This new contract with the U.S. Department of Energy aims to develop accessible, efficient reduced order model software to advance the goals of major modeling and simulation initiatives. The workflow envisioned will enable users to generate parameterized proper orthogonal decomposition (POD) bases from high fidelity computational fluid dynamics (CFD) simulations. The novel parameterization technique employed by this technology will address stability concerns often associated with POD while assisting users in successful production of fast running models with large eddy simulation (LES) quality.
Computational fluid dynamics methods have seen increasing use for high-fidelity modeling of complex systems, especially nuclear power systems; however, the high accuracy of computational fluid dynamics methods incurs increased computational cost. Thus, methods that provide high quality results at a reduced cost are in high demand. The goal of this work is to develop software to enable the production of high quality, reduced order models using a parameterized proper orthogonal decomposition approach. The technology will enable high fidelity, fast running models for arbitrary transients to be constructed from a set of steady state simulations. The fast running models will then be used to compute boron concentrations in the lower plenum and core region of a reactor or density variations in stratified pools of sodium cooled fast reactors.
