Instabilities, Turbulence and Coupling


Biological fluid flows (pulmonary and cardiovascular)

Flows for magnetic fusion - ITER


Instabilities, Turbulence and Coupling

The team develops a multidisciplinary expertise centered around numerical modeling and the study of neutral or ionized (plasma) fluid flows for the optimization of industrial or technological systems in four major fields with a strong societal impact: energy, urban planning and development, transportation, and health.
The physics of these systems is that of out-of-equilibrium and coupled phenomena, with instabilities leading to turbulence, and interactions between fluid and structure, mixing and transfers, turbulence and transport, ... which require the development of original methods and simulation codes. These studies often carried out in regimes of parameters relevant to the application are done in the context of strong collaborations with our socio-economic partners (AIRBUS, SAFRAN, IRSN, CEA, ITER, AP-HM ...) which are in the DNA of the team.

The team currently has 12 researchers and teachers, and structures its activity around 3 major families of flows.

Team leader

x >

Annuaire personnel permanent

x >

Doctorants, Post-Doctorants et CDD

x >

Dernières publications de l'équipe

  • Elena Alekseenko, A.A. Sukhinov, B. Roux. Modeling of multi-fractional suspended particle pathways in a shallow water basin under influence of strong winds. Regional Studies in Marine Science, 2024, 73, pp.103477. ⟨10.1016/j.rsma.2024.103477⟩. ⟨hal-04515082⟩ Plus de détails...
  • Uwe Ehrenstein. Generalization to differential–algebraic equations of Lyapunov–Schmidt type reduction at Hopf bifurcations. Communications in Nonlinear Science and Numerical Simulation, 2024, 131, pp.107833. ⟨10.1016/j.cnsns.2024.107833⟩. ⟨hal-04408097⟩ Plus de détails...
  • Jingtao Ma, Lincheng Xu, Jérôme Jacob, Eric Serre, Pierre Sagaut. An averaged mass correction scheme for the simulation of high subsonic turbulent internal flows using a lattice Boltzmann method. Physics of Fluids, 2024, 36 (3), ⟨10.1063/5.0192360⟩. ⟨hal-04514161⟩ Plus de détails...
  • Raffael Düll, Hugo Bufferand, Eric Serre, Guido Ciraolo, Virginia Quadri, et al.. Introducing electromagnetic effects in Soledge3X. Contributions to Plasma Physics, 2024, pp.e202300147. ⟨10.1002/ctpp.202300147⟩. ⟨hal-04474339⟩ Plus de détails...
  • Frédéric Schwander, Eric Serre, Hugo Bufferand, Guido Ciraolo, Philippe Ghendrih. Global fluid simulations of edge plasma turbulence in tokamaks: a review. Computers and Fluids, 2024, 270, pp.106141. ⟨10.1016/j.compfluid.2023.106141⟩. ⟨hal-04352255⟩ Plus de détails...
x >

Dernières rencontres scientifiques

Soutenances de thèses et HDR

26 juin 2024 - Hydrodynamics and permeable boundaries: instabilities, filtration, numerical methods - HDR defense Denis Martinand
Held on: Wednesday June 26 at 10:00 am ; amphi 3 Centrale Méditerranée

Abstract: In many cases of fluid mechanics, boundary conditions have such an impact that it could be argued they are more important than the dynamic equations themselves to understand and predict fluid flows. For fifteen years at M2P2, following this line of thought, I have studied analytically and numerically the couplings between hydrodynamics and transfers across permeable and semi-permeable boundaries.

On the one hand, this research is motivated by improving membrane separation (filtration) processes, the efficiency of which abates as the retained materials accumulate. Dynamic filtration techniques make use of hydrodynamic instabilities and turbulent flows to remix these materials and their effectiveness relies upon accurate understanding and models of the couplings between the phenomena associated with the trans-membrane flows and hydrodynamics. On the other hand, these couplings are also fundamentally and theoretically intriguing by their peculiar mechanisms of instability and mixing, and the specific numerical methods they require.

