A new wavelet-based method for detection of high impedance faults

Detecting high impedance faults is one of the challenging issues for electrical engineers. Over-current relays can only detect some of the high impedance faults. Distance relays are unable to detect faults with impedance over 100 Omega. In this paper, by using an accurate model for high impedance faults, a new wavelet-based method is presented. The proposed method, which employs a 3 level neural network system, can successfully differentiate high impedance faults from other transients. The paper also thoroughly analyzes the effect of choice of mother wavelet on the detection performance. Simulation results which are carried out using PSCAD/EMTDC software are summarized.