Reliable High Impedance Fault Detection with Experimental Investigation in Distribution Systems

An approach for high-impedance fault detection is introduced in this paper. This technique discriminates between high-impedance faults and switching conditions by utilizing changes in the third harmonic current/voltage magnitude in conjunction with the conventional wavelet algorithm. The concept of discrimination is based on the observation that switching conditions typically do not involve changes in the zero-sequence third harmonic magnitude. However, in the case of high-impedance arcing faults, a noticeable change in the third harmonic current or voltage magnitude occurs. The performance of the proposed technique is examined through a detailed simulation of an actual medium voltage radial distribution feeder. In this simulation, high-impedance faults are represented by an arc model. The simulation, conducted in MATLAB for different fault cases, reveals that all fault cases are detected using the proposed technique. Furthermore, the experimental validation of the reliability of the proposed technique is accomplished using the same actual distribution feeder.