Post-doc: CNN neural network – managing data uncertainty in the learning database.
The aim is to develop algorithms able to take into account the uncertainty in the learning database of neural networks. The project fits into the context of the dynamic state estimation of liquid-liquid extraction and benefits of its knowledge-based simulator as well as industrial data. Indeed, the status of an industrial chemical process is accessible through operating parameters and available monitoring measures. However, the measures being inherently associated with uncertainty, it is necessary to make the data consistent with process knowledge. Therefore, the goal is to find the best data set of operational parameters (input of the knowledge-based simulator) to provide the model to estimate the real process state known through monitoring measures (output of the knowledge-based simulator). A convolutional neural network (CNN) is being developed in another postdoctoral project to solve the inverse problem to find the best input thanks to the measured output. A consistent set of operating parameters is going to be obtained and state of the process is going to be known during the dynamic regime of the liquid-liquid extraction process. This first step is to evaluate the impact of the uncertainty of operational parameters on the outputs of the knowledge-based model. This step will need to connect the knowledge-based model to URANIE, internal platform developed by CEA ISAS. This knowledge must be taken into account in the second part of the project. The uncertainty observed on the outputs should be taken into account in the learning loop to improve the estimation of the operational parameters by the CNN. The impact of these uncertainties on the CNN computed results must be assesed in order to trust the ability of the CNN to estimate the state of the process.
Through this project, we are at the heart of the thematic of digital simulation for the best control of complex systems.
Optimal Multi Agent System management of smart heat grid using thermal storage
The aim of this work is a major contribution to a software framework based on coupling of Modelica/Jade environments that will allow to model, to simulate and to optimise the control of smart heat grid through dedicated thermal storage models development: interface specification to control the storages in the grid, simplified models design of heat grid’s most crucial components to be integrated in Agents (production, distribution/storage, consumption) and design of consumption and production forecast models in order to manage anticipation and improve the overall efficiency. The evaluation of performance is based on the test case build in Modelica simulation environment.
Multiscale approach of f elements aqueous solutions modeling
A post-doctoral position is available for one year at CEA-Marcoule
The study will be the modeling of concentrated aqueous phases of heavy metal salts using both microscopic and mesoscopic modeling.
Separation processes for heavy metals recycling usually use liquid-liquid extraction with the transfer of ionic species from a concentrated aqueous phase to an organized organic phase.
This post-doctoral research subject relates to the chemical properties of these processes, and especially to the characterization of the aqueous phase using as accurate as possible models. The goal is to understand the various effects (solvation, electrostatic and van der waals forces, entropy…) influencing the structural and energetic properties of these solutions. A multi-scale approach will be used to study some systems of interest for both fundamental and industrial point of view, the aim being the characterization of these solutions from their molecular structure to their thermodynamic properties. The tools and the approach used here have to be be valid for separative chemistry overall.
Modelling of interstitial cluster evolution in body-centered cubic metals after helium implantation
Under irradiation, structural materials inside nuclear reactors undergo changes in mechanical properties, which result from the formation of point defect clusters, such as cavities (clusters of vacancies) and interstitial dislocation loops (clusters of self-interstitial atoms). Understanding the formation processes of such clusters is thus of prime importance. Recently, three-dimensional interstitial clusters, known as C15 clusters, have been shown theoretically to be highly stable in iron. In order to detect such clusters experimentally, an idea is to make them grow, as shown for dislocation loops after helium implantation. This approach will be carried out experimentally in various bcc metals in the framework of the ANR project EPigRAPH, in collaboration with Chimie ParisTech, GEMaC and LPS.
In this project, the following modelling tasks will be performed by the postdoc:
- Electronic structure calculations will be done to obtain the energetic properties of point defects and point defect clusters in the bcc metals envisaged in the project.
- These data will then be used to parameterize a kinetic model based on cluster dynamics. This formalism is particularly well adapted to simulate the evolution of point defect clusters over long physical times.
Fabrication and characterization of high thermal conductivity SiCf/SiC composites
SiCf/SiC ceramic matrix composites are foreseen candidates for structure materials and claddings in fast neutron reactor of 4th generation. However, their use may be limited because of their too low thermal conductivity in the operating conditions (< 10 W/mK).
SiCf/SiC ceramic matrix composites are now elaborated by chemical vapour infiltration (CVI). In order to improve their thermal conductivity (reduced porosity), it is planned to develop a hybrid elaboration process combining CVI and liquid routes.
The objective of this study is to determine the conditions of elaboration of a SiC matrix by liquid routes and then to characterize the thermo-mechanical behaviour of the hybrid composites, particularly in relation to CVI references.
Large-area processing and design of functional piezoelectric nanomaterials for flexible sensors and systems
CEA LETI develops innovative highly flexible strain sensors which exploit the piezoelectric properties of self-organized gallium nitride nanowires. The fabrication steps are basically: i) nanowire growth, ii) nanowire assembly, iii) encapsulation, iv) contacting. First demonstrators with small active area (1.5 cm²) have already been achieved using the Langmuir Blodgett (LB) technique for the assembly of nanowires. The present project is concerned with the scaling-up of the assembly process over large surface areas, as well as controlled patterning of nanowire assemblies in 1D and 2D by using an innovative CEA LITEN roll-to-roll technology called Boostream® which has the same functionalities as LB in its basic function.
