A Practical and Innovative Workflow to Support the Numerical Simulation of CO2 Storage in Large Field-Scale Models

Abstract

Numerical simulation of the CO2 storage process in porous media, such as in hydrocarbon (gas or oil) depleted reservoirs and in saline aquifers, has been the most indicated tool due to its ability to represent CO2 capacity and the different trapping mechanisms that retain CO2 in the subsurface. Given the complexity of the physicochemical phenomena involved, the modeling needs to incorporate multiphase flow, complex representation of fluids, rock, and rock-fluid interaction properties. These include CO2 reactions with aqueous species and with reservoir rock minerals, in addition to the structural and stratigraphic aspects of the reservoir heterogeneity. These phenomena need to be represented on suitable temporal and spatial scales for accurate predictions of their impacts. Currently, many studies are focused on simulating submodels or sectors of the reservoir, where using finer grids is still practical. This level of grid refinement can be prohibitive, in terms of simulation times, for modeling the entire reservoir. To address this challenge, we propose a new and practical workflow to simulate CO2 storage projects in large field-scale models. When the proposed workflow is applied in both synthetic and real field cases, simulation time is reduced by up to 96% compared to that of the fine-grid model, preserving the same results in representing the aforementioned mechanisms. The workflow is based on classical and standard approaches to handle the high simulation time, but in this study, they are structured and sequenced in three steps. The first one considers the most relevant mechanisms for CO2 storage, ranked from a high-resolution sector model. With the mechanisms prioritized in the previous step, a single-phase upscaling of petrophysical properties can be applied in the field-scale model, followed by adopting a grid with dynamic sizing. The proposed methodology is applied to saline aquifer models in this study, but it can be extended for storage in depleted hydrocarbon reservoirs.