Abstract
Dynamic Mode Decomposition (DMD) was used to produce surrogates for the evolution of spatially-varying nuclide compositions in a TRIGA reactor. First, the method was explored by applying it to a single fuel element in which the data used was produced from the Polaris sequence in SCALE-6.2. Then the method was applied to a full core for which the data used to produce the surrogate was generated from a two-dimensional TRITON/NEWT model of a 78-element TRIGA core depleted over six full-power years. The surrogates have explicit time dependence and exhibited high accuracy for major nuclides in the problem studied. Then the analysis was extended to explore the propagation of time-dependent uncertainties in isotopic composition. The fuel element case showed potential where the full core generalization results revealed that the surrogates eigenpairs are sensitive to the perturbations in the data and hence failed to capture the whole space. Suggestions to address these difficulties were presented in the conclusion section.
| Original language | English |
|---|---|
| Pages (from-to) | 280-288 |
| Number of pages | 9 |
| Journal | Annals of Nuclear Energy |
| Volume | 129 |
| DOIs | |
| State | Published - Jul 2019 |
| Externally published | Yes |
Keywords
- Dynamic mode decomposition
- Nuclide evolution
- Surrogate modeling
- Uncertainty analysis
Fingerprint
Dive into the research topics of 'Modeling isotopic evolution with surrogates based on dynamic mode decomposition'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver