NUREG/IA-0185 "Development and Validation of a Transition Boiling Model for the RELAP5/MOD3 Reflood Simulation"

The heat transfer model of the RELAP5/MOD3.1 (R5M3) code was extensively reviewed and assessed. The most important deficiency of the current version of the code was attributed to its treatment of the transition boiling heat transfer regime. The current transition boiling model significantly underpredicts the heat transfer rate to the liquid phase. Since at low quality conditions the liquid boiling part is a major fraction of the total heat transfer, the current model underpredicts the quench temperature and the quenching rate under most conditions relevant to LOCA and degraded core analysis. Therefore, a new model has been developed and implemented in the R5M3 code for predicting the transition boiling'heat transfer. The new model is based on an extension of the phenomenological formulation suggested originally by Chen. It uti lizes only local state variables calculated by the R5M3 code and does not require other history parameters, such as quench position or CHF and minimum film boiling temperatures, which are not available at actual time step. A number of separate effect and bundle tests are analyzed with the modified code version. The predictions are compared with those obtained by the frozen code version and with available experimental data. Several variables, such as wall temperatures, vapor and liquid velocities, void fraction etc., are examined in order to evaluate the general prediction capability of the code in modeling boil-off and reflood transients. In addition, the current and the modified stand-alone transition boiling models are tested against a large sample of the available data-base on steady-state post dryout heat transfer. In all cases, the predictions of the modified model fit the measured data better. The temperature curves are physically and conceptually more sound than those predicted by the frozen code version. This is achieved by introducing a more realistic modeling of the transition boiling heat transfer which affects only one subroutine of the R5M3 code. Preliminary results of the FZK transition boiling model validation against both the full-length FLECHT-SEASET bundle test 31701 and the integral test LOFT-LP-LB-1 have shown a better prediction of the quench process than the frozen R5M3 version.