A Neural Network Model for Predicting the Maximum Water Inrush Quantity of the Roof on Working Face

As the prediction of the maximum water inrush quantity on working face is essential and difficult, and water inrush from aquifers overlying coal seams is widespread in China, factors that influence the maximum water inrush quantity of the roof are analyzed in this paper. It is found that there are multiple factors and their influence mechanisms are complex. Considering the high non-linearity between each of these factors and the water inrush quantity, based on the artificial neural network (ANN) theories and methods, the inclination angle of the working face, working face length, mining thickness, aquifer thickness, aquifuge thickness, and mudstone thickness in the aquifuge, are defined in this paper as the main factors for the case study area. An ANN-based model for predicting the maximum water inrush quantity of the roof is proposed. Training and actual test results show that the proposed model's computation results approximate to the ground truth, thus demonstrating the reliability of the proposed model.