Hybrid CFD and machine learning analysis of CO2 enhanced oil recovery in naturally fractured reservoirs

Abstract

Abstract CO 2 based Enhanced Oil Recovery (EOR) in unconventional reservoirs is an emerging technology. Scientific research efforts are directed towards understanding the propagation of CO 2 front due to the complex interplay between CO 2 injection and saturation, and reservoir’s constitutive relationships. Conventional methods for characterising CO 2 -EOR rely on high-fidelity numerical solutions that often result in over or under prediction of CO 2 geosequestration. In this study, we develop a novel hybrid Computational Fluid Dynamics (CFD) and Machine Learning (ML) framework that allows for rapid CO 2 geosequestration prediction and its optimal injection. Very low or high injection rates have been shown to result in low sweep efficiency or excessive entry pressure, while an intermediate injection rate offers the best balance between the two. CFD data-driven Gaussian Process Regression (GPR) and Extreme Gradient Boosting (XGBoost) models have been developed, trained and tested for predicting CO 2 saturation in the reservoir. Comparative analysis indicates that GPR outperforms XGBoost in terms of its predictive performance and robustness. Through the analysis of layer-resolved CO 2 front displacement and development of data-driven surrogate models, this study contributes a novel framework for CO 2 -EOR predictive modelling and optimising injection strategies in naturally fractured reservoirs.