Study Of Prediction Of The Friction Resistance Coefficient Of Mine Ventilation Based On Support Vector Machine

The frictional resistance coefficient of mine ventilation is one of the most important technical parameters that reflects the resistance the mine shaft and.It is the base of technology which includes designing mine ventilation,reforming ventilation system and strengthening the technology management.Usually,the coefficient is mainly gotten through looking up table and measuring on the spot.On the one hand,the frictional resistance coefficient table used in China is made in the 1980 s,which can not meet the use of current complex mine ventilation;On the other hand,the measurement is more complicated on the spot and it needs too much work.Therefore,the prediction of the friction coefficient of mine roadway brings great theoretical significance and great value in practical application for designing mine ventilation and managing mine ventilation safety.In this paper,it uses support vector machine(SVM)to predict the friction resistance coefficient of mine ventilation.After theoretical analysis,it is proved to be feasible to use the support vector machine(SVM)to predict frictional resistance coefficient of mine ventilation.By analyzing the nature of the kernel function of support vector machine(SVM),it has the advantages over several other kernel function based on the radial basis kernel function.Finally,it adopts radial basis kernel function as support vector machine(SVM)to predict the frictional resistance coefficient of mine ventilation.The research object is the coefficient of frictional resistance of bolting and shotcrete roadway,bolt-supported roadways and the anchor net supporting of roadway in cold water coal mine,which uses LIBSVM in MATLAB environment.The grid search method,genetic algorithm and particle swarm optimization(PSO)method were carried out in the research.It then gets the optimal radial basis kernel function parameters and punish combination.By using selected parameter combination,it sets up forecast model to forecast the validation set.Fromthe prediction results,we can see that the forecasting models,which are set up by combining the parameters of the three methods,can all effectively predict the frictional resistance coefficient of coal mine.It also proves that it is feasible to predict the frictional resistance coefficient of coal mine by using the support vector machine(SVM).Among them,the comprehensive performance of the prediction results of the models is the best,which adopts combination of genetic algorithm(GA)to select parameters.In the research,it adopts two kinds of ways to predict the frictional resistance coefficient of coal mine.One is support vector machine(SVM),and the other one is commonly used method-BP neural network.it respectively predict frictional resistance coefficient of coal mine in cool water coal mine.the result shows that the frictional resistance coefficient is more precised by adopting the support vector machine(SVM),and it has more superiority compared with the BP neural network to predict mine l frictional resistance coefficient of coal mine.