Comparing Empirical Models with ANN in Estimation of Vibrations Resulted from Blasting, Dareh-Baq Dam

Blast hole drilling and blasting are from among cost effective and economic methods of crushing the rock in civil projects, tunneling, as well as surface and underground mining. Ground vibration is the most important undesirable effect of blasting and if not controlled, it can lead to many damages. The paper is aimed at studying and prediction of effects of vibrations resulted from blasting on structure of dam on Dareh-Baq River. For this purpose, four empirical equations along with Artificial Neural Network (ANN) have been used to the aim of achieving a highly accurate model to predict vibrations of ground. Also, level of vibrations created would be compared through existing standards. According to the above goals, 73 blasting cases in Dareh-Baq River Dam area have been studied and required parameters as for prediction have been measured. From 73% of information related to blasting has been used to obtain empirical equation and also provide appropriate model in ANN; and, the remaining 27% of information have been used to specify performance and evaluate accuracy level of various models, in comparison to real values. After evaluation of the results, it became clear that ANN is of highest accuracy for prediction of vibrations resulted from blast. Also, in consideration of recorded vibrations and their comparison to existing standards, as well as distance of dam on Dareh-Baq River from location of blasting, energy from vibrations created will be dissipated and no undesirable effect would be imposed on dam structure.