基于小波包和KPCA的时频域故障检测方法  

Time and frequency domain fault detection method based on wavelet packet and KPCA

作　　者：吴胜强[1,2] 姜万录[1] 刘思远[1] 

机构地区：[1]燕山大学机械工程学院,河北秦皇岛066004 [2]邢台职业技术学院机电工程系,河北邢台054035

出　　处：《沈阳工业大学学报》2011年第2期172-176,187,共6页Journal of Shenyang University of Technology

基　　金：国家自然科学基金资助项目(50775198;51075349);河北省自然科学基金资助项目(E2008000812)

摘　　要：针对故障检测技术中存在的非线性和信息遗漏问题,在深入分析核主元分析法的基础上,提出了一种新的基于小波包和核主元分析法(KPCA)的时频域故障检测方法.利用小波包对原始信号进行预处理,提取包含时域和频域特征参数构成的特征向量,应用KPCA进行故障检测,同时对液压泵也进行了故障检测.试验结果表明,时域和频域特征参数构成的特征向量很好地反映了故障的特征,与PCA相比,KPCA的主元数目可选择范围宽,该方法对液压泵故障检测有良好的效果.For the nonlinear problem and information missing in fault detection,a new time and frequency domain fault detection method based on wavelet packet and kernel principal component analysis（KPCA） was proposed.The original signal is preprocessed with wavelet packet,the eigenvector containing both time-domain and frequency-domain characteristic parameters was extracted,and the fault was detected using KPCA.At the same time,the fault detection was also performed for the hydraulic pump.The experimental results show that the eigenvector can better reflect the fault features,and KPCA is able to pickup more principal components than PCA.The proposed method is valid to the fault detection of the hydraulic pump.

关 键 词：故障检测 核主元分析 核主元模型 主元数目 小波包 时频域 故障特征提取 液压泵 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]