Research on Single-Phase High Impedance Fault Identification of Neutral Ungrounded System

摘要

Most of 6-10kv distribution networks in China are neutral ungrounded systems. High impedance fault (HIF) is a common complex fault in distribution network. With the current is small and unstable, it is difficult to detect. In this paper, a method based on wavelet analysis and support vector machines(SVM) optimized by particle swarm optimization(PSO) is proposed to identify HIF in ungrounded systems. Firstly, a simulation model was established to simulate the high impedance fault. Then, discrete wavelet transform (DWT) was used as the feature extractor to decompose the HIF current signal by three-layer wavelet, and the signal energy of each frequency band was extracted as the characteristic quantity. After normalization, the feature vectors were input into PSO-SVM for training. Finally, the field high impedance fault signal is used to verify. Compared with the traditional method of SVM recognition after Fourier transform extraction, the recognition accuracy and speed are improved.