Distribution System Fault Detection through Feature-Summarizing Matrix: A Case Study

For power distribution systems, a fast and accurate fault detection method is imperative to guarantee power supply equipment and personal safety. To this end, a novel feature-summarizing matrix based fault detection method is proposed in this paper. Taking into account some well-known tricky conditions leading to imperceptive faults, specifically, that are long-transmission, non-linear loading, noises, and high impedance, the proposed method centers on a feature-summarizing matrix which is devised towards the tricky factors, and based on voltages RMS information delivered by Hilbert transform, and harmonics information by discrete Fourier transform, and a special detection method by Hilbert-Huang transform. A typical 10.5kV radial distribution network with three feeders is emulated with the APT software, for demonstrating the performance of the proposed fault detection method in comparison with several classical methods.