Influence of geological conditions on the powder factor for tunnel blasting

Abstract: The main purpose of this study was the application of the neural network (NN) model to thedetermination of the optimal powder factors, based on a series of observations and numerical experiments.The input parameters were 14 geological conditions. The data for the NN application in this study werecollected in a highway tunnel under construction in Korea. An optimum NN model was determined bytraining and testing models with the collected data. It was shown that the NN model could predict the powderfactor depending upon the selected input parameters. In addition, the strength of the relationship between thepowder factor and the 14 input parameters was evaluated by three different sensitivity analysis methods, i.e.the analysis of relative strength effect, the cosine amplitude method and the fractional factorial design. Fromthese analyses, not only the dominant factors among the input parameters, but also their interactions could bedetermined. Keywords: Powder factor; Neural network; Tunnel blasting; Relative strength effect; Cosine amplitudemethod; Fractional factorial design