Source clustering impact on Euclid weak lensing high-order statistics
In the coming years, the Euclid mission will provide measurements of the shapes and positions of billions of galaxies with unprecedented precision. As the light from the background galaxies travels through the Universe, it is deflected by the gravity of cosmic structures, distorting the apparent shapes of galaxies. This effect, known as weak lensing, is the most powerful cosmological probe of the next decade, and it can answer some of the biggest questions in cosmology: What are dark matter and dark energy, and how do cosmic structures form?
The standard approach to weak lensing analysis is to fit the two-point statistics of the data, such as the correlation function of the observed galaxy shapes. However, this data compression is sub- optimal and discards large amounts of information. This has led to the development of several approaches based on high-order statistics, such as third moments, wavelet phase harmonics and field-level analyses. These techniques provide more precise constraints on the parameters of the cosmological model (Ajani et al. 2023). However, with their increasing precision, these methods become sensitive to systematic effects that were negligible in the standard two-point statistics analyses.
One of these systematics is source clustering, which refers to the non-uniform distribution of the galaxies observed in weak lensing surveys. Rather than being uniformly distributed, the observed galaxies trace the underlying matter density. This clustering causes a correlation between the lensing signal and the galaxy number density, leading to two effects: (1) it modulates the effective redshift distribution of the galaxies, and (2) it correlates the galaxy shape noise with the lensing signal. Although this effect is negligible for two-point statistics (Krause et al. 2021, Linke et al. 2024), it significantly impacts the results of high-order statistics (Gatti et al. 2023). Therefore, accurate modelling of source clustering is critical to applying these new techniques to Euclid’s weak lensing data.
In this project, we will develop an inference framework to model source clustering and asses its impact on cosmological constraints from high-order statistics. The objectives of the project are:
1. Develop an inference framework that populates dark matter fields with galaxies, accurately modelling the non-uniform distribution of background galaxies in weak lensing surveys.
2. Quantify the source clustering impact on the cosmological parameters from wavelet transforms and field-level analyses.
3. Incorporate source clustering in emulators of the matter distribution to enable accurate data modelling in the high-order statistics analyses.
With these developments, this project will improve the accuracy of cosmological analyses and the realism of the data modelling, making high-order statistics analyses possible for Euclid data.
Simulation of heterogeneities in battery cells using materials with lower environmental impact
The electrification of vehicles to decarbonize our activities faces a dilemma concerning batteries, their environmental impact and the supply of materials needed to manufacture them. The low-environmental-impact materials being considered today to meet these needs (LF(M)P, sodium-ion technology, etc.) have specific electrochemical characteristics that need to be anticipated before they can be used in large-capacity batteries. These two- or multi-phase materials have an electrical potential that is only slightly dependent on the state of charge. This characteristic favors the appearance of a highly heterogeneous state of charge in the cell. The complex mechanism is linked in particular to fast charging, which is very important for vehicles, and which creates significant heating at the heart of the cells. These heterogeneities limit battery performance and shorten their lifespan. In addition, the flat voltage profile and heterogeneities make it extremely difficult to diagnose the cell's state of charge and state of health. Yet this information is crucial for battery management that maximizes battery life.
Our laboratory is developing advanced modeling tools that enable us to simulate these phenomena. Using a highly detailed numerical model of a large cell, applied to realistic cycling conditions, the candidate will highlight the internal state of cells, which is difficult to access experimentally, and show how cycling, thermal management or diagnostic strategies need to be adapted for the more sustainable chemistries envisaged today. To do this, he will use CEA's software platforms and supercomputers, and draw on CEA/LITEN's expertise covering all technological stages, from materials to real-life cell testing.
Predicting thermodynamic properties of defects in medium-entropy alloys from the atomic scale through statistical learning
The properties and behaviour of materials under extreme conditions are essential for energy systems such as fission and fusion reactors. However, accurately predicting the properties of materials at high temperatures remains a challenge. Direct measurements of these properties are limited by experimental instrumentation, and atomic-scale simulations based on empirical force fields are often unreliable due to a lack of precision.This problem can be solved using statistical learning techniques, which have recently seen their use explode in materials science. Force fields constructed by machine learning achieve the degree of accuracy of {it ab initio} calculations; however, their implementation in sampling methods is limited by high computational costs, generally several orders of magnitude higher than those of traditional force fields. To overcome this limitation, two objectives will be pursued in this thesis: (i) to improve active statistical learning force fields by finding a better accuracy-efficiency trade-off and (ii) to create accelerated free energy and kinetic path sampling methods to facilitate the use of computationally expensive statistical learning force fields.
For the first objective, we improve the construction of statistical learning force fields by focusing on three key factors: the database, the local atomic environment descriptor and the regression model. For the second objective, we will implement a fast and robust Bayesian sampling scheme to estimate the anharmonic free energy, which is crucial for understanding the effects of temperature on crystalline solids, using an adaptive bias force method that significantly improves convergence speed and overall accuracy.We will apply the methods developed to the calculation of free energy and its derivatives, physical quantities that give access to the thermo-elastic properties of alloys and the thermodynamic properties of point defects. To do this, we will use algorithmic extensions that allow us to sample a specific metastable state and also the transition paths to other energy basins, and thus to estimate the free energies of formation and migration of vacancy defects. The thermodynamic quantities calculated will then be used as input data for kinetic Monte Carlo methods, which will make it possible to measure the diffusion coefficients in complex alloys as a function of temperature.
