Prediction And Study Of Coal Mine Gas Monitoring Based On PSO-SVM

Although the coal mine accidents has declined recently, there are still some gaps compared with foreign countries and gas coal mine disaster still haunt our system, especially gas explosion accidents which highlight a serious threat to the safety of mine workers as well as underground equipment.This paper systematically analyzes the influence of various factors on the emission of gas emission including its volume and size, which are several most important factors, compares different forecasting methods, and chooses support vector machine prediction method for gas emission size modeling study. Learning ability and generalization ability of SVM is determined by its parameter selection,parameter selected in a number of parameters in the blind search takes too much time, and not necessarily for the best, so particle swarm algorithm is used to achieve the SVM parameter optimization. PSO algorithm is simple, and highly efficient, and of advantage to the search algorithm, the algorithm does not require the optimization function continuously differentiable, the various methods of universal strong field. This algorithm is better in solving multivariable, nonlinear problems.An optimization support vector machine prediction method of gas emission based on particle swarm algorithm is proposed in this paper, the method using the particle swarm algorithm optimizes the parameters which influence support vector machine to forecast the simulation. A real simulated coal gas collection system is build with sensor、MCU and PC, and the algorithm was applied to the system of gas concentration prediction. Simulation and experimental test results show that the particle swarm algorithm can efficiently find the optimal values of parameters, SVM model system established with these optimal parameters is effective in predicting the gas emission volume, the method could be used in mine gas monitoring and can provide certain theoretical reference in mine gas prediction.