| Autor: |
Zhang Qiantu, Fang Liqing, Zhao Yulong, Lv Yan |
| Jazyk: |
čínština |
| Rok vydání: |
2015 |
| Předmět: |
|
| Zdroj: |
Jixie chuandong, Vol 39, Pp 144-148 (2015) |
| Druh dokumentu: |
article |
| ISSN: |
1004-2539 |
| DOI: |
10.16578/j.issn.1004.2539.2015.12.031 |
| Popis: |
A new approach for mechanical fault diagnosis based on local characteristic- scale decomposition（ LCD） information entropy feature and support vector machine（ SVM） is proposed. Firstly,the fault mechanical vibration signal is decomposed by using the LCD to obtain a certain number of intrinsic scale component（ ISC）. Secondly,combined with information entropy theory,the singular spectrum entropy in time domain,power spectrum entropy in frequency domain,feature space entropy,marginal spectrum entropy and momentary energy entropy in time- frequency domain are defined and used as the feature vector. At last,the feature vectors are put into SVM classifier to recognize different fault type. The results of experiment of bearing fault diagnosis demonstrate that the method based on LCD information entropy feature and SVM is able to identify the bearing faults accurately and effectively,and the diagnosis effect is better than the method based on empirical mode decomposition（ EMD） information entropy and SVM. |
| Databáze: |
Directory of Open Access Journals |
| Externí odkaz: |
|
