| Spontaneous combustion of goaf coal is one of the serious disasters and accidents that pose a threat to the safety of coal mines.It occurs frequently and poses a serious threat,resulting in a large number of casualties and property losses.The prevention and control of coal spontaneous combustion in goaf is an important work content for coal mine safety production.The monitoring and prediction of coal spontaneous combustion in goaf is the basis for the prevention and control measures of coal spontaneous combustion in goaf,and is of great significance for the safety production of mines.At present,research on predicting a single indicator of coal spontaneous combustion is relatively complete.However,the collection of data on coal spontaneous combustion indicators in goaf areas is prone to errors caused by equipment,personnel,production conditions,etc.It is difficult to eliminate errors solely by predicting a single indicator time series.On the basis of studying the mechanism of coal spontaneous combustion and gas generation law in goaf,this paper takes the 6201 working face of Dongpang Mine as the research background,and uses methods such as programmed temperature rise experiment and LSTM time series prediction to study the gas generation law in the low-temperature oxidation and spontaneous combustion process of coal.The critical temperature of coal spontaneous combustion is analyzed,and a mathematical model for predicting the gas concentration and temperature of multiple indicators of coal spontaneous combustion is established.The main research content and achievements are as follows:(1)A prediction gas index system for coal spontaneous combustion has been established.Based on the results of programmed heating experiments,the gas concentration indicators for coal spontaneous combustion are optimized.Carbon monoxide,carbon dioxide,methane,ethane,ethylene,acetylene,chain to alkane ratio(propane/ethane),olefin to alkane ratio(ethylene/ethane),and carbon dioxide/carbon monoxide are selected as the multiple indicators for coal spontaneous combustion.The temperature nodes corresponding to gas indicators at different stages have been clarified.(2)A LSTM time series prediction model with optimized parameters was established.Sparrow search algorithm and Tent chaotic sequence were used to optimize LSTM time series prediction,obtaining LSTM prediction parameters and reducing prediction errors caused by manual parameter selection.The experiment shows that the root mean square error of ARIMA model,LSTM model and optimized SSA-LSTM model is 0.092,0.083 and 0.061 respectively,the absolute error is 0.074,0.065 and 0.049 respectively,and the absolute percentage error is 33.99%,33.24%and 17.51% respectively.The prediction results of optimized SSA-LSTM model have higher accuracy.(3)A mathematical model for predicting the temperature of coal spontaneous combustion gas with multiple indicators was established.The optimal gas indicators during the process of coal spontaneous combustion were selected,and the indicators were sorted according to the size of prediction error to obtain the prediction advantage indicators.The predicted control temperature was obtained by comparing the experimental data of program heating.(4)Based on the prediction results of LSTM and sparrow search algorithm,the dominant indicator for predicting the 6201 working face of Dongpang Mine was determined to be carbon monoxide,and the predicted control temperature was between 30℃ and 40℃.Based on measured data,no signs of coal spontaneous combustion were predicted within a 10 day range.There are 22 figures,14 tables and 77 references in this paper. |
