(218a) A Generalized Superstructure-Based Framework for Process Synthesis

Process synthesis is arguably the fundamental problem in chemical engineering: how to synthesize a process that meets specific goals while satisfying all relevant constraints. There are, in general, two types of approaches to process synthesis. First, sequential/hierarchical approaches¹ which rely on decomposing the problem into subproblems that are solved sequentially. While various decomposition strategies have been proposed, they all follow the same sequence: (1) reaction network, (2) separation network, and (3) heat exchanger network synthesis. Hierarchical approaches are easier to implement and can quickly generate reasonable base-case designs. However, they are limited in that they do not account for interactions among subproblems, which frequently leads to suboptimal designs. Alternatively, process synthesis problems can be solved using superstructure-based approaches where the process flowsheet includes a large number of useful unit operations and relevant interconnections². Compared to hierarchical approaches, superstructure-based approaches can, in principle, simultaneously consider complex interactions between design decisions. However, as we will illustrate, despite the extensive studies on superstructure-based approaches, most proposed methods either focus on developing superstructures for subsystems (e.g. reaction or separation) or embed a set of predefined technologies, strategies, or synthesis routes. Therefore, to some extent, existing superstructure-based approaches have to this day been employed in a hierarchical manner.

To address this limitation, thereby maximizing the advantages derived from a superstructure-based approach, we propose a generalized framework where all subsystems are considered simultaneously, thus accounting for their interactions. The proposed generalized framework starts, importantly, with redefining the problem statement for reactor, separation, and heat exchanger network synthesis (and, if necessary, a utility plant subproblem). Next, we discuss how superstructure-based approaches can be developed for these redefined, and broader, subproblems; and then present how these superstructures can be richly connected to capture the strong interactions among the subsystems. The resulting framework is a major departure from all previous approaches because it allows a number of new features. Specifically, the reaction superstructure is modeled to allow variable inlet component flows and variable limiting components for each reaction³. The connectivity between the reaction and separation superstructures is represented by one or more effluent streams that can be sent to different units in the separation superstructure. The flowrates of some components in these streams can be zero, which not only affects the destinations of the streams in the separation superstructure, but also affects the modeling of unit operations (e.g. distillation)⁴. The products/outlets of the separation superstructure are not necessarily pure and can be recycled. In the heat exchanger network subproblem, a set of candidate streams, some of which might not be selected, is defined based on the process streams in the reaction and separation superstructures. These candidate streams can be unclassified (i.e. unknown hot/cold identity) and their inlet/outlet temperatures and flowrates can be variables⁵. We show how, employing the methods for the three subsystems, we can formulate a single optimization problem that minimizes the annualized cost of the entire process. Finally, we demonstrate the applicability of our framework using a vinyl chloride production example.

Reference

1. Douglas JM. A hierarchical decision procedure for process synthesis. AIChE J. 1985;31(3):353-362.
2. Chen Q, Grossmann IE. Recent Developments and Challenges in Optimization-Based Process Synthesis. Annu Rev Chem Biomol Eng. 2017;8(1):249-283.
3. Ramapriya GM, Won W, Maravelias CT. A superstructure optimization approach for process synthesis under complex reaction networks. Chem Eng Res Des. 2018;137:589-608.
4. Kong L, Maravelias CT. Expanding the scope of distillation network synthesis using superstructure-based methods. Comput Chem Eng. 2020;133:106650.
5. Ryu J, Maravelias CT. Simultaneous Process and Heat Exchanger Network Synthesis Using a Discrete Temperature Grid. Ind Eng Chem Res. 2019;58(15):6002-6016.