Numerical Modeling for the Prediction of Residual CO2 Trapping in Water-Wet Geological Porous Media

Abstract

Residual trapping of CO2 has been identified as a reliable and rapid way to dispose large CO2 quantities. Several experimental investigations have been completed where residual trapping in sandstone was measured; these programmes identified that initial CO2 saturation and rock porosity are significant parameters which influence the residual saturation and thereby residual trapping capacity and effectiveness. In order to further improve fundamental understanding a computational tool need to be developed with which residual CO2 saturations can be predicted. Pore-scale two-phase fluid flow simulation is performed based on the integration of x-ray micro-tomography images (which provide a detailed description of the rock's pore space) and Navier-Stokes equations. X-ray micro-tomography (approximately (6µm)3 voxel size) images of F42 sand pack were used. The extracted pore morphology of each medium was obtained by segmenting the images based on their greyscale contrast using image processing software AVIZO Fire. These binary images were converted initially into surface and volume meshes which were then fed into a commercially available computational fluid dynamics code (ANSYS-CFX). Three dimensional transient, laminar flow fields were obtained by solving the continuity and Navier-Stokes equations using an Eulerian-Eulerian multi-phase flow approach. To incorporate the effect of capillary forces, free surface model was used which solved the pressure gradient at the interface correctly. The model assumes isothermal condition with no mass transfer between the brine and CO2. The inlet and outlet boundary conditions include CO2 flow rate and the pressure respectively. We simulated the drainage condition in this paper. Approximately 1.5 million tetrahedral elements were used to generate volume mesh, and the convergence criterion for all variables was set to 10-3. Initially all pore space was filled with brine, and then CO2 was injected from one inlet side at constant flow rate, obtained from the experiments. When the system was at connate water saturation, we stopped our simulation. The residual saturation depends on the flow rate of super critical CO2. The computations described here are a rapid, cost-effective and can reveal vital information for the planning of carbon geo-sequestration projects and associated risk and capacity assessments.