The Research Based On The Coal Mine Gas Outburst Prediction System Of Support Vector Machine(SVM)

Coal mine safety monitoring system is a way to automatically collect and process data and corresponding control system. In the most central place in modern coal production operations. Thus, real-time monitoring data to master, timely and accurate data to predict failures, improve coal mine safety monitoring capability to become an indispensable aspect of coal mine safety, providing a powerful guarantee stable production.Monitoring data for mine is the use of real-time data processing for reliable forecasts. Gas Emission predict at this stage in the most common network BP neural network, it will be the basis of the theoretical prediction of gas and neural network analysis technology organically combined with the theoretical basis of modern research methods, some of the gas can be predicted, but in the data processing deficiencies still exist some learning phenomena. To address this issue this paper wavelet analysis and immune support vector machine gas data prediction algorithm, based on a combination of hardware part STM32coal mine monitoring data sampling to improve the accuracy of data sampling, gas prediction accuracy problems, enhance the security of underground operations. See the extensive literature on the basis of thesis research work as follows:First, this paper presents a based on immune support vector machine (SVM) of coal mine gas outburst prediction model. Because when the gas outburst can be collected by the feature vector is limited, this for needs a large number of training samples of BP network, the training of the network has some limitations, such as local minimum point and learning problems. Therefore, the paper introduced the support vector machine (SVM) based on small sample learning. In order to further improve the prediction accuracy, the use of wavelet analysis of sampling data, refactoring, denoising, optimize the sampling data. Antibodies the immune evolutionary algorithm is applied to the principle of maximum matching of antigen to seek the optimal solution of parameters c and g, avoid the blindness of selecting optimization model. By comparing the simulation results of general SVM model, it will be based on the immune support vector chance for gas outburst prediction, greatly improving the prediction system of promotion ability, shortened the time prediction, further reflects the application value of the prediction system.Second, as a result of current prediction methods used in numerous parameters, some parameter acquisition is difficult, use is very inconvenient. To solve this problem, this article adopts the STM32of good cost performance and high stability as the core processor to realize collection, with the mine gas, temperature sensor, wireless radio frequency communication, sampling of safety monitoring data in real time and high precision, improve the reliability of data processing.Last, since the parameters used in the current prediction methods of many, some of the parameters collection difficulties,the use is inconvenient.In order to solve this problem,this paper uses the high performance price ratio and good stability properties of high STM32as the core processor to achieve the acquisition,with the underground gas, temperature sensor, wireless radio frequency communication,the sampling of real-time and high precision on safety monitoring data to improve the reliability of data processing.