Common Cause Failure Model of System Reliability Based on Bayesian Networks

摘要

Common cause failure is an important phenomenon for a system with failure dependent parts. In this paper, several common cause failure models are analyzed and compared. A new common cause failure model for system reliability estimation is presented based on Bayesian Networks. Examples of series system, parallel system and series-parallel system are given to explain how to use the model to evaluate the reliability of system, through which the weak parts of the system can be identified. Also, Monte-Carlo simulation method is used to estimate the reliability based on Bayesian Networks, the results of which are compared with the system reliability under failure independence assumptions. The simulation results show that the reliability model based on Bayesian Network is consistent with the traditional qualitative analysis which proves that the Bayesian Networks model is accurate and valid.