Lessons Learned in the Development of Source Term Surrogate Models for Repository Performance Assessment
Abstract not provided.
Publications
Surrogate Modeling of Fuel Dissolution
Abstract not provided.
High fidelity surrogate modeling of fuel dissolution for probabilistic assessment of repository performance
International High-Level Radioactive Waste Management 2019, IHLRWM 2019
Two surrogate models are under development to rapidly emulate the effects of the Fuel Matrix Degradation (FMD) model in GDSA Framework. One is a polynomial regression surrogate with linear and quadratic fits, and the other is a k-Nearest Neighbors regressor (kNNr) method that operates on a lookup table. Direct coupling of the FMD model to GDSA Framework is too computationally expensive. Preliminary results indicate these surrogate models will enable GDSA Framework to rapidly simulate spent fuel dissolution for each individual breached spent fuel waste package in a probabilistic repository simulation. This capability will allow uncertainties in spent fuel dissolution to be propagated and sensitivities in FMD inputs to be quantified and ranked against other inputs.
3 Results