相关熵的降噪机理及其在轴承故障诊断中的应用  

Denoising Mechanism of Correntropy and Its Application in Bearing Fault Diagnosis

作　　者：李辉 LI Hui(School of Mechanical Engineering, Tianjin University of Technology and Education, Tianjin 300222, China)

机构地区：[1]天津职业技术师范大学机械工程学院,天津300222

出　　处：《轴承》2021年第3期36-44,共9页Bearing

基　　金：国家自然科学基金项目(51375319);河北省自然科学基金项目(E2013421005);天津市支持京津冀科技成果转化项目(17YFCZZC00270)。

摘　　要：针对传统相关函数和谱相关密度难以有效处理强非高斯噪声干扰的问题,提出了一种基于循环平稳相关熵的故障诊断方法。以理论分析和几何图解等方式系统分析了相关熵的降噪机理,以余弦信号和仿真调幅信号为例解释了相关熵以及循环平稳相关熵的降噪机理并验证了其良好的噪声抑制能力;应用循环平稳相关熵方法对轴承内、外圈局部裂纹故障振动信号进行了分析和处理,试验结果表明,循环平稳相关熵谱密度具有解调功能,能准确刻画轴承局部裂纹故障的频谱特征,可有效提取淹没在强噪声环境中的微弱信号。The traditional correlation function and spectral correlation density are difficult to deal with strong non-Gaussian noise interference effectively,and a fault diagnosis method is proposed based on cyclostationary correntropy.The denoising mechanism of correntropy is systematically analyzed by theoretical analysis and geometric diagram description.The denoising mechanism of correntropy and cyclostationary correntropy are interpreted in detail by taking cosine signal and simulative amplitude modulation signal as examples,and its good ability of noise suppression is verified.The vibration signals of local crack faults of inner and outer rings of bearings are analyzed and processed by using cyclostationary correntropy method.The test results show that the cyclostationary correntropy spectral density has demodulation function,which can accurately describe frequency spectrum characteristics of local crack faults of bearings and effectively extract weak signal submerged in strong noise environment.

关 键 词：滚动轴承 故障诊断 相关熵 降噪 循环平稳信号 

分 类 号：TH133.33[机械工程—机械制造及自动化] TN911.72[电子电信—通信与信息系统]