Fs. Wen et Cs. Chang, A NEW APPROACH TO FAULT-DIAGNOSIS IN ELECTRICAL DISTRIBUTION NETWORKSUSING A GENETIC ALGORITHM, Artificial intelligence in engineering, 12(1-2), 1998, pp. 69-80
In this paper, a new approach to fault diagnosis in electrical distrib
ution network is proposed. The approach is based upon the parsimonious
set covering theory and a genetic algorithm. First, based on the caus
ality relationship among section fault, protective relay action and ci
rcuit breaker trip, the expected states of protective relays and circu
it breakers are expressed in a strict mathematical manner. Secondly, t
he well developed parsimonious set covering theory is applied to the f
ault diagnosis problem. A 0-1 integer programming model is then propos
ed. Thirdly, a powerful genetic algorithm (GA) based method for the fa
ult diagnosis problem is developed by using information on operations
of protective relays and circuit breakers. The developed method can de
al with any complicated faults, and simultaneously determine faulty se
ctions and any hidden defects in the feeder protection systems. Test r
esults for a sample electrical distribution network have shown that th
e developed mathematical model for the fault diagnosis problem is corr
ect, and the adopted GA based method is efficient. (C) 1997 Elsevier S
cience Limited.