Study Of Gas Emission Prediction Based On LLE-WPA-BP Algorithm

China is rich in natural coal resources.According to the accident statistics released by the state,the number of gas accidents is still in the forefront of the total number of coal mine accidents.In order to ensure the safety of underground work,it is necessary to carry out gas extraction in coal seams with high gas content.The gas drainage work is determined according to the gas emission of the working face in the coal mining area.Therefore,it is necessary to accurately predict the amount of gas emission in order to improve the effect of gas extraction and ensure the safety of mine operation.BP neural network is selected as the prediction model used in this thesis.Aiming at the problems that BP neural network is easy to fall into local minimum,slow convergence speed and random generation of weight and threshold,the LLE algorithm is used to reduce the dimension of the original input data.In the process of LLE algorithm parameter setting,six rounds of comparative experiments are carried out to obtain the best parameter value.Eight functions are used to test the optimization performance of WPA algorithm,PSO algorithm,FSA algorithm,and WPA algorithm is selected to optimize the parameter value of BP neural network.Three algorithms are substituted into Matlab software to construct LLE-WPA-BP prediction model.The model index system is established by analyzing the gas emission law,influencing factors and the standard of index selection.According to the index system,the monitoring data of Mabao Coal Mine is selected and substituted into LLE-WPA-BP prediction model for data analysis.It is also substituted into BP model,GA-BP model and WPA-BP model for comparative experiment.The experimental results of numerical simulation show that the fitting degree between the predicted value of LLE-WPA-BP model and the real value of gas emission from Mabao Coal Mine is the best,and the average absolute error,average relative error and average relative change value are the lowest,which are 0.0695m~3/t,0.0093 and 0.0193 respectively.It can more effectively predict the mine gas emission accurately and has a wider application range.The study of gas emission prediction based on LLE-WPA-BP algorithm has stronger prediction performance than some prediction models,and can better deal with the prediction of gas emission under complex conditions.So as to reduce the frequency of gas accidents and improve the safety of coal mine operation.The thesis has 29 figures,21 tables and 73 references.