Fuzzy rule-based expert system for power system fault diagnosis

Abstract

The paper demonstrates a novel component oriented fuzzy expert system (COFES) developed in PROLOG for power system fault diagnosis. This 'expert system' assesses faults on power systems using intelligent techniques that can take account of bad/missed SCADA data. Incorrect operation of protective relays and/or circuit breakers during single as well as multiple faults and corresponding uncertain incoming information render proper fault diagnosis a very involved task. To handle these uncertainties and rank various fault hypotheses a fuzzy signal model based on fuzzy information theory has been developed. The model measures degree of correctness of received and nonreceived input data. The proposed method incorporates fuzzy symbol classification through an enhanced knowledge-base which includes network model, predefined subnetworks, relaying schemes and fuzzy diagnostic rules. This expert system has been applied to a sample power system. The results obtained along with their evaluations are completely reported.