Application of EWT and PSO-SVM in Fault Diagnosis of HV Circuit Breakers

In order to improve the recognition rate of mechanical vibration signals of high voltage circuit breakers, a feasible new fault diagnosis method is proposed in this paper. Firstly, the empirical wavelet transform (EWT) is adopted to decompose the original multi-component signals into a series of intrinsic mode functions (IMF). Secondly, the envelop energy entropies of these IMF components are calculated as signal features. Finally, establishing the optimal support vector machine (SVM) classifier by particle swarm optimization (PSO) method. Using this EWT-PSO-SVM model to identify the unknown samples, the results show that the EWT method can effectively reduce modal aliasing problem, and the recognition rate of EWT-PSO-SVM model is higher than EMD-PSO-SVM model, these results verify the feasibility and superiority of the proposed EWT-PSO-SVM fault diagnosis method.