A new approach to fault diagnosis in electrical distribution networks using a genetic algorithm

Abstract

In this paper, a new approach to fault diagnosis in electrical distribution network is proposed. The approach is based upon the parsimonious set covering theory and a genetic algorithm. First, based on the causality relationship among section fault, protective relay action and circuit breaker trip, the expected states of protective relays and circuit breakers are expressed in a strict mathematical manner. Secondly, the well developed parsimonious set covering theory is applied to the fault diagnosis problem. A 0-1 integer programming model is then proposed. Thirdly, a powerful genetic algorithm (GA) based method for the fault diagnosis problem is developed by using information on operations of protective relays and circuit breakers. The developed method can deal with any complicated faults, and simultaneously determine faulty sections and any hidden defects in the feeder protection systems. Test results for a sample electrical distribution network have shown that the developed mathematical model for the fault diagnosis problem is correct, and the adopted GA based method is efficient. © 1997 Elsevier Science Limited.