Hide/Show Apps

Comparative study on ground vibrations prediction by statistical and neural networks approaches at Tunçbilek coal mine, panel byh

Akeil, Salah
In this thesis, ground vibrations induced by bench blasting from the Tunçbilek Coal Mine, Panel BYH, were measured to find out the site-specific attenuation and to assess the structural damage risk. A statistical approach is applied to the collected data, and from the data analysis an attenuation relationship is established to be used in predicting the peak particle velocity as well as to calculate the maximum allowable charge per delay. The values of frequencies are also analyzed to investigate the damage potential to the structures of Tunçbilek Township. A new approach to predict the peak particle velocity is also proposed in this research study. A neural network technique from the branch of the artificial intelligence is put forward as an alternative approach to the statistical technique. Findings of this study indicate, according to USBM (1980) criteria, that there is no damage risk to the structures in Tunçbilek Township induced by bench blasting performed at Tunçbilek coal mine, Panel BYH. Therefore, it is concluded that the damage claims put forward by the inhabitants of Tunçbilek township had no scientific bases. It is also concluded that the empirical statistical technique is not the only acceptable approach that can be taken into account in predicting the peak particle velocity. An alternative and interesting neural network approach can also give a satisfactory accuracy in predicting peak particle velocity when compared to a set of additional recorded data of PPV.