Numerical simulations of a turbulent flow in a fuel assembly

In the present paper turbulent flow in a geometry of 5 x 5 fuel assembly is investigated using open-source Computational Fluid Dynamics (CFD) software OpenFOAM. Simulated fuel assembly does not contain any flow deflector, hence it represents a region where flow is attached to the walls. There is also no heat transfer since turbulence alone is put under investigation. Simulations are using two different approaches: Unsteady Reynolds Averaged Navier-Stokes simulations (URANS) and Large Eddy Simulations (LES). The capability of various turbulent models are estimated: two low-Reynolds eddy viscosity models (isotropic k-omega SST and anisotropic v2f), a high-Reynolds Launder-Reece-Rodi Reynold's Stress Transport Model (LRR RSTM) and Smagorinsky LES model. Each model was performed on two meshes to investigate effects of different mesh elements (e.g. hexa vs tetra and prisms) on the results. Although all results are assessed against measurements from MATiS-H experiment (performed at KAERI in 2011), a special focus is put to a near-wall velocity profile, where measurements were not provided. In that region LES simulation could provide an additional data for assessing URANS results. Generally, the simulation results are in a reasonable agreement with measurements. Velocities are slightly overestimated whereas fluctuations are underestimated for up to 30%. URANS predicted quite similar velocity profiles, however fluctuation predictions differ more from each other. Discrepancies among the results obtained on different meshes (hexa vs tetra+ prisms) with the same model are similar in magnitude as results obtained with different turbulent models.