Sujet These Negulescu Crouseilles (1)

Sujet de stage M2 et thèse 2025-2028
Towards more efficient kinetic and hybrid (kinetic-fluid) models and codes
Responsables de la thèse:
Claudia Negulescu1 , Nicolas Crouseilles2
(1) Laboratoire IMT, Université Paul Sabatier, Toulouse
118 route de Narbonne, F-31062 Toulouse, France
e-mail : [email protected]
(2) INRIA de l’Université de Rennes, IRMAR Laboratory
e-mail : [email protected]
Figure 1: Left: Tokamak; Right: Turbulences in the density of the plasma
Summary
This PhD-thesis will focus on the theoretical as well as numerical aspects of the
turbulent plasma dynamics in a tokamak (see Fig. 1). The subject is related to
the magnetically confined, thermonuclear fusion processes (ITER project) with the aim
to contribute to the research-field of the development of new sources of energy, able to
satisfy the requirements of safety, low environmental impact and unlimited availability
of resources. The controlled thermonuclear fusion, based on an efficient magnetic con-
finement of very hot plasmas, is at the moment one of the most promising concepts.
Classical plasmas, characterized by low densities and high temperatures, are the core
of this physical field. The mathematical framework of this PhD project will be es-
sentially based on a kinetic description of the plasma dynamics. Discussions with the
IRFM-team at the CEA-Cadarache will permit to remain close to reality.
Context
Plasmas are much more than a gas constituted of charged particles. Collective ef-
fects play an important role and the underlying physics is very different from that of
neutral gases. The behaviour of a plasma is very complex, the main reasons being the
nonlinear and self-consistent nature of the coupled system charged-particles ↔ fields
and the multiscale nature of the plasma dynamics.
Fusion plasmas are weakly collisional, due to the high temperatures and the low
densities, such that the kinetic framework is the most appropriate approach for their
detailed description. In particular this amounts to solve a coupled system of two Vlasov
or Fokker-Planck equations for the ion/electron distribution functions fi,e
 q
 ∂t fi + v · ∇x fi +
 (E + v × B) · ∇v fi = Qii (fi ) + Qie (fi , fe )
mi
q
 ∂t fe + v · ∇x fe −
 (E + v × B) · ∇v fe = Qee (fe ) + Qei (fe , fi ) ,
me
coupled to Poisson’s equation for the electrostatic potential (or Maxwell’s equations in
the electromagnetic case)
q
−∆φ = (ni − ne ) , E = −∇φ .
ε0
The difficulty with a fully kinetic treatment is however the high-dimensionality of the
phase-space (6D). Furthermore the presence of multiple spatio-temporal scales makes
the problem even more complicated, rendering it inaccessible for numerical simula-
tions. In our particular case, apart the strong magnetic field, it is the small mass
ratio me /mi ≈ 10−4 of the particles which induces disparate scales and hence diffi-
culties; for a typical tokamak plasma with similar electron and ion temperatures the
electron dynamics is faster than
q the ion dynamics, the ratio of the thermal velocities
mi
being given by vth,e /vth,i = me
≈ 102 . This fact poses rather restrictive time-step
constraints related to the fast electron motion, when a standard discretization of the
bi-kinetic system is used, meaning that the numerical stability requires a CFL-condition
of the type vth,e ∆t ≤ ∆x.
Another phenomena introducing multiple scales in the problem, is the presence of
highly energetic particles in the plasma gas, such as α-particles (ions) or runaway elec-
trons. The “core region” of thermonuclear tokamak is mainly constituted of thermal
ions and electrons, characterized by Maxwellian distribution functions (and called in
the following bulk plasma). Due to instabilities one observes the occurrence of highly
energetic particles, which are characterized by the fact that their kinetic energy is much
higher than the thermal energy of the bulk plasma and their density is much lower,
namely one has
1
Eh = m |vh |2 ≫ kB Tc , nh ≪ nc ,
2
where the index ”h” stands for the energetic (hot) population and the index ”c” for
the thermal (cold) population. These energetic particles interact with the thermal bulk
2
plasma and have various impacts on the overall behaviour of the fusion plasma. This
PhD will focus on this specific energetic particle problematic.
Outline of the PhD
In this PhD thesis, we will investigate (from a theoretical as well as numerical point
of view) new mathematical models as well as numerical algorithms for a more efficient
plasma dynamics description, starting from a kinetic approach. As already mentioned,
a fully kinetic description of the whole electron-ion plasma is very precise, however
for the moment still out of reach in the full coupled 6D phase-space. Furthermore,
such a fully kinetic system contains too many irrelevant spatio-temporal scales for the
study of the plasma processes we intend to investigate. To redress this situation, a
hybrid macroscopic-kinetic approach has to be adopted, eliminating the unnecessary
fast dynamics and keeping the complete low-frequency physics. Such a reduced model is
obtained via an asymptotic analysis, letting some specific parameters tend towards zero.
The aim of this PhD project is to introduce a mathematical multiscale description

of the electron/ion dynamics of a magnetically confined tokamak plasma, separating the
thermal and the energetic particles, which behave differently. The distribution function
of the thermal particles is close to a Maxwellian distribution in velocity (fluid approach),
whereas the energetic particles are far from being in a thermal equilibrium (kinetic ap-
proach). The first aim is to obtain via an adequate scaling procedure and a subsequent
asymptotic limit, a hybrid kinetic-fluid model, the kinetic equation describing the fast,
collisionless particles, and the fluid description describing the slow particles, which had
the time to reach an equilibrium, through the collisions with the bulk. The second
step is then to construct a performant numerical scheme, being able to follow at the
discrete level this asymptotic limit. This should result in much cheaper simulations,
as fluid models are far less expensive than using full kinetic models. During all this
procedure, mathematical rigorous proofs shall sustain and permit to better understand
the methodology.
Required knowledge
To cope with this thematic, the PhD student needs some basic knowledge of physics,
functional analysis, PDEs, numerical methods, simulation languages.

Sujet These Negulescu Crouseilles (1)

Uploaded by

Copyright:

Available Formats

Sujet These Negulescu Crouseilles (1)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Sujet These Negulescu Crouseilles (1)

Uploaded by

Copyright:

Available Formats

Sujet de stage M2 et thèse 2025-2028

Figure 1: Left: Tokamak; Right: Turbulences in the density of the plasma

Outline of the PhD

The aim of this PhD project is to introduce a mathematical multiscale description

You might also like