Insights into site evaluation for underground hydrogen storage (UHS) on gas mixing-the effects of meter-scale heterogeneity and associated reservoir characterization parameters

Abstract

Underground Hydrogen Storage (UHS), as an emerging large-scale energy storage technology, has shown great promise to ensure energy security with minimized carbon emission. A set of comprehensive UHS site evaluation criteria based on important factors that affect UHS performances is needed for its potential commercialization. This study focuses on the UHS site evaluation of gas mixing. The economic implications of gas mixing between injected hydrogen gas and the residual or cushion gas in a porous storage reservoir is an emerging problem for Underground Hydrogen Storage (UHS). It is already clear that reservoir scale heterogeneity such as formation structure (e.g. formation dip angle) and facies heterogeneity of the sedimentary rock may considerably affect the reservoir-scale mechanical dispersion-induced gas mixing during UHS in high permeability braided-fluvial systems (a common depleted reservoir type for UHS). Following this finding, the current study uses the process-mimicking modeling software to build synthetic meandering-fluvial reservoir models. Channel dimensions and the presence of abandoned channel facies are set as testing parameters, resulting in 4 simulation cases with 200 realizations. Numerical flow simulations are performed on these models to investigate and compare the effects of reservoir and metre-scale heterogeneity on UHS gas mixing.

Through simulation, channel dimensions (reservoir-scale heterogeneity) are found to affect the uncertainty of produced gas composition due to mixing (represented by the P10-P90 difference of hydrogen fraction in a produced stream) by up to 42%. The presence of abandoned channel facies (metre-scale heterogeneity), depending on their architectural relationship with meander belts, could also influence the gas mixing process to a comparable extent (up to 40%). Moreover, we show that there is no clear statistical correlation between gas mixing and typical reservoir characterization parameters such as original gas in place (OGIP), average reservoir permeability, and the Dykstra-Parsons coefficient. Instead, the average time of travel of all reservoir cells calculated from flow diagnostics shows a negative correlation with the level of gas mixing. These results reveal the importance of 3D reservoir architecture analysis (integration of multiple levels of heterogeneity) to UHS site evaluation on gas mixing in depleted gas reservoirs. This study herein provides valuable insights into UHS site evaluation regarding gas mixing.