针对滚动轴承故障诊断的新时频特征提取方法

发布时间：2018-03-26 00:13

  本文选题：故障诊断　切入点：时频特征　出处：《机械传动》2016年07期

【摘要】：针对滚动轴承的故障诊断,提出了一种基于词包模型和短时傅里叶变换的特征提取方法。根据轴承故障的产生机理,不同轴承的振动信号在频域上会有相应的能量分布规律,然而在实际现场中,信号干扰或者生产环境等因素会弱化这种规律性,使得在频谱上难以准确看到相应分布特征。当采用词包模型时,把每一时间帧下能量在频率维度上的分布看成一个单词,则每段信号就表示成了由各个单词组成的一篇篇文档,这就可以直接从数据的角度去揭示能量分布的这种规律性。然后,以词包模型处理后的结果作为特征向量,用SVM分类算法诊断出结果。最后用无锡某汽车生产线SQI-MFS实验平台和美国凯斯西储大学的轴承振动数据进行了实验,实验验证了该方法比时域特征(RMS)和时频域特征(WEWEE)的诊断结果精确,可以在滚动轴承故障诊断领域展开应用。
[Abstract]:According to the fault diagnosis of rolling bearing, a feature extraction method based on word packet model and short time Fourier transform is proposed. According to the mechanism of bearing fault, the vibration signal of different bearing will have corresponding energy distribution law in frequency domain. However, in the actual field, the signal interference or the production environment will weaken the regularity, which makes it difficult to see the corresponding distribution characteristics accurately in the spectrum. Considering the distribution of energy in the frequency dimension of each time frame as a word, each signal is represented as a document of each word. This can reveal the regularity of energy distribution directly from the point of view of data. The diagnosis results are obtained by using SVM classification algorithm. Finally, the experimental data of bearing vibration of a certain automobile production line in Wuxi and the bearing vibration data of case Western Reserve University in the United States are carried out. The experimental results show that this method is more accurate than that of time domain feature (RMS) and time-frequency domain feature (WEWEE), and can be applied in the field of rolling bearing fault diagnosis.
【作者单位】： 江南大学自动化研究所;西门子中国研究院;
【基金】：国家自然科学基金(61104121,61202211) 江南大学自主科研计划重点项目(JUSRP51407B)
【分类号】：TH133.33
，

本文编号：1665486