A Novel Alarm Processing and Fault Diagnosis Expert System Based on BNF Rules

Abstract

In this paper a novel power system alarm processing and fault diagnosis expert system (AFDES) is presented. For the application of expert system (ES), there is a bottle-neck problem to obtain a maturing rule-base. Backus-Naur Form (BNF) is used to design a kind of expert rule frame which operator can write and add the rules with his own defining language to rule-base in ES. By this way a maturing rule-base can be set up gradually in the ES of power system. After brief introduction of using the BNF to construct the rule-base of fault diagnosis, the system configuration and alarm processing and fault diagnosis are described. The tactics of alarm messages pretreatment also introduced to accelerate the efficiency of reasoning engine, in which the alarm message filtration and classification according priority and synthesis level. At last, a case had been studied and analyzed to show how to forecast the key following fault which may happen right after the former trip according the occurred alarm messages for precaution