Turbulence Modeling Issues in Ads Thermal and Hydraulic Analyses

Accelerator Driven nuclear reactor Systems (ADS) have in several respects a prototypical character of the flow and cooling conditions combined with narrow operating conditions due to the materials engaged. E.g. the high local thermal load in the liquid metal cooled spallation target requires a very careful analysis by experimental and numerical means. Some of the main goals of the numerical analyses of the thermal dynamics of those systems and of required experiments are discussed. The prediction of locally detached and recirculating flows suffers from insufficient turbulence modeling; this has to be compensated by using prototypical model experiments, e.g. with water, to select the adequate models and numerical schemes. Some sensitivities and model uncertainties are discussed; some of them are reduced by so-called layered models like in the SST turbulence model or the DES. The well known problems with the Reynolds analogy in predicting the heat transfer in liquid metals requires prototypic liquid metal experiments to select and adapt the turbulent heat flux models. The uncertainties in liquid metal experiments cannot be neglected; so it is necessary to perform CFD calculations and experiments always hand in hand and to develop improved turbulent heat flux models. One contribution to an improved 3 or 4-equation model is deduced from recent Direct Numerical Simulation data. Of course, the ADS community would need such extended heat flux models, but even realizing standard 3or 4-equation ASM heat flux models in the commercial CFD codes would allow for an improved heat transfer modeling, especially when buoyancy is involved.