42 Mine Fire Early Prediction Method with Multi-Parameters Based on Neural Network Optimized by Levenberg-Marquardt Algorithm
Download citation file:
In view of the early prediction of mine fires, a new method had been studied. The parameters, the coal's spontaneous combustion tendency, the concentration of some gases, were comprehensively analyzed according to different mine condition, which could be used to predict mine fires. The fire prediction model was set up based on neural network optimized by Levenberg-Marquardt algorithm. This was benefit to speed up the convergence rate and improve forecasting rate of identification. Simulation results show that the rate of correct prediction was further improved. This algorithm propound a new way on fire information identification which achieves the most early...