I have addressed these couplings in a Taylor-Couette setup, where the centrifugal instabilities are accurately modelled and controlled. This Taylor-Couette cell specifically presents one or two cylinders permeable to the solvent and impermeable to solutes, modelled by specific boundary conditions. The presentation will focus on two aspects of these couplings. The first one deals with centrifugal instabilities, the dynamics of which evolves as the fluid flows downstream and is also extracted through the permeable cylinder. These instabilities can be described and predicted in the framework of non-linear global modes, but Direct Numerical Simulations also reveal more complex dynamics. The second one considers the coupling by osmotic pressure between concentration boundary layers and centrifugal instabilities. Both stability analyses and Direct Numerical Simulations show that this coupling promotes the instabilities and increase the flux across the membrane.


Dr. Laurette TUCKERMAN, CNRS-Sorbonne Université, Rapporteure
Pr. Uwe HARLANDER, Brandenburgische Technische Universität Cottbus, Rapporteur
Pr. François GALLAIRE, Ecole Polytechnique Fédérale de Lausanne, Rapporteur
Pr. Eric CLIMENT, INP Toulouse, Examinateur
Pr. Richard M. LUEPTOW, Northwestern University, Examinateur
Dr. Eric SERRE, CNRS-Aix-Marseille Université, Examinateur
Pr. Marc MEDALE, Aix-Marseille Université, Tuteur
27 mars 2024 - Study of the dynamics and passive control of heavy ions produced by plasma-wall interaction: toward the elaboration of a predictive model in the suite of codes SOLedge2D-EIRENE from experiments on the tokamak WEST / PhD Defense Luca Cappelli
Doctorant : Luca CAPPELLI

Date : le 27 Mars 2024 à 15h00 ; CEA, IRFM, F-13108 Saint-Paul-Lez-Durance, bâtiment 506, salle René Gravier

Abstract: Fusion, in the context of tokamak devices, is suggested as an alternative for sustainable and high performance energy production. To date, one of the issues preventing its industrial development, concerns plasma-wall interactions. Power discharged from the plasma to the tokamak vessel has two main consequences: gradual erosion of mechanical components and contamination of the plasma by eroded species. Tungsten (W) has emerged as a promising material to enhance the lifespan of mechanical components in tokamaks. However, managing W concentrations in the plasma, especially in comparison to lighter species like carbon, poses a challenge. Ongoing fusion research aims for reliable modeling of erosion and transport of impurities. Modeling is fundamental to support experiments, design and prepare optimized scenarios for future reactors as ITER. In particular fluid models are a promising tool for this end, but they need proper boundary conditions. This Ph.D. thesis delves into the modeling of Werosion and redeposition through Monte Carlo techniques and analytical models in the limit of a collisionless plasma sheath and a smooth eroded surface. The research involves simulating W erosion and transport in the plasma region adjacent to a solid wall, encompassing the combined Debye sheath and Chodura sheath (referred to as the plasma sheath) and a segment of the SOL with dimensions on the order of the Larmor radius of the eroded species. The plasma sheath plays a pivotal role in the erosion and redeposition of W. It is characterized by the presence of a dominant electric field and strong electron density gradients, making kinetic models the only way to solve transport within it. A portion of this research is dedicated to building a model for the sheath, which serves as a plasma background where eroded particle transport is solved using Monte Carlo techniques. The Monte Carlo model computes redeposition probability maps, offering an alternative method to particle tracking for the calculation of redeposition. The study also explores the impact of different impinging energy distributions of ions in both erosion and redeposition. Results indicate that approximating the energy distribution of impinging ions on the eroded material with the average energy underestimates erosion and redeposition only if the average energy is close to the sputtering energy threshold. Otherwise, even using the average energy is accurate. A large database regarding W redeposition was rapidly created using the redeposition probability maps. Data was then used to train a Neural Network (NN), capable of estimating redeposition as a function of local plasma parameters. Additionally, an analytical model is provided to explain the main mechanisms of redeposition and how to roughly estimate redeposition through integration. This approach is less accurate but it is more flexible because it does not need the use of probability maps, nor it is valid for a set of fixed conditions, as the NN. The analytical model allows to rapidly estimate other important parameters such as temperature and average charge state of sputtered and not-redeposited W. Experimental testing validates the sheath model assumptions regarding the electric potential drop, showing good agreement between modeling and experimental measures. Furthermore, the analytical model is applied to the WEST database to gain insights into W sources at the lower divertor. This Ph.D. thesis contributes in plasma-material research providing fundamental insights into the physics of local W erosion and redeposition in the collisionless limit. Beyond its theoretical contributions, this work has practical implications since it provides different modeling tools to estimate local W erosion. Such tools can be integrated into multispecies plasma solvers, paving the way for innovative time-dependent simulations. This in turn could be valuable for the design and optimization of future tokamak reactors incorporating W.

