基于小波包层次熵的电主轴振动信号特征提取方法

Feature extraction method of motorized spindle vibration signal based on wavelet packet transform and hierarchical entropy

为了对电主轴运转时振动信号特征进行更有效的提取,提出了基于小波包层次熵的电主轴特征提取方法.首先对电主轴振动信号利用小波包变换进行分解、重构,得到若干节点信号;然后计算不同节点信号的样本熵值,构成其小波包层次熵值;最后通过小波包层次熵值提取电主轴振动信号特征.计算结果表明:与傅里叶变换方法相比,基于小波包层次熵的分析方法不仅考虑了时间序列的低频成分,同时也考虑了其高频成分,可有效地提高特征提取的准确率,更精确和完整地描述电主轴振动信号的特征,为提取电主轴振动信号特征提供了一种快速有效的新方法.

In order to extract the characteristics of the vibration signal of motorized spindle more effectively,an extraction approach was proposed based on wavelet packet and hierarchical entropy.Firstly,the vibration signals of the motorized spindle were decomposed and reconstructed with wavelet packet transform to obtain several nodal signals.Then the sample entropies of different nodes were calculated to form their wavelet packet hierarchical entropy.Finally,according to the entropy value,the characteristics of vibration signal were extracted.Their calculation result showed that compared to the Fourier transformation method,the proposed analysis method not only the low frequency component of the time series would take into account,but also the high frequency component,so that the accuracy of feature extraction could be improved,the characteristics of motorized spindle vibration signals could be described more precisely and completely,and a fast and valid novel method for extraction of vibration signal characteristics of motorized spindle would be offered.