Equivalent Pore Channel Model for Fluid Flow in Rock Based on Microscale X-ray CT Imaging

Abstract

Pore-scale modeling with a reconstructed rock microstructure has become a dominant technique for fluid flow characterization in rock thanks to technological improvements in X-ray computed tomography (CT) imaging. A new method for the construction of a pore channel model from micro-CT image analysis is suggested to improve computational efficiency by simplifying a highly complex pore structure. Ternary segmentation was applied through matching a pore volume experimentally measured by mercury intrusion porosimetry with a CT image voxel volume to distinguish regions denoted as “apparent” and “indistinct” pores. The developed pore channel model, with distinct domains of different pore phases, captures the pore shape dependence of flow in two dimensions and a tortuous flow path in three dimensions. All factors determining these geometric characteristics were identified by CT image analysis. Computation of an interaction flow regime with apparent and indistinct pore domains was conducted using both the Stokes and Brinkman equations. The coupling was successfully simulated and evaluated against the experimental results of permeability derived from Darcy’s law. Reasonable agreement was found between the permeability derived from the pore channel model and that estimated experimentally. However, the model is still incapable of accurate flow modeling in very low-permeability rock. Direct numerical simulation in a computational domain with a complex pore space was also performed to compare its accuracy and efficiency with the pore channel model. Both schemes achieved reasonable results, but the pore channel model was more computationally efficient.