Keywords: Tungsten, Erosion, Redeposition, Plasma-wall interactions, kinetic models, Monte Carlo

Jury :
David TSKHAKAYA - Czech Academy of Sciences / Rapporteur
Andreas KIRSCHNER - Institute of Energy and Climate Research / Rapporteur
Richard PITTS - ITER organization / Examinateur
George TYNAN - University of California San Diego / Examinateur
Clarisse BOURDELLE - IRFM, CEA / Examinatrice
Tiberiu MINEA - University Paris-Saclay / Président du jury
Eric SERRE - M2P2, CNRS / Directeur de thèse
Nicolas FEDORCZAK - IRFM, CEA / Co-encadrant de thèse
Yannick MARANDET - PIIM, CNRS / Membre invité
4 mars 2024 - Simulation et modélisation des propriétés hydroacoustiques des antennes SONAR / Soutenance de thèse Thibaut Rossi
Doctorant : Thibaut ROSSI

Date : le lundi 4 mars 2024 à 14h00 au M2P2, dans l’amphi 3 / Centrale Méditerranée au 38 Rue Frédéric Joliot Curie, 13013 Marseille

Résumé : Les nouvelles générations de sous-marins militaires sont de plus en plus discrètes et donc deviennent de plus en plus difficiles à détecter. Une limitation majeure du seuil de détection est induite par le pseudo-bruit de couche limite généré à proximité des antennes SONAR. La connaissance du spectre en nombre d’onde-fréquence des fluctuations de pression pariétale s’avère indispensable pour estimer ce pseudo-bruit et ainsi améliorer la prédiction des performances et optimiser l’intégration des antennes SONAR sur les porteurs. De plus, une meilleure connaissance de la nature du bruit perçu sur les antennes dans le domaine nombre d’onde-fréquence permet de quantifier les effets des différents traitements du signal utilisés. En effet, les nouveaux traitements SONAR sont sensibles aux corrélations entre capteurs et peuvent dans certains cas présenter une forte amélioration du plancher de bruit perçu en sortie de traitement. Classiquement, on dispose de modèles empiriques pour estimer rapidement ce terme source. Mais le champ d’application de ce type de modèle est circonscrit par les données expérimentales et numériques qui ont servi à leur élaboration, en général des écoulements de plaque plane. En outre, ces modèles ne parviennent pas à estimer correctement les composantes spectrales à l’origine de la majeure partie du pseudo-bruit perçu, et ne sont pas adaptés pour traiter les écoulements turbulents de couche limite avec gradient de pression ou décollement. Dans le cadre de ces travaux de thèse, on propose de développer une méthode d’évaluation du pseudo-bruit généré par des couches limites turbulentes attachées et décollées. Tout d’abord, des simulations des grandes échelles basées sur la méthode de Boltzmann sur réseau sont réalisées pour évaluer la capacité de ce type de solveur à estimer directement le pseudo-bruit d’une couche limite turbulente. Pour cela, des ingrédients numériques particuliers sont utilisés pour stabiliser et réduire le coût numérique des simulations. Ensuite, un modèle analytique est revisité et une procédure numérique est proposée pour calculer le spectre en nombre d’onde-fréquence des fluctuations de pression pariétale de couches limites attachées et décollées. La méthodologie est validée pour une couche limite turbulente sur plaque plane et pour l’écoulement en aval d’une marche descendante. Des données expérimentales sont notamment exploitées pour valider les résultats de la marche descendante.

Mots clés : pression pariétale, équation de Poisson, méthode de Boltzmann sur réseau, simulation des grandes échelles, simulation RANS, écoulement décollé.

Jury :
Xavier Gloerfelt,          Professeur des universités, ENSAM, Rapporteur                        
Marc C. Jacob,           Professeur des universités, École Centrale de Lyon, Rapporteur / Président du jury
Véronique Fortuné,   Maître de conférences, Université de Poitiers, Examinateur 
Lionel Larchevêque, Maître de conférences, Aix-Marseille Université, Examinateur
Pierre Sagaut,            Professeur des universités, Aix-Marseille Université, Directeur de thèse
Raphaël Lardat,          Senior scientific, Thales DMS - UWS, Co-Encadrant