Forecast of Small Structure for Mine Based on BP Neural Networks

the paper elaborates the building, training and simulation method of BP networks which can predict the horizontal extending length of small faults in coal seam. Based on the summary of the influence factors including drop height for small scale faults, fault strike and fault dip, dip direction and other factors, combined with the typical sample data of 7^# coal seam which is mainly in Zhaoguan Mine Field, using Matlab software to construct a new network prediction model, and the horizontal extension length of small faults in working surface No. 2713 and No. 2712 of coal seam are predicted, it is obvious that the measured results got by using the model system is basically agree with the actual measurement result.