Journal of Computer Applications ›› 2018, Vol. 38 ›› Issue (8): 2185-2191.DOI: 10.11772/j.issn.1001-9081.2018020345

Previous Articles     Next Articles

Batch process monitoring based on k nearest neighbors in discriminated kernel principle component space

ZHANG Cheng, GUO Qingxiu, LI Yuan   

  1. Research Center for Technical Process Fault Diagnosis and Safety, Shenyang University of Chemical Technology, Shenyang Laoning 110142, China
  • Received:2018-02-07 Revised:2018-03-19 Online:2018-08-10 Published:2018-08-11
  • Supported by:
    This work is partially supported by the National Natural Science Foundation of China (61490701, 61573088, 61673279), the General Project of Liaoning Provincial Department of Education (L2015432), the Natural Science Foundation of Liaoning Province (2015020164).


张成, 郭青秀, 李元   

  1. 沈阳化工大学 技术过程故障诊断与安全性研究中心, 沈阳 110142
  • 通讯作者: 李元
  • 作者简介:张成(1979-),男,辽宁沈阳人,副教授,博士研究生,主要研究方向:过程故障诊断分析;郭青秀(1992-),女,内蒙古呼和浩特人,硕士研究生,主要研究方向:过程故障检测;李元(1964-),女,辽宁沈阳人,教授,博士,主要研究方向:基于数据驱动复杂过程故障诊断。
  • 基金资助:

Abstract: Aiming at the nonlinear and multi-mode features of batch processes, a fault detection method for batch process based on k Nearest Neighbors (kNN) rule in Discriminated kernel Principle Component space, namely Dis-kPCkNN, was proposed. Firstly, in kernel Principal Component Analysis (kPCA), according to discriminating category labels, the kernel window width parameter was selected between within-class width and between-class width, thus the kernel matrix can effectively extract data correlation features and keep accurate category information. Then kNN rule was used to replace the conventional T2 statistical method in the kernel principal component space, which can deal with fault detection of process with nonlinear and multi-mode features. Finally, the proposed method was validated in the numerical simulation and the semiconductor etching process. The experimental results show that the kNN rule in discriminated kernel principle component space can effectively deal with the nonlinear and multi-mode conditions, improve the computational efficiency and reduce memory consumption, in addition, the fault detection rate is significantly better than the comparative methods.

Key words: discriminated kernel principle component, k Nearest Neighbors (kNN), batch process, fault detection, semiconductor

摘要: 针对批次过程非线性、多模态等特征,提出一种基于判别核主元k近邻(Dis-kPCkNN)的故障检测方法。首先,在核主元分析(kPCA)中,高斯核的窗宽参数依据样本类别标签在类内窗宽和类间窗宽中判别选取,使得核矩阵能有效提取数据的关联特征,保持数据的类别信息;其次,在核主元空间中引用k近邻规则代替传统的T2统计方法,k近邻规则可以有效处理主元空间非线性和多模态等特征的故障检测问题。数值模拟实例和半导体蚀刻工艺过程仿真实验表明:基于判别核主元k近邻方法可以有效地处理具有非线性和多模态结构特征的故障检测问题,提高计算的效率,减少内存的占用,并且故障检测率明显优于传统方法。

关键词: 判别核主元, k近邻, 批次过程, 故障检测, 半导体

CLC Number: