Bayesian Network for Composite Power Systems Using Hybrid Mutual Information Measure

Abstract

The development of effective reliability based evaluation approaches for assessment of power system component importance is very crucial in the planning and operational decision-making processes of power systems. Bayesian network (BN) is one of the most powerful tools that have been used for this purpose. Generally, a BN may be constructed based on expert beliefs, casual effect, or learning methods. In this chapter, as a contribution to the previous literature, a new learning-based hybrid mutual information-oriented measure is developed for constructing the BN model for a composite power system (CPS) with emphasis on the involvement of the transmission components. In the previous literature, because of the lower failure probability of transmission components compared to generating units, transmission components have not been accurately involved in the BN model. The presented approach is implemented on IEEE 24-bus reliability test system. The analysis shows that the constructed BN of the case study based on the proposed hybrid mutual information measure provides the importance evaluation of transmission system components more precisely.