(11e) Carbon Capture Process Design Under Variable Operating Conditions with High-Performance Computing

Previous works have focused extensively on developing accurate first-principle models for carbon capture units, with various methodologies proposed for modeling key components such as the absorber column and heat exchanger [2,3,4,5]. There have been several modeling approaches proposed in the literature to model the CO2 absorptions into the MEA solvent. Most notably, enhancement factor models have been proposed to simultaneously capture the reaction kinetics and mass transfer within the columns [3]. The enhancement factor method incorporates CO₂ reactions within the liquid film by adjusting the mass transfer driving force. More classical models, such as the Kent-Eisenberg model, assumes the chemical equilibrium and fits the temperature dependency of the CO2 solubility using the fitted model [2]. While this model does not rigorously capture the reaction kinetics of CO₂ with water and MEA through predefined reaction equilibria, they can potentially provide better numerical robustness for simulation and optimization [3,5]. In addition to steady-state models, dynamic models have also been introduced to capture transient behaviors. Dynamic models account for mass and energy accumulation, which is essential for analyzing process control strategies [6,7,8]. Computational platforms like IDAES have also emerged, providing powerful tools for simulation and optimization of these systems [9]. Despite these advancements, existing design practices are typically optimized for deterministic conditions, which fail to account for fluctuations in feed composition caused by variable power plant operations.

Our work presents an optimization framework that addresses the challenge of feed condition uncertainty in carbon capture systems by leveraging GPU-accelerated tools. Building upon the steady-state carbon capture unit first principle model developed by Akula et al and using the Kent-Eisenberg model [2,5], we formulate a two-stage stochastic programming problem to minimize the capital and the operating costs. The explicit consideration of multiple scenarios significantly increases the problem size, making efficient computation a key challenge. Integrating state-of-the-art solvers with GPU-accelerated optimization techniques provides the necessary computational power to handle this increased complexity efficiently [10,11]. Additionally, this work demonstrates the capability of simulating chemical processes in an equation-oriented scheme using GPU, which is well-suited for large and complex flowsheets due to the parallelizable nature of GPU computing, enabling faster and more efficient computations. By addressing feed uncertainty, our framework aims to improve the resiliency and adaptability of carbon capture systems. The use of stochastic programming allows for a more accurate representation of operational variability, ultimately leading to better design and operation choices in the face of uncertainty. In turn, our GPU-accelerated approach enables faster and more scalable computations, paving the way for resilient and cost-effective carbon capture unit design and operation.

Reference

[1] Rao, A. B.; Rubin, E. S. A technical, economic, and environmental assessment of Amine-Based CO2 capture technology for power plant greenhouse gas control. Environmental Science & Technology 2002, 36 (20), 4467–4475. https://doi.org/10.1021/es0158861.

[2] Haji-Sulaiman, M. Z.; Aroua, M. K.; Benamor, A. Analysis of equilibrium data of CO2 in aqueous solutions of diethanolamine (DEA), methyldiethanolamine (MDEA) and their mixtures using the modified Kent Eisenberg model. Process Safety and Environmental Protection 1998, 76 (8), 961–968. https://doi.org/10.1205/026387698525603.

[3] Gaspar, J.; Fosbøl, P. L. A general enhancement factor model for absorption and desorption systems: A CO 2 capture case-study. Chemical Engineering Science 2015, 138, 203–215. https://doi.org/10.1016/j.ces.2015.08.023.

[4] Akula, P.; Eslick, J.; Bhattacharyya, D.; Miller, D. C. Modelling and parameter estimation of a plate heat exchanger as part of a Solvent-Based Post-Combustion CO2 capture system. In Computer-aided chemical engineering/Computer aided chemical engineering; 2019; pp 47–52. https://doi.org/10.1016/b978-0-12-818597-1.50008-4.

[5] Akula, P.; Eslick, J.; Bhattacharyya, D.; Miller, D. C. Model Development, Validation, and Optimization of an MEA-Based Post-Combustion CO2 Capture Process under Part-Load and Variable Capture Operations. Industrial & Engineering Chemistry Research 2021, 60 (14), 5176–5193. https://doi.org/10.1021/acs.iecr.0c05035.

[6] Kvamsdal, H. M.; Jakobsen, J. P.; Hoff, K. A. Dynamic modeling and simulation of a CO2 absorber column for post-combustion CO2 capture. Chemical Engineering and Processing - Process Intensification 2008, 48 (1), 135–144. https://doi.org/10.1016/j.cep.2008.03.002.

[7] Modekurti, S.; Bhattacharyya, D.; Zitney, S. E. Dynamic Modeling and control studies of a Two-Stage Bubbling Fluidized Bed Adsorber-Reactor for Solid–Sorbent CO2 Capture. Industrial & Engineering Chemistry Research 2013, 52 (30), 10250–10260. https://doi.org/10.1021/ie400852k.

[8] Thierry, D.; Biegler, L. T. Dynamic real‐time optimization for a CO2 capture process. AIChE Journal 2018, 65 (7). https://doi.org/10.1002/aic.16511.

[9] Lee, A.; Ghouse, J. H.; Eslick, J. C.; Laird, C. D.; Siirola, J. D.; Zamarripa, M. A.; Gunter, D.; Shinn, J. H.; Dowling, A. W.; Bhattacharyya, D.; Biegler, L. T.; Burgard, A. P.; Miller, D. C. The IDAES process modeling framework and model library—Flexibility for process simulation and optimization. Journal of Advanced Manufacturing and Processing 2021, 3 (3). https://doi.org/10.1002/amp2.10095.

[10] Pacaud, F.; Shin, S.; Schanen, M.; Maldonado, D. A.; Anitescu, M. Accelerating condensed Interior-Point methods on SIMD/GPU architectures. Journal of Optimization Theory and Applications 2023, 202 (1), 184–203. https://doi.org/10.1007/s10957-022-02129-5.

[11] Pacaud, F.; Schanen, M.; Shin, S.; Maldonado, D. A.; Anitescu, M. Parallel interior-point solver for block-structured nonlinear programs on SIMD/GPU architectures. Optimization Methods & Software 2024, 39 (4), 874–897. https://doi.org/10.1080/10556788.2024.2329646.