Pore Surface Roughness Quantification: From 2D to 3D

Abstract

Abstract Surface roughness is a persistent challenge that may significantly reduce NMR T2 relaxation time, leading to the misinterpretation of formation and fluid properties. Of particular interest to accurately characterizing pore structures from NMR T2 responses, it is critical to define and quantify surface roughness before modeling its impact on NMR T2 relaxation. Some sophisticated microscope techniques can provide accurate surface roughness measurements; however, these methods probe the surface roughness of selected 2D regions on the rock surface, which may not fully represent surface roughness in 3D space. As the most commonly used 3D imaging technology, micro-computed tomography (micro-CT) digitizes core plugs into voxelized images. The actual surface roughness cannot be visualized as the length scale of surface textures is smaller than the image resolution. However, when modeling NMR T2 relaxation in digitized pore structures, the irregular pore shape increases surface relaxation. From the macroscopic point of view, the irregularity of solid-pore interfaces serves as the visible "roughness" at the given resolution. Thus, in this study, we treat it as surface roughness, and this raises a practical question of how to characterize surface roughness from micro-CT images. We proposed an image-based pore surface roughness characterization method, called 3D PSR, which utilizes a plurality of 2D cross-sectional images to approximate the surface roughness of a 3D volumetric pore structure. The accuracy and reliability of 3D PSR are influenced by pore morphology, as well as the number and orientation of cross-sectional planes. To resolve these issues, in this study, we upgrade the workflow by directly characterizing pore surface roughness in 3D space, which is named 3D PSR 2.0. The key to the success of 3D PSR 2.0 is calculating spherical harmonic (SH) coefficients to reconstruct the pore surface and then building the reference surface to characterize surface roughness. The proposed workflow consists of five steps, including image segmentation, pore separation and diagnosis, surface reconstruction, roughness evaluation, and roughness parameterization. Numerical results demonstrate the successful application of spherical harmonics for 3D pore surface roughness characterization.