Effects of two-phase mixing and void drift models on subchannel void fraction predictions in vertical bundles

The evaluation of the subchannel code ASSERT against the OECD/NEA BFBT benchmark data demonstrated that at low pressures, the void fraction in the corner and side subchannels of a vertical bundle was over-predicted. Preliminary results suggest that this was due to the use of Carlucci's empirical correlation for void drift beyond its applicable range of pressure. Further examination indicates that the choice of the mixing and void drift models has a negligible effect on the error of the subchannel void fraction predictions. A single, isolated subchannel was simulated and results suggest that the root cause behind the over-prediction is inadequate mixing at the sides and corners of the bundle. Increasing the magnitude of the void drift coefficients in Carlucci's model at low pressure was found to improve the overall accuracy of the predictions. A simple correlation relating Ω to the outlet pressure was found to increase the number of points falling within experimental error by 1.0%. (author)