Gear fault diagnosis based on the structured sparsity time-frequency analysis
发布时间：2025-04-30
点击次数：
- 发布时间：
- 2025-04-30
- 论文名称：
- Gear fault diagnosis based on the structured sparsity time-frequency analysis
- 发表刊物：
- Mechanical Systems and Signal Processing
- 摘要：
- Over the last decade, sparse representation has become a powerful paradigm in mechanical fault diagnosis due to its excellent capability and the high flexibility for complex signal description. The structured sparsity time-frequency analysis (SSTFA) is a novel signal processing method, which utilizes mixed-norm priors on time-frequency coefficients to obtain a fine match for the structure of signals. In order to extract the transient feature from gear vibration signals, a gear fault diagnosis method based on SSTFA is proposed in this work. The steady modulation components and impulsive components of the defective gear vibration signals can be extracted simultaneously by choosing different time-frequency neighborhood and generalized thresholding operators. Besides, the time-frequency distribution
with high resolution is obtained by piling different components in the same diagram. Thediagnostic conclusion can be made according to the envelope spectrum of the impulsivecomponents or by the periodicity of impulses. The effectiveness of the method is verified by numerical simulations, and the vibration signals registered from a gearbox fault simulator and a wind turbine. To validate the efficiency of the presented methodology, comparisons are made among some state-of-the-art vibration separation methods and the traditional time-frequency analysis methods. The comparisons show that the proposed method possesses advantages in separating feature signals under strong noise and accounting for the inner time-frequency structure of the gear vibration signals.
- 合写作者：
- Ruobin Sun, Zhibo Yang, Xuefeng Che, Shaohua Tian, Yong Xie
- 卷号：
- 102
- 页面范围：
- 346-363
- 是否译文：
- 否
- 发表时间：
- 2018-04-14