Simulating fluid flow at pore scale with carbon dioxide in digital rock

Abstract

Carbon dioxide capture and storage technologies will be required to reduce emissions into the atmosphere and limit global warming in the coming decades. Geo-sequestration involving the injection of CO2 directly into the pore space in sedimentary rocks, saline formations or abandoned oil fields may present a permanent storage solution. More scientific research is needed to understand and optimize CO2 injection process and long-term storage at the pore scale of these underground geological formations. In our research, we model the rock pore space as a network of capillaries with spatially varying radii and the flow rate in each capillary is modelled as laminar flow. The capillary network representation is extracted from high-resolution X-ray microtomography images of suitable rocks [1]. This fined-grained capillary network representation allows for both single- and two-phase flow simulations with a high level of geometric accuracy at microscale. The two-phase flow simulations employ a time-dependent interface tracking approach for simulating displacement experiments that results in a system of differential-algebraic equations. Solving these time-dependent equations on the high-resolution three-dimensional geometrical representation of the rock sample obtained from the X-ray microtomography remains very time and computationally intensive. Alternatively, analysis is carried out on the aggregate results of multiple two-phase flow simulations, each applied to a different simplified capillary network model of the rock sample. These simplified network models are generated algorithmically and allow for a higher level of control on the number of capillaries and their properties. The resulting networks are optimized to match the physical properties of permeability and porosity of the original sample at a significantly lower computational cost. In this work we present results from the accuracy benchmark of the single-phase flow simulations against other known network methods from the literature and with respect to measurements of permeability performed at lab scale in the same rock samples. We then apply the methodology described to simulate two-phase flow on capillary network representations of porous sandstone rock samples extracted from the geometric boundaries of the connected pore space from x-ray microscale computer tomography data. In order to understand the physics of carbon dioxide trapping at the pore scale, we simulate two-phase fluid flow scenarios under varying conditions of temperature, pressure and fluid properties, and present our findings in optimizing parameters that maximize fluid saturation of the rock. [1] Neumann, R.F., Barsi-Andreeta, M., Lucas-Oliveira, E. et al. High accuracy capillary network representation in digital rock reveals permeability scaling functions. Sci Rep 11, 11370 (2021). https://doi.org/10.1038/s41598-021-90090-0