Computational Modelling in Nuclear Science and Engineering: The Role of Artificial Intelligence
View/ Open
Published version
Embargoed until: 5555-01-01
Reason: Version not permitted.
Embargoed until: 5555-01-01
Reason: Version not permitted.
Volume
19
Pagination
42 - 50
Journal
Nuclear Future
Issue
ISSN
1745-2058
Metadata
Show full item recordAbstract
Artificial Intelligence (AI) is proving transformative in other sectors – with due caution, the nuclear sector should actively explore how their community can gain from these new techniques. The AI4PDEs approach can unify physics-based and data-driven approaches, which will facilitate the application of data assimilation, uncertainty and coupling. This approach also enables running code on any Python-compatible computing resource, including the most recent energy-efficient computer The motivation for future use of AI in modelling is: to obtain models that can run as fast, and are as accurate as, current models; to have models that can run on different computer architectures; to form digital twins with relative ease; to quantify uncertainties; to design and control optimisation approaches; to build AI subgrid-scale models; to obtain rapi data-driven surrogates; to produce agile code; to couple physics and observations or multi-physics models relatively easily; and to integrate physics-based approaches with AI workflows. The adoption of these AI capabilities is expected to transform nuclear engineering as well as other fields.