مقايسه مدل هاي تجربي و شبكه عصبي در برآورد لرزش ناشي از آتشكاري در حضرياب سد مسجد سليمان

A Comparison Between Empirical Models And Neural Network Method In Estimation Of Blast Induced Vibration In Masjed Solaiman Dam underground excavation

Drilling and blasting operation is performed simultaneously with mechanical and electrical jobs in phase one of the development project of Masjed Solaiman Dam power plant. To avoid any damage to structures and facilities, precise estimation of land vibration is required. This would help a proper design for blasting operation, too, due to which the level of ground vibration must remain within the permitted limits set up by some well-accepted standards. in this investigation, based on maximum instantaneous charge and the distance from blast locations, vibration is estimated by means of both empirical models and neural network technique. The analysis of recorded data revealed that application of five empirical models resulted in more or less similar results whereas the neural network method showed to be much more accurate technique. Due to non-linearity, flexibility and the ability of accepting variation in the number of impacting factors, the neural network method was observed to be, by far, more capable technique. It is, therefore, concluded that the latter method can be considered as a proper tool for ground vibration estimation.