Contrôle Actif de tourbillons de Görtler en écoulements compressibles

[#Offre de #thèse] au laboratoire IUSTI : « Contrôle Actif de tourbillons de Görtler en écoulements compressibles » ; directeur de thèse Pierre Dupont, co-directeur Sébastien PIPONNIAU

Context
The PhD proposal is part of a global research program that involves 3 laboratories for 4 years, through
the ANR Benefit. This program involves the laboratories IUSTI (Marseille), Pprime (Poitiers) and Dyn-
fluid (Paris),respectively, experts in high speed flows, machine learning approaches, and linear stabi-
lity analysis). The plasma actuators studies are developed in collaboration with the French company
Safran Aircraft Engines.
Subject
This PhD project is part of the BENEFIT programme and focuses on the experimental investigation
and implementation of active flow control strategies for turbulent boundary layers developing over
concave walls. The research will be conducted jointly between IUSTI and Institut Pprime, with a strong
articulation between advanced experimental measurements, flow actuation technologies, and real-
time adaptive control strategies.
The main objective of the thesis is to experimentally study and control centrifugal instabilities of Gört-
ler type in compressible turbulent boundary layers, and to assess the performance of adaptive control
laws derived from reduced-order modelling and data-driven approaches developed within the broa-
der project. The work will concentrate on the generation of high-quality experimental databases, the
integration of plasma-based actuation systems, and the implementation of data assimilation and rein-
forcement learning techniques for adaptive control in realistic wind tunnel conditions.

The first part of the thesis will be devoted to the design, development, and exploitation of an extensive
experimental database of turbulent flows over concave walls. Experiments will be carried out in the S8supersonic wind tunnel at IUSTI, which allows continuous variation of Mach number (typically between
0.4 and 0.8) and unit Reynolds number over a significant range. The objective is to characterize the
onset and development of centrifugal instabilities under varying flow conditions.
High-resolution velocity measurements will be performed using stereo Particle Image Velocimetry
(PIV) in crossflow planes, complemented by unsteady wall-pressure measurements. Statistical and
stochastic analyses of synchronized velocity and pressure signals will be conducted to extract space–time
dynamics of the flow, identify dominant instability modes, and quantify mixing and momentum redistri-
bution mechanisms.
In a second stage, Dielectric Barrier Discharge (DBD) plasma actuators will be integrated upstream
of the curved wall in order to introduce controlled perturbations into the boundary layer. The actuation
system will be designed to allow flexible spatio-temporal forcing, consistent with the structure of in-
terpretable control laws developed within the project. The candidate will contribute to the integration,
calibration, and characterization of the plasma actuators under compressible flow conditions. Experi-
ments will then be conducted to assess the ability of different forcing strategies to trigger, suppress,
or modulate Görtler vortices. Phase-locked PIV measurements synchronized with the actuation si-
gnal will enable reconstruction of phase-averaged flow fields and detailed space–time analysis of the
controlled dynamics. Energy efficiency of the actuation strategies will be evaluated, with the objective
of identifying minimal-energy perturbations capable of inducing significant flow modifications.
In parallel, reinforcement learning strategies will be adapted by the partner Pprime for experimental
operation. The experimental platform will thus serve as a validation environment for adaptive control
frameworks, enabling comparison between purely model-based adaptation and data-driven online op-
timization. Robustness with respect to variations in Mach number, Reynolds number, and curvature
will be systematically investigated.
Throughout the project, strong collaboration between IUSTI and Pprime will be maintained. The ex-
perimental database generated in Marseille will feed reduced-order modelling and machine learning
developments in Poitiers, while adaptive control algorithms designed at Pprime will be implemented
and tested experimentally at IUSTI. The PhD candidate will spend research periods at both institutions
to ensure effective integration between experimental and data-driven components.
Overall, the thesis aims to provide a comprehensive experimental assessment of stability-informed,
machine-learning-assisted active flow control strategies. By combining high-fidelity measurements,
plasma-based actuation, and adaptive control in compressible turbulent flows, the work will contri-
bute to the development of energy-efficient and physically interpretable control approaches leveraging
intrinsic flow instabilities. The results are expected to advance both the fundamental understanding
of centrifugal instabilities in turbulent boundary layers and the practical implementation of adaptive
control strategies in realistic aerodynamic configurations.
Profile
We are looking for a motivated candidate with a strong background in fluid mechanics. Background
knowledge in compressible flows and related experimental techniques would be appreciated but is not
mandatory.

Le laboratoire est centré sur les sciences de l’ingénieur autour de recherches en mécanique et énergétique et aborde des problèmes en lien avec de nombreuses applications dans l’industrie, l’environnement ou la santé.

Profil du candidat :
Vous êtes titulaire d'un hashtag#master en hashtag#dynamique des hashtag#fluides ou avez suivi une formation approfondie en mécanique des fluides, avec notamment des connaissances avancées sur les hashtag#écoulements hashtag#turbulents.
 

Compétences recherchées :
- Indispensable : expérience en dynamique des fluides
- Souhaitable : expérience ou connaissances en écoulements compressibles et méthodes expérimentales.