Numerical investigation of gas mixing resulting from dispersion across scales during Underground Hydrogen Storage (UHS)

Underground Hydrogen Storage (UHS) offers a relatively safe and economical option to store hydrogen seasonally, where the gas mixing process poses considerable impacts. Therefore, accurate modeling of gas mixing during UHS in depleted gas reservoirs has become a hotly-debated topic recently. Previously, researchers have simulated UHS gas mixing from two perspectives: one employs realistic 3D models but overlooks small-scale processes and uses theoretical simulation inputs; the other uses simplified models capturing multi-scale processes but misses reservoir-scale heterogeneity. This lack of comprehensive evaluation limits understanding of UHS reservoir-scale gas mixing. In this study, we perform compositional numerical simulation with various molecular diffusion, dispersivity, heterogeneity, and reservoir grid cell sizes to present a comprehensive reservoir-scale UHS gas mixing evaluation framework considering various investigation scenarios. By incorporating the realized simulation input of these components, we demonstrate their relative importance from various aspects. Specifically, heterogeneity and coarse grid cells ( > 40 m) contribute orders of magnitude greater gas mixing than molecular diffusion and dispersivity. Coarse reservoir grid cells result in at least 10% and 1.8% of uncertainty in terms of hydrogen fraction in the produced stream and hydrogen mass recovery, respectively. Despite the minor contribution, dispersivity causes considerable effects on the hydrogen plume geometry. We demonstrate that it is computationally prohibitive to consider every gas-mixing related physics at every scale, and modelers must make trade-offs between simulation accuracy and affordability based on specific simulation targets. We provide insights into simulation strategies and data acquisition for modeling reservoir-scale UHS gas mixing.