Computational fluid dynamics (CFD) codes have become promising tools for the investigation of thermal hydraulics in revolutionary reactor concepts in the last decade. In Reynolds-averaged Navier-Stokes calculations, the CFD codes (for example, the ANSYS CFX code used here) use turbulence modeling, wall functions, and other approaches. Therefore, the accuracy of CFD codes for water flow under supercritical conditions has to be examined. The first aim of this work is to investigate the effects of different material property definition methods on the numerical results obtained with CFX code. The second aim is to assess the accuracy of the conventional turbulence models (such as k-
, k-
, and SST) under supercritical water conditions. The results and comparison of three independent validations for supercritical water flow in vertical smooth-bore tubes with upward flow direction are presented in this paper. It is well known that the material properties strongly depend on the temperature and the pressure near and above the thermodynamic critical point. It is demonstrated that rather than analytical or discrete point methods, the IAPWS-IF97 material table best represents the strongly changing material properties. A nonaxialsymmetric effect on result fields was not found based on the three validations; therefore, a rotational periodic or two-dimensional grid approach is recommended for further validations of homogenously heated, vertically installed, smooth-bore straight tubes cooled by supercritical water. The calculation results have been compared with measurements, and the computational errors for the three validations were found to be in the ranges of 0 to 25%, 0 to 18%, and 2 to 40% for the Swenson, Yamagata, and Herkenrath experiments, respectively. The results of the three validations indicate the need to improve a turbulence model to take into account the buoyancy effect on the turbulence for thermal-hydraulic calculations of the supercritical water.
