Blanc SIMI 9 - Sciences de l'information, de la matière et de l'ingénierie : Sciences de l'ingénierie, matériaux, procédés, énergie

GYrokinetic high Performance Simulation for ITER – GYPSI

Submission summary

The ANR project GYPSI, GYrokinetic high Performance Simulation for ITER, is dedicated to the understanding and control of turbulent transport in thermonuclear plasmas in magnetic confinement devices. This aspect of first principle physics plays a key role in achieving the level of performance expected in fusion reactors. In the ITER design, the latter was estimated by extrapolating an empirical law. The simulation and understanding of the turbulent transport in fusion plasmas remains therefore an ambitious endeavour. At the very high temperature achieved in these fusion plasmas the effect of collisions is small so that one must contemplate a kinetic description of the plasma media and therefore investigate plasma turbulence in the full phase space since the neutral fluid treatment is no longer appropriate. A reduction technique allows one to restrict the analysis to the gyrokinetic framework in a 5-dimension phase space. Still, the problem remains formidable from the numerical point of view.
This proposal follows the ANR project EGYPT ending in 2010. The idea of a cross-disciplinary research remains at the basis of the project since we regard the joint effort in physics, applied mathematics and computer science to be mandatory to develop a high performance simulation tool. In the GYPSI project, we have chosen to focus the effort on the GYSELA code, GYrokinetic SEmi-LAgrangian code, and ITER relevant simulations. The team has also evolved and new tasks have been defined based on the progress we have achieved in mastering the problem of gyrokinetic simulations. From the physics and mathematical point of view, we have identified four main tasks that are complemented by three tasks dedicated to applied mathematics and computer science.
(1) Simulation of multi-species plasmas proves to be extremely challenging in particular with respect to the ion-electron mass ratio. The required computer power to achieve such simulations will not be available before years from now. However, it is clearly important to develop the code so that it can handle such a problem with an efficient numerical scheme. (2) Furthermore, one can also address similar issues such as transport in two species plasmas, main ion and impurity, or by introducing reduced modelling of the electron response. This issue drives most of the effort dedicated to increasing the parallelisation performance of the code and especially the task of achieving the full separation of slow and fast variables by considering field aligned coordinates. (3) Another task of growing importance is associated with the small scales generation and the means to handle them. Various physical aspects will bound the scale that can be reached, such as collisions or the gyro-average effect. A precise analysis of the impact of coarse graining mechanism on the conservation properties and transport properties will be addressed from the physical and mathematical point of view. (4) From the numerical point of view, this effort will be backed by developing a geometrical algorithm closely related to the Hamiltonian formulation of the problem that should allow us to determine the coarse graining effects of the numerical scheme. (5) Reduction of the phase space, both the gyrokinetic reduction, and possible reductions to handle the electron dynamics, remains a matter of debate. For deeply trapped electrons, we have the required tools to compare full simulations to simulation with phase space reduction that will provide a first validation of the reduction procedure. Elements to address full six-dimension simulation to test the gyrokinetic reduction will also be analysed. (6) Finally, a task is dedicated to the global aspect of kinetic simulations, in particular that related to the plasma heating and flux driven simulations. (7) This task will be completed by a numerical activity dedicated to an international benchmark activity, regarding both the numerical performance and the validation of the codes.

Project coordinator

Monsieur Philippe GHENDRIH (COMMISSARIAT A L'ENERGIE ATOMIQUE ET AUX ENERGIES ALTERNATIVES - CENTRE D'ETUDES NUCLEAIRES SACLAY) – philippe.ghendrih@cea.fr

The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.

Partner

IRMA - Université de Strasbourg (UdS) UNIVERSITE DE STRASBOURG
CNRS CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE - DELEGATION REGIONALE PROVENCE
CEA COMMISSARIAT A L'ENERGIE ATOMIQUE ET AUX ENERGIES ALTERNATIVES - CENTRE D'ETUDES NUCLEAIRES SACLAY

Help of the ANR 600,000 euros
Beginning and duration of the scientific project: - 48 Months

Useful links