Uncertainty Quantification for Fusion Plasma-Surface Interactions
- Scientific Achievements
- Performed UQ and global sensitivity analysis (GSA) with the coupled multiphysics code Xolotl-Ftridyn, used to predict material properties of plasma-facing components in future fusion reactors
- Constructed a higher-order surrogate model that can replace evaluations of the expensive coupled code in the full Plasma-Surface Interactions modeling hierarchy
- Methodology
- Constructed polynomial chaos expansion (PCE) surrogates to render GSA computationally tractable
- PCEs are constructed using Bayesian compressive sensing with adaptive basis growth, as implemented in the FASTMath software UQTk
- Results
- In He plasmas under ITER-like conditions, the migration energies of small He clusters that impact the diffusion coefficient have a significant impact on both surface growth and He retention
- In conditions similar to the PISCES linear plasma device, the incoming beam energy and the surface binding energy of tungsten are found to be important
This work was supported by the FASTMath SciDAC Institute and FES/ASCR partnership project
“Plasma Surface Interactions 2 (PSI2): Predicting the Performance and Impact of Dynamic PFC Surfaces”
Surface growth and He retention predicted by the coupled code
Total sensitivity of each parameter as a function of time
Pieterjan Robbe, Tiernan Casey, Khachik Sargsyan, and Habib N. Najm, Sandia National Labs, Livermore CA