Experimental Investigation, Porosity-Permeability Modelling, and Artificial Neural Network Prediction of CO2 Injectivity Change for Sequestration

Abstract

CO2 sequestration in saline aquifer is a promising approach to effectively secure the anthropogenic CO2 gas. Complex fluid-rock interaction processes take place during the injection of CO2 would disrupt the thermodynamic equilibrium of CO2 injectivity at near wellbore. In this study, a comprehensive investigation on the CO2 injectivity change of different injection flow rates and brine salinity was performed using core flooding experiments, permeability change prediction using (Kozeny-Carman and Hagen-Poiseuille models) and artificial neural network model (ANN). Core flooding experiments revealed CO2 injectivity impairment increased with increasing brine salinity, with Hagen-Poiseuille being the most fitted model with R2 of 0.935. However, all porosity-permeability models failed to give a good prediction at changing injection flow rate with R2 is well below 0.4. The adopted ANN model showed good agreement with the experimental data at varying brine salinity and injection flow rates. The utilization of ANN for such prediction procedure can reduce the number of experiment, operating cost and provide reasonable predictions compared to existing analytical models.