文献类型：期刊文献

中文题名：基于改进PCA的故障检测与诊断方法

英文题名：A Method of Fault Detection and Diagnosis Based on Wavelet De-noising and Principal Component Analysis

作者：赵成燕[1];刘爱伦[1]

机构：[1]华东理工大学自动化研究所,上海200237

年份：2006

卷号：42

期号：1

起止页码：41

中文期刊名：石油化工自动化

外文期刊名：Automation in Petro-chemical Industry

语种：中文

中文关键词：主元分析;小波去噪;故障检测;故障诊断

外文关键词：principal component analysis（PCA） ;wavelet de-noise; fault detections fault diagnosis

摘要：提出了一种基于小波去噪和主元分析的故障检测和诊断方法。该方法利用小波分析先对正常工况下的数据进行处理,然后运用T2统计、Q统计方法,结合主元得分图和变量贡献图对一模型进行了仿真分析,结果表明,该方法是有效的。
A method of fault detection and diagnosis based on wavelet de-noise and principal component analysis is proposed. The data collected from the normal industry condition are processed by means of the wavelet analysis. The fault detection and diagnosis simulation to a model is performed by means of statistical method like Hotelling T^e and Q. The principal component scores charts and variables contribution charts are used to undertake fault diagnosis. Simulation results show that it is fairly effective.