Jury composé de :
Etienne Gravier, Examinator, Prof IJL (Nancy)
Bruce Lipschultz, Reviewers Prof York University (York)
Alberto Loarte, Reviewers, Dr ITER Organization(St Paul lèz durance)
Eric Serre, Advisor, DR CNRS CNRS-M2P2 (Marseille)
Frédéric Schwander, Advisor, MdC Ecole Centrale Marseille-M2P2
Patrick Tamain, Advisor CEA, Dr CEA-IRFM (St Paul lèz durance)
Abstract: The possibility to produce power by using magnetically confined fusion is a scientific and technological challenge. The perspective of ITER conveys strong signals to intensify modeling effort on magnetized fusion plasmas. The success of the fusion operation is conditioned by the quality of plasma confinement in the core of the reactor and by the control of plasma exhaust on the wall. Both phenomena are related to turbulent cross-field transport that is at the heart of the notion of magnetic confinement studies, particle and heat losses. The study of edge phenomena is therefore complicated by a particularly complex magnetic geometry.
This calls for an improvement of our capacity to develop numerical tools able to reproduce turbulent transport properties reliable to predict particle and energy fluxes on the plasma facing components. This thesis introduces the TOKAM3X fluid model to simulate edge plasma turbulence. A special focus is made on the Verification and the Validation of fluid code. It is a necessary step before using a code as a predictive tool.
Then new insights on physical properties of the edge plasma turbulence are explored. In particular, the poloidal asymmetries induced by turbulence and observed experimentally in the Low-Field-Side of the devices are investigated in details. Great care is dedicated to the reproduction of the MISTRAL base case which consists in changing the magnetic configuration and observing the impact on parallel flows in the poloidal plane. The simulations recover experimental measurements and provide new insights on the effect of the plasma-wall contact position location on the turbulent features, which were not accessible in experiments.
At the present time, a link between simulations and experiments is still missing. This issue is due to the geometrical complexity of the edge plasma and to the complexity of experimental measurements themselves which are often indirect (using models) and sometimes perturbative. One way to improve this point is to implement synthetic diagnostics in numerical simulations. In this work, we focus on the main diagnostic used in edge plasma measurements which is the Langmuir probe (LP). Various synthetic probes have be implemented in the 2D fluid turbulence code TOKAM-2D in order to clarify the physical variables actually measured by LPs and the perturbations induced by the presence of the probe on turbulence.