Fluid Property Characterization for Underground Gas Storage

Abstract

Numerical simulation is a widely employed technique for investigating the storage of carbon dioxide (CO2) or hydrogen (H2) in porous media. Consequently, it is crucial to have an accurate characterization of the properties of CO2, H2, and other gases such as methane (CH4), which necessitates the use of experimental data to support the numerical modeling studies. These data are utilized for calibrating empirical model parameters. In this study, laboratory data for pure components (H2, CH4, and CO2) and their mixtures were collected from various sources. The data encompassed density and viscosity measurements across a wide range of pressures, compositions, and temperatures. Measurements of H2 and CO2 solubility in pure water and brine with different salinities were also included. The Peng–Robinson Equation of State (PR EoS) was used to calculate the density while different models were calibrated to match viscosity and solubility. The results demonstrate significant disparities between the properties calculated using default model parameters and those after parameter tuning, particularly for viscosity and solubility values, as well as the density at high pressures. The models with tuned parameters had accuracy comparable to the uncertainties of the measurements. For density, the simulations had an accuracy within 1% of the measured values, while for viscosity, the accuracy was within 5% of the measured values. In terms of solubility, the models could also provide reliable predictions (errors lower than 5% for different brine salinities). These calibration steps were crucial in generating representative fluid models for the numerical simulation of gas storage in geological formations.