An Adaptive Fault Diagnosis System Framework for Aircraft Based on Man-in-loop

In the field of aviation, the traditional fault feature extraction is directly related to experience of persons, the selected features are always fixed in use, and the fault feature updates need the upgrading of the fault diagnosis system which include complex technologies and high cost. Additionally, the fault information collections are insufficient. It lacks fault confirmation feedback to fault diagnosis system when airplanes return to factory maintenance. An adaptive fault diagnosis system framework for aircraft is proposed in this paper. The man-in-loop information feedback method can make up the incomplete information collection. The fault data in flight and the new fault mode verified in the feedback are taken as inputs and triggers of fault feature optimization. The fault features can be extracted and reduced adaptively. And then, the fault features in the fault diagnosis system can be updated automatically, which improves the aircraft fault diagnosis ability in self-learning way.