摘要
提出基于核函数主元分析的轴承故障分类方法。该方法通过计算轴承振动信号原始特征空间的核函数来实现原始特征空间到高维特征空间的非线性映射。通过振动测试仪获取轴承在正常、外圈破损和保持架损坏状态下的实验数据,比较主元分析与核函数主元分析的故障分类效果。实验表明,核函数主元分析更适合提取故障信号的非线性特征,对故障特征状态有更好的分类效果,并对分类器有较强的鲁棒性。
An approach to bearing fault classification is presented based on kernel principal component analysis. In this approach, the integral operator kernel functions is used to realize the nonlinear map from the raw feature space of bearing vibration signals to the high dimensional feature space. The experimental data sets in the condition of new ball-bearing, the outer ring completely bro- ken, damaged cage with four loose elements were obtained from vibration testing instrument. The classification effect of Kernel Principal Component Analysis （KPCA） is compared based on the principal component analysis and kernel principal component analysis. The experimental results indicate that the method based on KPCA is more suitable for nonlinear feature extraction from fault signals. It can perform better fault classification ability and robust ness for various classifiers.
出处
《现代制造工程》
CSCD
北大核心
2015年第7期149-153,共5页
Modern Manufacturing Engineering
基金
重庆市基础与前沿研究项目(cstc2013jcyja6002)
重庆市教委科学技术研究项目(KJ1400908)
关键词
轴承
故障分类
核函数主元分析
特征提取
bearing
fault classification
Kernel Principal Component Analysis （KPCA）
feature extraction
