There are for now 18 projects.

AdHex : Adapted Hexdominant Meshes.
Heads of the project : Hervé Guillard and Youssef Mesri ;
Funding : IFPEN
Meshes currently used in different scientific domain (reservoir simulation, plasma physics, CFD)
can be qualified hexdominant, which means they consist in a large set of hexahedrons and a few other kind
of elements (prisms, pyramids ...). While mesh adaption is well understood in the context of simplicial meshes,
this is not the case for hexdominant meshes and a lot of questions in this domain are still open.
This project aims to study some techniques of adaption in hexdominant meshes.
See the Full abstract

MARS : Adaptive Mesh Resolution for multiscale
hyperbolic problems using P4EST library. Application to twophase flows.
Heads of the project : Marc Massot, Adam Larat ;
Funding : EM2C, Fédération de Mathématiques Ecole Centrale paris, Maison de la Simulation
This project is concerned with the numerical simulation of multiscale physics,
for which the properties of the system may experience large variations across
thin transition regions, which are localized in space in the computational do
main.
See the Full abstract

CONGAPIC : COnforming/Nonconforming GAlerkin solvers coupled with ParticleInCell schemes.
Head of the project : Martin Campos Pinto ;
Funding : AMIES
The objective of this project is to study a new class of nonconforming solvers for the Maxwell equations
that follow closely the structure of curlconforming Finite Element method (FEM) but involve fully discontinuous spaces,
like DG schemes. These solvers will be coupled with particleincell (PIC) methods in such a way that
the resulting scheme is compatible with the Gauss laws.
See the Full abstract

C1PKH : Simulation of parallel KelvinHelmholtz instabilities with C1 discretization.
Head of the project : Hervé Guillard, Boniface NKonga ;
Funding : INRIA
Due to the existence of nearsonic flows in the scrapeoff layer (SOL) of tokamaks,
a parallel flow instability, similar to the Kelvin–Helmholtz instability known in neutral fluids can develop in the vicinity of the limiter or divertor plates.
The simulation of this instability has been recently succesfully realized with 2nd order finite difference and finite volume codes.
The goal of this study is to consider the same physical problem and geometry with a
highorder numerical scheme, namely a $C^1$ discretization based on the use of PowellSabin elements.
The comparison of these numerical methods will allow to assess the relevance of C1 highorder method, in term of accuracy, CPU time and robustness for
edge plasma turbulence simulations.
See the Full abstract

C4ADD : CEDRES++ for Automated Divertor Design.
Head of the project : Holger Heumann ;
Funding : FZ Jülich and CASTOR INRIA SophiaAntipolis
This project combines CEDRES++ (a Newton solver for the socalled freeplasma boundary equilibrium problem) with reduced plasma edge models to implement various optimization methods for the design of divertor currents.
See the Full abstract
 DEEPOmP : Calcul parallèle par directives OpenMP et application à l’outil de simulation de la dynamique de structures filaires DeepLines.
Heads of the project : Boniface Nkonga and Philippe Dufrancatel ;
Funding : Principia
 FuMHDMoX : Full MHD 3D modeling for Xpoint geometry.
Heads of the project : Boniface Nkonga
This project is to develop, in the finite element context, a numerical strategy for the resolution of Magneto Hydrodynamic (MHD) equations.
See the Full abstract
 CouPhoMom : Coupling hot electrons and photons transport with moment methods.
Heads of the project : Stéphane Brull, Roland Duclous ;
Funding : CEA
This project deals with the coupled evolution of photons and electrons in a
plasma.
See the Full abstract

HOIS : High Order Implicit Schemes for hyperbolic systems.
Head of the project : Hervé Guillard ;
Funding : EuroMéditerranée 3+3 MedLagoon project
Hyperbolic systems are characterized by the coexistence of several distinct waves. When the velocities of these waves are
very different, the system become stiff and the time step necessary to capture one wave can be totally unsuitable for the computation of the other waves.
In these cases, implicit schemes are needed. However, implicit schemes require the solution of linear systems, a procedure that is generally costly and delicate.
To deal with these difficulties, the Defect Correction method (DeC) is a strategy that builds a high order time discretization using only the repeated
solving of firstorder accurate linear systems coming from firstorder spatial discretization.
However, the DeC method has only been evaluated for secondorder accurate time and space discretizations. The aims of this project is to study its suitability for more accurate
discretizations. Precisely, the project will investigate the construction of implicit DeC discretization based either on discontinuous Galerkin methods
or C^1 methods based on Bezier polynomials.
See the Full abstract

