Quantitative Comparison of Magnetic Pore Fabric and Pore Fabric Characterization Based on High-Resolution X-Ray Computed Tomography in Synthetic Samples and Sedimentary Rocks

Abstract:

Pore fabrics characterize the pore size, shape, distribution and connectivity in rocks, which are crucial to predict the preferred direction of fluid flow. The measurement of those properties using X-ray computed tomography (XRCT), the most common method to describe 3D pore fabrics, is challenging due to the trade-off between resolution and field of view. Magnetic pore fabrics (MPF), determined by measuring the anisotropy of magnetic susceptibility after impregnating samples with ferrofluid, show promising empirical correlations to the pore fabric in terms of pore orientation and anisotropy degree. However, these empirical relationships show large variability between studies, possibly related to the ferrofluid susceptibility and measurement frequency. This study aims at a more complete and systematic understanding of MPFs. Pore fabrics determined from XRCT have a voxel resolution of 15 μm, and are compared to MPFs, supposedly capturing pores down to 10 nm, for synthetic and natural samples. The synthetic samples have the advantage that their pore size can be controlled, and is more uniform than in natural rocks. At the same time, synthetic samples are prepared to replicate reservoir rock characteristics by using naturally occurring materials, e.g, quartz sand and calcite powder. Natural rock samples used in this study include upper marine molasse (OMM) with parallel or cross bedding, i.e., anisotropic sedimentary rock. The OMM sandstone is widely distributed in the Swiss molasse basin (SMB), and forms aquifers with 5-20% porosity and up to 650 mD permeability. From the XRCT data, the shape of each pore is simplified as external best-fit ellipsoid, mathematically described by a second order tensor, and average total shape ellipsoid is calculated by tensor addition. This representation reduces the influence of resolution-related artefacts. Bootstrapping is used to evaluate the statistical properties of the total shape ellipsoid. The orientation is described by the eigenvectors of the tensor. The anisotropy degree is the ratio of maximum and minimum eigenvalue of tensor, and the shape is also calculated from the eigenvalues and distinguishes between oblate and prolate ellipsoids. Then the preferred orientation, anisotropy degree and shape of the total pore shape ellipsoids are compared with MPFs, which are by definition second-order tensor properties. Initial results suggest that the maximum and minimum susceptibility axes of the MPF are sub-parallel to the major and minor axes of the total pore shape ellipsoid at 95% confidence. Conversely, there is no distinct correlation between the MPF and the orientation density function of maximum axes of the individual pore shape ellipsoids. This might be related to the presence of non-resolved porosity in the XRCT-data, i.e. the presence of pores < 4x4x4 voxels³. The MPF anisotropy degree depends on the intrinsic susceptibility of the ferrofluid, and is lower than the anisotropy degree of the total shape ellipsoid. The anisotropy shapes for both MPF and average pore shape are similar. The quantitative relationships between MPFs and pore fabrics presented in this study will result in a more robust interpretation of magnetic data, leading to more efficient pore characterization and the potential to analyze large datasets in a time-efficient and cost-effective manner. This will allow characterizing regional-scale pore fabric variations often crucial in geo-engineering.