(116i) Learning-Enabled Hybrid Modeling and Economic Predictive Control of Carbon Capture Process for Ship Decarbonization

Projections indicate that CO₂ emissions from shipping transport will increase to 1.6 billion tons by 2050 [1]. The International Maritime Organization (IMO) has implemented mandatory measures to reduce greenhouse gas (GHG) emissions within international shipping and has revised its strategy for GHG emission reduction in 2023 [2]. One promising strategy is to use post-combustion carbon capture (PCC) on-board ships to reduce ship carbon emissions due to its ability to be easily retrofitted into existing systems without significant changes [3]. Various studies have investigated post-combustion carbon capture for ships [4,5]. Within the PCC process, regenerating captured CO₂ demands a substantial amount of heat energy [6]. In this research, we consider the adoption of this emerging technology, and propose to use a separate diesel gas turbine and a waste heat recovery (WHR) system to provide heat energy for regenerating CO₂ from the rich solvent.

The complex structure and large scale of the integrated shipboard PCC plant present challenges in attaining safe and energy-efficient process operation while maintaining a high level of carbon capture rate. These considerations highlight the importance of developing and implementing advanced control solutions for consistent and efficient operation of shipboard carbon capture processes. Model predictive control (MPC) is one of the most widely used advanced control methodologies [7]. In the existing literature, MPC designs have been proposed for land-based post-combustion carbon capture processes [6, 8-10]. The significant disparities from the land-based counterpart in terms of process designs, component sizes, manner of heat supply, among other factors [5], pose challenges in transferring these existing control methodologies from land-based to shipboard applications. Additionally, there have been no results on dynamic modeling and control of shipboard post-combustion carbon capture plants.

Based on the above observations, in this work, we aim to address the dynamic modeling and optimal control problem for the shipboard post-combustion carbon capture process, for the first time. While the existing results for land-based facilities may offer insights into the dynamic modeling of shipboard carbon capture plants, we recognize the limitations of these methods when considering the application to shipboard plants: 1) first-principles modeling can encounter challenges due to the unavailability of the values of some crucial model parameters; 2) system identification often relies solely on input-output data, leading to models that overlook the dynamic behaviors of key state variables. To handle the large scale and complex structure and dynamics of the integrated plant, a promising modeling framework is hybrid modeling, which seamlessly integrates the available first-principles knowledge and data information. Hybrid models have the potential to offer high prediction accuracy, strong generalization capabilities, and good interpretability [11,12]. Based on a dynamic model, advanced control can be developed within the framework of economic model predictive control (EMPC) [13,14], which explicitly incorporates economic operational considerations into a control objective function to ensure safe operations, maintain efficient carbon capture, and minimize energy.

In this work, we present a complete design for the post-combustion carbon capture process on-board ships which incorporates ship engines, a diesel gas turbine, a waste recovery system, and a solvent-based carbon capture plant. Given the dynamic differential algebraic equations representing the first-principles model of this plant may have inaccurately determined parameters, two neural networks are established and integrated with the imperfect first-principles model to form a high-fidelity hybrid dynamic model for the shipboard post-combustion carbon capture process. An economic predictive control scheme is developed based on the hybrid model of the shipboard PCC process. The effectiveness and performance of the proposed framework are comprehensively evaluated under three ship operational conditions. The proposed approach is also compared with a solely neural network-based model to illustrate its superiority in terms of data efficiency and model robustness.

References

• Capros, A. De Vita, L. Hoeglund-Isaksson, W. Winiwarter, P. Purohit, H. Bottcher, S. Frank, P. Havlik, M. Gusti, and H. P. Witzke. EU energy, transport and GHG emissions trends to 2050-Reference scenario 2013. European Commission, 2013.
• 2023 IMO strategy on reduction of GHG emissions from ships. Marine Environment Protection Committee, Resolution, 2023.
• T. Aakko-Saksa, K. Lehtoranta, N. Kuittinen, A. Järvinen, J. Jalkanen, K. Johnson, H. Jung, L. Ntziachristos, S. Gagné, C. Takahashi, P. Karjalainen, T. Rönkkö, and H. Timonen. Reduction in greenhouse gas and other emissions from ship engines: Current trends and future options. Progress in Energy and Combustion Science, 94:101055, 2023.
• Feenstra, J. Monteiro, J. T. van den Akker, M. R. Abu-Zahra, E. Gilling, and E. Goetheer. Ship-based carbon capture onboard of diesel or LNG-fuelled ships. International Journal of Greenhouse Gas Control, 85:1–10, 2019.
• Luo, and M. Wang. Study of solvent-based carbon capture for cargo ships through process modelling and simulation. Applied Energy, 195:402–413, 2017.
• Decardi-Nelson, S. Liu, and J. Liu. Improving flexibility and energy efficiency of post-combustion CO2 capture plants using economic model predictive control. Processes, 6(9):135, 2018.
• Morari, and L. J. Lee. Model predictive control: past, present and future. Computers & Chemical Engineering, 23(4-5):667–682, 1999.
• Zhang, R. Turton, and D. Bhattacharyya. Development of model and model-predictive control of an MEA-based postcombustion CO2 capture process. Industrial & Engineering Chemistry Research, 55(5):1292–1308, 2016.
• Panahi, and S. Skogestad. Economically efficient operation of CO2 capturing process. Part II. Design of control layer. Chemical Engineering and Processing: Process Intensification, 52:112–124, 2012.
• D. Patrón, and L. Ricardez-Sandoval. An integrated real-time optimization, control, and estimation scheme for post-combustion CO2 capture. Applied Energy, 308:118302, 2022.
• Shah, M. Z. Sheriff, M. S. F. Bangi, C. Kravaris, J. S. I. Kwon, C. Botre, and J. Hirota. Deep neural network-based hybrid modeling and experimental validation for an industry-scale fermentation process: Identification of time-varying dependencies among parameters. Chemical Engineering Journal, 441:135643, 2022.
• S. F. Bangi, and J. S. Kwon. Deep hybrid modeling of chemical process: Application to hydraulic fracturing. Computers & Chemical Engineering, 134:106696, 2020.
• B. Rawlings, D. Angeli, and C. N. Bates. Fundamentals of economic model predictive control. 51st IEEE conference on decision and control, 3851–3861, 2021, Maui, Hawaii, USA.
• Ellis, H. Durand, P. D. Christofides. A tutorial review of economic model predictive control methods. Journal of Process Control, 24(8):1156–1178, 2014.