The Use of Time-Frequency Methods in Rotor/Stator Impact-Rubbing Detection

At present, detailed rubbing diagnostics is implemented as off-line analysis of measured data. During the machine operation, the overall level of vibration and the rotational frequency (1X) phasor are traced. On the basis of amplitude and phase changes, the presence of rub is identified. Although unmonitored, changes in sub synchronous spectral components are also caused by rubbing. The new method presented herein utilizes measurements of these changes for better online rubbing detection. The paper presents the authors work on rotor/stator rub detection in steam turbines. For automatic rubbing detection, an on-line tool, called cumulative full-spectra method and based on full-spectra computation, is proposed. The proposed automatic detection algorithm is based on characteristic value, which is computed using special norming of tracked frequencies in the cumulative full-spectra.© 2011 ASME