English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

In magnetic confinement fusion research, the achievement of high plasmapressure is key to reaching the goal of net energy production. Themagnetohydrodynamic (MHD) model is used to self-consistently calculate theeffects the plasma pressure induces on the magnetic field used to confine theplasma. Such MHD calculations serve as input for the assessment of a number ofimportant physics questions. The VMEC code is the most widely used to evaluate3D ideal-MHD equilibria, as prominently present in stellarators. However,considering the computational cost, it is rarely used in large-scale or onlineapplications. Access to fast MHD equilbria is a challenging problem in fusionresearch, one which machine learning could effectively address. In this paper,we present artificial neural network (NN) models able to quickly compute theequilibrium magnetic field of W7-X. Magnetic configurations that extensivelycover the device operational space, and plasma profiles with volume averagednormalized plasma pressure $\langle \beta \rangle$ ($\beta$ = $\frac{2 \mu_0p}{B^2}$) up to 5% and non-zero net toroidal current are included in the dataset. By using convolutional layers, the spectral representation of the magneticflux surfaces can be efficiently computed with a single network. To discoverbetter models, a Bayesian hyper-parameter search is carried out, and 3Dconvolutional neural networks are found to outperform feed-forwardfully-connected neural networks. The achieved normalized root-mean-squarederror ranges from 1% to 20% across the different scenarios. The model inferencetime for a single equilibrium is on the order of milliseconds. Finally, thiswork shows the feasibility of a fast NN drop-in surrogate model for VMEC, andit opens up new operational scenarios where target applications could make useof magnetic equilibria at unprecedented scales.

Proof of concept of a fast surrogate model of the VMEC code via neural networks in Wendelstein 7-X scenarios