CE46 - Modèles numériques, simulation, applications

SImulations with high-order schemes of tranSport and TurbulencE in a tokaMak – SISTEM

SISTEM

Simulations with high-order schemes of tranSport and TurbulencE in tokaMak

Objectives

The SISTEM project aims to successfully achieve the strong scaling-up of plasma simulations in view of the fusion operation in a tokamak of unprecedented size, and with stringent plasma conditions.

The effort will be twofold : enhance numerical performance and capability of solvers resolving fluid models of high-fidelity (3D), in order to tackle a much larger range of spatio-temporal scales than in current machines, and so, the inherent increase in the number of degrees of freedom and enhance the reliability of low-fidelity models (2D ensemble averaged equations) that will remain the only ones able to perform routine simulations prior to experiment, allowing us to vary engineering plasma parameters (power, pumping, …) as well as geometries of the magnetic equilibrium. On one side, the enhance accuracy and geometrical flexibility of the Hybrid Discontinuous Galerkin (HDG) method has the potential to satisfy a certain number of numerical issues, so as to progress towards predictive ITER simulations. New techniques will be developed to handle the strongly anisotropic equations describing a rapid compressible dynamics in the parallel direction to the magnetic field, and a slower incompressible-turbulence-like dynamics in the transverse direction. Specific nonlinear boundary conditions at the wall for the plasma and the magnetic equilibrium will be also addressed. An original implicit-explicit time-discretization scheme will also be developed in order to exploit HDG capabilities while satisfying HPC requirements for parallelization and memory management to tackle ITER size problems. On the other side, we will explore the development of various data assimilation techniques to improve the reliability of the turbulence modelling, which remain a major challenge nowadays for low-fidelity models. We will use experimental and numerical data from tokamak measurements and high-fidelity simulations, respectively, to reduce uncertainties on the free parameters inherently occurring in the models.

- A 2D high-order non-isothermal ?uid solver including neutral dynamics to simulate turbulent transport in ITER size edge plasma and coupled to realistic source of energy in the core.

- A1D minimisation algorithm s able to ?nd the right parameters such that the model ?ts a certain observation (coming from an experience or a more accurate model for instance).

- 3D solver in realistic geometry
- Application of the optimization algorithm in 2D to realistic cases from experimental measurements of WEST
- Parameters sensitivity analysis

- A. Elarif, B. Faugeras, F. Rapetti Tokamak free-boundary plasma equilibrium computation using ?nite elements of class C0 and C1 within a mortar element approach, J. Comp. Phys 2021 (accepted)

- S. Baschetti, H. Bufferand, G. Ciraolo, Ph. Ghendrih, E. Serre, P. Tamain, and WEST team, Self-consistent cross-?eld transport model for core and edge plasma transport, Nucl. Fus. 2021 (submitted)

- M. Scotto, G. Giorgiani, E. Serre, H.Bufferand, G. Ciraolo, P. Tamain, Ph. Ghendrih, F. Schwander, Development and application of a Hybrid Discontinuous Galerkin scheme to high temperature plasmas in variable magnetic con?gurations, 8th European Congress on Computational Methods in Applied Science and Engineering (ECCOMAS 2020, virtual) Jan 11–15, 2021, Paris, France

- G. Giorgiani et al., Paving The Path Towards Predictive Plasma Modeling For Fusion Applications, 8th European Congress on Computational Methods in Applied Science and Engineering (ECCOMAS 2020, virtual) Jan 11–15, 2021, Paris, France

The International Tokamak Experimental Reactor (ITER) currently underconstruction in South France has been designed as the key step between today's fusion research machines and tomorrow's fusion power plants. Regarding the expected thermonuclear plasma performance, ITER will require an unprecedented effort on the way to controlling plasmas heat and particle fluxes. This will call for the design of optimized plasma scenarios during ITER operation to control the heat flow from the thermonuclear source to the wall. The difficulty to get global experimental measurements in a nuclear environment in ITER, will require complementary numerical simulations based on fluid models to fine tune the magnetic configuration and adjust accordingly the edge plasma conditions. However, the capability of current solvers to perform such simulations, both for magnetic equilibrium and turbulence transport accounting for plasma-wall interactions, is still acknowledged by the international community as being largely insufficient.
The SISTEM project aims to successfully achieve the strong scaling-up of plasma simulations in view of the fusion operation in a tokamak of unprecedented size, and with stringent plasma conditions. The effort will be twofold :
- enhance numerical performance and capability of solvers resolving fluid models of high-fidelity (3D), in order to tackle a much larger range of spatio-temporal scales than in current machines, and so, the inherent increase in the number of degrees of freedom.
- enhance the reliability of low-fidelity models (2D ensemble averaged equations) that will remain the only ones able to perform routine simulations prior to experiment, allowing us to vary engineering plasma parameters (power, pumping, …) as well as geometries of the magnetic equilibrium.
On one side, the enhance accuracy and geometrical flexibility of the Hybrid Discontinuous Galerkin (HDG) method has the potential to satisfy a certain number of numerical issues, so as to progress towards predictive ITER simulations. New techniques will be developed to handle the strongly anisotropic equations describing a rapid compressible dynamics in the parallel direction to the magnetic field, and a slower incompressible-turbulence-like dynamics in the transverse direction. Specific nonlinear boundary conditions at the wall for the plasma and the magnetic equilibrium will be also addressed. An original implicit-explicit time-discretization scheme will also be developed in order to exploit HDG capabilities while satisfying HPC requirements for parallelization and memory management to tackle ITER size problems.
On the other side, we will explore the development of various data assimilation techniques to improve the reliability of the turbulence modelling, which remain a major challenge nowadays for low-fidelity models. We will use experimental and numerical data from tokamak measurements and high-fidelity simulations, respectively, to reduce uncertainties on the free parameters inherently occurring in the models. The techniques will concern an automative feed-back loop model to a variational approach based on the minimization of a cost function by direct calculations of the derivatives, the number of free parameter being reduced. This way has never been explored in the fusion community.
Finally, using the same grids and jointly developed numerical schemes for low and high fidelity models are important assets of the project to prepare future work, either via code-coupling or code-merging.
All these challenging issues will be addressed by 3 teams from Ecole Centrale Marseille, CEA Cadarache and University of Nice, which share a multidisciplinary expertise around the same numerical tools.
The combined development and use of a chain of codes based on low and high-fidelity models together with the operation of WEST in Cadarache puts our teams in a quasi-unique position in the fusion community and is one of the major assets of the project.

Project coordinator

Monsieur Eric Serre (Laboratoire de Mécanique, Modélisation et Procédés Propres)

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

M2P2 Laboratoire de Mécanique, Modélisation et Procédés Propres
IRFM Institut de Recherche sur la Fusion par Confinement Magnétique
UNS - LJAD Université Nice Sophia Antipolis - Laboratoire Jean-Alexandre Dieudonné

Help of the ANR 277,776 euros
Beginning and duration of the scientific project: September 2019 - 48 Months

Useful links

Explorez notre base de projets financés

 

 

ANR makes available its datasets on funded projects, click here to find more.

Sign up for the latest news:
Subscribe to our newsletter