Safety analysis of complex systems based on Bayesian Networks

This paper presents a novel approach for safety analysis of complex systems. The Bayeisan Networks has been modeled based on failure scenarios contained a large number information of Failure Modes, Effects and Criticality Analysis accumulated in the production and use stages of complex systems. The model´s constrution, probabilistic inference, sensetivity annlysis is analysised through a case study on controllable-pitch propeller and shafting system. The results show the validity and convenience of the method and great advantages and applications prospects of Bayesian Networks in safety analysis of complex systems.