Prognosis of Catastrophic Failures in Electric Power Systems

This paper describes a methodology for identification of most probable cascading sequence of events which may lead to catastrophic failures in power systems. Failure sequences are identified based on the selection of contingencies which are severe as well as most probable. These sequences are identified and stored for different possible loading conditions. Artificial intelligence is used to predict the most probable catastrophic failure events for any new loading condition from the stored knowledge base.