SMuK 2023 – wissenschaftliches Programm

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

P: Fachverband Plasmaphysik

P 11: Poster I

P 11.22: Poster

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

Neural Networks for the analysis of Langmuir probe characteristics — •Jasmin Joshi-Thompson and Mirko Ramisch — IGVP, University of Stuttgart, Germany

Developed in the early 1920s, Langmuir probes continue to be one of the most widely used plasma diagnostic tools. Theoretical curves are fitted to measured current-voltage (I-V) characteristics in order to obtain parameters such as electron density (n_e) and temperature (T_e). For extensive discharge conditions and comprehensive spatial profiles, measuring plasma parameters becomes more challenging and would best be addressed via automation, with manual checks for specific samples. In this work, deep neural networks are used for associating I-V characteristics to plasma parameters and are tested for robustness. Data is collected from the stellarator TJ-K for training and testing the networks, covering magnetized low-temperature plasmas in a broad parameter space. These networks are assessed as an adaptable, automated plasma characterisation method without the need for further control processes.