Identification of catastrophic failures in power systems using pattern recognition

This paper presents a new approach for online identification of sequences of events (worms) leading to catastrophic failures using pattern recognition method. Chains of events leading to catastrophic failures are recognized using risk indices. Risk index condenses both probability of occurrences of any contingency and its consequences i.e. severity in terms of load loss, overloads, voltage violations etc. Worms for different operating conditions (i.e. for different loading conditions and topologies) are identified offline and stored in the database. Probable worms for any new operating condition are recognized from the stored knowledge of similar operating conditions using pattern recognition scheme