A coupled numerical model to assess the caprock geochemical integrity and porosity change of an underground hydrogen gas storage system

Hydrogen gas during subsurface storage in deep geological formations is prone to losses due to its unique hydrodynamic and chemical properties. These losses mostly induced by hydrogeochemical interactions between the stored hydrogen gas, brine and caprock minerals amount to significant energy losses. The interactions in turn result in alterations to the geochemical integrity and porosity of the caprock formation, which has a direct impact on the caprock sealing capacity. It is established that the major objective of a gas storage project is confinement of the stored gas, which can be achieved by storage in a formation with a relatively high caprock sealing capacity. The understanding of the impact of hydrogen gas-brine interactions on caprock geochemical integrity and porosity is currently limited and hence further research with novel outcomes is required. We therefore develop a numerical geochemical model and use it to quantify and investigate the change in porosity of a caprock formation of a depleted gas reservoir and establish the reason for the porosity change. Due to the scarcity of experimental or field data, our work is realised by expanding upon an earlier numerical modelling effort from the published literature. The initial modelling code is the PHREEQC geochemical software package, which is used to investigate the changes in the geochemical parameters that influence hydrogen gas losses during the underground storage of the gas. IPhreeqc, an interfacing version of PHREEQC is next coupled with the Matlab Reservoir Simulation Toolbox (MRST) and used to verify the accuracy and performance of the developed geochemical model. The hydrogen gas-caprock mineral-brine interactions result in a decrease in the caprock's porosity despite the dissolution of all the primary minerals in the caprock formation. The precipitation of albite, a secondary mineral was observed to be the likely reason for the decrease as it countered the net effect of the primary minerals' dissolution on the caprock porosity. Good matches in the comparisons of the PHREEQC geochemical modelling results with those of the coupled IPhreeqc-Matlab model proved the accuracy and performance of the developed model. The novelty of our study lied in the quantification of the decrease in the caprock's porosity from 5.0% to 3.82% and the identification of the mechanism responsible for the porosity decrease.