详细信息
文献类型:期刊文献
中文题名:基于KPCA-SVC的复杂过程故障诊断
英文题名:Fault diagnosis of complex chemical process based on KPCA-SVC
作者:刘爱伦[1];袁小艳[1];俞金寿[1]
机构:[1]华东理工大学自动化研究所,上海200237
年份:2007
卷号:28
期号:5
起止页码:870
中文期刊名:仪器仪表学报
外文期刊名:Chinese Journal of Scientific Instrument
收录:CSTPCD;;EI(收录号:20072410652944);Scopus;北大核心:【北大核心2004】;CSCD:【CSCD2011_2012】;
语种:中文
中文关键词:核主元分析(KPcA);支持向量机分类(SVC);故障诊断
外文关键词:kernel principal component analysis (KPCA) ; support vector classification (SVC) ; fault diagnosis
摘要:本文提出了一种将核主元分析方法与支持向量机分类相结合进行故障诊断的方法,运用该方法对连续搅拌釜式反应器(CSTR)进行实时的故障诊断,实验结果表明KPCA-SVC故障诊断方法既充分利用了KPCA的特征提取能力和SVC的良好的分类能力,又避免了复杂的计算,有利于提高故障诊断模型的实时性。
Support vector machine (SVM) is an effective fault diagnosis method, but a number of data may lead to a more complicated structure of SVM classifier. By integrating the characteristics of kernel principal component analysis (KPCA) and SVM, a new fault diagnosis method is presented in this paper. The new method was applied to continuous stirred tank reactor(CSTR) model and the results show that this new method avoids complex computation and improves the real-time property of the fault diagnosis model.
参考文献:
正在载入数据...
