An extremum-seeking control method driven by input–output correlation. (October 2017)

• Record Type: Journal Article
• Authors: Salsbury, Timothy I.; House, John M.; Alcala, Carlos F.; Li, Yaoyu
• Is Part Of: Journal of process control. Volume 58(2017)
• Page Start: 106
• Page End: 116
• Publication Date: 2017-10
• Subjects: Extremum-seeking control -- Optimization -- HVAC -- Buildings -- Energy systems -- Feedback control; Process control -- Periodicals; Fabrication -- Contrôle -- Périodiques; Process control; Periodicals; Electronic journals; 660.281
• DOI: 10.1016/j.jprocont.2017.06.015
• Languages: English
• ISSNs: 0959-1524

Abstract

Highlights: Traditional ESC is interpreted from a statistical perspective. Correlation coefficient is proposed as the ESC feedback variable. Tuning requirements are minimized to just two parameters. Tests are presented on simulations of simple and complex systems. Full algorithm is presented in a simple form that would be easy for practitioners to implement. Abstract: This paper presents a new extremum-seeking controller (ESC) derived from a statistical interpretation of conventional ESC. The proposed ESC replaces the gradient feedback variable with a normalized correlation coefficient. The algorithm, as presented, can be fully configured with knowledge of only the open-loop response time of the system being optimized and the amplitude of the dither signal. Simulation tests show that the new algorithm has fast convergence times compared with other types of ESC algorithms. The paper also shows that the proposed method is not limited to a periodic dither signal and that it can also utilize a stochastic signal to similar efficacy. The purpose of the paper is to describe the new algorithm and present a practical implementation along with results from simulation tests.