Sliding window denoising K-Singular Value Decomposition and its application on rolling bearing impact fault diagnosis. (12th May 2018)

• Record Type: Journal Article
• Authors: Yang, Honggang; Lin, Huibin; Ding, Kang
• Is Part Of: Journal of sound and vibration. Volume 421(2018)
• Page Start: 205
• Page End: 219
• Publication Date: 2018-05-12
• Subjects: K-SVD -- Sparse representation -- Sliding window -- Inner product -- Rolling bearing; Sound -- Periodicals; Vibration -- Periodicals; Son -- Périodiques; Vibration -- Périodiques; Sound; Vibration; Periodicals; Electronic journals; 620.205
• DOI: 10.1016/j.jsv.2018.01.051
• Languages: English
• ISSNs: 0022-460X

Abstract

Abstract: The performance of sparse features extraction by commonly used K-Singular Value Decomposition (K-SVD) method depends largely on the signal segment selected in rolling bearing diagnosis, furthermore, the calculating speed is relatively slow and the dictionary becomes so redundant when the fault signal is relatively long. A new sliding window denoising K-SVD (SWD-KSVD) method is proposed, which uses only one small segment of time domain signal containing impacts to perform sliding window dictionary learning and select an optimal pattern with oscillating information of the rolling bearing fault according to a maximum variance principle. An inner product operation between the optimal pattern and the whole fault signal is performed to enhance the characteristic of the impacts' occurrence moments. Lastly, the signal is reconstructed at peak points of the inner product to realize the extraction of the rolling bearing fault features. Both simulation and experiments verify that the method could extract the fault features effectively. Highlights: A sliding window denoising K-SVD method is proposed for bearing fault detection. An optimal pattern which contains high-frequency oscillating information is developed. The effects of noise level and starting point on patterns learned by K-SVD are studied. The method is verified by simulated and experimental results.