Hydrogeochemical modeling of hydrogen storage in depleted gas reservoirs: Insights from local and global sensitivity analysis

Underground hydrogen storage (UHS) offers a solution for large-scale energy storage by addressing the challenges of location dependence, fluctuation, and forecasting inherent in renewable energy sources. To increase renewable energy storage capacity from GWh to TWh, sites with large capacity and easy accessibility need to be investigated. Hydrogen (H2) is a reductant in redox reactions with subsurface minerals and aqueous species. Therefore, modeling three-phase (gas–brine–mineral) reaction networks is crucial for understanding H2(g) behavior during underground storage. To meet this need, we performed time-dependent batch simulations of subsurface H2(g) loss using PHREEQC. Subsequently, the static and kinetic equilibrium reaction results were integrated into multiphase reactive transport models in PFLOTRAN, providing a complete simulation of H2(g) loss and hydrogen sulfide (H2S(g)) formation under storage conditions. The resulting suite of simulations reflects an abiotic reaction network among H2(g) and subsurface minerals in the H2–H2O–CO2–SiO2–Ca–Al–Fe–S system. The simulations highlight that quartz-rich sandstone reservoirs are ideal for hydrogen storage, with minimal H2(g) reactivity and dissolution in brine leading to around 1% total loss over a year. Reduction of sulfur (S) and ferric iron (Fe(III)) primarily controls subsurface H2(g) loss, contributing an additional 1 %–4 % loss and showing potential evidence of an autocatalytic effect. Finally, we conducted local and global sensitivity analysis over extensive simulations and a broad parameter space to identify water saturation and temperature as the key parameters influencing H2(g) loss and H2S(g) formation. Results underscore the importance of reducing uncertainty in minerals’ kinetic rate expression and reactive surface area.