Super-resolution Reconstruction of Shale Digital Core Images Guided by Fluid Flow Mechanisms

Digital core techniques play a crucial role in analyzing reservoir rock parameters for seepage simulation, but high-resolution images are often difficult to obtain due to scanning costs and image matching issues. Image super-resolution (SR) reconstruction techniques show great promise in restoring digital rock image features, yet previous studies have primarily evaluated the reconstruction effects from a computer vision perspective, rather than in the context of computational fluid dynamics, and hence hinders the application of this method in flow simulations. This study proposes an improved method based on very deep super-resolution (VDSR), leveraging velocity field to correct the pore structures in shale images. The velocity field is used to adjust the pore structures, aligning the flow parameters and pore properties in the SR images more closely with those in the original high-resolution images. We validate our approach using real paired scanning electron microscopy (SEM) images from five distinct regions and three-dimensional (3D) digital rock images of matrix shale, fractured shale and fracture with proppant. Our approach efficiently reproduces the fine-scale fluid flow in porous media, as demonstrated by comparisons of the porosity, velocity fields and equivalent permeability. Additionally, our method exhibits superior applicability in seepage simulation due to its ability to resolve finer pore structures. By addressing the gap between image reconstruction and flow simulation, our study provides a more accurate and efficient way to analyze reservoir rock parameters, ultimately enhancing the predictive capabilities for oil and gas production.