TY - JOUR
T1 - The Potential of Machine Learning for Enhancing CO2 Sequestration, Storage, Transportation, and Utilization-Based Processes: A Brief Perspective
AU - Gupta, Surojit
AU - Li, Lan
PY - 2022/2/1
Y1 - 2022/2/1
N2 - In the paper, we present a review of different types of CO 2 capture, storage, transportation, and utilization (CCSTU) processes. We have also reviewed their further development by using machine learning (ML) methods. Some examples of carbon capture or separation technologies (CCT) are absorption, adsorption, membranes, chemical looping, pyrogenic carbon capture and storage (PyCCS), hydrates, and mineral sequestration, which we review here. We have also classified hybrid processes where multiple methods can be synergistically used for CO 2 capture and utilization. ML methods have also been successfully utilized in CCSTU to enhance the efficiency of CCSTU by process optimization and incorporating new materials design. Based on the review, we have outlined some recommendations for future research, namely consideration of the carbon impact of AI-driven models, and more interactions among ML experts, experimentalists, and computational modelers, which will lead to a holistic approach rather than a competing attitude for rapid commercialization of low-carbon technologies. We also give examples of databases for future research.
AB - In the paper, we present a review of different types of CO 2 capture, storage, transportation, and utilization (CCSTU) processes. We have also reviewed their further development by using machine learning (ML) methods. Some examples of carbon capture or separation technologies (CCT) are absorption, adsorption, membranes, chemical looping, pyrogenic carbon capture and storage (PyCCS), hydrates, and mineral sequestration, which we review here. We have also classified hybrid processes where multiple methods can be synergistically used for CO 2 capture and utilization. ML methods have also been successfully utilized in CCSTU to enhance the efficiency of CCSTU by process optimization and incorporating new materials design. Based on the review, we have outlined some recommendations for future research, namely consideration of the carbon impact of AI-driven models, and more interactions among ML experts, experimentalists, and computational modelers, which will lead to a holistic approach rather than a competing attitude for rapid commercialization of low-carbon technologies. We also give examples of databases for future research.
UR - https://scholarworks.boisestate.edu/mse_facpubs/561
UR - https://doi.org/10.1007/s11837-021-05079-x
M3 - Article
JO - Materials Science and Engineering Faculty Publications and Presentations
JF - Materials Science and Engineering Faculty Publications and Presentations
ER -