Integrated importance analysis with Markov Bayesian networks

This paper proposed a Markov Bayesian network (MBN) model to facilitate the calculation of integrated importance measure (IIM) for component state in dynamic multi-state systems. First, the concept of IIM is introduced and its computation formula is decomposed into 3 parts to simplify the analysis process. Then, the MBN is proposed on the basis of Bayesian network and state transition diagram to describe the variable state distributions and state transition matrixes comprehensively. Third, the modeling method of MBN based on multi-state fault tree analysis (MFTA) is discussed to initialize the practical MBN. Finally, according to the MFTA of a power system, a case study is implemented to demonstrate the MBN modelling and IIM calculation process. The MBN of power system proves that the proposed modelling method could build the model from the MFTA equivalently. The IIM calculation process shows that MBN provide a convenient analysis framework. The IIM value of each component state also verifies its meaning and worth in system maintenance.