Physics informed deep learning for non-destructive testing

This PhD project lies within the field of Non-Destructive Testing (NDT), which encompasses a range of techniques used to detect defects in structures (cables, materials, components) without causing any damage. Diagnostics rely on physical measurements (e.g., reflectometry, ultrasound), whose interpretation requires solving inverse problems, which are often ill-posed.

Classical approaches based on iterative algorithms are accurate but computationally expensive, and difficult to embed for near-sensor, real-time analysis. The proposed research aims to overcome these limitations by exploring physics-informed deep learning approaches, in particular:

* Neural networks inspired by traditional iterative algorithms (algorithm unrolling),
* PINNs (Physics-Informed Neural Networks) that incorporate physical laws directly into the learning process,
* Differentiable models that simulate physical measurements (especially reflectometry).

The goal is to develop interpretable deep models in a modular framework for NDT, that can run on embedded systems. The main case study will focus on electrical cables (TDR/FDR), with possible extensions to other NDT modalities such as ultrasound. The thesis combines optimization, learning, and physical modeling, and is intended for a candidate interested in interdisciplinary research across engineering sciences, applied mathematics, and artificial intelligence.

Reducing the complexity of France's building stock to better anticipate anticipate energy demand flexibility and the integration of solar solar resources

The aim of this work is to respond to the current challenges of energy transition in the building sector, France's leading energy consumer. French public policies are currently proposing far-reaching solutions, such as support for energy-efficient home renovation and incentives for the installation of renewable energy production systems. On a large scale, this is leading to structural changes for both building managers and energy network operators. As a result, players in the sector need to review their energy consumption and carbon impact forecasts, integrating flexibility solutions adapted to the French standard. Some flexibility levers are already in place to meet the challenges of energy and greenhouse gas emission reduction, but others need to be anticipated, taking into account long-term scenarios for energy renovation and the deployment of renewable energy sources, particularly photovoltaic energy, across the whole of France. The issue of massification is therefore an underlying one. That's why this thesis proposes to implement a methodology for reducing the size of the French installed base based on previously defined criteria. In particular, the aim will be to define a limited number of reference buildings that are statistically representative of the behavior resulting from the application of flexibility strategies that meet the challenges of energy efficiency and limiting greenhouse gas emissions. To this end, the CSTB (Centre Scientifique et Technique du Bâtiment) is developing and making available a database of French buildings (BDNB: Base de Données Nationale des Bâtiments), containing information on morphology, uses, construction principles and energy consumption and performance.

Optimization of transports in the Gas Diffusion Layers of Proton Exchange Membrane Fuel Cells: Artificial Intelligent as a support to define optimal porous structures and usage

The design and manufacturing of innovative materials with required properties is a key objective for developing advanced technologies in the field of energy, such as Hydrogen and Alcaline fuel cells and Electrolysers. These improvements will contribute to propose even more attractive low-carbon electrical energy systems, with reduced pollution and green-house effects.
This thesis focuses on the Gas Diffusion Layer (GDL) which plays a crucial role on the performance and durability of Proton Exchange Membrane Fuel Cell (PEMFC).
Your main aim will be to set-up a numerical approach so as to propose improved porous structures to optimize the different transports inside a GDL, for given targets and constraints. To do so, you will make the bridge between advanced modeling of (electrical, heat, liquid, gas) transports in 3D porous media and artificial intelligence. You will then analyze the influence of the operating conditions on such optimal structures and propose design recommendations.
This work will be conducted in close relationship between world-renowned scientific actors : the fuel cells and the modelling teams of CEA/LITEN (Grenoble), the specialists of transports in porous media at CNRS/IMFT (Toulouse), and the specialists of GDL, modeling and AI at FZJ (Juelich, https://www.fz-juelich.de/en).
Scientific publications are expected and patents could also be proposed.

Towards real-time simulation of thermal scenes in a tokamak to support plasma operations.

Monitoring the surface temperatures and heat fluxes of the walls in nuclear fusion devices is crucial for the operation of fusion machines. To ensure the reliability of these measurements, particularly through infrared imaging, CEA is developing a digital twin capable of modeling the entire infrared (IR) measurement chain, from the thermal source to the sensor.
The objective of this thesis is to create a thermal model that can predict heat fluxes and surface temperatures across the entire machine wall, with a goal of real-time computation. This approach is based on two key developments:
1)Development of a Monte Carlo statistical method: This method will solve the heat equation over large geometries in a complex environment, including a variety of heat sources and materials.
2)Acceleration of calculations on graphics processing units (GPU): Utilization of the Kokkos environment to optimize calculation performance while ensuring portability across all high-performance computing (HPC) platforms.
These developments will be validated and quantitatively evaluated on two experimental platforms: the laboratory test bench MAGRYT and the WEST tokamak, used as a demonstrator machine. The thesis will be conducted in a collaborative framework between CEA/DRF/IRFM and CEA/DES/ISAS. The developments will be integrated into the IR digital twin developed by CEA/IRFM for fusion machines and within a dedicated ray-tracing application for CEA/DES.

