Thermal-hydraulic analysis of a representative Westinghouse lead-cooled fast reactor fuel bundle using CFD

Lead-cooled fast reactors (LFRs) are Generation IV reactor technologies that use molten lead as the primary coolant. Whilst lead offers advantages for economics, safety, and sustainability, its low Prandlt number and challenging experimental characteristics pose difficulties for thermal-hydraulic modelling and validation. To support LFR development, this study aims to advance modelling capabilities and understanding of the relevant physical phenomena through a series of Computational Fluid Dynamics (CFD) simulations of a Fuel Pin Bundle Simulator (FPBS) that shares design features with the Westinghouse LFR fuel assembly. Three geometrical configurations of the FPBS have been modelled, using the Reynolds-averaged Navier-Stokes (RANS) approach: a bare pin bundle (without spacer grids), a T-junction upstream of the main test section, and the full-length 360 degrees main FPBS test section including spacer grids and instrumentation wires. The sensitivity of the results to modelling choices, including turbulence models and approaches for the turbulent Prandtl number, is explored. The original contributions of this study are in the assessment of different RANS models of the Reynolds stresses, the assessment of different values and functions of the turbulent Prandtl number for the modelling of the turbulent heat fluxes, the exploration of the entry conditions on the flow and thermal development along the fuel bundle and the determinations of the effects of the intrusive instrumentation on the measured quantities. The bare bundle simulations showed only minor sensitivity to the turbulence model and produced friction factors in excellent agreement with existing correlations. Predictions in the upstream T-junction indicated the generation of significant swirl that enters the main test section, but the spacer grid acts as an effective flow straightener. Nusselt number predictions in the main FPBS test section showed good agreement with established correlations for liquid metal rod bundles. Instrumentation wires had only a minor effect on the temperature field and increased the pressure drop by 2.7 %. A sensitivity analysis of the turbulent Prandtl number (Prt) showed that Kay's correlation produced Nusselt numbers that were closest to the empirical correlation of Ushakov et al. (1977), with a mean deviation of 1.1 %. In contrast, a constant Prt= 0.9 resulted in an overprediction of 19 %.