One aim will be to try to relate the atomic transport properties to the complexity of the alloy. Since our approach is considerably faster than standard methods, we will be able to apply it to complex alloys comprising the elements W, Ti, V, Mo and Ta at temperatures and compositions that have not been studied experimentally.
Design of asynchronous algorithms for solving the neutron transport equation on massively parallel and heterogeneous architectures
This PhD thesis work aims at designing an efficient solver for the solution to the neutron transport equation in Cartesian and hexagonal geometries for heterogeneous and massively parallel architectures. This goal can be achieved with the design of optimal algorithms with parallel and asynchronous programming models.
The industrial framework for this work is in solving the Boltzmann equation associated to the transportof neutrons in a nuclear reactor core. At present, more and more modern simulation codes employ an upwind discontinuous Galerkin finite element scheme for Cartesian and hexagonal meshes of the required domain.This work extends previous research which have been carried out recently to explore the solving step ondistributed computing architectures which we have not yet tackled in our context. It will require the cou-pling of algorithmic and numerical strategies along with programming model which allows an asynchronousparallelism framework to solve the transport equation efficiently.
This research work will be part of the numerical simulation of nuclear reactors. These multiphysics computations are very expensive as they require time-dependent neutron transport calculations for the severe power excursions for instance. The strategy proposed in this research endeavour will decrease thecomputational burden and time for a given accuracy, and coupled to a massively parallel and asynchronousmodel, may define an efficient neutronic solver for multiphysics applications.
Through this PhD research work, the candidate will be able to apply for research vacancies in highperformance numerical simulation for complex physical problems.
Multiphysics modeling of fission gas behavior and microstructure evolution of nuclear fuels
The climate crisis demands urgent action and a rapid shift towards carbon-free technologies. This requires the development of advanced materials for more efficient electricity production and storage, including innovation in nuclear reactor fuels. To enhance the safety and efficiency of both current and future nuclear power plants, it is crucial to understand and predict fuel behavior under operating and accidental conditions.
A critical issue is related to fission gases produced upon nuclear fissions. These gases have low solubility and form small bubbles that grow from nanoscale to microscale during fuel operation, significantly impacting the fuel's overall properties. While experimental characterization is essential, numerical simulations complement this work by modeling bubble formation and growth, as well as the consequences in terms of changes in fuel properties. This approach is key to the design of next-generation, high-performance nuclear fuels.
This PhD project aims to advance simulation models for fission gas behavior within the polycrystalline structure of nuclear fuels, with a particular focus on uranium oxides. The PhD student will develop a physical model using the phase-field method, compute necessary input parameters, and conduct numerical simulations that replicate irradiation experiments performed in our department. Direct comparison between simulation results and experimental data will enable deeper insights into the underlying physics of gas behavior, including bubble formation, gas release, and fuel swelling. Additionally, this project will serve as validation for the INFERNO scientific code that will be used for these simulations on national supercomputers.
The research will be conducted at the Nuclear Fuel Department (DEC) of the IRESNE Institute(CEA-Cadarache), in collaboration with CEA fuel modeling and experimental characterization experts. The PhD student will have opportunities to share their findings through scientific publications and presentations at international conferences. Throughout the project, they will develop expertise in multiphysics modeling, numerical simulations, and scientific computing. These highly transferable skills will prepare them for a successful career in academic research, industrial R&D, or materials engineering.
References :
https://doi.org/10.1063/5.0105072
https://doi.org/10.1016/j.commatsci.2019.01.019
Multiphysical modeling of a dual-frequency induction-heated metallothermic reactor
The recycling of uranium extracted from spent fuel (reprocessed uranium or URT) is of major strategic interest as regards both closure and economics of the cycle as well as for national sovereignty. France has initiated the development of a reprocessing route for this URT, involving an entire production chain relying on SILVA laser enrichment technology.
In this context, the CEA is in charge of developing all the processes in this chain, in particular the steps involved in the conversion of uranium oxide into uranium metal required for laser enrichment. For this purpose, the “Laboratoire d'étude des technologies Numériques et des Procédés Avancés” (LNPA) is studying the transposition of the historical metallothermy process to a cold crucible type reactor. This dual-frequency inductive furnace is designed to melt a two-phase charge consisting of a fluorinated slag and a metal produced in situ by the metallothermic reaction.
Alongside a multi-year technology development program on reduced-scale inactive pilot plants, numerical modeling studies of the reactor are undertaken in order to consolidate the change in working scale and enable system parameters to be optimized before deployment of the technology in active operation on depleted uranium for validation tests. The aim of the proposed thesis work is to develop the magneto-thermo-hydraulic (MTH) multiphysical model of the cold crucible metallothermic furnace.