CCS (carbon dioxide capture and storage) is a promising option for mitigating climate changes. To predict the fate of injected CO2 in a deep reservoir, relative permeability (the ratio of the effective permeability of a fluid at a given saturation to absolute permeability) of supercritical CO2 and water of the reservoir rock is considered to be one of the most fundamental and influential properties. For determining the relative permeability, the steady-state method has been widely used as a reliable method, although it needs a complex apparatus and is time-consuming. An alternative method is the unsteady-state method, in which the relative permeability is determined based on history matching of transient monitoring data (fluid pressure and production rates of CO2 and water) with a multi-phase flow model. The unsteady-state method is relatively simple and short, but obviously its accuracy strongly depends on the flow model employed in the history matching. In this study, we conducted relative permeability measurements of supercritical CO2-water system for Berea sandstone with the unsteady-state method under a reservoir condition at a 1km depth (pressure = 9.5MPa, temperature = 44°C). Automatic history matching was performed with an inversion simulator iTOUGH2/ECO2N for multi-phase flow system of supercritical CO2, NaCl, and water. Our experiment also focuses on the impact of injection rate on the estimates of relative permeability, as it is known that the injection rate could have a significant effect on fluid distribution such as viscous fingering with changes in the ratio of the viscous to the capillary forces (i.e., capillary number). Our results indicate that the inversion estimates of relative permeability from the unsteady-method are well comparable with that from the steady method acquired in previous studies, although it is also indicated that careful considerations must be made on the dependence of the injection rate.

This content is only available via PDF.
You can access this article if you purchase or spend a download.