Hydrogen storage in fractional wet reservoirs: A pore-scale perspective

The subsurface provides a viable option to store large quantities of hydrogen to cater to the growing green energy demands. In this regard, it is essential to understand the fundamental nature of the fluids displacement that may occur in the subsurface to optimize the storage and recovery efficiency of hydrogen. We investigated the hydrodynamics part of the problem using direct numerical simulations by studying the nature of fluid displacements when hydrogen is injected into a tight porous medium totally saturated by brine and vice-versa in a mixed-wet system. From our analysis, it is evident that the wettability distribution morphology plays a significant role in governing the nature of the invasion patterns. Consequently, the displacement efficiency and fractal dimension are also seen to drastically vary for different mixed-wet systems when compared to the homogeneous wet system. For the investigated set of cases, we notice that the clustered wet systems which possess zones of strong imbibition and strong drainage are capable to accommodate large quantities of hydrogen through invasion and by trapping too. On the other hand, while flushing out the stored hydrogen we noticed approximately similar recovery efficiency on clustered wet systems either with intermediate or large wettability contrasts. However, unclustered wettability distributions showed transition flow regimes from viscous fingering to capillary fingering. We noticed that both the displacement efficiency and fractal dimension were much lower than compared to clustered wet systems. The complex nature of fluid flow dynamics in a mixed-wet system cannot be approximated by a uniform contact angle at the reservoir-scale. In reservoir simulators, the effect of mixed wettability must be taken into account in the relative permeability and capillary pressure terms. Results presented in this work can be a starting point in the direction to improve the existing pore network models. This can improve the predictive capabilities of the flow patterns in mixed-wet systems and thereby extract more reliable relative permeability and capillary pressure terms. Furthermore, results presented in this study can also be used to train machine learning algorithms to predict flow patterns and essential multiphase flow parameters on complex mixed-wet systems.