摘要
为了有效地提高支持向量机(SVM)在工业过程中的故障检测性能,提出一种基于滑动窗口的核熵成分分析(KECA)和支持向量机(SVM)结合(MWKECA-SVM)的非线性过程故障检测方法。运用核熵成分分析(KECA)提取包含非线性特征信息的得分向量作为SVM的输入。运用正常和故障数据的非线性特征向量训练SVM模型获得判别分类函数。建立模型之后,运用滑动窗口对模型进行动态更新。将MWKECA-SVM方法应用于田纳西-伊斯曼过程中,并与核主元分析(KPCA)、滑动窗口KPCA(MWKPCA)、KECA和SVM方法进行比较。结果表明,MWKECA-SVM方法能够动态地提取过程变量的特征信息,有效地提高故障检测率,在一定程度上增强了信息的动态提取和实时监控能力。
To effectively improve the fault detection performance of support vector machine(SVM)in industrial process,a fault detection method of nonlinear process based on moving window kernel entropy component analysis and support vector machine(MWKECA-SVM)was proposed.The kernel entropy component analysis(KECA)was used to extract the score vector containing the nonlinear feature information as the input of SVM.The nonlinear feature vectors of normal and fault data were used to train the SVM model to obtain the discriminant classification function.After the model was built,the moving window was used to dynamically update the model.The MWKECA-SVM method was applied to the Tennessee-Eastman process and compared with the kernel principal component analysis(KPCA),moving window KPCA(MWKPCA),KECA and SVM methods.The results showed that the MWKECA-SVM method could dynamically extract the feature information of process variables,effectively improve the fault detection rate and enhance the ability of dynamic extraction and real-time monitoring of information to a certain extent.
作者
郭金玉
李涛
李元
GUO Jinyu;LI Tao;LI Yuan(College of Information Engineering, Shenyang University of Chemical Technology, Shenyang 110142, China)
出处
《沈阳大学学报(自然科学版)》
CAS
2022年第1期37-44,共8页
Journal of Shenyang University:Natural Science
基金
辽宁省教育厅资助项目(LJ2019007)。
关键词
滑动窗口
核熵成分分析
支持向量机
非线性过程
故障检测
实时监控
moving window
kernel entropy component analysis
support vector machine
nonlinear process
fault detection
real-time monitoring