Design, implementation and simulation of an MPC algorithm for switched nonlinear systems under combinatorial constraints. (September 2019)

• Record Type: Journal Article
• Authors: Bürger, Adrian; Zeile, Clemens; Altmann-Dieses, Angelika; Sager, Sebastian; Diehl, Moritz
• Is Part Of: Journal of process control. Volume 81(2019)
• Page Start: 15
• Page End: 30
• Publication Date: 2019-09
• Subjects: Model predictive control -- Switched dynamic systems -- Mixed-integer nonlinear programming -- Optimal control -- Approximation methods and heuristics; Process control -- Periodicals; Fabrication -- Contrôle -- Périodiques; Process control; Periodicals; Electronic journals; 660.281
• DOI: 10.1016/j.jprocont.2019.05.016
• Languages: English
• ISSNs: 0959-1524

Abstract

Highlights: We present a warm-started MPC algorithm for constrained switched nonlinear systems. Combinatorial constraints such as dwell times and maximum switchings can be treated. The utilized decomposition approach for MINLP solution is generally applicable. The introduced software package pycombina facilitates easier use of the algorithm. Performance is assessed on a simulation study for a solar thermal climate system. Abstract: Within this work, we present a warm-started algorithm for Model Predictive Control (MPC) of switched nonlinear systems under combinatorial constraints based on Combinatorial Integral Approximation (CIA). To facilitate high-speed solutions, we introduce a preprocessing step for complexity reduction of CIA problems, and include this approach within a new toolbox for solution of CIA problems with special focus on MPC. The proposed algorithm is implemented and utilized within an MPC simulation study for a solar thermal climate system with nonlinear system behavior and uncertain operation conditions. The results are analyzed in terms of solution quality, constraint satisfaction and runtime of the solution steps, showing the applicability of the proposed algorithm and implementations.