A Prognostics and Health Management Strategy for Complex Electronic Systems

Bing Long,Houjun Wang,Qiang Miao,Michael Pecht

2012-01-01

Abstract:It is important to find the potential failure of subsystems or components for complex electronic systems such as radar, aviation systems, and then repair or replace them before the whole system fails. To repair or replace potential failure subsystems or circuits before a whole system fails, a new online prognostics and health management (PHM) system has been developed based on Mahalanobis distance (MD), the multisignal model, and the least squares support vector machine (LS-SVM). MD is used as a health index for monitoring each test parameter in the system. If health indices indicate that the test parameters will fail soon, the multisignal model is triggered to locate the potential failure subsystems at the system level. Then, LS-SVM is triggered to identify the failure modes or components in the potential subsystems. This paper presents an experiment with a radar transmitter system, which shows that the new PHM strategy can find potential failure circuits or components before the whole system fails. This method is useful for condition-based maintenance (CBM) of complex electronic systems.