Keywords: cracked shaft, rotor, crack detection, wavelet transform, transient response, frequency sweep, local resonances, residual imbalance, FFT analysis.

Rotating machinery operation can be much dangerous when a fatigue and-or growing crack develops in the shaft. Longer time intervals between overhauls and maintenance cost reduction programs require effective on-line monitoring of the machine condition. Previous works show great difficulty in detecting incipient cracks merely monitoring and studying machine vibration during steady state operation; from there analyzing transient response during a frequency sweep such as start up-stopping the rotor has been highly recommended by the first author of this work as well as several other authors [1,2].

In this paper, a comparison between the Fourier waterfall plots and wavelet transforms has been carried out during the process of identifying cracked shafts. A cracked unbalanced Jeffcott rotor realizing a machine start up with two different time intervals are studied. Crack-imbalance interaction yields involved responses complicating the diagnosis. For the first time conventional Fourier Transform, FFT, cascades, as well as wavelet analyses are both applied to start up cracked-imbalanced rotors and their analyzing capabilities are compared. Due to the transient nature of this analysis, typical FFT reduces its efficiency while wavelet transform seems more adequate.

Some mayor conclusions are that variation, depending on crack-imbalance angle, of the response amplitudes at the resonance time can occur [1,3]; yet this work a single relative angle orientation through a frequency sweep as that provided by a start-up has been analyzed. Results suggest that such kind of behaviour can reveal the presence of a crack in the shaft.

Finally sensitivity of the system response to the duration of the start-up time along with a wavelet pattern response analysis can provide a useful method to obtain behaviour patterns and detect a damaged shaft. At a frequency sweep, the wavelet approach seems to be superior compared to the typical FFT and cascade analysis, the latter due to higher precision in time details. Further work is carried on post-processing wavelet results combined by neural network identification methods.

1

J. Gómez-Mancilla, J.J. Sinou, V.R. Nossov, F. Thouverez, A. Zambrano, "The Influence of Crack-Imbalance Orientation and Orbital Evolution for an Extended Cracked Jeffcott Rotor", COMPTES-RENDUS Mecanique, R.C., Vol. 332, 955-962, 2004.

2

Y. Ishida, T. Inoue, "Detection of a rotor crack using a harmonic excitation and nonlinear vibration analysis", J. of Vibr & Acoust, Vol. 128(6), 741-749, 2006. doi:10.1115/1.2346693

3

J. Gómez-Mancilla, J.M. Machorro-López, V.R. Nosov, "Crack Breathing Mechanisms In Rotor-Bearings Systems, Its Influence On System Response and Crack Detection", paper 418, 3nd ISCORMA-3, Cleveland St. University, Ohio, Sept. 2005.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £145 +P&P)