Efficient nested three-phase isenthalpic flash calculations with a hybrid Newton and Brent algorithm for water-CO2-hydrocarbon mixtures in CO2 storage simulations

Abstract

Carbon dioxide storage into depleted reservoirs/saline aquifers requires robust numerical compositional simulation tools that reveal the correct physics of the CO₂-inclusive fluid in the injection and subsurface flows. When injecting CO₂ for storage in the pure content, it is highly possible to encounter the boiling/narrow-boiling behavior, in which pure CO₂ may exist in two phases at equilibrium. To capture such behaviors correctly, a compositional simulator with pressure-enthalpy specifications must be used. Moncorgé et al., (2022) established such a simulator embedding a nested isenthalpic (PH) flash calculation algorithm applying Broyden’s method. In this work, to improve the computational efficiency and robustness of the nested PH flash calculations, we developed a new hybrid Newton and Brent algorithm. In the new algorithm, Brent’s method is implemented to maintain the robustness, while Newton method is switched to reach fast convergence if the monotonic convergence criterion is satisfied. Moreover, an extrapolation strategy based on the sensitivity vector of phase compositions (in mole numbers) in terms of temperature, together with an accelerated isobaric-isothermal (PT) equilibrium calculation code, is proposed to further improve the computational efficiency of the PH calculations. Several case studies which encompass a variety of mixtures, including mixtures that exhibit narrow-boiling behavior. Our results demonstrate that the hybrid algorithm is more efficient and robust than the previous nested PH algorithms. The hybrid algorithm is a strong candidate to be included in PH phase equilibrium packages and compositional simulators for CO₂ storage.