A Comparison of Load Models for Composite Reliability Evaluation by Nonsequential Monte Carlo Simulation

Abstract

This paper presents a comparison of Markov load models for composite reliability evaluation by nonsequential Monte Carlo simulation. The proposed models represent the whole system load curve. The first model (M1) is an aggregated Markov model that represents all different states present in the load curve, without using any clustering technique. The second model (M2) consists of a hybrid Markov model, where all different levels of the load curve are also represented but it tries to preserve some chronology of the load curve. The third model (M3) consists of a non-aggregated Markov model. The frequency and duration (F&D) indices are calculated by the conditional probability method for all models. The indices calculated using these models are compared with the indices obtained when the usual clustered aggregated Markov model (M0) is used. The indices obtained by sequential Monte Carlo simulation with a chronological system load curve are used as comparison reference in order to validate the presented models.