基于危机征兆的煤矿内因火灾预测模型的研究

摘要

To effectively prevent coal spontaneous combustion from causing heavy losses, the occurrence symptoms of coal spontaneous combustion were selected as its prediction indicators before fire disaster, and through the statistical analysis of the accident cases,6 second-level indicators and 40 third-level indicators for coal spontaneous combustion were identified. BP neural network was selected as the prediction model,in which,the data establishment was realized by the probability theory analysis. In the Matlab-based environment, study was carried out on the factors affecting the prediction model performance, the optimal parameter values were determined,and the advantages of this method such as the generality,expansibility and high efficiency were summarized from the experiments.%为有效预防煤炭自燃引发的内因火灾造成重大损失，选取煤炭自燃发生的征兆为预测指标，通过统计分析事故案例，确定出煤炭自燃二级指标6个，三级指标40个。选用BP神经网络作为预测模型，其中数据的建立采用概率理论分析得到，并在Matlab的环境下研究影响预测模型性能的因素，确定最佳参数值，从实验中总结出该方法具有通用性、扩展性和高效性等优点。