Göttingen 2025 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

P: Fachverband Plasmaphysik

P 3: Magnetic Confinment Fusion/HEPP II

P 3.4: Vortrag

Montag, 31. März 2025, 17:35–18:00, ZHG102

Neural Networks as Solution Ansatz for the Ideal Magnetohydrodynamic Equilibrium Problem — •Timo Thun¹, Andrea Merlo², and Daniel Böckenhoff¹ — ¹Max-Planck-Institute for Plasma Physics, Wendelsteinstraße 1, 17491 Greifswald, Germany — ²Proxima Fusion, Am Kartoffelgarten 14, 81671 Munich, Germany

Quick and accurate solvers for the fixed-boundary ideal magnetohydrodynamic (MHD) equilibrium problem in non axisymmetric magnetic fields can accelerate stellarator optimisation, facilitate high-fidelity real-time control and enable other data-driven algorithms. Unfortunately, current MHD equilibrium solvers either require high computational wall-time or suffer from a lack of accuracy. Solvers based on Neural Networks (NN) enable very fast inference by transferring the bulk of computational load to model training and the creation of datasets, possibly overcoming this dilemma. Recent work presented a fast NN based ideal MHD surrogate model in the magnetic configuration space defined by the stellarator research device Wendelstein 7-X, using a dataset calculated by conventional solvers and the ideal MHD equilibrium force-residual. Training without a dataset removes implicit biases of its solution strategy and avoids computational costs associated with its creation. We present simple NN models trained solely on the physics-based force residual that achieve comparable or better flux surface averaged force residuals than conventional solvers.

Keywords: Machine Learning; Ideal Magnetohydrodynamic Equilibrium Problem; Vmec; Neural Networks; MHD Equilibrium