Improving the numerical accuracy of sharp interface treatment of evaporation modeling

Earlier attempts at computationally solving evaporation problems relied on using single-fluid formulation. In this formulation, the hydrodynamic equations are solved by treating the liquid and gaseous phases as a single fluid with spatially varying properties. Single-fluid formulation provides several numerical advantages: lower memory usage, faster turnaround times and the ability to easily extend an existing multiphase code to handle evaporation problems. Recent studies, however, have shown that single-fluid formulations smear the jump across the interface in quantities like fluid properties, temperature and velocity. This smearing introduces significant errors if the problem of interest is driven by interfacial gradients, e.g., in Stefan problems. Therefore, newer studies on evaporation are preferring a sharp treatment at the interface, for which several two-fluid formulations have been proposed. In the present work, we identify some features that are still lacking in the existing two-fluid formulations. We also show how these features can result in improved accuracy and how they can be introduced in an existing sharp interface code.