Modeling Permeability Distributions in a Sandstone Core for History Matching Coreflood Experiments

Abstract

Summary Saline-aquifer storage of carbon dioxide (CO2) has become recognized as an important strategy for climate-change mitigation. Saline aquifers have very large estimated storage capacities, are distributed broadly across the globe, and have the potential for geologic-scale retention times. Many of these storage sites are not well characterized, and it is critical to conduct detailed experiments and analysis to understand how features such as heterogeneity can influence the theoretical storage capacity, spatial extent of plume migration, and secondary trapping processes. Coreflooding experiments are used routinely by the oil and gas industry for such analysis and provide a very useful tool for studying saline-aquifer formations also. Numerical simulations of these coreflooding experiments can provide insight beyond the experimental measurements themselves, such as numerically studying how properties such as relative permeability and capillary pressure affect CO2 distribution in these systems under various flow conditions. However, accurate subcore-scale simulations of these experiments have remained a challenge, and the issue of how to represent subcorescale permeability has not been resolved previously. Laboratory coreflooding experiments injecting CO2 into a saline-water-saturated Berea sandstone core have been conducted at reservoir conditions. Computed-tomography (CT) scans of the core show large spatial variations of CO2 saturation, even within a relatively homogeneous core. Numerical simulations of the experiment have been conducted to study the effect of subcorescale heterogeneity and the role of permeability in determining the subcore-scale CO2 distribution in the core to explain these very large spatial variations in CO2 saturation. Numerical simulations of the experiment consistently showed that use of traditional methods for estimating subcore-scale permeability, typically based solely on porosity distributions, results in subcore-scale saturation distributions that do not match experimental measurements. In this paper, we develop a new method for calculating subcore-scale permeability distributions on the basis of capillary pressure measurements and porosity distributions as an alternative to the traditional porosity-only-based models. Using experimentally measured saturation and porosity distributions and capillary pressure data to calculate permeability, simulations based on this new method show a substantial improvement both in the absolute value and in the spatial distribution of predicted CO2-saturation values. With this technique for accurately calculating permeability distributions, it is possible to study subcore-scale multiphase flow of brine and CO2 to understand how small-scale heterogeneities influence the spatial distribution of CO2 saturation and to improve our ability to predict the fate of stored CO2.