矿山安全态势预测预警研究

Research on mine safety situation forecast and early warning

基于物联网技术获取矿山安全大数据并加以充分挖掘、利用,有利于实现矿山安全态势预测预警。以瓦斯爆炸事故为例,通过分析事故致因,构建了矿山安全态势评价指标体系,并对各评价指标进行了量化。基于长短期记忆(LSTM)网络和贝叶斯网络构建了矿山安全态势预测模型,根据矿山安全监测数据,通过LSTM得到矿山安全态势评价指标预测值,由贝叶斯网络根据评价指标预测值推理得出矿山安全事故风险概率,实现矿山安全态势预测。基于安全态势预测结果建立了预警机制,根据警情划分4级预警级别及响应部门,制定了相应的预警措施。以某煤矿某次瓦斯爆炸事故为例进行反演,结果表明基于LSTM和贝叶斯网络的矿山安全态势预测结果与实际情况吻合。

Based on the Internet of Things technology,obtaining the mine safety big data and making full use of the data are helpful to realize the forecast and early warning of mine safety situation.Taking the gas explosion accident as an example,by analyzing the cause of the accident,a mine safety situation evaluation index system is constructed,and each evaluation index is quantified.Based on the long and short-term memory(LSTM)network and the Bayesian network,a mine safety situation forecast model is proposed.According to the mine safety monitoring data,the mine safety situation evaluation index forecast values are obtained through the LSTM.The risk probability of mine safety accidents is inferred from Bayesian networks based on the evaluation index forecast values to obtain mine safety situation forecast.Based on the safety situation forecast results,an early warning mechanism is established.4 warning levels and response departments are classified according to the warning situation,and corresponding early warning measures are established.An inversion of a gas explosion accident in a coal mine is used as an example,and the results show that the forecast results of mine safety situation based on LSTM and Bayesian network are consistent with the actual situation.