Numerical analysis of hypersonic boundary layer transition sensitivity to gas models in flight conditions

Implicit/explicit transition for numerical simulation of Fluid-Structure Interaction problems treated by immersed boundary techniques

In many industrial sectors, rapid transient phenomena are involved in accident scenarios. An example in the nuclear industry is the Loss of Primary Coolant Accident, in which an expansion wave propagates through the primary circuit of a Pressurized Water Reactor, potentially vaporizing the primary fluid and causing structural damage. Nowadays, the simulation of these fast transient phenomena relies mainly on "explicit" time integration algorithms, as they enable robust and efficient treatment of these problems, which are generally highly non-linear. Unfortunately, because of the stability constraints imposed on time steps, these approaches struggle to calculate steady-state regimes. Faced with this difficulty, in many cases, the kinematic quantities and internal stresses of the steady state of the system under consideration at the time of occurrence of the simulated transient phenomenon are neglected.

Furthermore, the applications in question involve solid structures interacting with the fluid, undergoing large-scale deformation and possibly fragmenting. A immersed boundary technique known as MBM (Mediating Body Method [1]) recently developed at the CEA enables structures with complex geometries and/or undergoing large deformations to be processed efficiently and robustly. However, this coupling between fluid and solid structure has only been considered in the context of "fast" transient phenomena treated by "explicit" time integrators.

The final objective of the proposed thesis is to carry out a nominal regime calculation followed by a transient calculation in a context of fluid/immersed-structure interaction. The transient phase of the calculation is necessarily based on "explicit" time integration and involves the MBM fluid/structure interaction technique. In order to minimize numerical disturbances during the transition between nominal and transient regimes, the calculation of the nominal regime should be based on the same numerical model as the transient calculation, and therefore also rely on an adaptation of the MBM method.

Recent work defined an efficient and robust strategy for calculating steady states for compressible flows, based on "implicit" time integration. However, although generic, this approach has so far only been tested in the case of perfect gases, and in the absence of viscosity.

On the basis of this initial work, the main technical challenges of this thesis are 1) to validate and possibly adapt the methodology for more complex fluids (in particular water), 2) to introduce and adapt the MBM method for fluid-structure interaction in this steady-state calculation strategy, 3) to introduce fluid viscosity, in particular within the framework of the MBM method initially developed for non-viscous fluids. At the end of this work, implicit/explicit transition demonstration calculations with fluid-structure interaction will be implemented and analyzed.

An internship can be arranged in preparation for thesis work, depending on the candidate's wishes.

[1] Jamond, O., & Beccantini, A. (2019). An embedded boundary method for an inviscid compressible flow coupled to deformable thin structures: The mediating body method. International Journal for Numerical Methods in Engineering, 119(5), 305-333.

Improving phase field damage models - Application to vitroceramic materials subjected to self-irradiation

The vitrification of nuclear waste is a solution currently adopted for the storage of nuclear waste. The vitroceramic materials considered for this application consist of a glass matrix and inclusions of crystalline phases. Rich in radioactive elements, these inclusions undergo self-irradiation resulting in their swelling, which may cause cracking of the glass matrix. It is necessary to know the maximum amount of inclusions below which the material does not crack. An experimental study on radioactive materials, produced and monitored over time, is excessively expensive and the development of a numerical approach could make it possible to better target the materials to be studied.
Following Gérald Feugueur's thesis work on the subject, which highlighted the difficulty of current models in dissociating crack initiation and propagation, the main goal is to develop and test an improved phase field model incorporating an elasticity-independent crack nucleation criterion, based on regularized models of softening plasticity. The model will be implemented using the finite element method (FEniCS code) and an alternative method using Fourier transforms (AMITEX code). Following cross-validation, the most efficient implementation will be selected for application to large-scale 3D microstructures. Close exchanges with CEA Marcoule will enable us to characterize the microstructure of the materials, and an experiment currently underway should enable us to analyze the potential cracking of these materials under self-irradiation.

Characterization methods for LMJ’s layered cryogenic targets

Inertial Fusion on the Laser Megajoule facilty requires to form a spherical DT layer at cryogenic temperature. A major topic of interest for fusion experiments is the characterization of the layer quality and thickness. The characterization will be done using two technics : optical shadowgraphy and X-ray phase contrast analysis. A cryostat developed by CEA is already available to work on future target designs and layer characterization.
The objectives of the PhD are to understand and model (theorically and numerically) the physics of the layer observation and to develop the characterization test bench in the cryostat’s environment and the image analysis for the 3D description of the layer.
The student will have to learn to use the cryostat, its command system and its simple actual characterization system. After a bibliography research, he will have to study the physics governing the characterization (multiple reflections, refractions, phase contrast, …) and develop the acquisition and image analysis system allowing the 3D description of the layer using images obtained during experiments with the cryostat. Lastly, the coupling between the command of the cryostat and the characterization will be developed. For all these developments, the student will have access to extensive bibliographical data and the expertise of the host team

Compréhension et modélisation du transport des gaz dans un combustible UO2 présentant plusieurs familles de porosités

Sans objet (candidats français uniquement pour cette thèse)

High-Order Hexahedral Mesh Generation for HPC simulation

Top