基于小波包分解和PCA的轴承故障诊断  

A Wavelet Packet Decomposition and Principal Component Analysis Approach for Feature Extraction in Bearing Failure Vibration Signal

作　　者：杜振宁[1] 向春枝[2] 

机构地区：[1]杨凌职业技术学院信息工程学院,陕西杨凌712100 [2]河南广播电视大学理工学院,郑州450008

出　　处：《控制工程》2016年第6期812-815,共4页Control Engineering of China

摘　　要：轴承产生故障后产生非线性振动信号,传统的特征提取方式在非线性特征提取和非线性关系可视化上存在不足。提出了小波包分解和高阶累积量对振动信号进行特征提取,通过主成分分析法对特征数据进行了降维处理。该特征提取方式不仅可以揭示特征量之间的非线性关系,而且有利于提高分类速度和准确性;采用神经网络算法进行了故障分类。测试结果表明,该方法可以准确有效的识别出滚动轴承的故障类型。Bear failure could lead to nonlinear vibration signals. Traditional feature extraction methods have shortcomings in nonlinear feature extraction and data visualization. In this paper, wavelet packet decomposition and principal component analysis based approach for feature extraction in bear failure vibration signals is presented. Principle component analysis method is used to reduce the dimension of the feature data. This feature extraction method can not only reveal the characteristics of the non-linear relationship between amounts of features, but also help to improve the speed and accuracy of classification. Neural network algorithm fed by the features is for fault classification. Results show that our method can accurately and efficiently identify the type of bearing failures.

关 键 词：非线性信号 小波包分解 主成分分析方法 神经网络算法 

分 类 号：TH133.3[机械工程—机械制造及自动化]