Machine learning to predict effective reaction rates in 3D porous media from pore structural features

• DOI: 10.1038/s41598-022-09495-0
• Publication Date: 2022-03-31T16:05:08Z
• Authors: Min Liu; Beomjin Kwon; Peter K. Kang
• Keywords: Machine Learning; Science; Q; R; Medicine; Models, Theoretical; Porosity; Article; Culture Media

Large discrepancies between well-mixed reaction rates and effective reactions estimated under fluid flow conditions have been a major issue for predicting reactive transport in porous media systems. In this study, we introduce framework that accurately predicts directly from pore structural features by combining 3D pore-scale numerical simulations with machine learning (ML). We first perform fluid-solid hundreds of calculate concentration fields. then train Random Forests model 11 to quantify the importance determining rates. Based on information, artificial neural networks varying number demonstrate can be predicted only three features, which are specific surface, sphericity, coordination number. Finally, global sensitivity analyses using ML elucidates how affect The proposed enables accurate predictions few measurable is readily applicable wide range applications involving flows.