Optimized Phase Stability Algorithm in Three-Phase Systems with an Application to CO₂ Capture and Sequestration in Subsurface Reservoirs

In the efforts to optimize CO₂ capture and sequestration, the accurate prediction of phase behavior in subsurface reservoirs is critical. Our work introduces an efficient algorithm for phase stability testing coupled with multistage equilibrium multiphase negative flash calculations. Traditional phase flash calculations often encounter numerical issues when the number of moles in the vapor phase (nV) is outside the physically feasible range of 0 to 1. Our approach overcomes these limitations by employing a negative flash that allows for the mole fraction of components in each phase (xiV and xiL) to be positive and greater than zero, even if nV does not lie within the expected range. The study explores the stability of a three-phase system (V-L-W), utilizing the criterion of the negative flash which dictates that if number of moles in the vapor phase, liquid and water phase (nV, nL, and nW) fall within the range of 0 to 1, the system is stable. The presented stability testing algorithm starts the search with a three phase negative flash calculation, utilizing a set of diverse initial equilibrium K-values to detect multiple phases. The stability analysis of two-phase systems, including Vapor-Liquid (V-L), Vapor-Water (V-W), and Liquid-Water (L-W), follows the same criterion. Upon encountering instability, the algorithm shifts to the most probable stable two-phase system. This is achieved by removing the water phase and normalizing the mole fractions for the remaining components, followed by running two-phase flash calculations. A systematic approach is used for stability testing, involving both negative flash and tangent plane distance calculations for two- and three-phase systems respectively. The method determines the stable phases predicted based on the tangent plane method when the negative flash does not converge or align with the tangent plane solutions, suggesting that the algorithm is versatile and adaptable to varying conditions of multiphase systems. The algorithm's performance is validated through 0-D simulations of reservoir mixtures to showcase its ability to converge without any erroneous results, and is applied to a CO₂ storage scenario to showcase its physical accuracy predicting different trapping mechanisms of CO₂ in the subsurface.