Bridging 2D and 3D: High-fidelity pore network construction for sandstone transport property prediction

We characterize the complex pore architecture of four diagenetically altered sandstone samples using micro-CT imaging. By quantifying the pore size and pore connection distributions, we establish that the mean coordination number follows a power-law function of pore radius. Critically, we derive empirical scaling relations between 2D and 3D pore size distributions, as well as 2D and 3D coordination number distributions, both as a function of pore size. These novel relations allow for the 3D pore size and pore connectivity to be accurately characterized from high-resolution 2D imaging (e.g., SEM), even when the resolution or field-of-view limitations preclude detailed 3D micro-CT analysis. A new stochastic algorithm is presented in this work to generate a 3D pore network model that honors the derived connectivity function. This algorithm was validated by generating stochastic models equivalent to the four sandstone samples and successfully comparing the calculated single-phase and relative permeability results against those obtained from pore networks directly extracted from the 3D micro-CT images. This methodology offers a computationally efficient and scalable tool for predicting the macroscopic transport properties of complex reservoir rocks based on readily available 2D high rock resolution images.