Noise and weak symptom give rise difficulties to aero-engine vibration monitoring and faults diagnosis.The conventional methods in time or spectrum domain are not so effective for it is hard to detect some kinds of fault symptoms.After measuring the data in an aero-engine on the ground

its time-frequency characteristics under different working conditions were analyzed.First

the signals were analyzed using Morlet wavelet and then the frequency component could be extracted from their time-frequency plane.Some impact components were obvious in the time frequency plane which indicated that some faults might occur in the rotor system.They were located in the band of bearing characteristic frequencies.Second

the signal was decomposed using harmonic wavelet whose resolving capability is high in both time and frequency domain in the bearing characteristic frequencies.Then the components of different frequencies were obtained by harmonic wavelet.Energy of the signal was concentrated in some frequency band.From these results of the harmonic wavelet analysis in the bearing characteristic frequencies

the kind of fault can be determined easily.