Rapid Evaluation and Validation of AI-driven kinEtics for Accelerating Low-carbon technologies (REVEAL)

• Applicant: Dr.-Ing. Felix Döppel
• Subject Area: Technical Chemistry ; Methods in Artificial Intelligence and Machine Learning
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 569475041

The present-day chemicals industry heavily depends on fossil fuels, contributing significantly to the concerning rise in global CO2 emissions. However, for the chemical industry to adopt a sustainable carbon management strategy, large-scale CO2 valorization technology is needed. Dry reforming of methane holds promise in this direction, due to its ability to convert waste CO2 emissions back into valuable chemicals, using renewable energy. However, it currently lacks industrial adoption, due to the lack of sustainable reactor concepts that avoid catalyst deactivation. The rational design of such reactors requires detailed multi-scale modelling over time scales relevant for catalyst deactivation. Currently, this is far from feasible due to the prohibitively high computational cost of evaluating the chemical kinetics in sufficient detail. I will solve this problem by developing physically consistent and highly efficient machine learning models that drastically reduce the numerical effort without sacrificing on the detailed physical insights. I will then perform data-driven design of experiment to increase the accuracy of the models. This combination allows reliable simulations to identify reactor designs that avoid deactivation of the dry reforming catalyst even at long time-on-stream. However, the new methodology is not limited to dry reforming and will also facilitate the development of further CO2 valorization technologies. The potential project outcomes are: 1) accelerated development of CO2 valorization technologies; 2) achieving a wider adoption of multi-scale modelling as a tool for rational reactor design; and 3) establishing stronger collaborations between chemical engineering and machine learning communities to drive new innovations.

• DFG Programme: WBP Fellowship
• International Connection: Italy
• Host: Professor Dr. Matteo Maestri