Reduced Order Modeling for Coupled Neutronics/Thermal Fluids Analysis of Molten Salt Reactors in CRAB to Enable Uncertainty Analysis

We propose to implement and test reduced order methods for neutronics and thermal-fluids analysis, implemented into the MOOSE framework for use in molten salt fueled reactors, with an emphasis on MCFR. This capability will enable uncertainty quantification for transient safety analyses. A fission-matrix based methodology will be used for neutronics, and the thermalfluids will use Proper Orthogonal Decomposition (POD) and Galerkin projection relying on recent developments, in combination with a simplified turbulence model. The fission-matrix interpolation method is relatively novel, but has seen success modeling a diverse set of systems (including MCFRs) with temperature and control feedback. The fission matrix method will be implemented as a MultiApp plugin for Griffin, while the thermal-fluids reduced method will be incorporated directly into Pronghorn/SAM. These methods will complement other analysis capabilities at the NRC and, through their MOOSE implementation, they will be available within the CRAB suite. These fast running capabilities will be used to perform uncertainty quantification (due to uncertainties in, e.g., nuclear data, geometry/materials, and thermophysical properties) driven by stochastic sampling with DAKOTA, with the aim to enhance Best Estimate Plus Uncertainty capabilities.