High-impedance fault identification using a fuzzy reasoning system

A methodology is presented to detect high-impedance faults in radial distribution feeders by means of fuzzy logic reasoning. The proposed technique is based on the analysis of the feeder responses to impulse waves which are periodically injected at the feeder inlet. These responses are compared to standard responses which were previously stored in a database. Standard responses correspond to responses of the feeder operating in normal configurations. A supervisory system processes the information in the database and, using a fuzzy inference machine, indicates possible occurrences of abnormalities. The proposed system has been tested in a real feeder; the relevant results are presented.