基于全矢局部均值分解的齿轮故障诊断方法研究

Research of Gear Fault Diagnosis Method Based on Full Vector of Local Mean Decomposition

针对齿轮故障信号大多数是难分解的多分量的调幅-调频信号的问题,提出一种新的信号处理方法——全矢局部均值分解(FVLMD)方法。局部均值分解(LMD)可将多分量信号自适应地分解为多个单分量信号;全矢谱技术可以解决单通道信号不完整的问题。运用信息融合技术,将信号LMD分解得到的PF(Product Function)分量进行全矢谱融合分析,这样既可以将信号彻底分解,又可以保证其完整性。齿轮故障信号验证了该方法的有效可行性。

Aimed at the problem of most of the gear fault signals are the multi-component AM-FM signals which are very difficult to decompose, a new signal processing method of the full vector of local mean decomposition （FVLMD） was proposed. A multi-compo- nent signal was adaptively decomposed into a complex of series of single-component signals by using the local mean decomposition （LMD）. The problem which single-channel signal was usually not complete was solved by the full vector spectrum technology. By using the information fusion technology, the PF （ Product Function） component decomposed by LMD was fused with analysis by the full vec- tor spectrum. So the signals can not only be decomposed completely, but also their integrity can be ensured. Its effectiveness and feasi- bility are verified through the gear fault signals.