System-Level BITs Fault Diagnosis and Isolation Based on Diagnostic Tree and Bayesian Network

Large systems have complex structures and functions, which makes it more difficult in fault isolation. Given the design features of BITs (built-in test) of large complex systems, a method, based on the diagnostic tree and Bayesian network, has been proposed for fault diagnosis and isolation of System-level BITs. First, faults are rapidly isolated in accordance with the diagnostic tree as well as the information of periodic BITs and maintenance BITs. Then, a diagnostic method based on the Bayesian network has been put forward to solve the problems of ambiguity groups of three LRUs or more and to accomplish further diagnosis and isolation. Finally, a certain subsystem of radar has been exemplified to verify the effectiveness of the method proposed by this study.