Direct Numerical Simulation of Mixing Phenomena in the Upper Plenum of Advanced Reactors

Abstract The conceptualization and development of advanced nuclear reactors encompasses challenging fluid-flow concerns that significantly impact their operational safety and efficacy. The establishment of a comprehensive numerical database focused on high-fidelity data holds promising potential in facilitating the formulation of accurate and cost-effective reduced-resolution heat transfer models. These models can be designed based on a multiscale hierarchy developed as part of the recent U.S. Department of Energy-funded Center of Excellence for Thermal Fluids Applications in Nuclear Energy, which represents a significant stride toward resolving industry-specific challenges associated with the heat transfer behavior of advanced reactors. This paper considers direct numerical simulation (DNS) of the upper plenum with discharging jets of high-temperature gas-cooled reactors (HTGR). Three isothermal cases have been considered: one for the TAMU facility at Re = 10,622 and two for the MiGaDome facility at Re = 4097. The MiGaDome cases include single-jet and double-jet configurations, both operating at the same flowrate. The low Prandtl number fluid (Helium) is considered for all simulations. First- and second-order statistics are investigated, and the agreement with experimental data improvements has been observed compared to previous large eddy simulation (LES) studies. Furthermore, proper orthogonal decomposition (POD) analysis is employed to reveal and characterize the flow patterns in the upper plenum across various configurations and operating conditions. The energy transfer pathway was also discovered in the MiGaDome double-jet configuration. The generated high-fidelity DNS data will be utilized alongside data-driven methods to improve turbulence modeling closures.