| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Meeting Materials Challenges for the Future of Fusion Energy
|
| Presentation Title |
The role of AI in Advancing Materials Development and Testing for Fusion Energy Deployment |
| Author(s) |
Ross Allen |
| On-Site Speaker (Planned) |
Ross Allen |
| Abstract Scope |
Materials development for future fusion systems, requires down-selection in advance of long-term and expensive materials testing programmes, under large uncertainy. Explainable and probabilistic AI can accelerate testing and enhance confidence in decision-making in the down-selection process.
This talk explores digiLab’s collaboration with the UK Atomic Energy Authority (UKAEA) to develop probabilistic AI models for fusion materials, focusing on in-vessel structural materials. We demonstrate the application of novel AI methods, including physics-informed Gaussian Processes, to predict phenomena including creep-rupture under limited and uncertain data, improving early down-selection confidence. Probabilistic emulators for finite element analysis (FEA) models to enable efficient sensitivity analysis, identifying key material parameters driving system uncertainty. These emulators were then used to optimize physical testing by guiding sensor placement and experimental design while improving data analysis by employing inverse methods to determine physical parameters from complex sensor and data measurements. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, Modeling and Simulation, Characterization |