基于改进奇异值分解技术的齿轮调制故障特征提取

An improved method of detecting modulated gear fault characteristic based on singularity value decomposition

在介绍现有奇异值分离技术基本原理及其在故障诊断中的应用的基础上,研究了信号时间序列重构的吸引子轨迹矩阵奇异值分布与信号特征的关系,引入了自相关分析及信号叠加技术,改进了现有算法,使得吸引子轨迹矩阵的重构科学合理。研究表明:该方法能在强噪声背景下提取出所需的调制信号,并成功用于实测齿轮调制故障信号的提取,为齿轮箱故障诊断提供了一个新的思路。

The modulated fault characteristic is one of the most familiar gear faults, and the relevant research may decide the success of gear fault diagnosis. In this paper, the fundamental theory and application of singularity value decomposition in fault diagnosis are introduced. Based on the relations between the singular value distribution of the time series track matrix of attractor and the signal characteristics, the autoeorrelation analysis and signal stacking are introduced, which improve the existing method and make the method more logical in reconstructing the track matrix of attractor. The research shows that the improved method is successful in detecting the modulated fault signals of gearbox, even with strong noise, which provides a new idea on gear fault diagnosis.