Optimization of postblast ore boundary determination using a novel sine cosine algorithm-based random forest technique and Monte Carlo simulation
Zhi Yu, Xiuzhi Shi, Xianyang Qiu, Jian Zhou, Xin Chen, Yonggang Gou
Abstract
The accurate determination of postblast ore boundaries can significantly help to control ore loss and dilution in opencast mines. Determining the boundaries is difficult using methods other than direct and expensive blast-induced rock movement monitoring, so many mines directly use the preblast ore boundary to guide the shovel. A new postblast ore boundary determination method using a soft computing technique and stochastic modelling method is proposed. Based on a case study and performance comparison, a high-precision hybrid metaheuristic model combined with the sine cosine algorithm and random forest technique (SCA-RF) was developed and used in a Monte Carlo simulation to analyse the probability distribution and parameter sensitivity. Mining engineers can obtain a more accurate postblast ore boundary by moving the preblast ore boundary toward the free face by a certain distance after considering the probability distribution of blast-induced rock movement, which is significantly better than using the preblast ore boundary.