methodology/approach For this work, a total of 84 blasting events were considered and five effective factors on backbreak including spacing, burden, stemming, rock mass rating and specific charge were measured. To evaluate the accuracy of the proposed PSO model, multiple regression (MR) model was also developed, and the results of two predictive models were compared with actual field data. Findings Based on two statistical metrics [i.e. coefficient of determination (R-2) and root mean square error (RMSE)], it was found that the proposed PSO model (with R-2 = 0.960 and RMSE = 0.08) can predict backbreak better than MR (with R-2 = 0.873 and RMSE = 0.14). Originality/value The analysis indicated that the specific charge is the most effective parameter on backbreak among all independent parameters used in this study.