1.College of Safety Science and Engineering,Xi’an University of Science and Technology,Xi’an,China;2.National &Local United Engineering Research Center of Green Safety Efficient Mining,Xi’an ,China;3.Institute of Modern Coal Mining Technology,Shaanxi Coal Chemical Industry Technology Research Institute Company Limited,Xi’an ;4.College of Energy Engineering,Xi’an University of Science and Technology,Xi’an ,China;5.Key Laboratory of Western Mine Exploitation and Hazard Prevention Ministry of Education,Xi’an University of Science and Technology,Xi’an ,China;6.Safety Branch,China Coal Research Institute,Beijing ,China;7.State Key Laboratory of Coal Mining and Clean Utilization,Beijing ,China

Mine gas concentration monitoring is one of the most direct and effective means of prevention and control of gas accidents,in order to improve the utilization efficiency of monitoring information,a real-time gas concentration prediction model based on Lagrange-ARIMA is proposed.Firstly,PauTa criterion is applied to construct missing values of gas concentration monitoring.Secondly,sliding Lagrange interpolation method is used to predict missing values.Finally,based on ARIMA sequential learning and L1 norm minimization principle,Lagrange-ARIMA sequential learning window is determined to realize real-time prediction of gas concentration values.The simulation results show that the average error of Lagrange-ARIMA real-time prediction model is 1.397% when dealing with the missing value of gas concentration time series.When the width of sequential learning window is 85,the mean absolute error of predicted gas concentration series is 0.011 8.Compared with the traditional ARIMA static learning model,the width of learning window of Lagrange-ARIMA model is reduced by 90.3%,and the modeling complexity is significantly reduced.The mean absolute error is reduced by 16.3%,and the prediction accuracy can meet requirements in the field.