System Reliability Analysis Method Based on T-S FTA and HE-BN

For high-reliability systems in military, aerospace, and railway fields, the challenges of reliability analysis lie in dealing with unclear failure mechanisms, complex fault relationships, lack of fault data, and uncertainty of fault states. To overcome these problems, this paper proposes a reliability analysis method based on T-S fault tree analysis (T-S FTA) and Hyper-ellipsoidal Bayesian network (HE-BN). The method describes the connection between the various system fault events by T-S fuzzy gates and translates them into a Bayesian network (BN) model. Combining the advantages of T-S fault tree modeling with the advantages of Bayesian network computation, a reliability modeling method is proposed that can fully reflect the fault characteristics of complex systems. Experts describe the degree of failure of the event in the form of interval numbers. The knowledge and experience of experts are fused with the D-S evidence theory to obtain the initial failure probability interval of the BN root node. Then, the Hyper-ellipsoidal model (HM) constrains the initial failure probability interval and constructs a HE-BN for the system. A reliability analysis method is proposed to solve the problem of insufficient failure data and uncertainty in the degree of failure. The failure probability of the system is further calculated and the key components that affect the system's reliability are identified. The proposed method accounts for the uncertainty and incompleteness of the failure data in complex multi-state systems and establishes an easily computable reliability model that fully reflects the characteristics of complex faults and accurately identifies system weaknesses. The feasibility and accuracy of the method are further verified by conducting case studies.