Journée de Dynamique Non Linéaire

10h00   Volker Naulin   (Risoe National Laboratory Technical University of Denmark; presently at JET)
Issues in plasma edge turbulence: simulating complex behaviour

Résumé du séminaire de V. Naulin

Extremely important for the success of magnetised plasma as a means  
of making nuclear fusion a viable energy source is good energy  
confinement. A most important part in achieving the so-called high  
confinement regime, or H-mode, is the transport barrier at the edge  
of the plasma, where the confining magnetic field lines open to end  
on material surfaces.
Understanding, simulating and modelling the plasma edge, defined as  
the density gradient region together with the scrape off layer  
(SOL), are important steps in creating a predictive capability for  
present and future fusion devices. Transport into the SOL in the  
form of large events, like ELMs, is severely restricting the  
operating space of ITER, the new large fusion experiment presently  
constructed near Cadarache, France, and machines beyond.
Numerical simulations have in the last decade come to the point,  
where comparison with experiment begins to become realistic.  
Results from 3D flux simulations of electromagnetic turbulence  
using scale separation between background and fluctuations  
demonstrate the basic mechanisms of flow generation from  
turbulence. In realistic geometry, however, the flow energy never  
exceeds the energy content of the turbulence, demonstrating the  
lack of ingredients to form an edge transport barrier within such  
models.
Further outward in the SOL scale separation is not at all  
applicable. A large part of the observed dynamics in the SOL is  
determined by the lack of scale separation, resulting in  
fluctuations locally and temporarily exceeding the average by  
orders of magnitude.
Transport in these regions is carried by localised structures,  
leading to skewed PDFs and breaking the assumptions of diffusive,  
Fick type of transport law.
Plasma in the edge shows a number of properties of complex, self  
organised, critical systems. Thus modelling the plasma edge from  
first principles needs to surpass the traditional paradigms of  
transport models to provide technically meaningful data.