Prediction of over break phenomenon in tunnel excavation using intelligent optimization methods, Case study: Tazare coalmine

Knowadays tunnel excavation is an evident need according to the human requirement such as mineral supplement, transportation, underground storage etc. Over break is one of the unfortunate phenomenon’s that could be encountered in tunneling, particularly in the drill and blast tunneling method. Over break phenomenon, reduce the safety of the working environment and increase the operational costs. Over break prediction is the first step to decrease the damaging effects of this phenomenon in the tunnel construction process. The causing factors of over break are classified into two groups of uncontrollable factors (geological parameters) and controllable factors (blasting parameters) and all of the factors are nonlinearly correlated. In this study, 52 sets of causing factors and over break data were applied to the multiple linear and nonlinear regression, Artificial Neural Network (ANN), Fuzzy logic and Adaptive Network-based Fuzzy Inference System (ANFIS) to predict over break as output parameter. The determination coefficient (R2) values of multiple linear and nonlinear regression, Artificial Neural Network, Fuzzy and Adaptive Network-based Fuzzy Inference System Models have been calculated as 0.71, 0.73, 0.80, 0.95 and 0.91, respectively. The results of the prediction models illustrate that the fuzzy and ANFIS models, have done more appropriate prediction than other prediction models. Also, sensitivity analysis showed that burden value is the most effective parameter on the over break. With awareness of the over break occurrence, we can use controlling and preventing methods to reduce the harmful effects of this phenomenon, and ultimately improve project performance.