A fast neural network surrogate model for the eigenvalues of QuaLiKiz
Journal article, 2023

We introduce a neural network surrogate model that predicts the eigenvalues for the turbulent microinstabilities, based on the gyrokinetic eigenvalue solver in QuaLiKiz. The model quickly provides information about the dominant instability for specific plasma conditions, and in addition, the eigenvalues offer a pathway for extrapolating transport fluxes. The model is trained on a 5 × 106 data points large dataset based on experimental data from discharges at the joint European torus, where each data point represents a QuaLiKiz simulation. The most accurate model was obtained when the task was split into a classification task to decide if the imaginary part of eigenvalues were stable ( ≤ 0 ) or not, and a regression model to calculate the eigenvalues once the classifier predicted the unstable class.

Author

Emil Fransson

Chalmers, Space, Earth and Environment, Astronomy and Plasmaphysics

Andreas Gillgren

Chalmers, Space, Earth and Environment, Astronomy and Plasmaphysics

A. Ho

Dutch Institute for Fundamental Energy Research (DIFFER)

J. Borsander

Student at Chalmers

O. Lindberg

Student at Chalmers

W. Rieck

Student at Chalmers

M. Åqvist

Student at Chalmers

Pär Strand

Chalmers, Space, Earth and Environment, Astronomy and Plasmaphysics

Physics of Plasmas

1070-664X (ISSN) 1089-7674 (eISSN)

Vol. 30 12 123904

Borderline: developing an integrated core-edge modelling capacity for fusion relevant scenarios

Swedish Research Council (VR) (2020-05465), 2021-01-01 -- 2024-12-28.

Subject Categories

Subatomic Physics

Computational Mathematics

DOI

10.1063/5.0174643

More information

Latest update

1/10/2024