Fault discrimination through expert systems

This thesis presents a Knowledge Based Expert System (KBES) for faultdiagnosis in a power grid segment. This off-line system used rule_based systemtechnique. The developed ES identifies the most probable faulted section basedon the depicted performance indices; the calculation is based on the relayclassification categories. Full detailed explanations of protective relays andbreakers are obtained. Finally, the test results on a typical 220/66 kV substationprove the viability of the proposed system. Exactly correct diagnosis is obtainedeven for the fault conditions involving multiple faults and failure- operation ofprotective devices. Moreover, the fault on the dead zone ( zone between breakerand current transformer ) is edited and expected faulty length of faultytransmission line based on a special rules; are of its distinct features ..The system can be easily transferred from system to other with somemodifications. The proposed ES has been implemented using PROLOG facilities.