Prediction of Experimental Data on Heat Transfer to Supercritical Water with Two-Equation Turbulence Models

The present work provides information on the capabilities of presently available turbulence models in predicting supercritical water flow and heat transfer. The performance of eight low-Reynolds number two-equation models, including k-ε, k-ω and k-τ formulations, has been assessed against recently published experimental data on upward and downward flow in vertical circular pipes. The size of the pipes and the flow conditions are representative of sub-channels of tight short bundles being proposed for some reactor concepts. Both forced and mixed convection are involved in the experiments considered, providing a challenging workbench for the application of turbulence models. A CFD code has been purposely developed for studying heat transfer and flow distribution in different channel geometries adopting several approaches for dealing with turbulence. The results of the computations show strong differences in the capability of the adopted models to reproduce the observed heat transfer enhancement and degradation phenomena. In particular, remarkable differences in model performance are found in the cases of upward and downward flows, owing to the different mechanisms by which buoyancy affects turbulence in the two conditions. A detailed discussion on the effectiveness of the turbulence models tested is given in the paper, in terms of the models’ ability to reproduce the physical phenomena involved in the experimental data considered.