Research team
Expertise
His research activities comprise the study of plasma and plasma-surface interactions by means of experiments and machine learning, for various applications, mainly CO2 and CH4 conversion into value-added chemicals and fuels and plasma medicine.
What machine learning can bring to plasma catalysis: A generalized strategy to optimize methanol production from CO2 hydrogenation.
Abstract
Plasma-catalytic CO2 hydrogenation to methanol is a promising way to utilize CO2, owing to its potential to overcome traditional thermal barriers. It can help mitigate issues associated with climate change. However, one of the major challenges for plasma catalysis is to uncover the synergism between plasma and catalyst, to avoid trial-and-error experiments. Therefore, I will present a supervised learning (SL) framework to predict the methanol yield from experimental descriptors. Additionally, I will develop a so-called "methanol yield agent" by a reinforcement learning (RL) algorithm to perform effective actions on optimizing methanol production. Finally, I will experimentally validate the capability of my model to predict and regulate the methanol yield under varying catalyst compositions and reaction conditions to certify its implementation. I hope that my project will successfully demonstrate that a generalized, interpretable ML framework (SL model and RL strategy) offers a precious solution to guide plasma catalysis experiments and produce value-added compounds and renewable fuels for future large-scale industrial applications.Researcher(s)
- Promoter: Bogaerts Annemie
- Fellow: Li Jiayin
Research team(s)
Project website
Project type(s)
- Research Project