Reasoning over uncertainty in techno-economic assessments of fusion technology for commercialisation
File(s)
Author(s)
Griffiths, Thomas
Type
Thesis
Abstract
As fusion moves towards commercialisation, challenges arise due to considerable uncertainty in the technology. The scope of this thesis is to address these challenges through the use of probabilistic surrogate models. Specifically, we target the uncertainty in techno-economic assessments of spherical tokamak power plant concepts. The research takes an interdisciplinary approach, connecting uncertainty quantification and surrogate modelling with fusion engineering and design. Bayesian networks as surrogate models offer distinct advantages, such as bi-directional inference and a probabilistic representation of uncertainty. We argue that surrogate modelling through Bayesian networks therefore aids in decision-making processes, by offering the ability to reason over uncertain information in fusion designs. A universal framework has been developed and tested to implement these meta-models as tools for fusion developers in this way. Specifically, the network is used in an industry case study with Tokamak Energy. Through bi-directional reasoning, our results identify the feasible regions for plasma physics and engineering parameters that minimise capital expense and maximise heat and electricity production. By nature of being agnostic to the system they’re applied to, our meta-model framework can generalise for reasoning over uncertainty in systems across the fusion industry. By advocating for the wider use of uncertainty-aware frameworks such as ours, we propose that they can offer practical solutions to uncertain problems in this field.
Version
Open Access
Date Issued
2024-10-03
Date Awarded
01/04/2025
License URL
Advisor
Bluck, Michael
Sponsor
Engineering and Physical Sciences Research Council
Tokamak Energy (Firm)
Grant Number
EP/S023844/1
Publisher Department
Department of Mechanical Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)