(312e) Flexible Process Design and Operation with Distributed Control System Considering Uncertainties

Modern chemical and energy plants can be characterized by large complex network structures, which necessitate the design of efficient distributed control systems to overcome the limitations of centralized and decentralized control approaches. In such design, it is important to identify loosely coupled subsystems, while components in the same subsystem interact strongly. To this end, graph-theoretic methods have been proposed to identify such subsystems very efficiently using community detection algorithms [1]. In this talk, we will discuss an extension of these methods, where robust distributed control systems are designed considering the effects of uncertainty on the potential control performance (measured by modularity of community structures).

Also, for a general class of processes, there is a well-known trade-off between steady state economics and dynamic closed-loop control performance. Thus, to increase the robustness of distributed control systems, processes can be designed with more margins, while minimizing unnecessary expenses. This can be done by utilizing the concept of flexibility [2], identifying a flexible design which can handle all possible realizations of uncertainties. In this talk, we will discuss an effort to improve the existing methods for flexible process design, where multi-objective gradient descent-based method is used to efficiently estimate operable regions in the uncertainty space and to identify optimal flexible process design.

[1] P. Daoutidis, W. Tang, A. Allman, Decomposition of control and optimization problems by network structure: Concepts, methods, and inspirations from biology, AIChE J. 65, e16708, 2019.

[2] I.E. Grossmann, B.A. Calfa, P. Garcia-Herreros, Evolution of concepts and models for quantifying resiliency and flexibility of chemical processes, Comput. Chem. Eng. 70, 22-34, 2014.