基于高斯分量标准化的K近邻故障检测策略

K-NEAREST NEIGHBOR BASED ON GAUSSIAN COMPONENT STANDARDIZATION FOR FAULT DETECTION STRATEGY

下载PDF

导出

摘要针对复杂多工况工业过程故障检测问题,提出一种基于高斯分量标准化的K近邻(Gaussian Component Standardization K-Nearest Neighbor,GCS-KNN)故障检测策略。样本数据应用高斯混合模型(Gaussian Mixture Model,GMM)进行训练,将数据分解为多个高斯分量;通过每个高斯分量的均值和协方差对该分量内的数据进行标准化处理;应用K近邻(K-Nearest Neighbor,KNN)算法对标准化后的样本进行检测。GCS-KNN通过高斯分量标准化消除数据的多模态特性,提高传统基于KNN检测方法的检测率。利用数值例子和半导体工业过程仿真实验验证了该方法的有效性,并与传统的主元分析(Principal Component Analysis,PCA)、KNN、动态主元分析(Dynamic PCA,DPCA)和加权KNN(Weighted KNN,WKNN)等方法进行对比,结果证实此方法具有显著的优势。 To solve the problem of fault detection in complex industrial processes with multiple operating conditions, this paper proposes a K-nearest neighbor based on Gaussian component standardization(GCS-KNN) fault detection strategy. The sample data was trained by Gaussian mixture model(GMM), and the data was decomposed into multiple Gaussian components. The data in each Gaussian component was standardized by the mean and covariance of the component. K-nearest neighbor(KNN) algorithm was used to detect the standardized samples. GCS-KNN eliminated the multimodality of data through Gaussian component standardization, and improved the detection rate of traditional KNN-based detection methods. The effectiveness of this method was verified by numerical examples and semiconductor industrial process simulation experiments, and compared with traditional principal component analysis(PCA), KNN, dynamic PCA(DPCA) and weighted KNN(WKNN) method. The results show that this method has significant advantages.

作者张成赵丽颖郑百顺戴絮年李元 Zhang Cheng;Zhao Liying;Zheng Baishun;Dai Xunian;Li Yuan(Department of Science,Shenyang University College of Science,Shenyang 110142,Liaoning,China;Research Center for Technical Process Fault Diagnosis and Safety,Shenyang University of Chemical Technology,Shenyang 110142,Liaoning,China)

机构地区沈阳化工大学理学院沈阳化工大学技术过程故障诊断与安全性研究中心

出处《计算机应用与软件》北大核心 2023年第1期90-97,共8页 Computer Applications and Software

基金国家自然科学基金项目(61673279) 辽宁省自然科学基金项目(2019-MS-262) 辽宁省教育厅基金项目(LJ2019013)。

关键词高斯混合模型多模态故障检测 K近邻规则标准化半导体蚀刻过程 Gaussian mixture model Multimode fault detection K-nearest neighbor rule Standardization Semiconductor etching process

分类号 TP3 [自动化与计算机技术—计算机科学与技术]