ABSTRACT
A method is presented for interpreting the values of x-ray attenuation coefficients reconstructed in computed tomography of porous media, while overcoming the ambiguity caused by the multichromatic nature of x-rays, dilution by void, and material heterogeneity. The method enables determination of porosity without relying on calibration or image segmentation or thresholding to discriminate pores from solid material. It distinguishes between solution-accessible and inaccessible pores, and provides the spatial and frequency distributions of solid-matrix material in a heterogeneous medium. This is accomplished by matching an image of a sample saturated with a contrast solution with that saturated with a transparent solution. Voxels occupied with solid-material and inaccessible pores are identified by the fact that they maintain the same location and image attributes in both images, with voxels containing inaccessible pores appearing empty in both images. Fully porous and accessible voxels exhibit the maximum contrast, while the rest are porous voxels containing mixtures of pore solutions and solid. This matching process is performed with an image registration computer code, and image processing software that requires only simple subtraction and multiplication (scaling) processes. The process is demonstrated in dolomite (non-uniform void distribution, homogeneous solid matrix) and sandstone (nearly uniform void distribution, heterogeneous solid matrix) samples, and its overall performance is shown to compare favorably with a method based on calibration and thresholding.
ABSTRACT
X-ray micro-computed tomography (micro-CT) techniques for measuring the three-dimensional (3-D) distributions of diffusion-accessible porosity (φ(d)) and temporal tracer-concentrations (C(t)) within a dolostone sample subjected to solute diffusion are developed and tested in this work. The φ(d) and C(t) measurements are based on spatially resolved changes in X-ray attenuation coefficients in sequentially acquired 3-D micro-CT datasets using two (calibration and relative) analytical approaches. The measured changes in X-ray attenuation coefficient values are a function of the mass of X-ray absorbing potassium-iodide tracer present in voxels. Mean φ(d) values of 3.8% and 6.5% were obtained with the calibration and the relative approaches, respectively. The detection limits for φ(d) measurements at individual voxel locations are 20% and 36% with the calibration and the relative methods, respectively. The detection limit for C(t) are 0.12 M and 0.22 M with the calibration and the relative approaches, respectively. Results from the calibration method are affected by a beam-hardening artifact and although results from the relative approach are not affected by the artifact, they are subject to high detection limits. This work presents a quantitative assessment of micro-CT data for studies of solute transport. Despite limitations in precision and accuracy, the method provides quantitative 3-D distributions of φ(d) and C(t) that reflect solute diffusion in heterogeneous porous geologic media.
