Effect of Deck Length on Ground Vibration in Dragline Bench Blasting Using Artificial Intelligence Methods - Mining, Metallurgy & Exploration (2024)

Formulating a predictive model for blast-induced ground vibration (BIGV) has always been challenging due to the complex non-linear relationship between the BIGV and its influencing factors. Researchers have used empirical, artificial intelligence methods and numerical modelling methods to predict BIGV. In a dragline bench blasting, a massive quantity of explosives is used compared to other loading machinery, and the height of the bench depends on the size of the dragline. For effective utilization of explosives energy, decking is used to reduce the hazardous effects of blasting. In this paper, an attempt has been made for effective prediction of BIGV, and its mitigation on dragline bench blasting considering a deck length, stiffness ratio, hole spacing to burden ratio, hole depth to burden ratio, the distance of monitoring PPV from blasting site, Schmidt hammer rebound number of the bench to be blasted, and maximum charge per hole to total charge per blast. So artificial neural network (ANN) model and random forest (RF) model have been developed considering the parameters mentioned above. For the statistical performance of the predictive model, the coefficient of determination (R2) value is 0.8895, MSE is 22.84 for the ANN model, and the random forest (RF) model whose R2 is 0.8165 and MSE is 27.18. The sensitivity analysis results show that stiffness ratio, deck length, and Schmidt hammer rebound number are more influential parameters out of the selected parameters. The reduction of ground vibration with the introduction of decking led to a 3 to 15% reduction in ground vibration.