.png){kind=icon}
Return
Threshold variational mode decomposition cascade high concentrated time–frequency analysis algorithm
.png){kind=icon}
PRE
AI
DOI：10.1016/j.ymssp.2025.113337.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
To further improve the concentration performance of the time–frequency analysis (TFA) method and accurately capture the time–frequency (TF) spectral anomaly, we propose a threshold Variational Mode Decomposition cascade high concentrated TFA algorithm, termed improved Local Maximum Synchrosqueezing Generalized S transform based on threshold Variational Mode Decomposition (ILMSSGST). The LMSSGST method, combining the frequency self-adaptability of the Generalized S transform (GST) and the iteration operation of the Local Maximum Synchronosqueezing (LMSST) algorithm, achieves a TF representation with higher energy concentration. Then, the threshold variational mode decomposition (TVMD) algorithm is presented, which can adaptively select the optimal number of mode decompositions and realize signal denoising. By combining the TVMD algorithm with the LMSSGST algorithm, the anti-noise performance of the algorithm can be improved while maintaining high concentration performance. The improved ILMSSGST integrates TVMD for adaptive denoising and mode selection, with GST serving as the kernel for LMSST. The “improvement” lies in the cascade architecture that enhances both TF concentration and noise resistance, rather than modifying LMSST itself. The novel contributions include: 1) an adaptive TVMD for automatic optimal mode selection; 2) a cascade framework integrating TVMD and LMSSGST to enhance both concentration and noise resistance; and 3) multi-metric validation of TFA performance. To verify the effectiveness of the proposed method, the simulated signal and experimental signal of the bearing fault are tested, respectively. Experiments demonstrate that ILMSSGST achieves higher concentration improvement than state-of-the-art methods. The ILMSSGST can clearly describe the precise TF characteristics of the vibration signal, which provides an effective technical solution for anomaly detection.
Journal
IF:
8.9
Papers:
1.3W
Citations:
6.6W
Organization
Cited Papers
No cited papers available