Computational fluid dynamics prediction of subcooled boiling of water using a mechanistic bubble-departure model

This paper presents computational fluid dynamics (CFD) simulations of subcooled convective boiling flows. The simulation methodology is based on a Eulerian two-fluid approach. Heat transfer due to wall-boiling phenomena is captured with a well-established wall heat flux partitioning model. Bubble departure sizes are supplied to the wall-boiling model by calculating the lift-off diameter using a single formula derived from a mechanistic thermodynamic equilibrium model. In contrast to empirical correlations that typically are used in CFD modelling of boiling flows, the current bubble-departure model represents mechanistically the force balance on a bubble as it grows on a surface, up to its lift-off towards the bulk flow. A simplified approach is employed to estimate the relative velocity of a sliding bubble involved in the calculation of the lift off condition, based on a force balance in the flow direction. The model is extensively tested via comparison with an experimental database from the literature, in terms of vapour distribution along the flow, covering a wide range of conditions in terms of pressure from about 3 MPa to 15 MPa, mass flowrate (400-2100 kg/m2s) and heat flux (420-2200 kW/m2). It is demonstrated that predictions of the current method are in satisfactory agreement for low to moderate pressures up to around 7 MPa, without need for any model adjustment or use of fitting parameters. As the pressure increases, the model succeeds in tracking the basic trends, however, deviations of quantitative predictions of the vapour fraction distribution from the experimental values increase. The trend suggests that other closure terms, such as the nucleation site density, may also need to be modified, possibly using a mechanistic approach, to increase further the confidence in predictions of the current methodology in high-pressure conditions.