基于神经网络的煤层自然发火的非线性预测

煤炭自燃是一典型的非线性现象。笔者论述了煤炭自燃的危害 ,从非线性理论的角度分析了煤炭自燃的本质特征 ;应用神经网络中BP网络这一高度非线性关系映射建立了自然发火预测模型 ,克服了传统预测方法的不足并在山东枣庄矿业集团公司柴里煤矿进行了预测分析 ,预测结果与验证结果基本吻合 ,取得了满意的效果 ,为解决煤炭自燃的预测提供了一条良好的思路和方法 ,具有较大的理论意义和应用价值。

A Non-linear Prediction of Coal Self-ignition Based on Neural Network

Coal self-ignition is a typical non-linear phenomenon. The hazard of coal self-ignition was discussed;the essential characteristics of coal self-ignition were analyzed in term of non-linear theory. The BP neural network as a highly non-linear mapping was applied to build the self-ignition prediction model,which has been applied in Chaili Coalmine of Shandong Zaozhuang Mining Industry Group Corp. with satisfactory results. It also provides a scientific way of thinking and a route for solving the prediction of coal self-ignition,which is of great significance both theoretically and practically.