Application Research Of Time Series Analysis Based On R Language In Gas Concentration Prediction

Gas accidents in coal enterprises not only has a serious impact on normal production operations but may even lead to casualties.At present,every coal mine is equipped with a mine safety monitoring and control system for real-time monitoring information such as gas concentration.Statistical analysis of the large amount of data obtained by this monitoring platform can provide a basis for decision-making for preventing and controlling gas overruns.In R language environment,this paper studies the related theoretical knowledge and construction process of ARIMA model and TAR model.Based on the measured data,the modeling analysis is carried out,and the prediction effects of the two models are analyzed and compared.The main contents are presented as follows:Based on R language,the theoretical knowledge and model construction methods of time series analysis are analyzed.Taking the data generated by the working face as an example,the ARIMA prediction model is constructed by a series of means such as stationarity test,differential analysis,model parameters determination and model checking.The TAR model which is used to predict the gas concentration is constructed by steps of nonlinear test,parameter estimation,fitting test,etc.Moreover,both models show better prediction results.Based on the gas data of the upper corner of the 427 fully mechanized mining face in Chenjiashan Coal Mine,the sample data and test data were selected and applied to the model.The ARIMA model and the TAR model are used to predict the gas concentration time series in the next 24 hours.The predicted values range from 0.14%to 0.23%and 0.127%to 0.261%,which are within the distribution range of the actual value and match the original value basically.By comparing and analyzing the prediction results of the two models,it is found that the TAR model is more convenient than the ARIMA model and the error is smaller.At the same time,the prediction results of the two models are compared to obtain the average of the two models.The average absolute error and the average relative error of the TAR model are smaller,and the prediction effect is better than the ARIMA model.