Machine Learning-Based Modeling and Prediction of CO2-Rock-Fluid Interactions in Porous Media for Enhanced Oil Recovery and Carbon Storage

Faculty Development Competitive Research Grant Program (AI and Data Science) 2024-2026

This proposal aims to utilize machine learning (ML) and deep learning (DL) techniques to model and predict the interactions between carbon dioxide (CO2), rock, and fluids in porous media for both enhanced oil recovery (EOR) and carbon storage purposes. The study emphasizes the importance of capturing and mitigating CO2 emissions due to increasing energy demand and their detrimental effects on climate change and energy processes. By leveraging ML and DL algorithms, the project intends to develop accurate predictive models for CO2 interaction parameters, such as solubility, adsorption, minimum miscibility pressure (MMP), and capacity loading. The proposed models will be compared with empirical correlations and computational procedures, and their efficacy will be assessed through sensitivity analysis and validation against experimental data from oil/rock samples in Kazakhstan fields. The project also highlights the development of user-friendly web applications for stakeholders to access and utilize the models.