(371aa) Feasibility Study and Machine Learning-Based Optimization for Shipboard CO2 Capture Leveraging Available Energy Sources

Carbon dioxide (CO₂) emissions from fossil-fuel combustion are significantly responsible for climate change and global warming [1]. Among CO₂ emission sources, transportation is the second largest source of CO₂ emission, with marine transport contributing 3% of global CO₂ emission [2]. Under rapid economic development, the quantity of CO₂ emitted by ship transportation has been forecasted to rise to 1903 million tons of CO₂ by 2050 [3]. Consequently, the International Maritime Organization (IMO) has issued a mandatory strategy to achieve net-zero greenhouse gas emissions by 2050, aiming to cut at least 40% of CO₂ emissions by 2030 [4].

Among the available literature, solvent-based chemical absorption is the most commonly used method for shipboard CO₂ capture (SCC). Previous studies have focused on the feasibility investigation of Monoethanolamine (MEA)-based SCC processes leveraging available heating energy [5], [6], [7], [8]. It indicates that heat recovery from the ship engine’s flue gas is insufficient for a 90% of CO₂ capture rate. Some reports have proposed the combustion of additional fuel to supply the extra heating energy for capturing 90% CO₂ from engine flue gas [2], [7], [9]. However, combusting of fuel generates more CO₂. Consequently, the combustion of additional fuel to capture 90% of CO₂ is impractical for the SCC process in terms of energy, environment, and economy. Additionally, MEA-based CO₂ absorption releases a certain amount of MEA into the environment (140-600 ppm), leading to the installation of wash columns for MEA emissions of less than one ppm [10]. Addressing these critical research gaps, this study develops advanced designs for the SCC process at a low CO₂ capture cost and low MEA emission, which is achieved solely by leveraging available energy sources on the ship.

Upgrading a design typically requires comparing the optimal performance of the proposed design with the conventional design. However, due to the lack of a feasible design and the high computational cost of one simulation, few studies have been investigated about optimization for the SCC process leveraging available energy sources to capture 90% CO₂. The feasible design can be achieved by developing advanced designs and applying data analysis methods to select the most feasible one. Meanwhile, machine learning (ML), which confirmed the fidelity for predicting precisely and quickly the performance of various chemical processes [11], [12], [13], can significantly reduce the computing time of the SCC simulator. Therefore, we propose combining data analysis and ML-based optimization to determine the optimal performance of the most feasible SCC design.

This study proposes advanced designs and employs machine learning-based optimization to achieve cost-effective CO₂ capture from the flue gas of a diesel ship engine by leveraging available energy sources. Specifically, we develop CO₂ capture and ship engine simulators, which are validated and then applied to develop conventional and four advanced designs for the SCC process. A first deep neural networks (DNN) model is developed for the conventional design, serving as the basis for formulating two optimization problems to highlight the techno-economic limitations of the conventional design. Therefore, four advanced designs are analyzed to exhibit their potential for reducing CO₂ capture cost and heating energy compared to the conventional design. Pearson and Spearman correlation coefficient methods are then employed to estimate correlation values, ultimately selecting SCC using lean vapor compression (LVC-SCC) design as the most feasible design. A second DNN model is developed for the LVC-SCC design before being used to formulate the third optimization problem. Finally, at the optimal operating conditions, a techno-economic-environmental analysis that compares the LVC-SCC design with the conventional design and available energy quantities on the ship is conducted to highlight the feasibility of SCC leveraging available energy sources.

Reference

• Pathak, R. Slade, R. Pichs-Madruga, D. ̈Urge-Vorsatz, R. Shukla, and J. Skea, “Climate change 2022 mitigation of climate change: Technical summary,” 2022.
• Luo and M. Wang, “Study of solvent-based carbon capture for cargo ships through process modelling and simulation,” Applied Energy, vol. 195, pp. 402–413, 2017.
• “Second IMO GHG study,” 2009.
• “IMO’s work to cut GHG emissions from ships,” 2023.
• Sridhar, A. Kumar, S. Manivannan, S. Farooq, and I. A. Karimi, “Technoeconomic evaluation of post-combustion carbon capture technologies on-board a medium range tanker,” Computers & Chemical Engineering, vol. 181, p. 108545, 2024.
• V. D. Long, D. Y. Lee, C. Kwag, Y. M. Lee, S. W. Lee, V. Hessel, and M. Lee, “Improvement of marine carbon capture onboard diesel fueled ships,” Chemical Engineering and Processing Process Intensification, vol. 168, p. 108535, 2021.
• Oh, D. Kim, S. Roussanaly, R. Anantharaman, and Y. Lim, “Optimal capacity design of amine-based onboard CO₂ capture systems under variable marine engine loads,” Chemical Engineering Journal, vol. 483, p. 149136, 2024.
• Visona`, F. Bezzo, and F. d’Amore, “Techno-economic analysis of onboard CO₂ capture for ultra-large container ships,” Chemical Engineering Journal, vol. 485, p. 149982, 2024.
• Einbu, T. Pettersen, J. Morud, A. Tobiesen, C. Jayarathna, R. Skagestad, and G. Nysæther, “Energy assessments of onboard CO₂ capture from ship engines by MEA-based post combustion capture system with flue gas heat integration,” International Journal of Greenhouse Gas Control, vol. 113, p. 103526, 2022.