Dual-energy CT scan protocol optimization to monitor transient fluid saturation distributions during three-phase flow in sand columns

Abstract

Surfactant foam injection is a promising in situ remediation technique for aquifers impacted by Non-Aqueous Phase Liquids (NAPL). Though the viscous properties of foam can promote immiscible displacement, pore plugging by foam is a concern in an environmental remediation (ER) context, as maintaining flow conditions may require the application of pressures that exceed soil matrix failure thresholds. Optimization of the technology for ER may be furthered by gaining a better understanding of the relationship between foam morphology and NAPL desaturation. This relationship is poorly understood, partly because the interactions between organic phases and foam occurring at specific positions within the flow field at any given time are difficult to assess quantitatively. A Dual-energy X-ray Computed Tomography (CT) technique was developed to enable the measurement of sand column characteristics, as well as transient fluid saturation profiles within the porous medium during foam injection. The available settings of a medical CT scanner were systematically explored in order to identify the acquisition and reconstruction parameters best suited to image the fluids of interest: 0.4 mm slice thickness, a medium kernel (U40u), peak tube voltages of 140 kVp and 80 kVp, and the addition of a doping agent to the NAPL at a concentration of 20 wt%. Using datasets produced by imaging sand columns at various stages of foam injection experiments, an algorithm was developed in order to extract saturation profiles from CT images. The algorithm included multiple scan averaging to reduce image noise, image registration, and specific matrix operations required for calculating fluid saturation values in two- and three-phase systems. The accuracies of the CT-extracted porosity and fluid saturation data were compared to, and found to be in good agreement with, experimentally determined average porosity and saturation values. The scan parameter optimization method developed here can be applied to future studies involving any combination of site-specific natural materials.