Identifying internal fault from magnetizing conditions in power transformer using the cascaded implementation of wavelet transform and empirical mode decomposition

摘要

In this paper, by decomposing high-frequency band of differential current using wavelet transform and giving the extracted features to the empirical mode decomposition method, internal faults from magnetization conditions in power transformers are discriminated. The merit of the suggested method is validated by various simulation results using PSCAD and practical data set gathered from a laboratory 3-phase transformer. Internal faults could be precisely discriminated from magnetizing current in half cycle of the fundamental frequency; this is one advantage of the method. Also, an innovative technique is suggested to tackle current transformer saturation negative impacts on differential relays. At end, the method is assessed and compared with some other well-known discrimination methods by plenty of simulations and experiments.