Research On Fault Diagnosis Of Mine Ventilation System Based On SVM

The purpose of the fault diagnosis of the ventilation system is to determine the location of the fault source and the degree of fault,which is of great significance for improving the safety of the ventilation system.Considering the whole network of the ventilation system,the fault will inevitably cause changes in the air volume of the ventilation system.The root cause is the change in the equivalent wind resistance of the faulty roadway.This article defines it as a resistance change type fault.The fault diagnosis of the ventilation system studied in this paper is to determine the location of the branch of resistance change and its equivalent wind resistance according to the amount of air flow monitored by the roadway.The core issue is the inverse mapping relationship between air flow and wind resistance.This paper uses the SVM method to study the fault diagnosis of mine ventilation system,transforms the fault branch position diagnosis problem into multiple classification problems,and uses the regression method to predict the equivalent wind resistance value of the fault.Using the mine ventilation simulation system MVSS to generate a “wind resistance fault — air volume” relationship sample,with air volume as the input feature,resistance change position and equivalent wind resistance as output,using the SVM method to train and build a fault location classification model and fault branch equivalent wind resistance prediction Regression model,and finally use these two models to diagnose faults in the ventilation system.The results show that the SVM method characterized by the air volume can diagnose and predict the wind resistive fault position and the fault equivalent wind resistance that causes the abnormal change in the air volume of the ventilation system.The diagnostic accuracy of the fault location is related to the length of the input feature vector.The position of the fault equivalent wind resistance prediction result with a relative error less than 5% can reach more than 70%.