A novel ridge identification method integrating the multi-index optimal resonant frequency band extraction for bearing fault diagnosis under variable-speed conditions

View at publisher Summarise with VIRAT
Authors: Xu Zhang, Xuhao Zhang, Xin Wang, Tao Yang, Wei Li
Cited by: 0
References: 26

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Abstract Bearings often operate under variable-speed conditions, leading to instability in their characteristic frequencies. Variable-frequency characteristic signals manifest as multiple shifting energy bands in the time-frequency representation (abbreviated as TFR). The ridge identification method can extract time-frequency ridges from these bands to isolate varied characteristic frequencies. However, such a method faces the following challenges: (1) The random noise causes blurring effects in the TFR of raw signals. (2) The low-frequency band characteristic signals introduce interfering harmonics in the TFR, resulting in the ridge distortion. Aimed at noise reduction and TFR resolution enhancement, a novel ridge identification method integrating the multi-index optimal resonant frequency band extraction and differentiable short-time Fourier transform (abbreviated as DSTFT) is developed for identifying bearing faults under variable speeds. The tunable Q-factor wavelet transform (abbreviated as TQWT) is adopted to generate initial subbands, and weight vectors eliminating irrelevant subbands are then derived based on the subband Smoothness Index ( SI ). An index, abbreviated as KCS , is constructed by the integration of kurtosis ( Ku ), crest factor ( CF ), and SI , and then is combined with a proposed selective merging strategy to select an optimal resonant frequency band for fault frequency extraction. The DSTFT is used to obtain clear TFRs of fault characteristic frequencies (abbreviated as FCF ) and shaft rotation frequencies. Instantaneous fault characteristic frequency (abbreviated as IFCF) ridges and instantaneous shaft rotation frequency (abbreviated as ISRF) ridges are subsequently extracted using a fast path optimization algorithm, and the mean ratio of IFCF to ISRF is calculated to match the fault characteristic coefficient (abbreviated as FCC ). An SSF-DFP-100 bearing fault test rig is built to collect signals from faulty bearings under variable speeds, and these signals are analyzed by the ridge identification method with fused multi-index optimal resonant frequency band extraction. The results indicate that the constructed KCS effectively resists interference from high-frequency white noise and low-frequency harmonics, and the KCS -based dynamic selective subband merging strategy ensures that the frequency range of the selected optimal subband fully encompasses the optimal resonant frequency band. The developed ridge identification method successfully achieved the extraction of multiple distinct IFCF and ISRF ridges from the fault characteristic signal, thereby accurately diagnosing the fault types in the inner and outer races of the bearing.

Fields of Study & Subjects

Machine Fault Diagnosis Techniques Gear and Bearing Dynamics Analysis Fault Detection and Control Systems

Publication details

Journal / source
Journal of Instrumentation
Volume
21
Issue
03
Pages
P03041–P03041
Publication date
1 March 2026
Language
EN