(132b) Application of Machine Learning in Developing Membrane Technologies

Membrane-based technologies are advancing rapidly. They are used in a variety of critical applications such as gas separation (carbon dioxide separation, air separation, hydrogen recovery, and many more) and water treatment. The materials utilized in these systems, however, determine their performance. Membrane technology advancements, powered by recent advances, are currently capturing more data than ever before. However, the synthesis and development of next-generation materials takes time and money. To enhance membrane technology, data processing and interpretation must be both time and cost effective. Various theoretical methods and models have been developed to speed material discovery. Because of their assumptions and hypotheses, these models are unable to characterize complex materials and systems. The recent rise of machine learning as a powerful tool for combining multimodality and multifidelity data and finding correlations between interrelated phenomena has a unique promise in this field for tackling these difficulties. Thus, the objective of this work is to develop hybrid machine learning models to estimate membrane performance, including permeability, selectivity parameters, etc., for CO2 capture.