(243f) Decompositions to Handle Disparate Time-Scales in Economic Model Predictive Control

Economic model predictive control (EMPC) is an advanced predictive control technique that combines economic optimization and feedback control (e.g., [1], [3]). While there have been some reported EMPC systems that have been developed and deployed in industry [6], industrial application of EMPC to large scale systems is still difficult. One particular challenge faced in large scale applications is formulating the EMPC problem in a computationally efficient manner to cope with multiple time-scales. In particular, process/system economics and dynamics are often of disparate time-scales with the economics accounting for a much larger time scale than the control time scale. To address this, EMPC typically requires a large prediction horizon. Additionally, process/system dynamics may evolve over multiple time scales. Addressing all these time-scales in a single centralized EMPC system requires a model that is suitable across all time-scales. The resulting centralized problem may be very computationally expensive to solve as it requires both fine grain resolution for the shorter time scales and a horizon required to capture longer time scale effects.

A hierarchical EMPC framework capable of addressing disparate time-scales is proposed. The main component of the algorithm framework is the design and use of temporal decomposition strategies (e.g., singularly perturbed modeling and hierarchical control concepts). To perform the decomposition over a long horizon, a temporal decomposition strategy is adapted based on recent work presented in the literature (e.g., [2], [4], [7]). To handle the multiple process time-scales, a singular perturbed framework [5] is employed to develop a composite control structure to control the fast and slow dynamics. The closed-loop properties under the resulting hierarchical EMPC framework is analyzed. The approach is demonstrated on a chemical process network example.

[1] Angeli D, Amrit R, Rawlings JB. On average performance and stability of economic model predictive control. IEEE Transactions on Automatic Control. 2012;57:1615-1626.
[2] Deng H, Ohtsuka T. A highly parallelizable Newton-type method for nonlinear model predictive control. In: Proceedings of the 6th IFAC Conference on Nonlinear Model Predictive Control. Madison, Wisconsin, 2018:426--432.
[3] Ellis M, Durand H, Christofides PD. A tutorial review of economic model predictive control methods. Journal of Process Control, 2014;24:1156--1178.
[4] Shin S, Faulwasser T, Zanon M, Zavala VM. A parrallel decomposition scheme for solving long-horizon optimal control problems. Submitted, 2019.
[5] Kokotovic P, Khalil HK, O'Reilly J. Singular Perturbation Methods in Control: Analysis and Design. London, England: Academic Press, 1986.
[6] Wenzel MJ, ElBsat MN, Ellis MJ, Asmus MJ, Przybylski AJ, Baumgartner R, Burroughs JH, Willmott G, Drees KH, Turney RD. Large scale optimization problems for central energy facilities with distributed energy storage. In: Proceedings of the 5th International High Performance Buildings Conference. 2018; paper number: 3560.
[7] Zavala WM. New architectures for hierarchical predictive control. In: 11th IFAC Symposium on Dynamics and Control of Process Systems. Trondheim, Norway, 2016:43--48.