基于EMD的旋转机械振动信号Winger分布分析

The EMD-based Vibration Signals' Winger Distribution in Rotating Machinery

经验模态分解 (EmpiricalModeDecomposition)简称EMD ,主要思想是把一个时间序列的信号 ,分解成不同尺度的本征模函数 (IntrinsicModeFunction ,IMF)。把EMD和Winger分布分析相结合的时频分析方法引入了旋转机械振动信号分析领域。通过把振动信号序列进行EMD分解 ,然后对每个分解后的IMF进行Winger分布分析 ,可取得抑制频率干扰的效果 ,使时频谱图更清晰。首先 ,对一个有两个频率成分的仿真振动信号Winger分布时频图和信号经过EMD和Winger分布分析相结合产生的时频图进行对比。然后 ,对旋转机械油膜涡动故障振动信号进行同样的对比。仿真信号和真实信号的研究结果说明 ,用EMD和Winger分布分析相结合的时频分析方法对旋转机械的振动信号的时频分析比通常的Winger分布分析有效。

The main idea of empirical mode decomposition(EMD) is separating the series data into components with different scale, say intrinsic mode function(IMF). The EMD-based vibration signals' Winger distribution analysis was introduced to the field of vibration signals' analysis in rotating machinery. Firstly, separate the series data are seperated into intrinsic mode functions(IMFs) using empirical mode decomposition(EMD),Then,the Winger distribution transformation is applied to every IMFs. Last, every transformed IMFs is assembled into one time-frequency spectrum. An oil film whirling vibration signal and its emulation were used to compare the difference between normal Winger distribution and the introduced new method. The result show that the new method is better than the normal's.