airblast

Drilling and blasting procedures are essential components of hard rock mining. Although designed to fragment rock mass into smaller segments, these methods are linked to detrimental environmental effects, such as airblast, ground vibrations, noise, emissions, and flyrock, which may threaten the safety of lives and property. This paper proposes employing a hybrid particle swarm-optimized artificial neural network (PSO-ANN) to forecast airblast in quarry blasting. Additionally, 42 datasets encompassing critical metrics, such as hole depth (HD), maximum charge per delay (MCPD), distance between the blasting face and monitoring point (D), total explosive mass (TEM), and number of holes (NH), were documented at a granite quarry in Botswana. The efficacy of the PSO-ANN model was evaluated against artificial neural network (ANN) models, USBM predictor, and multivariate regression analysis (MVRA) using the coefficient of determination (R2), mean squared error (MSE), and mean absolute percentage error (MAPE). The PSO-ANN model featuring a network architecture of 5-34-1 (5 input parameters, 34 hidden neurons, and 1 output parameter) with radbas transfer function demonstrated superior performance relative to other models.