Abstract
Rolling element bearing faults account for main causes of rotating machine failures. It is crucial to identify the incipient fault before the bearing steps into serious fault condition. The Hilbert envelope spectrum has been proved powerful and with high practical value to detect transient components in vibration signal but sensitive to noise. Based on the conventional singular value decomposition (SVD) theory, accumulative component kurtosis (ACK) is introduced to de-noising of vibration signal processing. The proposed ACK-SVD emphasizes the accumulative components (ACs) rather than the single singular component (SC) to select the effective SCs to recover signal. The superiority of the ACK-SVD over traditional SVD de-noising is verified by both simulated signals and actual vibration data from two rolling element bearing rigs. The results demonstrate the proposed method can efficiently identify the rolling element bearing faults, especially the early ones with strong background noise.
Highlights
Rolling element bearings (REBs) are critical mechanical components in rotating machinery with extensively application fields especially in modern industrial areas and their working state directly affects the performance of the whole machinery and the production efficiency
Different to the work in ref [12], where Yu D. et al combines the empirical mode decomposition (EMD) with Hilbert transform together to propose the local marginal spectrum which is verified effective in fault diagnosis of rolling bearings by practical vibration signals, singular value decomposition (SVD) has been applied to feature matrix formed by the intrinsic mode function (IMF) and residue from EMD process, the singular value (SV) forms the feature vectors based on which the Mahalanobis distance to the normal state can be calculated to measure the fault condition of rolling bearings [19], while, the SVD has been used to form the condition feature vectors on the product function obtained by local mean decomposition (LMD) to vibration signals [20]
The accumulative component kurtosis (ACK) varies with the number of singular component (SC) added together as shown in Fig. 6, the peak arises after the fifth SC added, which is consistent with the kurtosis maximum rule, the singular order is 5
Summary
Rolling element bearings (REBs) are critical mechanical components in rotating machinery with extensively application fields especially in modern industrial areas and their working state directly affects the performance of the whole machinery and the production efficiency. Different to the work in ref [12], where Yu D. et al combines the EMD with Hilbert transform together to propose the local marginal spectrum which is verified effective in fault diagnosis of rolling bearings by practical vibration signals, SVD has been applied to feature matrix formed by the IMFs and residue from EMD process, the SVs forms the feature vectors based on which the Mahalanobis distance to the normal state can be calculated to measure the fault condition of rolling bearings [19], while, the SVD has been used to form the condition feature vectors on the product function obtained by local mean decomposition (LMD) to vibration signals [20].
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
