A comparative analysis of the flow in a rod bundle with two guide-tubes and mixing grids, URANS simulation and Wall-Modeled LES of the ALAIN experiment

Abstract

The flow characteristics in a 5 × 5 rod bundle representative of ALAIN experiment is investigated using Unsteady Reynolds-Averaged Navier–Stokes and Wall-Modeled Large Eddy Simulation methodologies. A review of the numerical methodology is first presented to highlight the impact of mesh resolution and computational domain size on simulation outcomes. Experimental measurements of the velocity along two traverses of the rod-bundle lattice are then employed as reference data to evaluate the accuracy of computational fluid dynamics simulations. The average error on the velocity statistics reveals substantial discrepancies between the numerical results and experimental measurements. The disturbance induced by the mixing grids and the presence of guide-tubes represent a challenging feature to reproduce numerically, both at close and long distance from the grids. Computation on the error made on the velocity statistics reveals that Wall-Modeled Large Eddy Simulation perform better than the $k-ϵ$ URANS model to predict the axial velocity at close distance from the grid and the transverse flow at large distance from the grid, but comes at a much higher computational cost. The present analysis enhances the understanding of flow dynamics in rod bundles and informs the selection of appropriate turbulence models for future simulations.