Embedded Files
Foremost among the numerous challenges that remain before ITER operation is the issue of power exhaust, i.e. the control of heat fluxes onto the tokamak walls at high heat confinement performance and for both steady-state and transient regimes.
The difficulty to get global experimental measurements in tokamak, particularly with a nuclear environment in ITER, requires complementary numerical simulations to well-designed the magnetic configuration and to tune accordingly the edge plasma conditions.
However, the capability of current solvers to perform such simulations is still acknowledged by the international community as being an issue that requires a strong scale-up of these latter for such a tokamak of a size as yet unequalled. Compared to current tokamaks ITER simulations lead to an increase in the number of degrees of freedom of two or three orders, depending on the conditions and plasma variables to simulate. Moreover, a precise description of plasma-wall interactions is mandatory to get realistic estimates of the power load on the wall, which requires in particular modeling plasma transport up to the tokamak wall. A precise description of the wall as well as the dynamics of all particles produced becomes therefore a requirement for engineering studies during the operation of ITER.
The SISTEM project is concerned by fluid models. Scaling-up such simulations towards ITER operation needs thus a drastic enhancement of 3D fluid solvers performances, both in terms of numerical performances and capability to deal with the physics of the power exhaust in ITER. However, despite the constant increase of computational power, the routine exploration prior to experiments requires the development of reliable reduced models corresponding to lower-order representations of the solution and thus a more modest computational cost.
To tackle all these issues, the SISTEM project aims over the four years to develop a new generation of algorithms in upgraded versions of three state-of-the art numerical platforms developed for many years by project partners :
TOKAM3X-EIRENE (Tamain et al. 2017),
SOLEDGE2D-EIRENE (Bufferand et al. 2013),
CEDRES++/FEEQS.M (Heumann et al. 2015).
To succeed all these challenges, the SISTEM project has been structured around three objectives all involving the three partners:
Objective 1: succeed the scale up of our three solvers to tackle simulations in a realistic ITER configuration and the inherent resolutions associated. On this path, high-order finite-element methods have the potential to satisfy a certain number of numerical issues later detailed, and that have already been clearly identified by the partners to progress towards predictive tokamak plasma simulations (Giorgiani et al. J. Comp. Phys. 2017). This objective is clearly associated to a strong effort of numerical optimization both in relation with supercomputers architectures (memory access, gpu/cpu, I/O,..), and on the development of more efficient algorithms working in elemental bricks of each code like the linear algebra solvers.
Objective 2: increase the reliability of reduced models based on the evolution of averaged quantitiesby exploring the development of optimal control algorithms for data assimilation. This is fully innovative in the fusion community. A major challenge for reduced models for power exhaust simulation is the improvement of the turbulence modelling related to the transport of heat in the perpendicular direction to magnetic flux surfaces from the core, where heat is produced by fusion, to the tokamak chamber, where the heat is extracted. The assimilation of experimental data from tokamak measurements has the potential to reduce uncertainties on the free parameters inherently occurring in the models used to close the averaged fluxes and stresses due to fluctuations (Bufferand et al. CPP 2015).
Objective 3 : promote an interdisciplinary expertise and strong synergies to succeed the scaling-up of edge plasma simulations towards ITER operation. The challenging issues related to the simulation of plasmas in a tokamak of the size of ITER requires an interdisciplinary expertise integrating skills and expertises in plasma physics, applied mathematics, computational fluid mechanics and computing science that are gathered in the three project partners on a regional territory neighbouring ITER. In complement to succeed the scaling-up of edge plasma simulations, the project offers the opportunity for the three partners to get a leading position in the international community on the path to perform simulations during ITER operation.
Page updated
Google Sites
Report abuse