The application of fuzzy clustering and algebra neural network in coal combustion forecasting and warning

For the forecasting and waring system of coal spontaneous combustion, the using of multi-index data fusion technology for predicting the dangerous degree of coal spontaneous combustion is more scientific and reliability than the single index parameter.Through extracting multi-index gases and gas-ratio based on sensitivity and regularity of index gases ,such as N₂, O₂, CO, CO₂, CH₄, C₂H₂ and so on, and according the nonlinear mapping relationship between temperature and gas index, the applications of neural network and cluster analysis in the previous forecast and real-time monitoring provide reliable guarantee for coal mine safety, and ensure the establishing of automatic alarming software system.