<![CDATA[In recent decades, generation of ground vibrations results from blasting activities in mining sector has been identified as a significant cause of extensive harm to nearby structures, vegetation, and individuals. Hence, it is imperative to closely monitor and accurately forecast the uncertain levels of vibration, and implement the appropriate steps to mitigate their potentially harmful impact. The objective of this study was to establish a correlation between the peak particle velocity and the various parameters that influence it. This study employed the deployment of the artificial neural network approach to assess and forecast the uncertain ground vibrations. In this study, a multilayer perception neural network with three layers and a feed-forward back-propagation architecture was employed. The network consisted of five input parameters, namely the distance from the blast face, maximum charge per delay, spacing, burden, and depth hole. The output of interest was the peak particle velocity. The neural network was trained using the Levenberg–Marquardt algorithm, and the training dataset comprised 29 experimental records and blast event data obtained from the limestone mine in Indonesia. In order to assess the effectiveness and the precision of the artificial neural network model that was created, a total of four conventional predictor models were utilized. These models were proposed by reputable sources such as the US Bureau of Mines, Ambraseys–Hendron, Langefors–Kihlstrom, and the Bureau of Indian Standards. The results collected from the demonstrate study show that the artificial neural network model suggested in this research has the ability to provide more precise estimations of ground vibrations in comparison to existing conventional prediction models. The artificial neural network model yielded a coefficient of determination (R2) of 0.9332 and a root mean square error (RMSE) of 0.4763.]]>
Copyrights © 2024
