20 / 2020-02-26 07:12:00

Practical Models for Blast-induced Flyrock Assessment

Flyrock prediction,Classification and regression tree,Gene expression programming,Logistic regression

Flyrock is known as one of the most dangerous side effects of blasting operation in open-pit mines, which can damage equipment, structures and mine workers. Therefore, the correct prediction of this phenomenon at the blast design stage not only can improve the safety of mining operation but also can enhance mining productivity. This study proposes novel and practical models using classification and regression tree (CART) and the hybrid gene expression programming-based logistic regression (GEP-LR) techniques to assess blast-induced flyrock based on a database compiled from a limestone open-pit mine in Iran. Several controllable parameters including burden, spacing, stemming, hole length, sub-drilling, powder factor, and charge per delay along with the geological strength index (GSI) are used as independent indicators for modelling, while the measured flyrock distance is introduced as the output parameter. Two new models are proposed for prediction of flyrock distance and its intensity. The prediction performance of the proposed models is evaluated using several statistical indices. The results prove the robustness and high accuracy of the models for flyrock prediction. These models can be used easily in practice by engineers/researchers without a need to implement any specific software or training for usage.