Saved in:
Bibliographic Details
Main Authors: Thun, Timo, Merlo, Andrea, Conlin, Rory, Panici, Dario, Böckenhoff, Daniel
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2507.03119
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908918152691712
author Thun, Timo
Merlo, Andrea
Conlin, Rory
Panici, Dario
Böckenhoff, Daniel
author_facet Thun, Timo
Merlo, Andrea
Conlin, Rory
Panici, Dario
Böckenhoff, Daniel
contents We present a novel approach to compute three-dimensional Magnetohydrodynamic equilibria by parametrizing Fourier modes with artificial neural networks and compare it to equilibria computed by conventional solvers. The full nonlinear global force residual across the volume in real space is then minimized with first order optimizers. Already,we observe competitive computational cost to arrive at the same minimum residuals computed by existing codes. With increased computational cost,lower minima of the residual are achieved by the neural networks,establishing a new lower bound for the force residual. We use minimally complex neural networks,and we expect significant improvements for solving not only single equilibria with neural networks,but also for computing neural network models valid over continuous distributions of equilibria.
format Preprint
id arxiv_https___arxiv_org_abs_2507_03119
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Improving ideal MHD equilibrium accuracy with physics-informed neural networks
Thun, Timo
Merlo, Andrea
Conlin, Rory
Panici, Dario
Böckenhoff, Daniel
Machine Learning
Artificial Intelligence
Plasma Physics
We present a novel approach to compute three-dimensional Magnetohydrodynamic equilibria by parametrizing Fourier modes with artificial neural networks and compare it to equilibria computed by conventional solvers. The full nonlinear global force residual across the volume in real space is then minimized with first order optimizers. Already,we observe competitive computational cost to arrive at the same minimum residuals computed by existing codes. With increased computational cost,lower minima of the residual are achieved by the neural networks,establishing a new lower bound for the force residual. We use minimally complex neural networks,and we expect significant improvements for solving not only single equilibria with neural networks,but also for computing neural network models valid over continuous distributions of equilibria.
title Improving ideal MHD equilibrium accuracy with physics-informed neural networks
topic Machine Learning
Artificial Intelligence
Plasma Physics
url https://arxiv.org/abs/2507.03119