Miralles Dolz, Enrique
(2024)
Uncertainty quantification in fusion power plant design
PhD thesis, University of Liverpool.
|
Text
201382220_Feb2024.pdf - Author Accepted Manuscript Download (12MB) | Preview |
Abstract
The characterisation and quantification of uncertainties are fundamental activities in engineering design, often carried out within the framework of probability theory. While probability theory is successful for capturing aleatory uncertainties, it imposes too strict assumptions when dealing with epistemic uncertainty. This thesis develops methodologies for incorporating epistemic uncertainty into the engineering design processes of fusion power plants, specifically for uncertainty quantification, sensitivity analysis, optimisation under uncertainty, and risk analysis. Conventional techniques for sensitivity analysis show limitations in handling epistemic uncertainty. This work introduces an interval-based global sensitivity analysis method that overcomes these limitations. This method only requires expressing the input parameter uncertainty as intervals, and can be combined with interval analysis for uncertainty propagation in a computationally efficient way. Nuclear data is an essential input for any nuclear calculation. However, nuclear data for some reactions relevant to fusion power plant design is subject to epistemic uncertainty due to lack of experimental data. This thesis develops a framework capable of incorporating nuclear data uncertainty into neutronics optimisation problems, and shows its usefulness for finding more reliable designs. Probabilistic risk assessment, including fault-tree analyses, employ logic gates that rely on precise distributions for failure rates based on available data. This becomes a notable limitation in contexts where data scarcity leads to imprecise failure rate estimations, such as in assessments for fusion power plants. In this work, logic gates are generalised to operate with precise probabilities, intervals and p-boxes, with any input correlation. An application on a pressure tank system is shown, suggesting that assumptions on the probability or dependence of the events can have a substantial impact on the outcome of the risk analysis, and these should be carefully addressed in any serious assessment. Incorporating epistemic uncertainty in engineering design activities is critical for advancing technology safely and efficiently. In fusion power plant design, where the technology is complex and the stakes are high, neglecting epistemic uncertainty can lead to designs that fail to comply with their requirements or, at worst, pose serious risks. The methodologies developed in this thesis overcome limitations of existing methods, offering an improvement for making better-informed decisions and an advance towards transparency and honesty.
| Item Type: | Thesis (PhD) |
|---|---|
| Uncontrolled Keywords: | nuclear fusion, optimisation under uncertainty, risk analysis, sensitivity analysis, uncertainty quantification |
| Divisions: | Faculty of Science & Engineering Faculty of Science & Engineering > School of Engineering |
| Depositing User: | Symplectic Admin |
| Date Deposited: | 16 Jan 2025 10:34 |
| Last Modified: | 19 Mar 2025 10:31 |
| DOI: | 10.17638/03185662 |
| Supervisors: |
|
| URI: | https://livrepository.liverpool.ac.uk/id/eprint/3185662 |
| Disclaimer: | The University of Liverpool is not responsible for content contained on other websites from links within repository metadata. Please contact us if you notice anything that appears incorrect or inappropriate. |
Altmetric
Altmetric