(615h) Marrying Materials and Processes: A Superstructure Inspired Optimization Approach for Pressure Swing Adsorption Processes for Pre-Combustion CO2 Capture

Adsorption based carbon capture is one of the main technologies to reduce greenhouse gas emissions resulting from increasing use of fossil fuels. Absorption based technologies are the preferred method for postcombustion capture due to the low level of CO₂ concentration in the flue gas (5-15 %). However, adsorption-based separation processes are an energy-efficient candidate for precombustion capture. The precombustion capture refers to the elimination of CO₂ from fossil fuel before completion of the combustion. For instance, in coal gasification, CO₂ is separated from H₂ once syngas undergoes the water-gas shift reaction (WGSR). Gas-fired power plants can be also modified to use H₂ fuel by steam reforming natural gas followed by WGSR combined with a precombustion technology.

Pressure swing adsorption (PSA) processes have been extensively studied for this purpose over the past decade.^1,2 There are two main players in any PSA process namely, the adsorbent and the process cycle configuration. It is therefore important to use the “right” process for a given adsorbent to yield the best possible performance,³ albeit this is usually not the case. Conventionally, every single cycle configuration is optimized for every single adsorbent to meet the purity/recovery requirements (one-by-one optimization) or the energy/productivity requirements. An alternative approach for this optimization problem is to design a “mother-cycle” and conduct the optimization in a single attempt. This approach would cover all possible cycle configuration candidates and based on a detailed optimization, the best cycle(s) can be chosen for any given adsorbent subjected to the process requirements.

The overarching goal of this work is two fold. First, to highlight the superiority of employing a superstructure inspired (mother-cycle) optimization approach when compared to a detailed one-by-one optimization approach for material screening and process design applications. Second, to highlight the need to “marry” every adsorbent to the process(s) to maximize its potential. To this aim, a binary system of CO₂/H₂ at syngas conditions was chosen as the system of interest. From the materials angle, several hypothetical materials that have equilibrium and kinetic characteristics similar to commonly employed porous materials like MOFs, zeolites, and activated carbon were investigated. From the process angle, a basic 4-step PSA, a 4-step PSA with light product pressurization (LPP), and a 5-step PSA with inert purge were investigated to meet various purity/recovery constraints for the given feed conditions. A “mother-cycle” incorporating the steps in aforementioned cycle configuration was subsequently developed.

To meet the goals of the study, a detailed PSA model was used to simulate the process. The system of coupled partial differential equations obtained from mass, energy and momentum balances were discretized using a high resolution finite volume method.⁴ Optimal decision variables of the process, i.e. the step times, the pressures, and the feed velocity were obtained by performing a rigorous optimization using a genetic algorithm routine on the detailed PSA model for every adsorbent considered. Unlike the conventional approach, the “mother-cycle” has both continuous and discrete variables that is optimized. The former constitutes process variables like pressure, step times, and feed velocity, while the latter constitutes the presence or absence of a given step in the “mother-cycle”.

The optimization routine produces purity-recovery Pareto fronts that would yield the best possible performance for a given material-process combination. The Pareto fronts obtained from the “mother-cycle”-based optimization was compared with the ones obtained from a one-by-one optimization for every single PSA cycle configuration considered here. It is shown that a similar, if not the same, Pareto fronts can be obtained from both the optimization approaches, but at a lower computational cost for the case of the “mother-cycle”-based optimization. Upon confirming the effectiveness of employing a “mother-cycle”-based optimization, the same technique was then employed to identify the process combinations that would maximize the process performance for the hypothetical adsorbents. It is shown that a given material might make use of different cycle configurations to maximize the process performance depending on what is expected from the material-process combination. In practice, this might translate to a scenario that leads to using one cycle to obtain high purities and another to obtain high recoveries. Finally, the effects of feed composition and purity-recovery constraints on the choice of “right” cycle configuration were also investigated.

To conclude, the findings of this study suggests that the “mother-cycle”-based optimization framework proposed here eliminates the need to conduct multiple time-consuming optimizations for each combination of adsorbent and cycle configuration. Additionally and most importantly, it provides a solid basis for exploiting a plethora of process configurations for a given adsorbent, thereby enabling every adsorbent to maximize its process performance.

References

1. Casas, N.; Schell, J.; Joss, L.; Mazzotti, M. A Parametric Study of a PSA Process for Pre-Combustion CO₂ Sep. Purif. Technol. 2013, 104, 183–192.
2. Subraveti, S. G.; Pai, K. N.; Rajagopalan, A. K.; Wilkins, N. S.; Rajendran, A.; Jayaraman, A.; Alptekin, G. Cycle Design and Optimization of Pressure Swing Adsorption Cycles for Pre-Combustion CO₂ Appl. Energy 2019, 254, 113624.
3. Sircar, S. Pressure Swing Adsorption. Ind. Eng. Chem. Res. 2002, 41 (6), 1389–1392.
4. Haghpanah, R.; Majumder, A.; Nilam, R.; Rajendran, A.; Farooq, S.; Karimi, I. A.; Amanullah, M. Multiobjective Optimization of a Four-Step Adsorption Process for Postcombustion CO₂ Capture Via Finite Volume Simulation. Ind. Eng. Chem. Res. 2013, 52 (11), 4249–4265.