Assessment of semi-mechanistic bubble departure diameter modelling for the CFD simulation of boiling flows

Eulerian-Eulerian two-fluid computational fluid dynamic (CFD) models are increasingly applied to predict multiphase and boiling flows in nuclear reactor thermal hydraulics. In these models, nucleate boiling is usually accounted for by partitioning the heat flux between the different mechanisms of heat transfer involved. Although structured in a mechanistic fashion, heat flux partitioning models are still forced to rely on mainly empirical closure relations. Between the numerous closures required, the bubble departure diameter in particular has a significant influence on the predicted interfacial area concentration and void distribution within the flow. There is now abundant evidence in the literature of the limited accuracy and reliability of the empirically-based correlations that are normally applied in CFD models. In view of this, in this work more mechanistic formulations of bubble departure have been introduced into the STAR-CCM+ code. The models are based on a balance of the hydrodynamic forces that act on a bubble at the nucleation site. Their performance, and compatibility with existing implementations in a CFD framework, are assessed against two different data sets for vertically upward subcooled boiling flows. In general, a significant number of modelling choices is required by these mechanistic models and some recommendations are made. The models are extended to include a more physically-consistent coupled calculation of the frequency of bubble departure. In general, predictions of the wall temperature reach a satisfactory accuracy, even if numerous numerical and modelling uncertainties are still present. In view of this, several areas for future work and modelling improvement are identified, such as the proper modelling of the local subcooling acting on the bubble cap.