Speaker
Geert Verdoolaege
Description
See the full Abstract at http://ocs.ciemat.es/EPS2018ABS/pdf/P2.1078.pdf
Statistical analysis of edge-localized mode timing in JET
G. Verdoolaege1,2 and JET Contributors*
1
Department of Applied Physics, Ghent University, Ghent, Belgium
2
Laboratory for Plasma Physics, Royal Military Academy, Brussels, Belgium
Probability distributions of edge-localized mode (ELM) properties in tokamaks under
stationary plasma conditions contain information about ELM variability that can potentially
be exploited for scenario optimization and ELM control purposes [1–4]. Such information
complements averages taken over multiple ELM bursts and enables calculation of
occurrence probabilities of specific ELM events. In this work, the distributions of inter-ELM
time and ELM duration are studied in a database of JET carbon wall (CW) and ITER-like
wall (ILW) plasmas. A new algorithm has been developed, based on gamma mixture
modelling for robustly extracting the ELM timing under a wide variety of plasma conditions,
with minimal user input. The Weibull distribution is seen to be a good model for the
probability density function of the inter-ELM time, as well as the ELM duration [1, 3]. It
allows to distinguish between type I ELMs (representing the majority of ELM activity in the
database) and small type III ELMs, for similar ELM frequencies. Different distribution
shapes are observed throughout the database, characterized by a variety of widths and tail
lengths. The typical width of the distribution quantifies the regularity of the ELM timing,
while long tails point to intermittent ELM activity. The distribution parameters are visualized
using a projection based on a faithful similarity measure between the distributions (Rao
geodesic distance), allowing to discern several clusters, including those corresponding to
plasmas from baseline and hybrid scenarios in JET ILW [5]. Several trends with the main
plasma parameters are observed and quantified. The database is being extended with
additional plasmas from JET and other devices, aiming at a systematic characterization of
statistical ELM properties and their dependences under various conditions.
[1] A. Shabbir, et al., Nucl. Fusion 57, 036026, 2017.
[2] A. Shabbir, et al., Rev. Sci. Instrum. 87, 11D404, 2016.
[3] B.J. Fritch, A. Marinoni and A. Bortolon, 59 th APS DPP 69(12), JP11.28.
[4] A.J. Webster et al., Phys. Rev. Lett. 110, 155004, 2013.
[5] G. Verdoolaege and J.-M. Noterdaeme, Nucl. Fusion 55, 113019, 2015.
*
See the author list of X. Litaudon et al., Nucl. Fusion 57, 102001, 2017.
