Geometrical optimization of PWR spacer grids using GeN-Foam and Genetic Algorithms

Abstract

This paper presents the results of Computational Fluid Dynamics (CFD) oriented geometrical optimization using the GeN-Foam solver applied to subchannels of the fuel assembly in a PWR-type nuclear reactor. GeN-Foam is a coarse mesh OpenFOAM solver designed to study nuclear engineering problems involving the coupled solution of thermohydraulics, neutronics and thermomechanics. However, the solver could be used for complex geometry simulations, enabling multi-scale coupled simulations. To use GeN-Foam under these conditions, the results of the code for complex geometry simulation had to be evaluated. This assessment involved comparing  the results obtained with the solver and those presented in a literature reference study. Despite the higher numerical diffusivity of the solver, this comparison demonstrated that GeN-Foam is capable of studying the fluid dynamics of fuel assemblies in nuclear reactors for both coarse and refined geometry conditions. After GeN-Foam was assessed, optimization was performed on subchannels of a fuel assembly using Genetic Algorithms (GA), evaluating the influence of geometric parameters of the spacer grids to minimize pressure drop and maximize secondary flow. Pareto Front solutions were assessed to identify a geometry that best balanced these two objectives. The optimized model showed better results than the reference study, as expected.  However, the results also highlight the need to incorporate thermal physics and neutronics to ensure that the optimized solution meets the subchannel´s flow and heat exchange requirements. All tools used in this work are well-established in the literature, free, and open-source.