Abstract
Transient feature caused by local defect is crucial for bearing health diagnosis and sparse learning has been widely used to enhance the desired features with the noise interference removed. In order to overcome the difficulties of conventional sparse approaches in dictionary construction and the limitations of sparse representation results, by considering the merits of time-frequency manifold (TFM) learning, a new self-enhancement method, shift-invariant manifold sparse learning (SI-MSL), is proposed in an adaptive way. Different from the conventional time-domain sparse methods, the envelope signal is isolated and self-rebuilt with the latent shift-invariant structure enhanced. First, the intrinsic manifold modes of a given signal can be adaptively mined by establishing TFM learning on a local segment, and the amplitude-phase separation technique is then applied to complete shift-invariant sparse on the raw envelope signal. Second, based on the reinforcement learning of shift-invariant manifold mode for the whole signal, a novel self-enhancement of bearing can be achieved by solving a convex optimization problem. Finally, by utilizing phase preserve and a series of inverse transforms, the transient feature of the given signal can be self-enhanced in an efficient way. Specially, envelope spectral entropy is used to adaptively output the optimal manifold basis function from the local TFM signature. Different from conventional sparse representation, this method utilizes feature mining of TFM learning to ensure more accurate sparse expression from dictionary construction to solution, providing more accurate diagnostic results. Experimental and comparison results show that the proposed method can achieve a good effect in interference suppression with the transient feature enhancement in a self-learning process, which is benefit for an accurate and intelligent rolling bearing health diagnosis.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: 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.