Segregation of close frequency components based on reassigned wavelet analysis for machinery fault diagnosis
Vibration signals of rotating machinery often contain many closely located frequency components. While Fast Fourier Transform (FFT) analysis of the signals can identify exact frequency components in the vibration spectrum easily, conventional wavelet analysis is generally incapable of discriminating...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Published: |
Springer Heidelberg
2015
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/59465/ http://dx.doi.org/10.1007/978-3-319-09507-3_50 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Universiti Teknologi Malaysia |
| Summary: | Vibration signals of rotating machinery often contain many closely located frequency components. While Fast Fourier Transform (FFT) analysis of the signals can identify exact frequency components in the vibration spectrum easily, conventional wavelet analysis is generally incapable of discriminating closely located frequency components in vibration signals due to overlapping and interference appearing in wavelet results. Wavelet transforms based on wavelet reassignment algorithm to improve time-frequency resolution display is presented in this chapter. The proposed reassigned (modified) Morlet wavelet was tested using simulated signal and experimental data obtained from a multi-stage blades rotor test rig. This study showed that this method was capable of segregating close BPF components which were otherwise lumped together in conventional wavelet analysis display. The reassigned Morlet wavelet analysis was shown to be useful for multi stage blade rubbing diagnosis as well as other general condition monitoring applications such as those for gear and bearing faults diagnosis. |
|---|