The application belongs to the technical field of on-board early warning of
ball bearing outer ring peeling faults in an aero-engine, and particularly relates to an on-board early warning method for
ball bearing outer ring peeling faults in an aero-engine, which fuses feature quantities extracted based on a
wavelet envelope spectrum and a specified
frequency band filtering method through fuzzy
processing, and constructs a feature quantity for fault warning again, compares the fused feature quantity with a fault threshold value to determine whether a
ball bearing outer ring peeling fault occurs, and performs early warning. The overall calculation amount is not large, the calculation speed is relatively fast, and the demand for on-board early warning of ball bearing outer ring peeling faults in an aero-engine can be well met.