Comparison of two different equations of state for application of carbon dioxide sequestration

Abstract

We suggest that different equations of state (EOS) algorithms can and frequently will provide very different predictions of CO 2 migration following injection for sequestration. Rather than carry out an exhaustive examination of all EOS algorithms available, we elected to evaluate this general hypothesis by making detailed comparisons of simulation results of two very common EOS algorithms. We simulated and compared CO 2 migration patterns using two fundamentally different EOS algorithms – Modified Redlich-Kwong EOS (MRKEOS) and Span and Wagner EOS (SWEOS). In general, the predictions of thermophysical properties for both algorithms are close, except for a contrast in the predicted fugacity coefficient of CO 2, which subsequently propagates to a contrast in predicted solubility in water/brine. Typically, MRKEOS underestimates solubility of CO 2 compared to both SWEOS and experimental solubility data. In simulations of CO 2 migration, dissolution rates of separate-phase CO 2 predicted from the two EOS algorithms were significantly different, even for small contrasts in predicted fluid properties from EOS algorithms, resulting in markedly different migration patterns. We also examined the potential disparities of simulating integrity of caprock using these two common EOS algorithms. To simplify the analysis and to isolate the roles of specific properties, we limited these simulations to one dimension. Simulation results from both EOS algorithms indicate that the distance that separate-phase CO 2 migrates through an unfractured caprock varies linearly with the amount of injected CO 2, logarithmically with permeability, and inversely with porosity. More general sensitivity analyses were conducted to investigate the roles of individual parameters with respect to various properties, including how brine density, viscosity, and CO 2 solubility in brine, affect CO 2 flow and transport. General results suggest that both brine density and CO 2 solubility are critical factors. The combined results of this study suggest that choice of EOS algorithm is critical, because even small differences in predicted density and solubility can lead to dramatic differences in migration patterns and timing of different processes such as gravity segregation and caprock penetration. We only evaluated two EOS algorithms, but the resulting disparities were great enough that we conclude that other EOS algorithms will also produce variable results in predicted migration and related processes.