Efficient ANN method for post-contingency status evaluation

A method based on radial basis function neural network (RBFNN) architecture is proposed for fast and accurate post-contingent information evaluation, contingency screening and ranking. Estimation of bus voltage magnitude is desired for voltage based contingency analysis whereas estimation of MW, MVA and Mvar line flows are required for power flow based contingency analysis. However, knowledge of voltage magnitudes and angles of all system buses are sufficient to determine the above quantities. Therefore, two neural networks; one for voltage magnitude and other for voltage angle estimation corresponding to normal as well as each contingent condition are proposed in this paper. These estimates are further used to compute two types of performance indices (PIs) for contingency screening and ranking. These PIs are comparable with those obtained by power flow analysis. The effectiveness of proposed method is demonstrated on two IEEE test power systems. Since accurate bus voltage magnitude and angle predictions are achieved quickly by the proposed method, the developed neural networks provide valuable information to operators in real-time operation.