High Impedance Arc Fault Detection Based on the Harmonic Randomness and Waveform Distortion in the Distribution System

High impedance arc faults (HIAFs) happening in the medium-voltage distribution system may result in damages to devices and human security. However, great difficulties exist in identifying these faults due to the much weaker features and the varieties when grounded with different surfaces. This paper presents an integrated algorithm to detect the HIAFs with high-resolution waveform data provided by distribution-level PMUs deployed in the system. An improved arc model is proposed, which can continuously imitate the randomness and intermittence during the "unstable arcing period" of arc faults. The integrated algorithm consists of two branches. First, the variations of HIAFs during unstable arcing period are identified with the unified harmonic energy and global randomness index, which can unify the scales of harmonic content in different fault situations and enlarge the disparities from non-fault conditions. Then, the waveform distortions of HIAFs during the stable arcing period are identified with discrete wavelet transform to extract the detailed distribution characteristics. The reliability and security of the proposed algorithm are verified with numerical simulations and field tests in a 10-kV distribution system.