Bearing fault diagnosis algorithm based on maximum correlated kurtosis feature mode decomposition and compound Gini index
The maximum correlated kurtosis feature mode decomposition (MCKFMD) method can effectively remove redundant information and enhance fault features, but its effect is affected by the number of decomposition modes, the number of initialized filters and the filter length. To address this problem, a bea...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | zho |
| Published: |
Editorial Department of Electric Drive for Locomotives
2023-07-01
|
| Series: | 机车电传动 |
| Subjects: | |
| Online Access: | http://edl.csrzic.com/thesisDetails#10.13890/j.issn.1000-128X.2023.04.002 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The maximum correlated kurtosis feature mode decomposition (MCKFMD) method can effectively remove redundant information and enhance fault features, but its effect is affected by the number of decomposition modes, the number of initialized filters and the filter length. To address this problem, a bearing fault diagnosis method based on compound Gini index (CGI) and MCKFMD was proposed. Firstly, a new sparse index called CGI was constructed by combining the squared Gini index in the time domain and frequency domain to quantify the abundance of periodic impulses in the time domain and frequency domain, and its performance was evaluated and verified. Secondly, CGI was used as the fitness function for the sand cat swarm optimization (SCSO) algorithm to quickly and accurately obtain the optimal parameter combination of MCKFMD and realize the adaptive decomposition of fault signals. Finally, CGI was used to select the optimal mode for Hilbert envelope demodulation to achieve fault feature extraction. The proposed method was verified for effectiveness by using analog and experimental signals, and comparative studies have shown that it is more effective in extracting periodic fault features compared with the parameter-optimized variational modal decomposition (VMD) and fixed-parameter MCKFMD. |
|---|---|
| ISSN: | 1000-128X |