Uncertainty quantification in three-dimensional plasma equilibrium reconstruction

Robert Köberl

2025-04-15 · PhD

Abstract

The ideal magnetohydrodynamic (MHD) equilibrium is central to analyzing magnetically confined plasmas. Equilibrium reconstruction involves inferring an MHD equilibrium from experimental data. For 3D geometries, such as stellarators, this can be computationally intensive. This work explores probabilistic approaches for uncertainty quantification in equilibrium reconstruction using data-driven methods like dimensionality reduction and surrogate modeling. This enables sample-based uncertainty estimates on equilibrium quantities while reducing computational costs.

@phdthesis{dissertation,
	author = {Köberl, Robert},
	
title = {Uncertainty quantification in three-dimensional plasma equilibrium reconstruction},
	year = {2025},
	school = {Technische Universität München},
	pages = {153},
	language = {en},
	abstract = {The ideal magnetohydrodynamic (MHD) equilibrium is central to analyzing magnetically confined plasmas. Equilibrium reconstruction involves inferring an MHD equilibrium from experimental data. For 3D geometries, such as stellarators, this can be computationally intensive. This work explores probabilistic approaches for uncertainty quantification in equilibrium reconstruction using data-driven methods like dimensionality reduction and surrogate modeling. This enables sample-based uncertainty estimates on equilibrium quantities while reducing computational costs.},
	keywords = {Plasma Physics;Ideal Magnetohydrodynamics;Equilibrium Reconstruction;Uncertainty Quantification;Bayesian Inference;Surrogate Modeling;Dimensionality Reduction},
	note = {},
	url = {https://mediatum.ub.tum.de/1763047},
	doi = {},
}