Impact of structural heterogeneity on upscaled models for large-scale CO_2 migration and trapping in saline aquifers

摘要

Structural heterogeneity of the caprock surface influences both migration patterns and trapping efficiency for CO_2 injected in open saline aquifers. Understanding these mechanisms relies on appropriate modeling tools to simulate CO_2 flow over hundreds of square kilometers and several hundred years during the postinjection period. Vertical equilibrium (VE) models are well suited for this purpose. However, topographical heterogeneity below the scale of model resolution requires upscaling, for example by using traditional flow-based homogenization techniques. This can significantly simplify the geologic model and reduce computational effort while still capturing the relevant physical processes. In this paper, we identify key structural parameters, such as dominant amplitude and wavelength of the traps, that determine the form of the upscaled constitutive functions. We also compare the strength of these geologic controls on CO_2 migration and trapping to other mechanisms such as capillarity. This allows for a better understanding of the dominant physical processes and their impact on storage security. It also provides intuition on which upscaling approach is best suited for the system of interest. We apply these concepts to realistic structurally heterogeneous surfaces that have been developed using different geologic depositional models. We show that while amplitude is important for determining the amount of CO_2 trapped, the spacing between the traps, distribution of spillpoint locations, large-scale formation dip angle affect the shape of the functions and thus the dynamics of plume migration. We also show for these cases that the topography characterized by shorter wavelength features is better suited for upscaling, while the longer wavelength surface can be sufficiently resolved. These results can inform the type of geological characterization that is required to build the most reliable upscaled models for large-scale CO_2 migration.