Predicting blast-induced outcomes using random forest models of multi-year blasting data from an open pit mine

Abstract

Rock fragmentation and movement are two important outcomes of the blasting process in open pit mines. They are influenced by blasting design parameters, as well as by the physical and geomechanical characteristics of the rock mass. This paper presents the analysis results of multi-year blasting data from an open pit mine in Canada and proposes a predictive model for the blast-induced outcomes that incorporates both rock mass properties and blasting parameters. The analysis employed the decision tree (DT) and random forest (RF) models to determine influential parameters, confirming that the blast-induced fragmentation and movement are influenced by rock mass characteristics (i.e. intact rock strength and rock quality designation, RQD), as well as by blasting design parameters. The decision tree model facilitates the visualization of geomechanical and blasting design parameters influencing the blast-induced outcomes. The robust random forest model provides prediction of blast-induced outcomes. The decision tree and random forest models make it possible to determine blasting design parameters that could be modified to achieve better blast-induced outcomes based on the rock mass conditions at the mine site.