文献类型：期刊文献

中文题名：基于复值小波模和幅角经验正交分解的奇异性特征提取

英文题名：Singular Signal Feature Extraction and Identification Based on Empirical Orthogonal Function for Module Maxim and Phase Matrix of Wavelet

作者：柳桂国[1];韩包海[2];黄道[1]

机构：[1]华东理工大学工业自动化国家工程中心分部,上海200237;[2]浙江工商职业技术学院,浙江宁波315020

年份：2007

卷号：33

期号：6

起止页码：855

中文期刊名：华东理工大学学报（自然科学版）

外文期刊名：Journal of East China University of Science and Technology

收录：CSTPCD;;EI(收录号：20080411059294);Scopus;北大核心:【北大核心2004】；CSCD:【CSCD2011_2012】；

语种：中文

中文关键词：特征提取;Hermitian复值小波;EOF分析;滚动轴承

外文关键词：eature extraction; Hermitian wavelet; empirincal orthogonal function（EOF）analysis;rolling bearings

摘要：提出了基于Hermitian复值小波模和幅角经验正交分解方法,采用这种方法可以提取信号奇异性特征。通过在滚动轴承故障诊断应用表明:小波模和幅角协方差矩阵的特征值向量反映了在时间-尺度平面上的分布结构,不受时间平移影响,便于信号的奇异性特征提取;用主成分重构信号小波模和幅角,能更清晰地反映信号的奇异性特征,便于分类识别.
The singular signal feature extraction and identification based on empirical orthogonal function for module maximum and phase matrix of wavelet is proposed. By using the method, the singular characteristic value vector and the plot of maximum and phase are obtained as the signal feature. The experiments on the application of fault diagnosis for rolling bearings shows that singular characteristic value vector has the merit of time translation invariability, the plot of denoised module maximum and phase matrix of wavelet clearly reflects the singular feature of vibration.