Simulating the peak particle velocity in rock blasting projects using a neuro-fuzzy inference system

Ground vibration is an adverse effect induced by rock blasting in civil and mining projects. Peak particle velocity is the most important descriptor to evaluate the ground vibration in the blasting sites. The present paper proposes an adaptive neuro-fuzzy inference system (ANFIS) for simulating the PPV in Shur River Dam area, Iran. For checking the ANFIS performance in simulating the PPV, a linear regression model is also used. To achieve the objective of this research, 90 blasting operations were monitored in the mentioned site and the values of weight charge per delay and distance between the blasting face and the installed seismograph, as the most effective parameters on the PPV, were measured. Using magnitude of three error indices, i.e., coefficient of correlation, variance account for and root mean square error, we proved that the proposed ANFIS model can simulate the PPV with a high degree of accuracy and reliability. The values of the coefficient of correlation obtained from the ANFIS and linear regression models were 0.983 and 0.876, respectively, that indicate the ANFIS outperforms the linear regression model.