Composite Power System Reliability Modeling, Evaluation and Reliability-Based Analysis by Bayesian Networks

Bayesian Networks (BNs) as a strong framework for handling probabilistic events have been successfully applied in a variety of real-world problems, but they have received little attention in the area of composite power systems reliability assessment. Reliability assessment by BN provides some additional capabilities in comparison to conventional methods, both at the modeling and at the analysis levels. At the modeling level, several restrictive assumptions, implicit in the conventional methods, can be removed. At the analysis level, a variety of applicable reliability-based analysis which is hardly achievable in conventional methods, can be conveniently performed. This paper proposes a methodology based on Minimal Cutsets (MCs) to apply BNs to composite power system reliability modeling, reliability assessment and reliability-based analysis.To have a more accurate BN model, a new method of MC determination for composite power system is proposed. Bayesian structure is extracted, based on the determined MCs. Bayesian parameters are defined based on the logical relationships of nodes. To make the proposed method applicable to large composite power systems, virtual nodes are proposed and combined with Bayesian model. Also, a variety of reliability-based analyses are presented which are hardly achievable in conventional methods. The proposed method is validated by applying to RBTS and comparing the results with other reliability analysis methods. The proposed methodology is applied to the Reliability Test System (RTS), to show its feasibility in large networks.