基于IITD和奇异值差分谱的滚动轴承故障诊断

发布时间：2018-04-22 13:39

本文选题：固有时间尺度分解 + 固有旋转分量　；参考：《矿山机械》2016年02期

【摘要】：基于ITD方法的线性变换和Akima插值,提出了一种改进的固有时间尺度分解方法 (Improve Intrinsic Time-scale Decomposition,简称IITD),将该方法与奇异值差分谱相结合,实现了滚动轴承故障的精确诊断。首先通过IITD方法将非平稳的原始加速度振动信号分解成若干个平稳的固有旋转分量(Proper Rotation Component,简称PRC);然后挑选包含故障特征信息最丰富的PR分量作为主PR分量;构造主PR分量的Hankel矩阵并进行奇异值分解,得到相应的奇异值差分谱,选择奇异值差分谱中的最大突变点来确定重构信号的奇异值个数,进而得到降噪后的主PR分量;最后对降噪后的主PR分量进行包络解调分析,提取滚动轴承的故障特征。实例分析表明,相比传统包络谱分析和基于经验模态分解和奇异值差分谱的方法,该方法能更有效地提取出滚动轴承的故障特征。
[Abstract]:Based on the linear transformation and Akima interpolation of the ITD method, an improved Intrinsic Time-scale decomposition method is proposed. The method is combined with the singular value difference spectrum to realize the accurate diagnosis of the rolling bearing fault. At first, the non-stationary original acceleration vibration signal is decomposed into several stationary inherent rotational components (Proper Rotation component) by IITD method, and then the PR component, which contains the most abundant fault characteristic information, is selected as the principal PR component. The Hankel matrix of the main PR component is constructed and the singular value difference spectrum is obtained. The number of singular values of the reconstructed signal is determined by selecting the maximum mutation point in the singular value difference spectrum, and then the principal PR component after the noise reduction is obtained. Finally, the main PR component after noise reduction is analyzed by envelope demodulation, and the fault feature of rolling bearing is extracted. Compared with the traditional envelope spectrum analysis and the method based on empirical mode decomposition and singular value differential spectrum, the method can extract the fault characteristics of rolling bearing more effectively.
【作者单位】：湖南信息职业技术学院;中南大学交通运输工程学院;
【基金】：湖南省科技计划项目“新型螺旋转子参数化设计与数字化制造技术研究”(2013FJ3019)
【分类号】：TH133.33

【相似文献】