Investigation on the effects of model uncertainties on subcooled boiling

Highlights: • Subcooled boiling was predicted with considering the model uncertainties. • Predicted confidence interval agreed well with experiment data with error bands. • Sample size should be increased to reach a statistical convergence results. • Bubble diameter, bubble departure diameter and frequency dominate the prediction. • Drag & lift forces and liquid-interface HTC are also important for the prediction. - Abstract: CFD codes have been used in the prediction of boiling heat transfer and two-phase flow in the past two decades. Successful validation cases have been developed to prove the capability of two-phase CFD codes, which employed more than ten auxiliary models and dozens of coefficients. However, the applicability and the reliability two-phase CFD codes have not been acknowledged universally yet since the accuracy of these codes depend on the choice on these auxiliary models and their coefficients. The uncertainties introduced from the selections of models and coefficients have not been investigated thoroughly. In the current work, the subcooled boiling phenomena were predicted with considering the uncertainties of closure models. The 95% coverage intervals with 95% confidence level were compared with the experimental results with error bands to assess the accuracy and reliability of the code. Besides, effects of model uncertainties on the subcooled boiling were analyzed. It is found that the uncertainties from bubble departure diameter, drag force, lift force, liquid-interface heat transfer coefficient and bubble diameter play predominant rules in the modelling of subcooled boiling process.