PNNL

Abstract

Packed columns are commonly used in post-combustion processes to capture CO2 emissions by providing enhanced contact area between a CO2-laden gas and a CO2-absorbing solvent. To study and optimize solvent-based post-combustion carbon capture systems (CCSs), computational fluid dynamics (CFD) simulations can be used to model the liquid–gas countercurrent flow hydrodynamics in these columns and derive key determinants of CO2-capture efficiency. However, the large design space of these systems hinders the application of CFD for design optimization due to its high computational cost. In contrast, data-driven modeling approaches can produce fast surrogates to study large-scale physics problems. We build our surrogates using MeshGraphNets (MGN), a graph neural network that efficiently learns and produces mesh-based simulations. We apply the model to a random packed column domain with over 160K mesh graph nodes and a design space consisting of three key design parameters: surface tension, contact angle, and liquid inlet velocity. Our models can adapt to a wide range of design parameters and accurately predict the complex interactions within the system at rates over 1600 times faster than CFD, affirming its practicality in downstream design optimization tasks. This underscores the robustness and versatility.