NAISS
SUPR
NAISS Projects
SUPR
Machine learning based reactor physics calculations
Dnr:

NAISS 2026/3-414

Type:

NAISS Medium

Principal Investigator:

Jan Dufek

Affiliation:

Kungliga Tekniska högskolan

Start Date:

2026-06-01

End Date:

2027-06-01

Primary Classification:

10399: Other Physics Topics

Webpage:

Allocation

Abstract

Machine Learning methods are becoming more and more widely used in nuclear reactor physics calculations for various purposes. In this project, we will use the computing resources to generate training data using Monte Carlo reactor physics codes, such as Serpent 2 and OpenMC, and create neural network models of nuclear reactors that allow for rapid fuel cycle optimisations, and optimisations of design and safety margins.