Investigation of the near-wall modelling strategy and the concept of the turbulent Prandtl number for low Prandtl number flows

The development of GEN IV nuclear reactors represents a tremendous challenge from an engineering point of view. In particular, the reliability of engineering tools to reproduce thermal hydraulic behaviour of heated liquid metal flows is doubtful. This thesis investigates if the standard models in Reynolds averaged CFD codes, such as the application of the turbulent Prandtl number concept and the Reynolds analogy, can accurately predict turbulent heat transfer for low Prandtl number flows. This assessment is based on the comparison of results obtained for a heated hannel flow using a near-wall resolved approach, with DNS and LES data from literature. The analysis results confirmed the poor performance of standard models to simulate turbulent heat transfer in liquid metal flows. Moreover, it was demonstrated that these weaknesses are relevant for design activities as they can lead to underestimations of crucial parameters such as the fuel cladding temperature during operation. However, this thesis proposes strategies to circumvent this problem using either a near-wall resolved or wall-function approach. These strategies were successfully tested for the sub-channel of a hexagonal fuel assembly, with a good agreement between calculated and experimentally measured Nusselt data.