Multi Label RBF Classification Method for Composite System Reliability Evaluation

In this paper a new approach for evaluation of composite system reliability of power systems is described. In this method Monte Carlo Simulation (MCS) is combined with Multilabel Radial Basis Function (MLRBF) classifier to compute system reliability indices. MLRBF is a classification technique in which target vector of each instance is assigned into multiple classes. In this study MLRBF is used to characterize states of a complete power system (failure or success) without requiring optimal power flow (OPF) analysis. As a result, the computational burden of the reliability evaluation analysis to perform OPF is eliminated. For illustration, the proposed method is applied to the IEEE Modified Reliability Test System (IEEE-MRTS). The obtained results demonstrate that MLRBF algorithm in reliability evaluation provides good performance in classification accuracy while reducing computation time dramatically.