On-line vibration analysis with fast continuous wavelet algorithm for condition monitoring of bearing

The newly developed technique of wavelet transform enables us to observe the evolution in time of the frequency content of a signal. This property makes it very suitable for the detection of vibration transients. However, current algorithms for vibration analysis either have low resolution of features for detailed analysis in the frequency-band scale or are very time consuming. In this paper, we present fast wavelet-based algorithms for vibration analyses. The approximated Morlet wavelet is used to develop an infinite impulse response causal filter with the error kept at an acceptable level. Using this filter, the continuous wavelet transform (CWT) can be computed. The rapid computation of the CWT together with autocorrelation enhancement, is developed for the detailed on-line vibration analysis. Thus, the raw vibration signal can be continuously processed and monitored, with warning or alarm signals being generated when pre-programmed levels are exceeded. Examples are used to illustrate the application of arbitrary fine frequency-band scale analysis and single specified central frequency in the time domain using the wavelet power spectrum, which enables small defect detection of bearings. Real-time condition monitoring, fault detection, tool wear monitoring, etc., which are related to vibration, can be achieved and implemented with the detailed analysis of vibration signals.