Importance and sensitivity of variables defining the performance of pre-split blasting using artificial neural networks

Blasting is used in mining and civil excavations to break the rockmass for ore production or creation of space. One of the disadvantages of the method is that it causes breakage beyond the proposed excavation lines. Pre-split is used to control such breakage in rockmass. This paper introduces the use of digital image analysis technique (DIAT) to measure rock breakage and half cast factor (HCF%). A new factor, undamaged area (AUD%), is introduced to improve the prediction, owing to the shortcomings of the HCF%. A comparison of significant data of AUD% and HCF% showed that HCF% has a lacuna of overpredicting, whereas AUD% can be used effectively to address the issue. Artificial neural networks are used to solve the complexity of multiple variables involved in predictions and to ascertain the importance and sensitivity of such variables in prediction regimes. This method will help users to evolve blast-induced damage objectively. A comprehensive scheme of things for future developments is included, which can be helpful to researchers in defining their research in this area.