Research On Prediction Of Gas Concentration Based On Support Vector Machine And Immune Genetic BP Neural Network

Coal accounts for the primary enery consumption in China,and is the main energy source in China,having an important strategic position in our country,so the safety of production is a major problem in coal mine.However,the coal geological structure is complex,and the coal seam gas content is large,so coal mine safety accident rate is much higher than the world's major coal producing countries.Among coal mine safety accident,the highest frequency and the most damage is gas disaster accident.Therefore,the prediction of gas concentration is of great significance to the safety production of coal mine and the personal safety of workers.In this paper,aiming at the coal mine safety production,based on the existing gas monitoring technology,combined with the actual coal mine,the gas data denoising algorithm based on support vector machine,and the prediction algorithm of gas concentration data based on immune genetic BP neural network are put forward,to study on the denoising and prediction of gas concentration data collected in underground mine.This paper mainly includes the following aspects:First of all,by analysing the characteristics of underground gas data in the actual production of coal mine,it is concluded that the influence of underground complex environment is generally noise.so the gas data denoising method is proposed based on support vector machine,and is used in the processing the collected gas data.The effectiveness of the proposed algorithm is verified by simulation experiments on gas data.Secondly,aiming at the problem of the lack of gas concentration prediction in advance,the immune genetic BP neural network algorithm for the prediction of underground gas concentration is put forward.In order to scientifically determine the structure of BP neural network,according to the characteristics of gas concentration data,a solution based on phase space reconstruction theory is proposed.By finding the best embedding dimension m,the structure of BP neural network is determined.In order to make up for the BP neural network algorithm for network training slow and easy to fall into the convergence,An immune genetic algorithm based optimization algorithm is proposed.The weights and thresholds of BP neural network are treated as the problem to be solved(the antigen),and produce initial antibody population.By introducing immune genetic mechanism,improve the efficiency of the algorithm and overcome the defects of BP neural network easily trapped into local extremum.By using the proposed algorithm on the prediction of coal gas concentration data,the effectiveness of the proposed algorithm is verified.Finally,the prediction algorithm is applied to the coal mining face gas acquisition system and our University's coal mine gas monitoring system.The practicability of the algorithm is verified.In addition,study the underground personnel positioning and wireless communication system to meet the needs of coal production.