资源环境科技发展态势分析平台

An accurate description of the relative permeability–saturation function is crucial for reliable predictions of multi-phase flow behavior in subsurface applications. Although extensive efforts have been put forth to investigate the relative permeability behavior in different types of porous media, only few studies have focused on rough-walled fractures. In this work, we present an entirely new, cost-effective, heavily-parallelized, dynamic pore-network modeling framework that is employed to conduct a systematic study of relative permeability curves under two-phase flow conditions in rough-walled fractures. We first build a two-dimensional (101.42 × 24.86 mm²) equivalent pore network of a Berea sandstone fracture from its x-ray images. Subsequently, dynamic primary drainage and imbibition simulations are conducted in the fracture. We show that the two-phase fluid occupancy maps predicted from the simulations agree well with the fracture fluid configurations obtained via x-ray computed tomography. Afterwards, the validated model is used to probe two-phase flow properties in a series of synthetic aperture fields generated with a broad range of geometric characteristics including aperture spatial correlation length (normalized correlation length varying from 0.05 to 0.95), anisotropy factor (0.25 ∼4), surface roughness (normalized fracture roughness varying from 0.05 to 0.4), and mean aperture size (50 ∼800 µm). The generated results provide novel insights into the effects of these features on two-phase flow properties such as fluid–fluid interfacial area, phase interference, and relative permeability. Moreover, based on the simulation results we propose two new correlations to describe the relative permeability curves for primary drainage and imbibition processes in rough-walled fractures.