抄録
The paper covers the project results obtained at ANSYS Germany in a 3-year R&D consortium “CFD Methods for the Prediction of Critical Heat Flux” funded by the German Ministry of Industry, where different research centers have been involved as well (Technical University of Munich, Karlsruhe Institute of Technology, Helmholtz Zentrum Dresden Rossendorf). The main goal was the improvement of predictive capabilities of ANSYS-CFD for the coverage of flow regimes from nucleate subcooled boiling up to flow conditions close to critical heat flux (CHF). The paper will provide an insight in the chosen Eulerian multiphase flow modelling concepts and their application to two test facilities and experiments with comparison to validation data. The chosen geometries and working conditions should help to simulate the real conditions in a Pressurized Water Reactor. Special interest has been paid to the prediction of wall temperature extrusions indicating the occurrence of CHF. The CFD simulation methodology involves the combination of conjugate heat transfer (CHT), the extended Rensselaer Polytechnic Institute (RPI) wall boiling model (for large vapor volume fraction) together with population balance and flow regime transition modelling, commonly known as the inhomogeneous MUSIG model. The first application case is the test facility operated by the Technical University of Munich for convective wall boiling of a Novec-649 refrigerant heated by a massive copper heater under atmospheric pressure conditions. As a second validation experiment the derived ANSYSCFD modeling has been applied to available experimental data from the COSMOS-L test facility (vertical annulus) at Karlsruhe Institute for Technology, for the investigation of CHF for a water-water vapor system under moderate pressure of up to 2-3 bar. For this case, we were able to compare the occurrence of CHF in the experiments with the ones obtained with the simulations. Both experiments complement each other and cover a wide range of pressure levels as well as working fluids. These parameters are the reason for calibration needs of the RPI submodels. This work provides an insight into the established state-ofthe-art of wall boiling predictions by means of CFD, its current predictive capabilities as well as an outlook for future development.
