| As the coal mine enters deep mining,the ground stress and gas pressure increase gradually,and the danger of coal and gas outburst increases gradually and becomes complicated,making prediction difficult.The current single prediction method is difficult to achieve real-time analysis and multi-index fusion to improve the prediction accuracy,and there is interference in the field test,which is easy to produce false alarms.Based on the application status of three signal signals of acoustic emission(AE),electromagnetic radiation(EMR)and gas concentration in coal and gas outburst danger monitoring and early warning,this paper constructs a coal and gas outburst monitoring and early warning system for AE EMR and gas in Faer coal mine,analyzes the correspondence between the conventional indicators of AE EMR and gas in Faer coal mine.Based on the analyzes of the interference characteristics of the AE and EMR signals,proposes singular values decomposition and noise reduction methods are used to filter the interference signals,and the prominent dangerous precursory characteristics of the AE EMR and gas are analyzed.Based on this,characteristic information of the AE EMR and gas are extracted,the early warning method of the trend and the critical threshold were determined,solved the problem of quantifying the extraction of acoustic and gas characteristics information.The AHP method was used to determine the weight of the AE EMR and gas in the early warning.A comprehensive early warning method for AE EMR and gas was proposed.The reliability is improved that the comprehensive warning accuracy is higher,and the warning effect is good.The main research results are as follows:(1)It is determined that the optimal time scale for the comparative analysis of AE EMR and gas indicators and conventional indicators is 24 hours.The 24 hours average value of AE EMR and gas is compared with the conventional index.It is found that the AE EMR and gas index has a good correspondence with the change of conventional index and coal thickness,The reliability of AE EMR and gas monitoring and early warning was descripted laterally.(2)There are mainly two types of AE and EMR interference signals:one is a"burr"type and the other is an"n"type interference signal.The optimal time scale for the singular value decomposition of AE and EMR signals is 8 hours.The singular value decomposition noise reduction method can remove the interference signal from the original signal effectively,and responds better to the changing trend of the useful signal.Compared with the five-point average method,the noise reduction effect is better.(3)Before the outbreak of danger occurs,the AE EMR and gas signals will have an increasing or decreasing trend.A linear fitting analysis was performed on the AE EMR and gas filter curve,and the slope K,the correlation coefficient R~2,and the change rate C of the fitted curve were determined as the characteristic information of the AE EMR and gas curve.Based on this,the trend warning method is proposed as a single indicator warning method,and the warning threshold is determined.(4)The weights of AE EMR and gas warnings was determined by the AHP are0.2,0.3,and 0.5;according to the sum of the weights of different situations,the comprehensive warning status of AE EMR and gas was divided into three types of early warning which are green without warning,yellow warning,red danger.(5)The comprehensive early warning method of AE EMR and gas has high reliability,the accuracy rate of danger warning is 90%,the rate of false alarm is 10%,there is no omission,and the accuracy rate of warning against non-dangerous situations is 100%.The research results have certain reference for the research on noise reduction and feature information extraction of AE EMR and gas signals,and have important guiding significance and application value for coal and gas outburst monitoring and early warning. |
PDF Full Text Request