Explicit Methods for Incorporating Bayesian Networks into Common Cause Analysis

摘要

Two explicit methods for incorporating common cause failures (CCF) into reliability analysis based on Bayesian networks are proposed in this paper. CCF are multiple failures due to a common cause (CC) which tend to increase the system unreliability. Two traditional methods are often used for CCF analysis, explicit and implicit methods, both of which have shortcomings especially in handling large fault trees. Our methods provide accurate and efficient reliability analysis by incorporating Bayesian networks (BN) into CCF analysis, which could avoid the disadvantages of traditional methods. Complex dependencies among common causes, which could not be dealt with in traditional methods, can also be solved by properly modifying the Bayesian networks.