Bearing Fault Feature Extraction via Adaptive Frequency Matching and Coupling Analysis
Weiyang Xu, Jialong He, Guofa Li, Chenchen Wu, Chenhui Qian, Wei Yang · IEEE Sensors Journal · 2024
Aiming to solve the problem of frequency coupling offset in rolling bearing fault diagnosis, we propose a bearing fault feature extraction method based on adaptive frequency matching (AFM). This method is based on the analytic signal discretization expression of signal envelope frequency. By constructing a dictionary matrix and deriving a coefficient vector solution formula, combined with a frequency grid iterative matching strategy, it achieves accurate extraction of the basic rotation frequency and fault feature frequency of rolling bearings. The significant features of this method include the following: direct analysis is performed on the signal envelope frequency, avoiding the errors that may be introduced by time-domain reconstruction; the coexistence of basic rotational frequency and fault characteristic frequency was fully considered, achieving simultaneous extraction of both; through AFM and iterative optimization, the frequency offset characteristics caused by the coupling between the basic rotational frequency and the fault characteristic frequency was effectively extracted and explained.