Dynamic Iterative Estimation of Three-Phase Saturation in Heterogeneous Core Using Single-Energy X-Ray Computed Tomography

Abstract

Single-energy X-ray computed tomography (CT) enables repeated in-situ imaging of multiphase flow in centimeter-scale cores, but oil/water/gas three-phase saturation estimation remains underdetermined because each voxel provides only one gray-value response. With this study, we develop a three-phase saturation estimation workflow for water-alternating-gas (WAG) displacement in a layered heterogeneous core by combining macroscopic material-balance constraints with dynamic iteration. The method first identifies the gas phase using the CT-response contrast between gas and liquid phases, and removes the gas contribution from the measured CT signal. Cumulative oil production is then used to constrain the whole-core hydrocarbon-equivalent remaining volume, while a dynamic oil/water threshold is determined within the liquid-voxel set, and the liquid-phase CT baseline is iteratively updated. Compared with the two-phase reduction, fixed-threshold (fixed-Tow), and static liquid-baseline workflows, the dynamic iterative method maintained low material-balance closure deviation while retaining relatively low CT forward-reconstruction residuals. Across the one-at-a-time parameter tests, the largest change in whole-core material-balance closure mean absolute error (MAE) was 0.0415, and the complete processed sequence required approximately 167 seconds, only 2.4% longer than the fixed-Tow workflow. Three-phase spatial distributions at representative displacement stages further show that the method can more continuously characterize preferential sweep, water invasion, and gas-channel development. The workflow was demonstrated in a layered artificial core with permeabilities of approximately 2.0–10.0 md; its applicability to substantially lower-permeability systems, stronger compositional mass transfer, or poorly separated oil/water CT responses remains to be independently evaluated. The proposed workflow provides a practical approach for quantitative three-phase saturation characterization under conventional single-energy CT conditions.