Artificial neural network approach for prediction and control of ground vibrations in mines

Blasting is a cheap, economical means of rock excavation for mining or civil engineering projects. Ground vibrations resulting from the use of huge quantities of explosives are of great interest to design engineers of surface and underground structures. Only a small fraction of explosive energy is effective in fragmenting the rockmass, the rest being dissipated as ground vibrations, air blasts, noise, fly rocks, backbreaks, etc. The peak particle velocity produced by ground vibration is commonly used to evaluate the risk of vibration damage due to blasts. Empirical equations have been derived to predict ground vibration but these are not generally applicable beyond the specific conditions used to produce the data. Ground vibration effects are complex and are influenced by numerous controllable and uncontrollable parameters. An attempt has been made to predict the ground vibration using an artificial neural network. It was found that the network provided better results than a conventional regression method based on correlation coefficient and error estimations.