Effect of bubble interactions on the prediction of interfacial area in TRACE

The conventional thermal-hydraulic nuclear reactor system analysis codes utilize a two-field, two-fluid formulation for two-phase flows. To provide closure, static flow regime transition criteria and algebraic relations are employed to estimate the interfacial area concentration (ai). To better reflect the dynamic evolution of two-phase flow, an experimental version of TRACE is developed to estimate the ai using the one-group interfacial area transport equation (IATE), which includes mechanistic models for bubble coalescence and disintegration. These models account for: (1) bubble break-up due to impact of a turbulent eddy, (2) bubble coalescence due to random collision driven by turbulent eddies, and (3) bubble coalescence due to the acceleration of a bubble in the wake region of a preceding bubble. To assess the impact of including bubble interaction models in TRACE, code predictions of experimental data measured by a multi-sensor conductivity probe are compared to both the IATE and flow regime based predictions. In total, 50 air-water vertical co-current upward and downward bubbly flow conditions in pipes with diameters ranging from 2.54 to 20.32 cm are evaluated. It is found that TRACE, using the conventional flow regime relation, always underestimates ai by predicting a larger bubble size than observed in the experimental data. Additionally, the axial trend of the ai prediction is always linear because ai in the conventional code is predominantly determined by the pressure. However, TRACE with the one-group IATE significantly improves the prediction results, yielding a ±13.0% difference with the data. It is found that the non-linear developments observed in the experimental data, which reflect bubble interactions, are predicted well using the IATE. Moreover, in several conditions dominated by bubble interactions, the ai trend displayed is opposite to the effect of pressure. In these cases, the conventional TRACE relation predicts an incorrect trend in ai, while the IATE predicts the experimental data well.