Modelling of blasting-induced air overpressure wave propagation under atmospheric conditions by using ANN model

Abstract

In this study, the propagation characteristics of blast-induced air overpressure waves in different atmospheric conditions were investigated. For this purpose, 158 blasting shots were observed at an excavation site in Istanbul, and 562 air overpressure events were recorded at 32 different measurement stations. The recorded air overpressure and weather conditions data were used together to train an artificial neural network (ANN) model to determine the air overpressure propagation characteristics that would occur in future studies. Besides, to provide a meaningful interpretation of the results of the study, the propagation trend of air overpressure waves was simulated by using the Kriging method. As a result of the study, it was found that there is a high correlated exponential relationship (r(2) =0.79) between air temperature and air overpressure propagation, while wind velocities under 6 m/s have no effect on the propagation. The developed ANN model allowed the estimation of the air overpressure that any blasting under different atmospheric conditions at any point in and around the excavation site. Atmospheric conditions and constructions have an impact on the blast-induced air shock propagation. Considering the atmospheric conditions and structure variation, it can be said that only dynamic models can represent the site for air shock propagation. Therefore, it is essential to carry out the application verified in this study.