摘要
针对滚动轴承振动信号的低信噪比、高复杂性及非平稳特性,提出基于经验模态分解、多尺度熵算法与支持向量机的故障诊断方法。对振动信号通过小波包降噪提高信噪比,然后利用经验模态分解得到多个本征模态函数分量,选择与降噪信号强相关的本征模态函数分量计算其多尺度样本熵,确认能区分故障类型的最佳尺度。将这一尺度下相应分量的样本熵作为特征向量,经过归一化处理后输入支持向量机进行故障分类。试验结果表明在小样本条件下可以准确识别滚动轴承故障类型,为滚动轴承的故障识别提供了一种高效诊断方法。
Aiming at the low SNR, high complexity and non-stationary characteristics of vibration signals of rolling bearings, a fault diagnosis method based on EMD, MSE and SVM was proposed. The SNR of the vibration signal is improved by wavelet packet noise reduction, and then several IMF components are obtained via EMD. The IMF component strongly correlated with the noise reduction signal is selected to calculate its SME in order to confirm the best scale that can distinguish the fault type. The sample entropy of the corresponding component of this scale is used as the feature vector, and the input SVM after normalization is used for fault classification. The test results show that the fault type of rolling bearing can be accurately identified under the condition of small sample, which provides an efficient diagnosis method for the fault identification of rolling bearings.
出处
《机械制造》
2018年第4期78-83,共6页
Machinery
作者简介
张文哲（1994-），男，硕士研究生，主要研究方向为机械制造及其自动化.;张为民（1965-），男，教授，主要研究方向为机械制造及其自动化。
