1. Feature extraction of rolling bearing multiple faults based on correlation coefficient and Hjorth parameter.
- Author
-
Yu, Mingyue and Fang, Minghe
- Subjects
ROLLER bearings ,STATISTICAL correlation ,FEATURE extraction ,SELF-adaptive software - Abstract
The paper has brought in the Hjorth Complexity parameter and combined it with Intrinsic time decomposition (ITD) algorithm as characteristic parameter index in order to implement accurate identification of multiple faults of rolling bearings. Firstly, concerning about the much uncertainty in manual setting of decomposition layer number in ITD, overdoing of automatic decomposition and the fact that a larger correlation coefficient of signal relates with the greater correlation of signals before and after decomposition and vice versa, the paper has carried out self-adaptive determination of the number of ITD decomposition layers. Secondly, regarding the insensitivity of Hjorth Complexity parameter to noise and the fact that with larger Complexity parameter, signals are simpler and it becomes more available to dig out characteristic information of fault from signals. With Complexity parameter as the index of characteristic parameter, option of optimal Proper rotation component (PRC) is made after ITD. Finally, through the comparison with other methods and the analysis of multiple faults of bearings, it indicates that correlation coefficient can self-adaptively determine the number of ITD decomposition layers and prevent from overdoing and underdoing of decomposition. The Hjorth Complexity parameter can be treated as index parameter to implement optimal PRC option, based on which multiple fault characteristics of bearings can be effectively extracted and the type precisely determined. • The decomposition layers of ITD are determined based on the correlation coefficient adaptive. • Hjorth parameter is introduced into the feature extraction of rolling bearing compound faults. • Complexity parameter is used as the index parameter of the optimal PRC selection. • Cross-correlation function is brought in to make mixed study on the acceleration signals from different sensors. • ITD is combined with Hjorth parameter. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF