Vous êtes ici : GDR - Version française > Annonces > Thèses et Post-docs
-
Partager cette page
Appel à candidatures
|
Recherche, Emploi
PhD Position (F/M): Data assimilation from stochastic reduced order models In turbulent fluid dynamics
Du 1 octobre 2024 au 30 septembre 2027
INRAE, Rennes, France
Contacts : valentin.resseguier@inrae.fr ; dominique.heitz@inrae.fr
Contacts : valentin.resseguier@inrae.fr ; dominique.heitz@inrae.fr
The aim of the project is to develop, analyze, validate and compare reduced data assimilation technique for 3D incompressible turbulent flows. The point of departure is the stochastic closure modeling procedure and the existing code coupled with the C++ solver Ithaca FV. The PhD student will acquire a broad vision of the data assimilation process and its limitation: from the theory, to the implementation, the data qualification and the experiments. Both synthetic and experimental data will be considered. We shall also consider the efficient treatment of unknown turbulent inflow conditions and the development of efficient hyper-reduction techniques to handle nonlinear terms.
INRAE presentation
The French National Research Institute for Agriculture, Food, and Environment (INRAE) is a major player in research and innovation. It is a community of 12,000 people with 272 research, experimental research, and support units located in 18 regional centres throughout France. Internationally, INRAE is among the top research organisations in the agricultural and food sciences, plant and animal sciences, as well as in ecology and environmental science. It is the world’s leading research organisation specialising in agriculture, food and the environment. INRAE’s goal is to be a key player in the transitions necessary to address major global challenges. Faced with a growing world population, climate change, resource scarcity, and declining biodiversity, the Institute has a major role to play in building solutions and supporting the necessary acceleration of agricultural, food and environmental transitions.closure modeling.
Work environment, missions and activities
This PhD thesis is funded by the ANR project RedLUM, in collaboration with Inria Bordeaux (Angelo Iollo, Tommaso Taddei), Sant’Anna School of Advanced Studies (Giovanni Stabile), and the companies Scalian DS and Weather Measures. The student will collaborate with the other partners of the team to ultimately devise an efficient simulation framework for agricultural frost control involving wind machines at an agricultural plot scale.
Keywords: model order reduction; data assimilation; observation model; stochastic
To monitor and control fluid system, we seek to estimate the air flow around system, though in fluid mechanics, simulations are generally very expensive in terms of calculation time. To tackle real-time applications, it is necessary to deduce from a data set a reduced-dimensional model, which approximates the original PDE in a specific application framework.
Despite the many recent advances, rapid and reliable approximation of parametric high-Reynolds turbulent flows remains an outstanding task for state-of-the-art model reduction techniques. First, due to the slow decay of the Kolmogorov n-width, projection-based (Galerkin or Petrov-Galerkin) reduced-order models (PROMs) based on proper orthogonal decomposition (POD) need to cope with significantly under-resolved representations of the solution field, which prevent accurate approximations of the full system dynamics. Second, the chaotic nature of the system challenges the predictive abilities of state-of-the-art projection methods.
To address these issues, several authors have proposed stochastic closures of PROMs: the distinctive feature of these approaches is to approximate the deterministic chaotic dynamics through a system of stochastic ordinary differential equations for the dominant POD modes; the prediction of the solution is then obtained through a PROM ensemble forecast. Eventually, these simulations are coupled to a measurement stream (data assimilation) to correct them on-the-fly.
The aim of the project is to develop, analyze, validate and compare reduced data assimilation technique for 3D incompressible turbulent flows. The point of departure is the stochastic closure modeling procedure and the existing code coupled with the C++ solver Ithaca FV. The PhD student will acquire a broad vision of the data assimilation process and its limitation: from the theory, to the implementation, the data qualification and the experiments. Both synthetic and experimental data will be considered. We shall also consider the efficient treatment of unknown turbulent inflow conditions and the development of efficient hyper-reduction techniques to handle nonlinear terms.
Training and skills
Master's degree/Engineering degree
The candidate should have a strong background in numerical methods for PDEs and good programming skills.
Background in computational fluid dynamics and statistical learning will be highly valued.
INRAE's life quality
By joining our teams, you benefit from (depending on the type of contract and its duration):
- up to 30 days of annual leave + 15 days "Reduction of Working Time" (for a full time);
- parenting support: CESU childcare, leisure services;
- skills development systems: training, career advise;
- social support: advice and listening, social assistance and loans;
- holiday and leisure services: holiday vouchers, accommodation at preferential rates;
- sports and cultural activities;
- collective catering.
How to apply
https://jobs.inrae.fr/en/ot-21789
The French National Research Institute for Agriculture, Food, and Environment (INRAE) is a major player in research and innovation. It is a community of 12,000 people with 272 research, experimental research, and support units located in 18 regional centres throughout France. Internationally, INRAE is among the top research organisations in the agricultural and food sciences, plant and animal sciences, as well as in ecology and environmental science. It is the world’s leading research organisation specialising in agriculture, food and the environment. INRAE’s goal is to be a key player in the transitions necessary to address major global challenges. Faced with a growing world population, climate change, resource scarcity, and declining biodiversity, the Institute has a major role to play in building solutions and supporting the necessary acceleration of agricultural, food and environmental transitions.closure modeling.
Work environment, missions and activities
This PhD thesis is funded by the ANR project RedLUM, in collaboration with Inria Bordeaux (Angelo Iollo, Tommaso Taddei), Sant’Anna School of Advanced Studies (Giovanni Stabile), and the companies Scalian DS and Weather Measures. The student will collaborate with the other partners of the team to ultimately devise an efficient simulation framework for agricultural frost control involving wind machines at an agricultural plot scale.
Keywords: model order reduction; data assimilation; observation model; stochastic
To monitor and control fluid system, we seek to estimate the air flow around system, though in fluid mechanics, simulations are generally very expensive in terms of calculation time. To tackle real-time applications, it is necessary to deduce from a data set a reduced-dimensional model, which approximates the original PDE in a specific application framework.
Despite the many recent advances, rapid and reliable approximation of parametric high-Reynolds turbulent flows remains an outstanding task for state-of-the-art model reduction techniques. First, due to the slow decay of the Kolmogorov n-width, projection-based (Galerkin or Petrov-Galerkin) reduced-order models (PROMs) based on proper orthogonal decomposition (POD) need to cope with significantly under-resolved representations of the solution field, which prevent accurate approximations of the full system dynamics. Second, the chaotic nature of the system challenges the predictive abilities of state-of-the-art projection methods.
To address these issues, several authors have proposed stochastic closures of PROMs: the distinctive feature of these approaches is to approximate the deterministic chaotic dynamics through a system of stochastic ordinary differential equations for the dominant POD modes; the prediction of the solution is then obtained through a PROM ensemble forecast. Eventually, these simulations are coupled to a measurement stream (data assimilation) to correct them on-the-fly.
The aim of the project is to develop, analyze, validate and compare reduced data assimilation technique for 3D incompressible turbulent flows. The point of departure is the stochastic closure modeling procedure and the existing code coupled with the C++ solver Ithaca FV. The PhD student will acquire a broad vision of the data assimilation process and its limitation: from the theory, to the implementation, the data qualification and the experiments. Both synthetic and experimental data will be considered. We shall also consider the efficient treatment of unknown turbulent inflow conditions and the development of efficient hyper-reduction techniques to handle nonlinear terms.
Training and skills
Master's degree/Engineering degree
The candidate should have a strong background in numerical methods for PDEs and good programming skills.
Background in computational fluid dynamics and statistical learning will be highly valued.
INRAE's life quality
By joining our teams, you benefit from (depending on the type of contract and its duration):
- up to 30 days of annual leave + 15 days "Reduction of Working Time" (for a full time);
- parenting support: CESU childcare, leisure services;
- skills development systems: training, career advise;
- social support: advice and listening, social assistance and loans;
- holiday and leisure services: holiday vouchers, accommodation at preferential rates;
- sports and cultural activities;
- collective catering.
How to apply
https://jobs.inrae.fr/en/ot-21789