LAUSR

Abstract Rising computational capabilities have allowed for the possibility of more complex thermal hydraulic safety codes. One of these branches of advancement is to start using three dimensional solvers in system codes instead of one dimensional. CATHARE-2 already included a 3D module, but within the EDF and CEA NEPTUNE project started in 2001 the 3D solver, as a part of CATHARE-3, has been modified to be more versatile and at the same time faster. Moving from one dimensional solvers to three dimensional poses issues due to the added degrees of freedom. In a 3D solver terms describing turbulence appear in the equations due to the time and space averaging inherent to the field equations which need to be taken into account. Within the NURESAFE program turbulence models have been added to the 3D module of CATHARE-3. In this paper the sensitivity of simulation results to the different turbulence models and level of detail in meshing have been studied. For this purpose simulations of boil-up and reflooding experiments performed at the PERICLES 2D facility were carried out. The results show that the turbulence models do cause change in the results. With sub channel level meshing they are visible, while with channel level meshing turbulence modelling has minimal effect.