Nuclear energy (NE) power promises to become a reliable, carbon-free resource capable of meeting the energy needs of our nation and the world. Numerical simulation has been an intrinsic part of nuclear engineering in research, design, and licensing of existing and proposed conventional nuclear power plants. Nuclear modeling and simulation tools available today, however, are mostly low dimensional, and empirically based, valid for conditions close to the original experiments, and in many cases represent only incremental improvements on decades-old legacy codes. The development, deployment, verification, and validation of higher-fidelity computational capabilities for analyzing, modeling, simulating, and predicting complex thermo-fluid phenomena will advance nuclear power capabilities by resolving technical, cost, safety, and licensing issues. In particular, Argonne National Laboratory’s close collaboration with Framatome and the US Nuclear Regulatory Commission (NRC) in the area of high performance computing (HPC) and computational fluid dynamics (CFD) provides an opportunity for further advancement of modeling and simulations for a scalable carbon-free energy quest.
In close collaboration with Framatome and the NRC, this project aims to validate open-source Nek5000 code for two sets of experiments: the ALAIN loop flow test and the PANDA facility Geometries. Framatome designs, manufactures, and installs components, fuel, and instrumentation and control systems for nuclear plants. The second partner, the US NRC, has a final say on licensing of new nuclear power plants and extending the operations of old ones in the US. Recent collaboration with the NRC Office of Nuclear Regulatory Research on Nek5000 CFD application to a range of problems shows great potential for support of some of the NRC’s confirmatory calculations, exploratory analysis of generic safety issues, and the extension of experimental data for developing nuclear system analysis code flow correlations, especially beneficial in specific regimes and geometries where the experimental data are hard to come by or are of insufficient resolution and accuracy. In particular, further Nek5000 validation work for PANDA facility geometries including the recent modifications of HYMERES-2 tests can be used to inform lower-fidelity fast-turn-around models from Reynolds-averaged Navier-Stokes (RANS) to reduced-order modeling approaches.