• CN：11-2187/TH
• ISSN：0577-6686

• 机械动力学 •

### 一种基于时频峭度谱的滚动轴承损伤诊断方法

1. 1.长安大学电子与控制工程学院；
2.长安大学汽车学院
• 出版日期:2015-06-15 发布日期:2015-06-15
• 基金资助:
国家自然科学基金(51175049，61201407)、陕西省自然科学基础研究计划(2014JM7246)资助项目

### A Ball Bearing Defect Diagnosis Method Using Time-frequency Kurtosis Spectrum

DUAN Chendong1, GAO Peng1, XU Xianfeng1, GAO Qiang2

1. 1.School of Electronic & Control Engineering, Chang’an University；
2.School of Mobile Engineering, Chang’an University
• Online:2015-06-15 Published:2015-06-15

Abstract: In order to find defect characteristic frequencies of ball bearing accurately and effectively, a new time-frequency kurtosis spectrum analysis method is proposed based on the frequency slice wavelet transform (FSWT). Firstly, Vibration signal of a monitored ball bearing is decomposed in time-frequency domain by applying the FSWT. Then, on the basis of the time-frequency decomposition result, a kind of kurtosis is calculated which is related to the amplitude of each frequency component in the time-frequency domain of the signal. By using the kurtosis sequence, a time-frequency kurtosis spectrum is obtained. Thereafter, several higher peaks on the spectrum curve are selected, frequencies related to these peaks on the spectrum are thought as the center frequencies of those resonance bands to be analyzed, and the analyzed resonance bands are also fixed yet at the same time. And then, according with the analyzed bands, the corresponding time-frequency slices are settled on the time-frequency decomposition plane of the signal. Afterwards, the components over these time-frequency slices are rebuilt by means of doing the reverse transform of the FSWT. To get envelops of these components, the demodulation processing is used. Finally, an equivalent power spectrum of these envelopes is established for discovering the defect characteristic frequencies of the monitored bearing. It is proved that the proposed method is able to extract the defect characteristic frequencies efficiently. In addition, because several frequency-bands are used, this makes enough energies of the signal to be used for performing the envelope analysis, when a ball bearing has several different kinds of defects it can identify the defect characteristic frequencies perfectly.