Composite System Adequacy Assessment Using Monte Carlo Simulation and Logistic Regression Classifier

Abstract

This paper presents a new method that combines Logistic Regression Classifier (LRC) and Monte Carlo Simulation (MCS) to evaluate the adequacy of a composite power system. LRC is used to pre-classify the system states as failure or success based on training data set provided by conventional MCS itself, but with a relaxed error tolerance level. The proposed method is applied to the IEEE Reliability test system (IEEE-RTS-79) to calculate the annualized and annual indices.The results thus obtained are compared with that of conventional MCS. In different cases, the simulation results provide a significant improvement in computational burden and indices calculation time while maintaining resonable accuracy.