For large scale seasonal storage, Underground Hydrogen Storage (UHS) can be used to balance fluctuating sustainable energy generation and energy demand. Similar to underground natural gas storage, depleted gas fields potentially allow for cost-efficient hydrogen storage. One of the major cost factors in UHS is the amount of cushion gas required and the purity of the hydrogen produced during the production cycle.
The hydrocarbon gas remaining in the reservoir can be used as cushion gas to significantly reduce UHS costs. To evaluate the composition of the gas produced during the production cycle of UHS, numerical simulation was applied. One of the important processes in UHS is molecular diffusion within the reservoir. The hydrogen recovery factor and methane to hydrogen production ratio were compared for cases with and without diffusive mass flux. Furthermore, a sensitivity analysis was carried out to identify important factors for UHS. The following parameters were investigated: permeability contrast, vertical to horizontal permeability ratio, reservoir heterogeneity, binary diffusion coefficient, and pressure dependent diffusion. In addition, the effects of numerical dispersion on the results were evaluated and are discussed.
The results of numerical simulation show the importance of diffusion on hydrogen storage in depleted gas reservoirs. Molecular diffusion plays a major role in case of heterogeneous reservoirs and large permeability contrasts. In low permeability zones, the diffusive mass transport becomes dominant over advective flux. Hydrogen propagates further into the low permeable layers of the reservoir when diffusion effects are considered compared with the cases neglecting diffusion. Similar effects are observed during the production cycle. Hydrocarbon gas in low permeability zones becomes more mobile due to diffusive transport. Thus, a larger amount of methane is back-produced with hydrogen for the cases when diffusion is simulated. It is shown that if molecular diffusion is ignored, the hydrogen recovery factor can be overestimated by up to 9% during the first production cycle and the onset of methane contamination can be underestimated by half of the back production cycle. Simulating pressure dependent diffusion might be important for specific configurations and should be covered in a sensitivity.
The results show that molecular diffusion within the reservoir has an impact on the onset of methane contamination when hydrocarbon gas is used as cushion gas in UHS. Also, the total amount of hydrogen produced is overestimated. For UHS operations, both, the composition and amount of hydrogen is important to design facilities and to determine the economics of UHS and hence diffusion should be evaluated in UHS simulation studies.