Inverse uncertainty quantification in thermal-hydraulic applications
Ensuring the safety of nuclear reactors is fundamental. Nuclear safety is guaranteed through the Defence-in-Depth principle, that establishes multiple independent barriers that protect the people and the environment from the harmful effects of radiation, along with the systems and measures needed to...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis de maestría |
| Fecha de publicación: | 2025 |
| País: | España |
| Institución: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | inglés |
| OAI Identifier: | oai:upcommons.upc.edu:2117/442761 |
| Acceso en línea: | https://hdl.handle.net/2117/442761 |
| Access Level: | acceso abierto |
| Palabra clave: | Nuclear reactors Reactors nuclears Àrees temàtiques de la UPC::Energies::Energia nuclear |
| Sumario: | Ensuring the safety of nuclear reactors is fundamental. Nuclear safety is guaranteed through the Defence-in-Depth principle, that establishes multiple independent barriers that protect the people and the environment from the harmful effects of radiation, along with the systems and measures needed to guarantee their integrity. Deterministic Safety Assessments are one of the main ways to demonstrate regulatory compliance, as they evaluate the plant’s capability to maintain barrier integrity under postulated initiating events and Design Basis Accidents. Best Estimate Plus Uncertainty (BEPU) methodologies are gradually establishing themselves as the favored way to perform Deterministic Safety Assessments of Nuclear Power Plants, because they account for uncertainties in plant states and physical behaviors while employing accurate- to-reality (best estimate) simulation codes. An essential step in applying BEPU methodologies is determining the uncertainty associated with physical models. To obtain the probability density functions of the physical model param- eters, Inverse Uncertainty Quantification (IUQ) methods are applied, which use the compari- son of simulated models to experimental data for this. This thesis applies an IUQ methodology within the context of the second exercise of the ATRIUM project, organized by the OECD/NEA, which seeks to test the guidelines provided by the SAPIUM project. The targeted phenomenon is post-critical heat flux heat transfer, specifically film boiling, due to its significance in many nu- clear accidents. Three separate-effect test databases were evaluated using the RELAP5 thermal- hydraulic code: Becker, Stewart, and THTF. The Stewart database was found unsuitable due to its procedural complexity, and the Becker database introduced more variability than was de- sired, thus the THTF dataset was chosen for detailed analysis. A comprehensive sensitivity analysis, involving over 5000 simulations, identified the most influ- ential uncertainty parameters affecting film boiling phenomena in RELAP5. Using these param- eters, IUQ was conducted employing a methodology developed at UPC involving the optimiza- tion of these uncertain parameters to a large number of cases. Given the limited experimental dataset, synthetic cases were generated to expand the generation database. The derived Proba- bility Density Functions (PDFs) for selected parameters were subsequently validated through forward propagation. The results show that the method followed is promising, but they also highlight limitations in modeling accuracy and experimental data availability. This master thesis presents solutions to these issues, and where future work could build on these results, bringing an opportunity for future research. |
|---|