On-line prediction of closest loadability margins using neural networks

Determining loadability margins to various security limits is of great importance for the secure operation of a power system. In this paper, a novel approach based on neural networks is proposed for online prediction of the closest loadability margin. For each operating point, the closest loadability margin is calculated using a pair of multiple power flow solutions. Radial Basis Function Network (RBFN) is trained with obtained closest loadability margins for different operating points in normal and contingency conditions. As an important result, online contingency ranking is obtained in addition to prediction of the closest loadability margin for current operating point. A clustering algorithm is used to speedup the RBFN training process. The simulation results for the IEEE 14-bus test system demonstrate the effectiveness of the proposed method.