Vulnerability Assessment of a Large Sized Power System Using Radial Basis Function Neural Network

Vulnerability assessment of a power system has been of great concern due to the continual blackouts in recent years which indicate that a power system today is too vulnerable to withstand an unforeseen catastrophic contingency. This paper presents a new approach to assess vulnerability of a power system based on radial basis function neural network. A new feature extraction method named as the neural network weight extraction is also proposed to reduce the number of input features to the neural network. The effectiveness of the proposed approach has been demonstrated on a large sized IEEE 300-bus system. Test results prove that the radial basis function neural network accurately predicts the vulnerability of the power system.