Bayesian Data Assimilation for Fuel Performance based on OFFBEAT

The EPFL and the PSI have been developing in recent years a new fuel performance tool named OFFBEAT. The tool is based on the OpenFOAM finite-volume library and allows for 1.5-D, 2-D and 3-D fuel performance analyses, both for base irradiation and for transients. OFFBEAT has been preliminary verified and validated and activities are being carried out in terms of uncertainty quantification and data assimilation. An objective is to guide the development of OFFBEAT according to a Best Estimate Plus Uncertainty logic, as well as to fully employ available integral experimental data for model calibration. The current paper focuses on the application of Bayesian Monte Carlo techniques for data assimilation. In particular, the geometry and experimental data from the NEA IFPE database have been employed and the focus is on low-burnup data. This allows, for this preliminary work, to exclude any significant effect related to the history of the pin.