Decentralized Fault Diagnosis of Large-Scale Processes Using Multiblock Kernel Partial Least Squares

• Published in: IEEE transactions on industrial informatics 2010-02, Vol.6 (1), p.3-10
• Main Authors: Yingwei Zhang, Hong Zhou, Qin, S.J., Tianyou Chai
• Format: Article
• Language: English
• Subjects: Algorithms; Artificial neural networks; Chemical industry; Covariance matrix; Decentralized; Fault diagnosis; Kernel; Kernels; Large-scale systems; Least squares method; Least squares methods; Monitoring; Monitors; multiblock kernel partial least squares (MBKPLS); Neural networks; nonlinear component analysis; Nonlinearity; Principal component analysis; process monitoring
• ISSN: 1551-3203; 1941-0050
• DOI: 10.1109/TII.2009.2033181

In this paper, a decentralized fault diagnosis approach of complex processes is proposed based on multiblock kernel partial least squares (MBKPLS). To solve the problem posed by nonlinear characteristics, kernel partial least squares (KPLS) approaches have been proposed. In this paper, MBKPLS algorithm is first proposed and applied to monitor large-scale processes. The advantages of MBKPLS are: 1) MBKPLS can capture more useful information between and within blocks compared to partial least squares (PLS); 2) MBKPLS gives nonlinear interpretation compared to MBPLS; 3) Fault diagnosis becomes possible if number of sub-blocks is equal to the number of the variables compared to KPLS. The proposed methods are applied to process monitoring of a continuous annealing process. Application results indicate that the proposed decentralized monitoring scheme effectively captures the complex relations in the process and improves the diagnosis ability tremendously.