Parameter and Grid Sensitivities of Aquifer Models for Underground Hydrogen Storage

The worldwide demand for low-emission hydrogen is expected to rise by ca. 500 % in the time span from 2023 to 2030 due to recent governmental mandates and incentives (IEA Global Hydrogen Review 2024). Aquifers are a widespread and resource-efficient underground hydrogen storage (UHS) possibility, which are - under certain conditions – easier to explore and scale than salt caverns and depleted gas fields. In this study, we quantify the sensitivities of an aquifer UHS model concerning uncertain parameters, the understanding of which is a prerequisite for optimizing operational conditions (e.g. rates and well configurations) and assessing risks for safety and revenue.

A model using Darcy’s law for flow is investigated that describes the motion of two phases (liquid and gas) and three components (water, methane, hydrogen). The model geometry, boundary conditions and parameter distributions are chosen based on real data of a multi-layer sandstone aquifer. The numerical model is implemented in DuMu^X and includes real gas behavior and the computation of gas mixture viscosities.

The following tendencies are observed in the simulations: (1) Dirichlet boundary conditions influence dynamic well pressures if the domain size is chosen too small; (2) converged gas plumes require very fine grid resolutions in some places, which constitutes a large computational cost using regular grids; (3) the caprock’s curvature and dip promote vertical and horizontal motion, respectively; (4) the containment of gases is predominantly controlled by entry pressures; (5) higher heterogeneity in porosity and permeability fields increases gas spreading in all directions and reduces hydrogen recovery. The uncertainty of all of these factors and of petrophysical parameters (porosity, permeability, anisotropy) need to be considered in a future global sensitivity analysis.