摘要
针对滚动轴承故障振动信号非平稳性与非线性的特点,提出将集合经验模态分解(ensemble empirical mode decomposition, EEMD)方法用于轴承信号处理.滚动轴承故障诊断的重要环节是特征提取,其直接关系到轴承故障诊断的正确率.因此,将熵知识应用到轴承特征提取步骤中,应用奇异熵与能量熵知识,提出一种峭度值与以上两种熵进行特征融合的特征提取方法,完成滚动轴承故障诊断.该方法首先对滚动轴承的振动信号进行EEMD模态分解为若干个本征模态函数(intrinsic mode function, IMF)之和,对每个含有故障特征的IMF进行奇异熵、能量熵与峭度值求取;其次,将求得的三种数据输入核主元分析(kernel principal component analysis, KPCA)中进行特征融合与特征提取;最后,将提取的特征作为支持向量机(support vector machine, SVM)的输入参数进行故障分类.试验结果表明此方法能够准确有效地识别出滚动轴承的工作状态,实现了滚动轴承故障分类的自动化.
In view of the characteristics of non-stationary and non-linear vibration signals of rolling bearing faults, an ensemble empirical mode decomposition(EEMD)method is proposed for bearing signal processing.Secondly, the important link of rolling bearing fault diagnosis is feature extracticn, which is directly related to the accuracy of bearing fault diagnosis.Therefore, the entropy knowledge is applied to the bearing featare extraction step.Using the knowledge of singular entropy and energy entropy, a feature extraction method of feature fusion between kurtosis value and the above two entropy is proposed to complete the rolling bearing fault diagnosis.Firstly, the vibration signal of rolling bearing is decomposed into the sum of several intrinsic mode functions(IMF) by EEMD mode decomposition, and the singular entropy, energy entropy and kurtosis value of each IMF containing fault characteristics are calculated.Secondly, the three data obtained are input into kernel principal component analysis(KPCA) for feature fusion and feature extraction.Finally, the extracted features are used as the input parameters of support vector machine(SVM) for fault classification.The experimental results show that this method can accurately and effectively identify the working state of the rolling bearing and realize the automation of the rolling bearing fault classification.
作者
高淑芝
李天池
GAO Shu-zhi;LI Tian-chi(Shenyang University of Chemical Technology,Shenyang 110142,China)
出处
《沈阳化工大学学报》
CAS
2022年第2期151-159,共9页
Journal of Shenyang University of Chemical Technology
基金
辽宁省自然科学基金项目(20170540725)
辽宁省高端人才建设项目-辽宁省特聘教授(〔2018〕3533)。
关键词
滚动轴承
集合经验模态分解
熵
峭度
支持向量机
rolling bearings
ensemble empirical mode decomposition
entropy
kurtosis
support vector machine
