Prediction of Dust Emission Due to Open Pit Mine Blasting Using a Hybrid Artificial Neural Network

This study developed a new perspective of artificial neural networks using dimensional analysis to be applicable to certain prediction problems. To this end, dimensional analysis was combined with multi-layer perceptron and radial basis function neural networks for application to prediction of blast-induced dust emission. The dimensional analysis concept was designed to improve the prediction performance in the neural network process. The results of the developed neural networks were validated by comparing them with the measured results. Insights indicated that the dimensional multi-layer perceptron neural network analysis was better than that of the dimensional radial basis function neural network. In addition, the region's winds and their impacts on dust emissions were investigated in detail to estimate wind velocity and direction. A radius of approximately 1 km was identified as the dust impact zone, where agricultural and residential areas may be affected adversely.

Automated summary

This study developed a new perspective of artificial neural networks using dimensional analysis to improve prediction of blast-induced dust emission. Additionally, it investigated the impact of region's winds on dust emissions in detail.