A computer assisted intelligent storm outage evaluator for power distribution systems

A knowledge-based approach for evaluating storm-related outages in distribution systems is proposed. At the outset, binary voltage sensors capable of transmitting the real-time voltage on/off symptoms are recommended to be installed at strategic locations in the distribution system. A semantic network model has been developed to represent the knowledge of the distribution system faults and their associated symptoms. The real-time voltage symptoms, transmitted by the sensors, serve as the basic input to the semantic network model in generating fault hypotheses. A rule base, incorporating temporal and topological knowledge of symptoms and system operators´ heuristics, is used by the inference mechanism to determine the actual faults from the hypotheses set. The results suggest that the fault types and their exact sections can be diagnosed accurately