• CFD codes
• supercritical fluids
• experimental data
• nuclear

This work is focused on the study of phenomena concerning the heat transfer between fluids with pressure higher than the critical value and an heated vertical pipe with small internal diameter, and on the ability not of turbulence models implemented in the commercial CFD codes to evaluate these phenomena.
The fluid with pressure above the critical value are characterized by the strong variation of fluid properties across a particular temperature called ‘pseudocritical’. The strong variation especially of specific heat (which shows a significant peak) and of density(which passes from liquid-like to gas like) across the pseudocritical temperature, impact heavily on the heat, which is indeed strongly altered by these factors. This changes are strongest near the critical pressure, and decrease with the increasing of the same.
The study was conducted using STARCCM+ v. 5.004 and in few cases FLUENT v.13, and are analyzed many experimental data regarding water and carbon dioxide flowing in vertical pipes with different diameter and length and with different operative fluids conditions. The comparison between experimental and numerical values provided by the codes have been conducted on the basis of the estimation of fluid temperature which is in contact with the inside of the solid wall.
Are also made analysis on the Nusselt number, the global heat transfer coefficient between the fluid and heated length and other parameters of interests as the coefficient of buoyancy, chosen from several formulations available in the literature (Jackson 2011). another important tool of investigation are the profiles of axial velocity and turbulent kinetic energy, because show the presence of buoyancy and laminarization effects.
The investigation regard also the ability of turbulence models to provide accurate values as possible in relation to the experimental data; the case studies are characterized by anisotropic turbulence because of the presence of the solid wall, but not all the turbulence model (adopting RANS system) in low-Reynolds mode are able to provide god results, infact the range of correct application of each model is quite limited respect to all the different experimental conditions that can be found in the paper present literature.
The turbulence models considered are the Lien k- and SST k- low-Reynolds which are that are implemented in STARCCM+, and are tested the turbulence models implemented in the last FLUENT version (v.13) with more than two constitutive equations.