Assessment of turbulence models for heavy liquid metals in computational fluid dynamics

Immagine di anteprima
Maciocco, Luca
Titolo del periodico
Titolo del volume
A summary and analysis of the results obtained from work-packages 1-6 of the ASCHLIM project is presented in this document, with the aim of defining the performances and shortcomings of CFD turbulence models currently adopted for the simulation of Heavy Liquid Metals flows in nuclear applications. Two classes of problems are analysed, one related to liquid metals physical characteristics (low Prandtl number) and one related to the flow morphology in typical ADS applications. Concerning the first class, some drawbacks were found in the use of wall-functions for HLM flows. In fact, thermal wall-functions currently implemented in commercial CFD codes are in general unsuitable for HLM flows, unless the first grid point lays in the thermal sublayer y+ < 70 ÷ 100. It was also proved that the Reynolds analogy is not applicable for very low Peclet number flows (~100). However, correct results were obtained for higher Pe (~1000), even using a constant value for the turbulent Prandtl number (0.9). Due to the lack of experimental measurements of turbulence quantities, it was not possible to draw strong conclusions concerning the second class of problems. However, results confirmed the capability of two-equation models to give a reasonably good prediction of the main flow characteristics in complex flow morphology typical of spallation-targets applications. Higher order models, both for momentum and heat turbulence transport, should be used in cases where turbulence anisotropy is important.
heavy liquid metals , CFD , computational fluid dynamics , HLM , accelerator driven system , ADS