Impact of Equation of States in Underground Hydrogen and CO2 storage and Material Balance Analysis

In the ongoing energy transition era, depleted gas reservoirs (DGR) are considered valuable candidates for underground hydrogen and carbon dioxide storage, offering large storage capacities and geological containment. However, the successful implementation of these technologies relies on the ability to accurately predict fluids’ dynamics behaviour at reservoir conditions. This work aims to assess the impact of different Equations of State (EoSs) on the results when simulating underground H2 and CO2 storage. Furthermore, a Material Balance Analysis (MBA) using the p/z method is performed in order to evaluate the accuracy of gas storage predictions by comparing the results obtained from the simulator with those derived from MBA. A simplified reservoir model was developed in tNavigator® (Rock Flow Dynamics) which is, to the best of our knowledge the only one commercial reservoir simulator which has natively implemented the GERG model. The scenarios simulated are the following: first CH4 production, followed in one case by CO₂ injection and in the other case by H2-CH4 mixtures injection. In the latter scenario three different injection with varying gas composition are simulated: 80% methane + 20% hydrogen, 50% methane + 50% hydrogen and 100% hydrogen. A sensitivity analysis was performed to quantify how the choice of EOS Peng–Robinson (PR), Redlich–Kwong- Soave (RKS), affects predictions in the compositional model respect to the GERG-2008, since prior studies have demonstrated the accuracy of the GERG-2008 in predicting fluid properties, although the high computational cost. Constant Composition Expansion (CCE) tests were simulated to assess thermodynamic properties of gas mixtures at reservoir thermodynamic conditions. In particular the gas compressibility factor (Z-factor) is obtained, and it is essential for applying the material balance method, relevant to evaluate the reservoir performance. However, during the modelling of the gas injection phases a critical issue raised: the CCE tests is applied to a mixture with constant composition, but during injection, the composition of reservoir fluids changes continuously. Therefore, the CCE method cannot predict the thermodynamic properties of the gas mixtures during such phase, leaving a critical gap in the ability to perform accurate material balance calculations for the injection phases. To overcome this limitation, an alternative analytical method for calculating the Z-factor and gas density during periods of non-constant composition is used. The method is based on a reformulation of the real gas law, utilizing analytical expressions for gas density and mixture molar mass. The methodology was validated against production phase data by comparing its results with the outputs from the CCE test. The comparison showed a very low Mean Absolute Percentage Error (MAPE) of less than 1.5% for all tested Equations of State, confirming the accuracy of the approach and supporting it’s use in the injection stages. Since during the injection, pressure increases, the p/z plot was adapted and allowed to retrieve the Max Injectable Gas (MIG) in volume and mass with low absolute percentage error (APE), compared to simulated data. The error analysis showed APE equal or below 5% for both mass and volume estimation in all the scenario tested. Overall, this study validates an analytical approach to estimate Z-factor from simulation outcomes during injection phase and evaluates the accuracy of the p/z method in prediction the MIG for a given maximum pressure.