A COMPREHENSIVE FAULT DIAGNOSTIC SYSTEM USING ARTIFICIAL-INTELLIGENCEFOR SUB-TRANSMISSION AND URBAN DISTRIBUTION NETWORKS

This paper describes an intelligent diagnostic system for an interconn ected distribution network developed to assist the system operator wit h fault identification during restoration. The intelligent process uti lizes only those data available in a standard SCADA system such as the post fault network status, the list of the tripped breakers, main pro tection alarm, and the conventional event log. The fault diagnostic sy stem is implemented by three independent mechanisms, namely the generi c core rule, the generic relay setting inference and the specific post -fault network matching and learning. The generic core rule generates various possible fault locations and the generic relay inference exami nes whether each possible fault location is logical and valid. The spe cific network matching compares whether the post fault network and the related tripped breakers are identical to a previous fault event. Tes t results obtained from two distribution networks confirm that the dev eloped system is practical, reliable and accurate.