Bereiche | Tage | Auswahl | Suche | Downloads | Hilfe
P: Plasmaphysik
P 16: Magnetischer Einschluß 1
P 16.2: Vortrag
Montag, 7. März 2005, 11:30–11:45, HU 3038
Pattern recognition methods for 2d turbulence characterization — •Henning Thomsen1, T. Klinger1, R. König1, A. Alonso2, and C. Hidalgo2 — 1Max-Planck-Institut für Plasmaphysik, EURATOM Association, Germany — 2Laboratorio Nacional de Fusion, Association EURATOM-CIEMAT, 28040 Madrid, Spain
Tremendous efforts in the field of plasma turbulence were undertaken in recent years in order improve the understanding, by theory with sophisticated numeric simulations in more and more realistic geometry and by the experimentalists with improved diagnostics like turbulence imaging techniques. Statistical turbulence characterization is mandatory for an objective comparison between theory and experiment. Since the experimental measurements predominantly took place in only one spatial dimension with a very good temporal resolution, the characterization was usually limited to spectral analysis and correlation techniques. Nowadays, extremly fast cameras are available, which allow turbulence imaging in two spatial dimensions with a good spatial and temporal resolution. Completely new data analysis tools are necessary for an objective evaluation of the data and a comparison to numerical results. In this paper we present such a data evaluation method based on pattern recognition and motion capture techniques, which allow for the automatic derivation of velocity fields, distributions of structure life times and other characteristic turbulence quantities. Turbulence imaging data from the TJ-II stellarator are presented and the capabilities of the new analysis methods are demostrated.