Radial basis function networks for fast contingency ranking

Abstract

This paper presents an artificial neural network-based approach for static-security assessment. The proposed approach uses radial basis function (RBF) networks to predict the system severity level following a given list of contingencies. The RBF networks are trained off-line to capture the nonlinear relationship between the pre-contingency line flows and the post-contingency severity index. A method based on mutual information is proposed for selecting the input features of the networks. Mutual information has the advantage of measuring the general relationship between the independent variables and the dependent variable as against the linear relationship measured by the correlation-based methods. The performance of the proposed approach is demonstrated through contingency ranking in IEEE 30-bus test system.