Dual least squares support vector machines based spatiotemporal modeling for nonlinear distributed thermal processes. (June 2017)

• Record Type: Journal Article
• Authors: Xu, Kang-Kang; Li, Han-Xiong; Yang, Hai-Dong
• Is Part Of: Journal of process control. Volume 54(2017)
• Page Start: 81
• Page End: 89
• Publication Date: 2017-06
• Subjects: System modeling -- Karhunen-Loève decomposition -- Least squares support vector machines -- Distributed parameter system -- Rademacher complexity; Process control -- Periodicals; Fabrication -- Contrôle -- Périodiques; Process control; Periodicals; Electronic journals; 660.281
• DOI: 10.1016/j.jprocont.2017.03.006
• Languages: English
• ISSNs: 0959-1524

Abstract

Abstract: In this paper, a dual least squares support vector machines (LS-SVM) is proposed to model the thermal process. The infinite-dimensional system is first transformed into a finite-dimensional system through space-time separation. Then, the dual LS-SVM model is to approximate the two nonlinearities embedded in the system. Through space-time synthesis, the dual LS-SVM based spatiotemporal model is able to approximate the complex DPS with inherent coupled nonlinearities. The generalization performance of the proposed model is discussed using Rademacher complexity. Finally, simulations on a curing process demonstrate the effectiveness of the proposed modeling method.