Evaluation of interfacial drag models for use in TRAC-M rod bundle components with validation against steady-state and transient boil-off experiments

Interfacial drag models for bubbly-slug flow were reviewed for use in rod bundle components in the new NRC consolidated thermalhydraulic code, TRAC-M. Void fraction predictions in rod bundles using standard (pipe) bubbly-slug interfacial drag models have been unsatisfactory for low mass fluxes. The EPRI Full Range Drift Flux model, the Bestion model, the Analytis modified Bestion model, and the current CATHARE rod bundle model are reviewed in comparison to the current TRAC-M (pipe) model. These interfacial drag models were evaluated analytically by using the drift-flux approach to calculate the axial void fraction distributions for various boil-off and level swell tests with a broad range of pressures and mass fluxes and comparing the results of these calculations to the experimental data of these test series to identify an appropriate candidate. The analytical evaluation of the models was based on comparisons of predictive quality over three ranges: the two-phase region, the region for which the bubbly-slug interfacial drag is the sole contributor to the net interfacial drag in TRAC-M, and the region for which the bubbly-slug interfacial drag contributes to the net interfacial drag in TRAC-M. The Bestion interfacial drag model was selected and implemented in a test version of TRAC-M. Simulations were performed of the THTF quasi-steady-state level swell tests and the G2/Westinghouse transient boil-off tests, tests with low mass fluxes and varying pressures, the range in which predictions with previous models were observed to be of poor quality. A high mass flux FRIGG test, for which previous void fraction predictions were acceptable, was also simulated to cover a comprehensive range of mass fluxes. The TRAC-M predictions using the updated interfacial drag model were validated against experimental data for all three test series and compared to both TRAC-M predictions using the current interfacial drag model and TRAC-B predictions, focusing mainly on the axial void fraction profiles. Updating the interfacial drag model led to remarkable improvements in the TRAC-M void fraction predictions. Based on results of this evaluation, a recommendation is made for updating the bubbly-slug interfacial drag model for rod bundle components in future versions of the TRAC-M computer code.