(178h) Practically Achievable Process Performance Limits for Pressure-Vacuum Swing Adsorption Based Post-Combustion CO2 Capture.

Adsorption-based post-combustion CO₂ capture is considered as a potential competitor to solvent absorption. Most developments in adsorption has taken place in the materials domain with hundreds/thousands of new materials being proposed as potential alternatives. However, it is not yet clear what are the actual process performance potentials of these materials. A significant roadblock for making such evaluations lies in the computational demands of simulating and optimizing adsorption processes. Recent innovations in machine learning to accelerate adsorption process simulations while maintaining prediction accuracy.

In this work, the machine-assisted process learning, and emulation (MAPLE) framework, a data-driven modeling framework trained using a detailed process model is used to explore the operation, performance, and materials limits of PVSA-based CO₂ capture. Various case studies are considered to evaluate performance indicators, such as CO₂ purity, recovery, minimum energy, and maximum productivity at various flue gas compositions. For each case study, the MAPLE model is coupled with a multi-objective genetic algorithm, which searches tens of thousands of unique combinations of isotherms and process operating conditions. This results in an optimized pair of isotherm and process conditions that are best suited for the separation condition providing the limits of achievable performance. We present such process performance limits for a. wide range of CO2 feed compositions. The results indicate that traditional fixed bed PVSA processes are not well suited for low CO₂ composition flue gas (CO₂ < 20%) even with the best possible adsorbent. This is mainly due to the requirement of very low regeneration pressure and the low efficiency of industrial vacuum pumps at the low-pressure levels. The affinity parameters for CO₂ and N₂ are mapped to the minimum energy and maximum productivity to identify trends. The results indicate a wide range of CO₂ isotherms can achieve the US Dept. of Energy (US-DOE) set separation targets. Results also indicate that the gap between the energy consumption of available adsorbents and the achievable limit with the hypothetical best adsorbent varies between 20 and 2.5% as the CO₂ feed composition changes between 0.05 and 0.4. This indicates a limited potential for the development of new adsorbents of PVSA-based CO₂ capture and that future studies should not neglect the importance of process engineering.