< Terug naar vorige pagina
Publicatie
Discriminant analysis for characterization of edge-localized plasma instabilities
Boekbijdrage - Boekabstract Conferentiebijdrage
Advanced data analysis techniques rooted in probabilistic learning, Bayesian inference, signal processing and pattern recognition have led to the evolution of a domain in nuclear fusion plasmas which can be described as U+201Csoft diagnosticsU+201D. Soft diagnostics are becoming increasingly important as they have demonstrated their useful potential in making inferences from complex and/or large amounts of data produced in fusion experiments. This work presents a discriminant analysis based soft diagnostic for the identification of different classes of edge-localized modes (ELMs) and for the determination of operational boundaries for ELMy regimes. ELMs are instabilities occurring in tokamak plasmas in the high-confinement regime (H-mode). Large, unmitigated ELMs can result in significant transient heat loads released onto the plasma-facing components. Hence, discrimination of ELM types and their control and mitigation are crucial for avoiding a significant reduction in the divertor lifetime. The developed diagnostic method has been applied for the classification of type I and type III ELMs in a set of JET carbon-wall plasmas by utilizing the statistics of inter-ELM time intervals and energy losses. This provides a quantitative classification of ELM types, complementary to phenomenological approaches. Further, for earmarking the boundary between type I and type III ELMy regimes a linear discriminant function is constructed using four global plasma parameters, namely toroidal field, plasma current, total input power and average triangularity, in addition to the distributions of inter-ELM time intervals. The function describes the separation boundary between type I and type III ELMy regimes and provides insight into the dependence of the boundary on the plasma parameters. Essentially, this work develops a simple recognition diagnostic with an aim of contributing to physical understanding of ELMs and optimization of control and mitigation schemes.
Boek: 1st EPS conference on Plasma Diagnostics
Aantal pagina's: 1
Jaar van publicatie:2015