±â°èÇнÀÀ» ÀÌ¿ëÇÑ ¸ÞźÀÇ Ã˸Šº¯È¯ °³¼±¿¡ °üÇÑ ¿¬±¸
Improving Catalytic Conversion of Methane Using Machine Learning
±èÇö¿ì
KRICT

Nowadays, artificial intelligence, which includes machine learning, becomes familiar and useful in various fields. Catalysis research community is not an exception and machine-learning (ML) approaches are helpful to understand and improve the performance of catalysts. For example, the oxidative coupling of methane (OCM) is investigated with several ML techniques as the accumulated data of catalyst compositions and experimental conditions are available. Moreover, methane is naturally abundant, and selective activation of C-H bonds is challenging. In this regard, we are also studying catalytic reactions to transform methane molecules to hydrocarbons such as ethylene and benzene in high-temperature and nonoxidative conditions. Here, we will present a ML method for improving catalytic conversion of methane by optimizing reaction conditions and catalyst compositions. First, we will show that ML models can reasonably predict the latent relationship in catalytic reactions of methane in both oxidative and nonoxidative conditions. Then, we will explain our ML strategies to propose better reaction conditions based on ML predictions. Finally, we will briefly introduce the data platform for catalysis research managed by our institute.