SVD和复合平方包络谱的滚动轴承故障诊断
Rolling Bearing Fault Diagnosis Based on Singular Value Decomposition and Composite Square Envelope Spectrum
陈琳 1陶涛 1李洪强1
作者信息
- 1. 扬州工业职业技术学院智能制造学院,江苏 扬州 225127
- 折叠
摘要
由于传统奇异值分解故障诊断方法难以选择子信号故障点以及Hankel矩阵行数,提出了一种基于奇异值分解和复合平方包络谱的滚动轴承诊断方法.首先分析了两种常用的子信号重构方法的区别.然后引入DR指标来确定奇异值分解序列,并通过数值模拟得到了轴承故障诊断的Hankel矩阵最佳行数范围.由于得到的子信号存在畸变和子信号间能量分布的奇异值分布,通过反对角平均法以及复合平方包络谱对轴承进行故障诊断.最后通过实验证明了提出方法在缺少先验知识情况下能够实现有效的轴承故障诊断.
Abstract
It was difficult to select the fault points of sub signals and the number of rows of Hankel matrix in traditional SVD fault diagnosis methods.A rolling bearing diagnosis method based on singular value decomposition and composite square enve-lope spectrum was proposed.Firstly,the difference between two commonly used sub signal reconstruction methods was analyzed.Then,DR index was introduced to determine the singular value decomposition sequence,and the optimal range of Hankel matrix rows for bearing fault diagnosis was obtained by numerical simulation.Due to the distortion of the sub signals and the singular value distribution of the energy distribution between the sub signals,the anti angle average method and the composite square en-velope spectrum are used to diagnose the bearing fault.Finally,experiments show that the proposed method can achieve effective bearing fault diagnosis without prior knowledge.
关键词
滚动轴承/故障诊断/奇异值分解/平方包络谱Key words
Rolling Bearing/Fault Diagnosis/Singular Value Decomposition/Square Envelope Spectrum引用本文复制引用
基金项目
2020江苏省产学研合作项目(20200673)
出版年
2024
NETL
NSTL
万方数据