Missions

The HR-EBSD consortium, sponsored by three leading French industrial groups (SAFRAN-Tech, EDF, and CETIM) in collaboration with the Georges Friedel Laboratory (LGF) at the École des Mines de Saint-Étienne, sets out to develop an open-source software for high-angular-resolution analysis of EBSD diffraction patterns (HR-EBSD) that is both reliable and applicable to their material challenges.


EBSD/TKD techniques (Electron Backscattered Diffraction / Transmission Kikuchi Diffraction) are standard analytical methods employed in Scanning Electron Microscopy. The diffraction patterns produced by this method are direct projections of the crystal lattice examined by the electron beam.
The HR-EBSD/TKD method is designed to analyze localized residual stresses by quantifying the elastic distortions within the crystal lattice, which manifest as distortions in the diffraction pattern. It is presently constrained by its inherent relative nature, as it depends on image correlation techniques between a so-called reference image and a current image.


The StrainCorrelator software, developed internally by LGF over approximately fifteen years, provides a comprehensive platform for the analysis of residual stresses in relative mode.
To address the existing bottleneck caused by lack of knowledge of the mechanical state at the reference point, an innovative technique utilizing a Radon transform is currently being developed. The proof of concept is conducted using simulated diagrams. These are considerably easier to interpret than experimental diagrams, as they do not account for physical effects that are currently difficult to model. It has been shown that the Radon transform exhibits all the essential properties required for the determination of all diffraction parameters: source point location, crystal orientation, and the deviatoric component of the elastic strain tensor.


The primary objective of the individual appointed to this postdoctoral position will be to adapt the Radon transform to the context of experimental diagrams.

To accomplish this, generative AI techniques will be examined, either within the context of the experimental diagram itself or during the Radon transform analysis phase.

Activités

- Conducting a bibliographic review of generative AI techniques with the aim of exploring their applicability to the specified problem.
- Assuming responsibility for and enhancing the performance of the existing software;
- Development of the Radon transform and its adaptation to experimental diagrams;
- Deployment and implementation of the software with industrial partners;
- Scientific analysis and interpretation of the results;
- Preparation of scientific publications and presentation at conferences.

Compétences

- Demonstrated expertise in developing complex software solutions (the current version of StrainCorrelator is implemented in C++, utilizing Qt for the graphical interface, alongside MPI and OpenMP for parallel processing). The door remains accessible for GPU programming or any other high-performance scientific language.
- Proficiency in image processing and analysis, along with an understanding of projective geometry.
- Proven expertise in generative artificial intelligence.
- Proficiency in SEM, EBSD, and continuum mechanics is desirable but may be developed during the postdoctoral tenure.
- Capacity to collaborate effectively within a team and engage with industry or academic stakeholders.
- Capacity to listen attentively and analyze expressed needs
- Ability to present one's findings through scientific publications in peer-reviewed journals and at international conferences.

Contexte de travail

The designated individual will be affiliated with the Georges Friedel Laboratory at the School of Mines (Saint-Etienne, France). The LGF laboratory comprises approximately one hundred permanent researchers, post-doctoral fellows, and doctoral students specializing in the disciplines of materials, mechanics, and processes.

The research activity on High Angular Resolution EBSD is conducted under the leadership of Claire Maurice, CRHC CNRS, and is incorporated within the "Physics and Mechanics of Materials" (PMM) team.


This postdoctoral position is part of a consortium supported by three leading French industrial companies. The chosen candidate will collaborate closely with members of this consortium, notably Frédéric Adamski (SAFRAN-TECH), Dominique Loisnard and Maxime Mollens (EDF), as well as Anthony Nakhoul and Daniel Maisonnette (CETIM).