Systems Reliability Analysis and Fault Diagnosis Based on Bayesian Networks

This paper presents the application of Bayesian networks(BN) to the reliability analysis and fault diagnosis of systems. For systems, it is essential to do reliability analysis. Also it is necessary to do fault diagnosis when a system failed, but the better way is to do fault diagnosis before the system has failed. Bayesian networks have many special characteristics, one of them is that they are parallel structures, so the time of solving is much shorter than that of many common methods. Bayesian networks permit not only computing the reliability indices of a system but also presenting the effect of each component or some components on the system reliability to distinguish the unsubstantial part of the system, so we can know which part is the weakest one of the system. Two examples proved the validity and superiority of the method in the application of the reliability analysis and fault diagnosis of system.