Fast contingency selection through a pattern analysis approach

Abstract This paper presents a method for automatic contingency selection and static security evaluation of electrical power systems. The method employs multi-layer perceptron neural networks whose inputs are power flows and injections, while the outputs compute performance indexes associated with post-contingency scenarios. Contingency ranking and selection are performed based on the artificial neural networks responses. Classifications of system operating state with respect to static security are also provided. The performance of the method is evaluated for different operating conditions using the IEEE 24-bus test system.