Evaluating the Influence of Surface Characteristics on Boiling and CHF Predictions with the MITB Model in Subcooled Flows

Abstract

The MIT Boiling Model (MITB) offers a detailed physics-based alternative to the empirical, correlation-based approaches of subchannel codes which are still the nuclear industry standard while limited in general applicability. In contrast, MITB is designed for integration into multiphase computational fluid dynamic (M-CFD) frameworks, enabling detailed resolution of three-dimensional effects introduced by features such as spacer grids and mixing vanes. When coupled with models for critical heat flux (CHF), MITB enhances predictive capabilities for CHF onset estimation. While its potential has been demonstrated in capturing boiling mechanisms and CHF behavior, an important challenge remains in establishing model sensitivity to surface characteristics. These characteristics play a central role the ebullition cycle, heterogeneously affecting bubble nucleation and the consequent bubble dynamics on the boiling surface. As dominant effects, surface roughness impacts nucleation site density and bubble formation, while wettability affects bubble contact angles and their nucleation and detachment dynamics.

In this study, the predictive capabilities of MITB were evaluated based on its ability to capture the effects of surface characteristics, such as roughness and wettability, on boiling behavior. The contribution of surface properties to the model's ability to predict CHF was also examined. The MITB model was first implemented in a point-averaged form to assess the sensitivity of individual closure models to varying flow conditions and surface attributes. Closure relations for bubble departure diameter (BDD), bubble departure frequency (BDF), and nucleation site density (NSD) were selected based on their physical consistency and accuracy in predicting boiling curves, as validated against experimental data from the MIT flow boiling loop at low-pressure conditions. The influence of localized flow conditions, particularly void fraction distributions, was also considered to assess the impact of the extension of the MITB formulation to a M-CFD framework.

In the work, two different CHF prediction methods were tested: a mechanistic and physically consistent surface-dependent model based on the work of Demarly and a more traditional empirical local void-fraction-based criterion. Comparative analysis reveals that both models tend to underpredict CHF, particularly for highly wettable rough surfaces. The mechanistically-based model demonstrates lower overall error, but has significant sensitivity to the treatment of the bubble growth time parameter. The void-fraction-based criterion method, though less accurate, offers improved computational robustness.