Artificial neural network for prediction and control of blasting vibrations in Assiut (Egypt) limestone quarry

摘要

According to the data scattering of ground vibrations, the traditional charge weight scaling law is unable to predict the PPV up to an acceptable limit. Three different models of ANN were used in predicting the peak particle velocity (PPV). It was found that the ANN model which is based on the large number inputs parameters gives better prediction of PPV over the ANN models that use single or two inputs parameters. Also, the ANN in which two-input parameters are used is better than one-single input. That is to say, increasing the number of input variables result in increasing the ability of ANN to learn and to be expert. Comparisons between the predicted PPV using the empirical scaling law regression and the neural network models were made. It was found that different models of neural networks give much better prediction of PPV than does the scaling law model.