The Technique Research About Prediction And Forecast Of Coal Spontaneous Combustion Based On Network

The paper introduces the current research of prediction for coal spontaneous combustion firstly, points out the advantages and disadvantages of the current used methods;It gets the non-linear characters of spontaneous combustion through it's mechanism and process .According the non-linear characters, and the advantages of network and time series, find the fit network for the spontaneous combustion fatalness rank prediction and for the spontaneous combustion prediction in mined zone, through the contrast experiments between three network models(BP ,RBF ,TD-BP) to predict time series. Especially study the prediction network for the spontaneous combustion in mined zone, and put into idiographic mines. It brings the dynamic method which according to the network and time series theories to predict the index gases consistence and the position of high temperature areas, on the base of analysis the index gases. It proves that using the traditional BP network can distinguish the rand of spontaneous combustion fatalness through the validated result. But the prediction of the coal spontaneous combustion trend it only has better effect for the own mine's or resemble ones no matter using which network model, because the particularity of mine. The BP network fit to predict earlier index gases consistence, and the TD-BP network fit to predict latter ones. The instaurations of two networks offer a new way and method to solve the prediction of coal self-ignition. Figure [14] Table[23] Reference[37]...