IMAGYNE : Improving Methods to Access realistic geometry and high performance in a GYrokinetic codE.
Heads of the project : Guillaume Latu and Virginie Grandgirard ;
Funding : G8Exascale NUFUSE et ANR GYPSI
The objectives of this project are the development of numerical methods to access realistic geometry and realistic physics in the gyrokinetic code GYSELA, and to
improve the performances of the associated solvers.
See the Full abstract

LESSIV : Lagrangian/Eulerian Solvers and SImulations for Vlasov.
Head of the project : Philippe Helluy ;
The objective of this project is to compare several semiLagrangian
and fullyEulerian high order solvers for plasma physics simulations
on complex geometries
See the Full abstract

MultiSplash : MultiSpecies Plasmas Simulations
Head of the project: Francis Filbet
This project is devoted to the construction of numerical methods for multispecies
Vlasov models, based on particle interactions (mean field interactions) issued from
applications on plasma physics.
See the Full abstract

MTFPT : Multiresolution techniques for fluid and plasma turbulence
Head of the project : Kai Schneider ;
Funding : ANR SiCoMHD
Turbulence in fluids and plasmas is characterized by a large multitude of spatial and temporal dynamically active scales and an inherent intermittency. Suitable decompositions, e.g. based on multiresolution, benefit from this nonhomogeneous distribution of active scales and allow for an efficient compression while controling the precision of the data representation.
See the Full abstract
 OPICV : Optimized PIC simulations for VlasovPoisson with strong magnetic field
Head of the project : Edwin ChaconGolcher, Sever Hirstoaga ;
Funding : INRIA
This project is concerned with solving numerically a stiff (multiscale) 4D VlasovPoisson system.
See the Full abstract
 PASSWORD : PlASma in Silico and Waves
Head of the project : Bruno Després ;
Funding : ANR Chrome
The project Password aims at studying the numerical solution of XModes equations by means of standard finite elements and finite difference techniques. XMode equations come from the time harmonic Maxwell’s equations with the cold plasma dielectric tensor. It can be used to model the heating of a magnetic fusion plasma in Tokamaks (ITER),
and is also encountered in reflectometry experiments to probe the plasma.
See the Full abstract

SeLHex : A semiLagrangian solver for the guiding center model on a regular hexagon mesh.
Head of the project : Eric Sonnendrücker ;
Funding : Max Planck Institute for plasma physics
The project SeLHex aims at developing a semiLagrangian solver for the guiding center model on a regular triangular mesh of a hexagonal domain coupled to a spline Finite Element Poisson solver. This will be used to validate to what extent such an approach could favourably replace the traditional topologically polar grids of the poloidal plane which suffer from a singularity at the center and an unequal resolution between the center and the outer par of the computational domain.
See the Full abstract

SAHS : Statistical Analysis of Hot Spots.
Head of the project : Antoine Bourgeade ;
Funding : CEA / CESTA
See the Full abstract

VORTICITY : Efficient solving of the vorticity equation in drift models of plasma turbulence
Head of the project : Hervé Guillard, Patrick Tamain ;
Funding : INRIA, CEA
Drift models use for the determination of the electric potential an equation that expresses that the plasma current is divergence free. In the electrostatic limit,
this equation becomes an anisotropic elliptic equation with variable coefficients and is extremely difficult to solve.
On the other hand, in electromagnetic models, this equation is much easier to solve but this option implies to consider the small time
scales associated with electromagnetic effects. This project intends to compare the two options and to study
the possible efficient ways to solve the vorticity equation for the limit where electromagnetic effects are small.
See the Full abstract