Off-campus UMass Amherst users: To download dissertations, please use the following link to log into our proxy server with your UMass Amherst user name and password.
Non-UMass Amherst users, please click the view more button below to purchase a copy of this dissertation from Proquest.
(Some titles may also be available free of charge in our Open Access Dissertation Collection, so please check there first.)
Base wavelet selection criteria for non-stationary vibration analysis in bearing health diagnosis
While the wavelet transform has been increasingly applied to dynamic signal analysis for system health monitoring in manufacturing, civil and mechanical structure, surveillance, and medical diagnosis, selection of the base wavelet has remained largely as an ad hoc process. The presented study addresses this issue by introducing a strategy to select base wavelets for analyzing non-stationary vibration signals measured from rolling bearings. Specifically, criteria based on energy, Shannon entropy, correlation, and information theoretic measures have been investigated for their appropriateness in decomposing bearing vibration signals and extracting signatures that are indicative of structural defects. Subsequently, two comprehensive criteria: the energy-to-Shannon entropy ratio and MinMax information measure, have been developed. The effectiveness of the developed criteria has been evaluated on a Gaussian-modulated sinusoidal signal, using both real-valued and complex-valued wavelets. Upon confirmation of their effectiveness, vibration signals measured from a ball bearing was used as input to further evaluate the applicability of the developed criteria. Finally, base wavelets selected from the above described criteria were used to devise multi-scale signal processing techniques for simultaneous decomposition of bearing vibration signals in the time-frequency domain and defect severity characterization. ^ Besides bearing vibrations, the developed wavelet selection criteria can be applied to enhancing signal processing for a broad range of dynamic signals, such as physiological wave forms or wind speed profiles. The presented study on base wavelet selection has also enabled insight into wavelet-based transformation techniques, which are of interest to research in multi-resolution analysis and object identification. ^
Yan, Ruqiang, "Base wavelet selection criteria for non-stationary vibration analysis in bearing health diagnosis" (2007). Doctoral Dissertations Available from Proquest. AAI3275786.