Application of a BP neural network in predicting destroyed floor depth caused by underground pressure

Abstract

Predicting the destroyed floor depth caused by the mining of coal seams is of great importance in judging whether the mining of a deep coal seam can be safely performed above a confined aquifer and to prevent the inrush of water from the floor. Thirty sets of coal mining data on destroyed floor depth were selected for study. A comprehensive analysis of the factors that influence the depth of destruction of coal seam floor strata was performed and combined with the ability of a BP neural network to address dynamic nonlinear information. Then, a set of test samples was assembled and used to construct a predictive model using a BP neural network. The model was then used to predict the destroyed floor depth of the 7105 working face of the Baizhuang Coal Mine in the Feicheng coal field. To verify the effectiveness of the model, the depth of the destroyed strata comprising the coal seam floor was measured using equipment called the “Double Sided Sealed Borehole Water Injection Device.” By comparing the predictions made by the BP neural network with actual measurements, the conclusion was reached that a BP neural network model can effectively be used to predict the destroyed floor depth caused by the mining of a coal seam.