Data-based fault diagnosis of traction converter and simulation study

A data-based fault diagnosis method is applied to fault diagnosis of traction converter in this paper. The wavelet transform is used to extract fault characteristics and support vector machine (SVM) is used to diagnose faults. The pros and cons of SVM and radial basis neural network (RBF NN) in fault model classification are also compared follow behind. The simulation results show that, SVM has a good reliability and better generalization capability than RBF NN for fault diagnosis, which verify the superiority of SVM.