Authors: M. Seneca
Year: 2025
Industry: Nuclear & SMR
Keywords: Mention FSE, Pebble Bed, Validation, Comparison, Verification, Sensitivity
ABSTRACT
This study presents a comprehensive verification and validation of the high-fidelity NekRS Large Eddy Simulation (LES) framework for modeling conjugate heat transfer in low-flow pebble bed reactor scenarios, with a focus on both conduction, convection, and radiative effects. The validation campaign is anchored by the SANA experiments, which represent a stringent benchmark for reactor-relevant conditions. The verification process began with a tall-cavity benchmark coupling the MOOSE Net Radiation Method with natural convection. Simulation results using MOOSE were compared to one-dimensional analytical solutions across varying surface emissivities, yielding Nusselt numbers within 1% of the analytical predictions. Subsequently, a single-pebble case was modeled using both MOOSE (via the surface-to-surface view-factor method in Cardinal) and NekRS (via the P1 approximation), revealing temperature and velocity distributions that agreed within 5%, despite differing radiation methodologies. The validation phase progressed to multi-pebble geometries. NekRS simulations of a 283-pebble bed experiment at Texas A&M achieved strong agreement with measured velocity profiles, confirming the code’s accuracy in capturing mixed-convection flows under low Reynolds number conditions. Additional validation was performed using the SANA 5 kW experiments for both helium–graphite and helium–alumina pebble beds. These simulations incorporated full conjugate heat transfer, including internal conduction through pebbles, natural convection of the interstitial gas, and radiative heat exchange. The best-performing simulation—achieved with an optimized chamfer thermal conductivity—yielded a mean temperature error of 1.56% in the graphite case, significantly improving upon the 4.3% error reported for Pronghorn [31]. Sensitivity studies revealed that chamfer conductivity plays a critical role in shaping the thermal field and flow structure, with conductivity values closer to that of helium promoting natural convection dominance, while higher values suppressed convective effects via increased conduction. Beyond validation, this study generated ten high-fidelity datasets for use in correlation development, encompassing five mixed convection simulations for a 127-pebble bed (with and without a reflector, under constant and variable properties) and five natural convection simulations for a 1568-pebble bed filled with FLiBe, covering Rayleigh numbers from 50,000 to 5,000,000. These simulations provided detailed Nusselt number data in both core and near-wall regions, contributing to the development of new heat transfer correlations for low-flow pebble bed reactor designs. Future work will include the iii implementation of advanced contact resistance models for improved chamfer conductivity representation, an extension of the SANA validation to higher-power cases with stronger thermal gradients, and the development of new heat transfer correlations for small pebble beds