作者: Joachim Moortgat , Zhidong Li , Abbas Firoozabadi
DOI: 10.1029/2011WR011736
关键词: Three-phase 、 Thermodynamics 、 Equation of state 、 Viscous fingering 、 Rate of convergence 、 Work (thermodynamics) 、 Physics 、 Finite difference method 、 Finite element method 、 Finite difference
摘要: [1] Most simulators for subsurface flow of water, gas, and oil phases use empirical correlations, such as Henry's law, the CO2 composition in aqueous phase, equations state (EOS) that do not represent polar interactions between CO2and water. Widely used are also based on lowest-order finite difference methods suffer from numerical dispersion grid sensitivity. They may capture viscous gravitational fingering can negatively affect hydrocarbon (HC) recovery, or aid carbon sequestration aquifers. We present a three-phase compositional model higher-order element incorporate rigorous efficient three-phase-split computations either three HC water-oil-gas systems. For phases, we Peng-Robinson EOS. allow solubility CO2in water adopt new cubic-plus-association (CPA) EOS, which accounts cross association H2O molecules, molecules. The CPA-EOS is highly accurate over broad range pressures temperatures. main novelty this work formulation reservoir simulator with EOS-based unique computations, satisfy both equalities fugacities all global minimum Gibbs free energy. provide five examples demonstrate twice convergence rate our method compared approach, compare experimental data other simulators. consider during CO2sequestration aquifers, water-alternating-gas injection, full modeling phases.