自适应奇异值分解的随机共振提取微弱故障特征

发布时间：2018-05-27 07:18

本文选题：振动 + 农业机械　；参考：《农业工程学报》2017年11期

【摘要】：针对农业机械设备在强背景噪声下微弱故障特征难以提取的问题,提出一种基于自适应奇异值分解的随机共振微弱故障特征提取方法。首先,将原始信号奇异值分解并重构得到分量信号,构建互信息差分谱,权衡各分量信号对原始信号的贡献率,自适应选取有效奇异值个数,以克服已有方法人为主观选择或仅考虑奇异值大小等不足;其次,对选取的有效奇异值对应的分量信号自适应随机共振,使其微弱故障特征增强;最后,对增强的分量信号统计学平均以提取微弱故障特征。仿真和轴承外圈故障试验结果表明,该方法不仅克服了强背景噪声下有效奇异值的选取困难,而且结合自适应随机共振,有效提取出仿真信号100 Hz和轴承外圈155.5 Hz的故障特征频率,因此,所提方法不仅能够更好的增强微弱故障特征,而且分析结果优于单纯的奇异值分解和随机共振方法。该文提出的方法不仅可适用于强噪声背景下轴承的故障诊断,同时为农业机械设备的轴承故障诊断提供参考。
[Abstract]:In order to solve the problem that it is difficult to extract the weak fault feature of agricultural machinery and equipment under strong background noise, a new method based on adaptive singular value decomposition (SVD) is proposed to extract the weak fault feature of stochastic resonance. Firstly, the singular value of the original signal is decomposed and reconstructed to obtain the component signal, and the difference spectrum of mutual information is constructed. The contribution rate of each component signal to the original signal is weighed, and the number of effective singular values is selected adaptively. In order to overcome the shortcomings of existing methods, such as subjective selection or only considering the size of singular value; secondly, adaptive stochastic resonance of the component signal corresponding to the selected effective singular value enhances the weak fault feature. The enhanced component signals are statistically averaged to extract weak fault features. The results of simulation and bearing outer ring fault test show that this method not only overcomes the difficulty of selecting effective singular value under strong background noise, but also combines with adaptive stochastic resonance. The fault characteristic frequency of simulation signal 100Hz and bearing outer ring 155.5 Hz is extracted effectively. Therefore, the proposed method can not only enhance the weak fault feature better, but also outperform the singular value decomposition (SVD) and stochastic resonance (SR) method. The method presented in this paper is not only suitable for bearing fault diagnosis in strong noise background, but also provides a reference for bearing fault diagnosis of agricultural machinery and equipment.
【作者单位】：北京科技大学机械工程学院;内蒙古科技大学机械工程学院;
【基金】：国家自然科学基金资助项目(51075029)
【分类号】：O324;S22

【相似文献】