Evaluation of different interface-capturing methods for cryogenic two-phase flows under microgravity

The distribution of the gas-liquid interface is crucial to the accurate calculation of the flow and heat transfer of in-orbit cryogenic propellants, for which the surface tension force overtakes the gravitational force. As an essential oxidant, liquid oxygen has a lower surface tension coefficient and viscosity than most room-temperature fluids, causing a greater possibility of interface instability and breakage. Conventional numerical methods have seldom been assessed in terms of cryogenic two-phase flows under microgravity, and commercial software cannot provide a consistent platform for the assessment. In this study, a unified code based on OpenFOAM has been developed for evaluating four interface-capturing methods for two-phase flows, namely, the algebraic volume of fluid (VoF), geometric VoF, coupled level set and VoF (CLSVoF), and density-scaled CLSVoF with a balanced force (CLSVoF-DSB) methods. The results indicate that the CLSVoF-DSB method is most accurate in predicting the interface motion, because it uses the level set function to represent the gas and liquid phases. The gas-liquid interface predicted by the CLSVoF-DSB method is the most stable because it adopts the scaling Heaviside function to weaken the effects of spurious currents and increases the stability. The numerical algorithm of the algebraic VoF method is the most simple, so it has the highest efficiency. The geometric VoF uses the isoface to locate the gas-liquid interface in a grid cell, so it can obtain the thinnest interface. In applications of liquid oxygen, the CLSVoF-DSB method should be used if the overall accuracy is required. Published under an exclusive license by AIP Publishing.

Automated summary

The distribution of the gas-liquid interface is crucial to accurate calculation of cryogenic propellant flow and heat transfer. A unified code based on OpenFOAM was developed to evaluate four interface-capturing methods, and the CLSVoF-DSB method was found to be the most accurate in predicting interface motion.