Acceleration of CO2 Solubility Trapping Mechanism for Enhanced Storage Capacity Utilizing Artificial Intelligence

Abstract

Abstract This paper conducts a thorough examination of the carbon dioxide (CO2) solubility trapping mechanism, a pivotal facet of Carbon Capture and Storage (CCS) technology crucial for enhancing storage capacity. Leveraging the capabilities of Artificial Intelligence (AI), our objective is to innovate and expedite the solubility trapping process. The overarching aim is to hasten the solubility trapping mechanism, thereby achieving heightened efficiency and storage capacity in CCS applications. To assess the potential acceleration of solubility during geological CO2 storage and appraise the field application of successful CO2 sequestration, a multitude of case studies is imperative. These case studies, encompassing various reservoir characteristics, are facilitated through the application of an artificial neural network (ANN). Specifically, we have developed an ANN model for geological CO2 solubility in saline aquifers. The training and testing of the ANN model were executed using data generated from a synthetic aquifer, focusing on solubility and its trapping index. Employing Python with TensorFlow, we conducted training and testing iterations, selecting the optimal model based on the calculated coefficient of determination (R2) and root mean square error (RMSE) values. The model successfully predicted the duration of the solubility trapping mechanism and storage efficiency. Our findings suggest that the ANN model serves as a valuable tool for forecasting storage effectiveness and evaluating the success of CO2 sequestration. In scenarios where conventional simulations fall short, our model may offer a viable solution.