A computationally efficient approach to model reactive transport during CO2 storage in naturally fractured saline aquifers

Abstract

Energy storage is critical to any decarbonization, energy sustainability, and energy security strategy. Natural gas, CO2, and hydrogen will compete for sizeable geological storage sites in the future, and society needs to consider all possible storage sites. Saline aquifers and depleted hydrocarbon reservoirs have emerged as key sites for storage. While carbonate rocks possess significant storage potential, as evidenced by successful CO2-enhanced oil recovery (CO2-EOR) projects, most operational and planned Carbon Capture and Storage (CCS) initiatives primarily focus on sandstones. Consequently, extensive fundamental and applied research is necessary to advance technologies for carbonate rock utilization and capitalize on the storage capacities of vast saline aquifers and infrastructure available in near-abandoned depleted carbonate reservoirs. However, the challenges posed by the fracture systems and the substantial heterogeneity inherent to such reservoirs necessitate a comprehensive understanding of rock integrity and the role of fracture networks in the overall storage process. Moreover, modeling CCS in carbonates introduces additional complexities due to reactive minerals, such as calcite and dolomite, which can undergo reactions with injected CO2. Consequently, a numerical formulation capable of producing accurate results within a reasonable computational time is required, considering computationally intensive processes like fracture flow and geochemical reactions. In this work, three approaches to modeling naturally fractured media were assessed, such as the explicit representation of fractures through local grid refinement (LGR), the traditional dual permeability/dual porosity model (DPDK), and the embedded discrete fracture model (EDFM). The combination of EDFM with dynamic gridding in the matrix gridblocks demonstrated superior computational efficiency compared to LGR and DPDK simulations. This novel approach was 75 times faster than LGR and 37 times faster than DPDK while yielding comparable results in terms of different CO2 retention forms (free supercritical fluid, dissolved in water, and residual trapped gas) in one of the geological models investigated. Therefore, it was proposed a novel combination of EDFM and dynamic gridding as a suitable and accurate numerical approach to simulate reactive CO2 transport in fractured saline aquifers. This approach offers a practical means of efficiently evaluating carbonate reservoir targets for CCS projects, enabling effective decision-making in energy storage endeavors.