Detection of crack in a transient stepped rotor system using continuous wavelet transform

The present work deals with the detection of transverse surface fatigue cracks in rotor bearing systems mainly used in turbo machines, aero engines etc. The detection of faults in high speed aero-engines is gaining more importance in these days. The aero-engines are run-up type rotor systems, which are started and stopped frequently. Thus the transient analysis of the run-up type rotor has proved to be a good crack detection tool in the early literature. But most of the works in these areas were focused on a simple jeffcott rotor with single acceleration transient analysis, which is an ideal case. In real time systems the acceleration is not constant in the transient period, the system starts with low acceleration as it reaches the zone of resonance and passes through the zone of resonance with a high acceleration. In the present work, the same has been exploited for a stepped shaft; also the effect of bearing stiffness on crack detection has been studied. The vibration data obtained in the transient analysis has been processed using a signal processing technique called continuous wavelet transform (CWT). The results obtained in the present analysis can be a good diagnostic technique for crack detection. The analysis has been done by developing a finite element code in MATLAB.