Assessing the Importance of Air-Decking Blasting in Controlling BIGV at an Open Pit Gold Mine in Tanzania using ANN Model

Abstract

Air-decking blasting technique is used to control blast induced ground vibration (BIGV) in several mines including open-pit gold mines located in Tanzania. While the importance of air-deck to control BIGV is practically evident, theoretical models such as BIGV prediction models cannot be used to assess the importance of air-decking. The main objective of this study was to assess the importance of the air-decking blasting technique to control BIGV using the artificial neural network (ANN) Model. To achieve this objective, ANNs were modeled and trained to learn the pattern of data using Multilayer Perception with Back Propagation to predict BIGV. The main results showed that the normalized importance of air-decking in predicting BIGV was 92.4%. Other important parameters were distance from blasting with a normalized importance of 100% and MIC which was relatively low with a normalized importance of about 46.2%. Parameters such as charge length, powder factor, bench height, charge per length, and stemming length were by far less important than air-decking. The ANN model developed in this study appeared to perform well by incorporating air decking parameters, which traditional BIGV predictors could not. The model also can predict BIGV with an error of about 1.8%. It was recommended that the air-decking technique used at the gold mine should be maintained and practiced to control BIGV for the sustainable development of the mining industry in Tanzania.