SMuK 2023 – wissenschaftliches Programm

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

P: Fachverband Plasmaphysik

P 11: Poster I

P 11.3: Poster

Mittwoch, 22. März 2023, 14:00–15:30, HSZ EG

Applying machine learning to the inverse scattering problem for experimental plasma profiles — •Ewout Devlaminck, Christos Vagkidis, Mirko Ramisch, and Alf Köhn-Seemann — IGVP, University of Stuttgart, Germany

This work proposes a novel method to study the spatially resolved electron density profile of experimental plasmas using machine learning. The approach, here applied to an atmospheric plasma torch, solves the so-called inverse scattering problem of recovering the plasma profile from non-invasive measurements of the scattered microwave field. The proposed multi-output neural network is trained on 1D scattered intensity profiles, obtained from full-wave FDTD simulations of a high-frequency microwave beam traversing the plasma torch setup with various plasma profile settings. As opposed to the conventional experimental diagnostic, which only provides information on the line-integrated plasma density, the neural network can use the same measurement data to predict multiple parameters describing the complete spatial density profile.