Abstract:
This article presents a novel approach for detecting high-impedance faults (HIFs) in active distribution networks that operate in resonant grounding mode. The method leve...View moreMetadata
Abstract:
This article presents a novel approach for detecting high-impedance faults (HIFs) in active distribution networks that operate in resonant grounding mode. The method leverages the impedance characteristics of the network. It builds on previous research showing that healthy feeder currents have lower amplitude and opposite phases compared to faulty feeder currents only within a characteristic frequency band (CFB). The proposed method utilizes empirical wavelet transform (EWT) with comprehensive voting theory. The CFB is first determined based on the variation of zero-sequence impedance, and EWT is then applied to extract transient current features within the CFB accurately. Three complementary criteria are then constructed to measure transient currents’ amplitude, polarity, and waveform variation trend: feature energy ratio, feature integral area, and feature cross-correlation coefficient. Finally, a comprehensive voting mechanism is employed, following the “minority obeying majority” principle to achieve fault detection. This method has been tested in active distribution networks with distributed generators (DGs) considering various conditions. Results show that the process is highly adaptable and effective even under significant noise interference (up to 3 dB). In addition, the method’s effectiveness has been validated through testing in a 10-kV real-model experiment and on-site recording data.
Published in: IEEE Transactions on Instrumentation and Measurement ( Volume: 73)