Underground hydrogen storage (UHS) in porous media is expected to play a key role in enabling large scale, long duration energy storage for renewable energy systems. The feasibility of UHS in depleted gas reservoirs depends strongly on multiphase flow behavior and gas trapping during injection and withdrawal cycles. This study experimentally investigates the co - flow of hydrogen, methane, and a 50–50% molar hydrogen methane mixture with 1 % w.t NaCl brine in a Berea sandstone core under UHS relevant conditions. Steady state horizontal core flooding experiments at 18 °C and 100 bar, combined with medical X-ray CT imaging, were used to determine relative permeability (kr) and capillary pressure (Pc) during drainage and imbibition cycles. The results showed that hydrogen exhibited slightly higher kr than methane during drainage, with near end point kr values of 0.0399 and 0.0233, respectively, at a gas fractional flow ( 𝑓𝑔 ) 𝑓𝑔 = 0.99, reflecting hydrogen’s lower viscosity and higher mobility. During imbibition, both methane and the mixture showed marginally higher kr than hydrogen, indicating improved producibility. The mixture displayed intermediate behavior: at 𝑓𝑔 = 0.8 the mixture’s drainage (kr=0.0022) and imbibition 2 ( kr=0.0015) kr values, and Pc (Pc = 0.18–0.41 bar) fell between those of hydrogen (kr = 0.0014 , Pc = 0.22–0.40 bar) and methane (kr = 0.0016, Pc = 0.14–0.33 bar). Hydrogen showed greater hysteresis than either methane or the mixture, with linear trapping coefficients (A) of 0.66–0.459 versus 0.62–0.35 for methane, and 0.15 for the mixture across successive drainage/imbibition cycles. The lower mixture hysteresis suggests higher recovery efficiency than hydrogen gas. The experimentally derived kr and Pc data provide valuable input for UHS modeling, reservoir simulation, and injection/withdrawal optimization. This study concludes that mixtures of hydrogen and methane will not exhibit significantly different kr/Pc under relevant UHS conditions, and simulations may treat the mixed systems similarly to the pure systems. However, experimental errors (including the potential for trapped residue gas from previous gas type experiments in the core) make conclusions from these outcomes problematic and future studies may be required to verify the results.

Geothermal energy is expected to play a major role in decarbonizing heat supply in the Netherlands. Knowing the geomechanical behavior of a reservoir and seal can be crucial when engineering in the subsurface[5]. That is why this thesis focuses on characterizing the geomechanical behavior of the upper Delft Sandstone. Due to limited prior research and known lithological heterogeneity, this study seeks to quantify the constitutive behavior of the upper Delft Sandstone Member. Another goal is to measure the static moduli of the upper Delft Sandstone Member and to try to predict these moduli in parts of the reservoir where no cores are available.
From 87 m of core from well DEL-GT-01, 39 suitable samples were tested. Petro-physical measurements and classification into sandstones and shales were followed by UCS (15 samples), triaxial (13 samples), and Brazilian Disc Tests (11 samples). These provided strength, yield stress, Young’s modulus, Poisson’s ratio, and tensile strength.
Results show weak correlations between porosity and strength, but stronger samples had higher Young’s modulus and yielded later. Coarse, organic-rich sands were weaker, while P- and S-wave velocities correlated with strength parameters. Triaxial tests confirmed increasing yield point with confining pressure and lower friction angles in shales.
A linear model predicted static Young’s modulus in parts of the reservoir where no cores have been taken from dynamic moduli with an average absolute error of 2.6 GPa. Uncertainty remains due to limited sample data and high reservoir heterogeneity (REV larger than sample scale). The results in this study were compared to a geomechanical research of the IJsselmonde formation, which is a very similar formation. The data was not found to be very comparable, due to slightly different formations and different test techniques. Nevertheless, this data was the most comparable data out there.
Answering the main research question is and will remain difficult, because one can never know the geomechanical behavior of the complete reservoir. This study however did make a good start. It is now known that the upper Delft Sandstone member is moderately weak to strong. A correlation within the sands was also found with the grain size, or the amount of organic grains. It is not certain which of the two had the greatest effect. The ranges of dynamic moduli are known and we know what we need to test if we want to reduce the uncertainty within certain correlations.
This study shows the importance of extensive geomechanical data collection since the geomechanical behavior is not that easily explained.