Study On Gas Concentration Prediction In Fully Mechanized Face Based On Information Fusion

In coal mine disasters,gas disasters are often accompanied by serious accidents.Th,erefore,accurate and reliable gas concentration prediction can provide effective control measures and decision-making basis for mine gas prevention and control work.In this paper,the monitoring data of gas concentration in the monitoring system of Chenjiashan Coal Mine for five consecutive days is analyzed.Combined with the simulation of the gas concentration field on the working face,the gas concentration distribution interval prediction of the coal mining machine area in the fully mechanized m,ining face is formed.The main research work of this paper as follows:The traditional time series prediction method and machine learning algorithm are used to analyze the gas concentration data.The mathematical characteristics of the mine gas concentration monitoring data are obtained by statistical methods,and the original gas concentration sequence is preprocessed to make the input data meet the input requirements of the algorithm.The autoregressive moving average(ARMA),cubic exponential smoothing(Holt-Winters)and support vector machine regression(SVR)algorithms were used to predict the gas concentration values at 8 time poiints every 10 min.The results show that all three algorithms can reflect the change of gas concentration.The value has a good fitting effect on the original data and is compared with the measured values of the mine.The mean absolute error are:0.013%,0.024%and 0.008%,respectively,and the three algorithms are combined to obtain the final combined model.The results show that the combined effect of SVR-ARM,A is better in the SVR-ARMA and SVR-Holt-Winters combination models.The comparison with the measured values shows that this combination can provide some technical support for mine gas prediction.By establishing a two-dimensional geometric model of the fully mechanized mining face,the distribution law of wind flow field and gas concentration field is obtained.By establishing a two-dimensional geometric model of the fully mechanized mining face,the distribution law of wind flow field and gas concentration field is obtained.The gas concentration field is corrected by using the predicted gas concentration of the upper corner,and 8 time points are obtained every 10 min.The gas concentration distribution interval of the coal machine area.By integrating the gas concentration information obtained by data mining and numerical simulation,the gas concentration distribution interval in some areas of the fully mechanized mining face can be predicted,which can provide a certain decision basis for mine gas prediction and early warning work.