基于神经网络的自然发火危险性评价与预测

Evaluation and Prediction of Coal Spontaneous Combustion Hazards Risk Based on Neural Network

对采空区煤炭自然发火进行了危险性评价,从自然发火条件入手,建立采空区自然发火事故模型,确定采空区发火现状危险等级为C级,较危险;利用采空区已有的自然发火预测指标建立BP神经网络的时间序列预测模型,对未来该采空区有无发火危险进行了预测,确定未来采空区发火可能性大小.结果表明:运用BP神经网络的时间序列预测模型对煤炭自然发火进行预测,采空区自然发火处于"有发火危险"程度,发火危险性较大,因此应做好采空区火灾预防工作.

Risk of coal spontaneous combustion in goaf was evaluated and the incident model of spontaneous combustion established to determine the risk grade. The result shows that this risk grade is C, which means ＂quite dangerous＂. In addition, the time series prediction model based on BP neural network was also set up by using of forecasting indexes of goal spontaneous ignition. By using this model, we forecast future risk of spontaneous combustion, showing that the goaf is in danger of spontaneous combustion. So prevention measures must be taken timely.