基于多核支持向量机的多模态过程故障检测  

Multi-modal Process Fault Detection Based on Multiple Kernel Support Vector Machine

作　　者：李元 李榕 LI Yuan;LI Rong(College of Information Engineering,Shenyang University of Chemical Technology)

机构地区：[1]沈阳化工大学信息工程学院

出　　处：《化工自动化及仪表》2022年第3期286-293,共8页Control and Instruments in Chemical Industry

基　　金：国家自然科学基金项目(61673279)。

摘　　要：针对具有多模态、非线性特征的复杂分布数据的工业过程,提出一种基于局部相对概率密度(LRPD)的多核支持向量机(MKSVM)故障检测方法LRPD-MKSVM。首先,计算训练样本的局部概率密度矩阵并进行标准化处理,来消除数据的多模态特性;其次,运用标准化后的概率密度矩阵训练多核SVM模型,获得判别分类函数;之后,将测试数据的概率密度矩阵作为多核SVM模型的输入,对其进行分类;最后,将该方法应用于TE多模态工业过程,分别与基于单核的高斯核函数SVM(RBFSVM)、多项式核函数SVM(POLYSVM)分类方法对比分析,结果表明:基于多核SVM方法的分类正确率明显优于单核SVM方法。Aiming at the industrial process which boasting of complex distributed data and multi-modal and nonlinear characteristics,a multiple kernel support vector machine(MKSVM)fault detection method based on local relative probability density(LRPD)was proposed.Firstly,having local probability density matrix of the training sample calculated and standardized to eliminate multi-modal characteristics of the data;secondly,having the standardized probability density matrix adopted to train multiple kernel SVM model so as to obtain the discriminant classification function;and then,having the probability density matrix of the test data taken as the input of the multiple kernel SVM model and to classify it;finally,having this method applied to the Tennessee-Eastman multi-modal industrial process,including having it compared with the single kernel Gaussian kernel function SVM(RBFSVM)and polynomial kernel function SVM(POLYSVM)classification method.The results show that,the classification accuracy rate based on the multiple kernel SVM significantly outperforms that of the single kernel SVM.

关 键 词：故障检测 复杂工业过程 非线性 多核支持向量机 TE多模态数据集 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]