Bearing Feature Extraction Method Based on the Time Subsequence
Although pure time-domain features have the advantages of fast extraction speed and clear physical meaning, the diagnostic accuracy is slightly inferior to other methods. To solve this problem, a new bearing feature extraction method based on the time subsequence (BOTS) is proposed, which combines w...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | zho |
| Published: |
Editorial Office of Journal of Mechanical Transmission
2023-11-01
|
| Series: | Jixie chuandong |
| Subjects: | |
| Online Access: | http://www.jxcd.net.cn/thesisDetails#10.16578/j.issn.1004.2539.2023.11.022 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841546985682763776 |
|---|---|
| author | Wang Dexue Nie Fei Zheng Zhifei Yu Yongsheng |
| author_facet | Wang Dexue Nie Fei Zheng Zhifei Yu Yongsheng |
| author_sort | Wang Dexue |
| collection | DOAJ |
| description | Although pure time-domain features have the advantages of fast extraction speed and clear physical meaning, the diagnostic accuracy is slightly inferior to other methods. To solve this problem, a new bearing feature extraction method based on the time subsequence (BOTS) is proposed, which combines word package model and time subsequence. First, the sliding window is used to slide in the vibration signal to obtain multiple continuous and non-stationary time series, which are regarded as a document. For each time series, multiple continuous subsequences of fixed length are randomly intercepted to obtain the time-domain or frequency-domain characteristics of subsequences. Then, the random forest algorithm is used to count the class votes of all subsequences in each time series, and a dictionary is constructed based on the class votes. Finally, the dictionary is used as a new feature and input into the random forest classifier for training and learning. A variety of experiments are carried out using the bearing data provided by the SQI-MFS experimental platform of Wuxi Innovation Center of SIEMENS China Research Institute, Southeast University and Institute of Mechanical Failure Prevention Technology. The experiments show that the features extracted by BOTS+ wavelet packet energy method have higher recognition. |
| format | Article |
| id | doaj-art-9023e50f520e4e08b1be90091221f09e |
| institution | Kabale University |
| issn | 1004-2539 |
| language | zho |
| publishDate | 2023-11-01 |
| publisher | Editorial Office of Journal of Mechanical Transmission |
| record_format | Article |
| series | Jixie chuandong |
| spelling | doaj-art-9023e50f520e4e08b1be90091221f09e2025-01-10T14:59:26ZzhoEditorial Office of Journal of Mechanical TransmissionJixie chuandong1004-25392023-11-014714615344815908Bearing Feature Extraction Method Based on the Time SubsequenceWang DexueNie FeiZheng ZhifeiYu YongshengAlthough pure time-domain features have the advantages of fast extraction speed and clear physical meaning, the diagnostic accuracy is slightly inferior to other methods. To solve this problem, a new bearing feature extraction method based on the time subsequence (BOTS) is proposed, which combines word package model and time subsequence. First, the sliding window is used to slide in the vibration signal to obtain multiple continuous and non-stationary time series, which are regarded as a document. For each time series, multiple continuous subsequences of fixed length are randomly intercepted to obtain the time-domain or frequency-domain characteristics of subsequences. Then, the random forest algorithm is used to count the class votes of all subsequences in each time series, and a dictionary is constructed based on the class votes. Finally, the dictionary is used as a new feature and input into the random forest classifier for training and learning. A variety of experiments are carried out using the bearing data provided by the SQI-MFS experimental platform of Wuxi Innovation Center of SIEMENS China Research Institute, Southeast University and Institute of Mechanical Failure Prevention Technology. The experiments show that the features extracted by BOTS+ wavelet packet energy method have higher recognition.http://www.jxcd.net.cn/thesisDetails#10.16578/j.issn.1004.2539.2023.11.022Rolling bearingFault diagnosisFeature extractionFault status identification |
| spellingShingle | Wang Dexue Nie Fei Zheng Zhifei Yu Yongsheng Bearing Feature Extraction Method Based on the Time Subsequence Jixie chuandong Rolling bearing Fault diagnosis Feature extraction Fault status identification |
| title | Bearing Feature Extraction Method Based on the Time Subsequence |
| title_full | Bearing Feature Extraction Method Based on the Time Subsequence |
| title_fullStr | Bearing Feature Extraction Method Based on the Time Subsequence |
| title_full_unstemmed | Bearing Feature Extraction Method Based on the Time Subsequence |
| title_short | Bearing Feature Extraction Method Based on the Time Subsequence |
| title_sort | bearing feature extraction method based on the time subsequence |
| topic | Rolling bearing Fault diagnosis Feature extraction Fault status identification |
| url | http://www.jxcd.net.cn/thesisDetails#10.16578/j.issn.1004.2539.2023.11.022 |
| work_keys_str_mv | AT wangdexue bearingfeatureextractionmethodbasedonthetimesubsequence AT niefei bearingfeatureextractionmethodbasedonthetimesubsequence AT zhengzhifei bearingfeatureextractionmethodbasedonthetimesubsequence AT yuyongsheng bearingfeatureextractionmethodbasedonthetimesubsequence |