Toggle Main Menu Toggle Search

Open Access padlockePrints

Randomized Kernel Principal Component Analysis for Modeling and Monitoring of Nonlinear Industrial Processes with Massive Data

Lookup NU author(s): Dr Jie ZhangORCiD



This is the authors' accepted manuscript of an article that has been published in its final definitive form by American Chemical Society, 2019.

For re-use rights please refer to the publisher's terms and conditions.


Kernel principal component analysis (KPCA) has shown its excellent performance in monitoring nonlinear industrial processes. However, model building, updating, and online monitoring using KPCA are generally time-consuming when massive data are obtained under the normal operation condition (NOC). The main reason is that the eigen-decomposition of high-dimensional kernelmatrix constructed from massive NOC samples is computationally complex. Many studies have been devoted to solving this problem through reducing the NOC samples, but KPCA model constructed from the reduced sample set cannot ensure good performance in monitoring nonlinear industrial processes. The performance of KPCA model depends on whether the results of eigen-decomposition of the reduced kernel matrix can well approximate that of the original kernel matrix. To improve the efficiency of KPCA based process monitoring, this paper proposes randomized KPCA for monitoring nonlinear industrial processes with massive data. The proposed method uses random sampling to compress kernel matrix into a subspace which maintains most of the usefulinformation about process monitoring. Then, the reduced kernel matrix is operated to obtain desired eigen-decomposition results. Based on these approximated eigendecomposition results, the proposed randomized KPCA can enhance the performance in monitoring nonlinear industrial processes. This is because the commonly used monitoring statistics are related to the eigenvalues and eigenvectors of kernel matrix. Finally, numerical simulation and the benchmark TE chemical process are used to demonstrate the effectiveness of the proposed method.

Publication metadata

Author(s): Zhou Z, Du N, Xu J, Li Z, Wang P, Zhang J

Publication type: Article

Publication status: Published

Journal: Industrial & Engineering Chemistry Research

Year: 2019

Volume: 58

Issue: 24

Pages: 10410-10417

Print publication date: 19/06/2019

Online publication date: 28/05/2019

Acceptance date: 28/05/2019

Date deposited: 30/05/2019

ISSN (print): 0888-5885

ISSN (electronic): 1520-5045

Publisher: American Chemical Society


DOI: 10.1021/acs.iecr.9b00300


Altmetrics provided by Altmetric


Funder referenceFunder name