Development of overbreak prediction models in drill and blast tunneling using soft computing methods

Abstract

According to the human requirement, such as mineral extraction, transportation, etc., tunnel construction is an evident need. Mining and tunneling activities are full of problems with uncertainty, which make these problems complex and difficult. Overbreak is one of these problems that we are encountering in the tunneling process, particularly in the drill and blast method. Overbreak causes to decreasing the safety and increasing the operational costs. Therefore, it must be minimized. Overbreak prediction is the first step to decreasing the damaging effects of this phenomenon. Overbreak influencing factors are classified into two groups of uncontrollable factors and controllable factors. In this study, 267 sets of causing factors and overbreak data were used to overbreak prediction model’s development using the multiple linear and nonlinear regression analysis, artificial neural network, fuzzy logic, adaptive neuro-fuzzy inference system, and support vector machine. The determination coefficient values of these models have been obtained as 0.84, 0.87, 0.93, 96, 0.97, and 0.87, respectively. The results illustrated that the fuzzy and adaptive neuro-fuzzy inference system models have done more appropriate prediction than other prediction models.