Modeling and detection of high impedance faults

High Impedance Faults (HIF) on distribution feeders in a power system are abnormal electrical conditions that cannot be detected and cleared by conventional protection schemes due to its low fault current. In this paper, a detection scheme utilizing the distortion around zero-crossing as a distinguishing feature of HIF is proposed. Documented data from an 18-day program of ignition testing undertaken by HRL at the TCA High Energy Facility (NSW, Australia) has been analysed to determine the distinctive features of HIF. These key features are summarised as arcing phenomenon, randomness and intermittency, and harmonics and distortion. Using these features, a model has been developed and applied in ATP/EMTP (Alternative Transients Program / Electro-Magnetic Transients Program). Discrete Wavelet Transform (DWT) has been used to extract these identifiable characteristics, using the Modulus Maxima (MM) of wavelet transform to detect the distribution of distortion and discriminate HIF from other faults in the network. The use of this method has performed well within a simulated environment and with field test data and is worthy of further investigation.