为实现超长距离易自燃双煤巷布置煤自燃隐蔽性火源的早期预判,提出在气体分析法的基础上,联合运用井下监测系统大数据的分析法,提高隐蔽性自燃火灾早期预判的可靠性。引入能够仅反映煤相对缓慢氧化的背景CO浓度的概念,为筛除井下柴油机动车尾气产生CO的波动干扰,先做出基本假设：在一个时间单元(如一天)内多台柴油机动车繁忙工作中总能找到同时间歇的瞬间(监测时间间隔),按一天中CO最小监测值确定背景CO浓度。以最短自然发火期的一部分为考查期,以背景CO浓度持续增长为判断方法准则,给出回归趋势法和积分法2种预判模型,将预判分三级,即重点巡查、加强巡查和查明出现原因等三级。结合红庆梁煤矿得到CO最低的平均递增率k_1^*= 0.607 ppm/d。基于上述可靠的预判方法,将超长距离巷道每天全面巡查,改为有针对性的重点巡查。

To achieve the early prediction of the latent fire source of coal spontaneous combustion in the long-distance spontaneous combustion coal seam double coal roadways, it is proposed to jointly utilize the Data Analysis Method of the coal mine monitoring system which can early predict and forecast the latent spontaneous combustion fire and improve the reliability, while it still keeps and ameliorates the Gas Analysis Method. In order to single out the fluctuations of CO generated by the exhaust gas of diesel vehicles in the coal mine, it is introduced the background of CO concentration, and the basic assumption is made: in one day, one or more diesel vehicles can finally be found intermittent moments in the busy work (monitoring time) ), then given the background CO concentration in a day which is determined by the minimum monitored value of CO. Taking the shortest natural combustion period as the examination period, and the continuous increase of CO concentration as the criterion of judgment method, we give two prediction and prediction models of regression trend method and integral method, furthermore,the forecasting and forecasting is supposed to divide into three levels, namely, key inspection, enhanced inspection and the reason query. In combination with Hong Qingliang Coal Mines, the lowest average rate of increase of CO, k1^*= 0.607 ppm/d is obtained. Based on the above-mentioned reliable predicting and forecasting methods, the long-distance roadway will be changed the comprehensive inspection every day to the targeted key inspection.

TD752.2

国家自然科学基金项目（面上项目，重点项目，重大项目）