The aim of the post doc is to develop a new building block for the Boostream® equipment enabling a controlled assembly of wires with a pre-defined design. The candidate will carry out studies to optimize the wire assembly, develop the process of film patterning and fabricate, integrate and characterize GaN nanowire piezoelectric transducers with dimensions of 15x15 cm².
More generally, this post doc will also provide the opportunity to develop a generic knowledge to manipulate micro or nano wires or fibers giving new solutions in various fields such as surface structuration, electronic skin, energy...
In situ analytical device based on the LIBS technique for the characterization of hard environment liquid media
The proposed research project aims at developing an in situ analytical device based on the LIBS technique for the characterization of hard environment liquid media such as high temperature melting materials or highly volatile liquid metals used for development of low carbon energy production. The project involves two CEA teams specialized in LIBS instrumentation, analytical developments and high temperature environments.
At high temperature, the molten metals have a high surface reactivity leading to processes of oxidation, slagging … Non-intrusive analysis of this surface by traditional LIBS tools leads to a non-representative results of the molten metal chemical composition. In this project, a new-patented concept based on a mechanical stirring coupled to the LIBS device is developed in order to have a renewable and stable surface of the liquid metal. The aim is to have an on-line representative composition of the metal during the treatment process. The developed demonstrator will be validated for the analysis of impurities (at ppmw ranges) in liquid silicon (T> 1450 °C) during the purification process and the crystallization one for photovoltaic applications. At the end of the project, recommendations for in-situ analysis of liquid sodium (used as cooling fluid in nuclear reactors) will be given.
Hydrothermal carbonization as a pretreatment of wastes before their thermochemical conversion by gasification
Gasification, a thermochemical transformation generally performed at about 850°C, produces a gas that can be valorised in cogeneration, or for the synthesis of chemical products or fuels. Some bottlenecks are still present mainly for the gasification of biogenic or fossil origin wastes: irregular feeding in the reactor due to the heterogeneity in form and composition; formation of inorganic gaseous pollutants (HCl, KCl, NaCl, H2S) or organic ones (tars), which are harmful for the process and/or decrease its efficiency, and must be removed before the final application.
The objective of the post-doctoral work will be to test and optimize a pre-treatment step of the resource based on hydrothermal carbonisation (HTC). This transformation is performed at 180-250°C, in a wet and pressurised environment (2-10 MPa). The principal product is a carbonaceous solid residue (hydrochar), that can be valorised by gasification. HTC aims to limit the release of inorganic and organic pollutants in gasification, and to homogenise and improve the physical properties of the resource.
The proposed approach will consist in: experimentations in batch reactors on pre-selected resources and model materials, together with quantification and analyses of products; analysis of results aiming at elucidating the links between the resource and the properties of the hydrochar, as a function of operating conditions; an evaluation of mass and energy balances for the HTC-gasification process.
Etudes sur la physique des gaz et des interactions matière/laser pour la démonstration à l’échelle laboratoire de l’épuration isotopique du palladium (naturel).
Le palladium est un métal rare dont la demande mondiale est en forte augmentation. Or, il est présent en tant que produit de fission dans les combustibles nucléaires usés qui sont retraités en France. Il serait donc intéressant de recycler ce métal. Pour cela, il est nécessaire de procéder à une épuration isotopique, afin de supprimer un des isotopes du palladium, le 107, qui est un radionucléide artificiel à vie longue émetteur béta. Dans le cadre d'un nouveau projet sur 4 ans construit en réponse à l'appel d'offre du Plan d'Investissement et d'Avenir de l’État, le Service d’Etude des Procédés d’Enrichissement propose un contrat post-doctoral ayant pour objectif la compréhension des interactions gaz/laser dans le procédé de séparation isotopique du palladium par Lasers actuellement en cours de développement. L’objectif principal du projet est la démonstration finale de la faisabilité de séparation de palladium naturel (et non radioactif) pour la phase suivante de développement d’un premier pilote.
Le post-doctorant devra en particulier assurer l’étude du mode de production de la vapeur atomique près du point de fusion du métal pur, des mesures de spectroscopie par laser dans l’UV afin d’affiner les séquences sélectives de photoionisation des isotopes désirés. Pour ce faire, il participera à la définition, au montage et au développement de l'évaporateur, et au couplage des lasers du procédé avec l’enceinte à vide. Des échanges seront mis en place sur ce sujet spécifique avec des spécialistes reconnus au sein de la Direction de la Recherche Fondamentale du CEA. Les mesures de diagnostics des lasers mais aussi les mesures provenant des interactions gaz/laser sont à développer. La programmation (en Python et/ou sous Labview) de ces outils est un point essentiel du poste proposé. Une attention particulière sera portée sur les publications à réaliser essentiellement dans le cadre des interactions gaz/laser (photoionisation sélective des atomes d’intérêt et extraction).
Proton conducting interpenetrating polymer networks as new PEMFC membranes
This subject takes place in the frame of the development of proton exchange membrane fuel cells (PEMFC) and the main objective is to increase their performance and durability for operation above 100°C at low relative humidity.
The current standard membranes for use in PEMFC applications remain perfluorosulfonated ionomers such as Nafion® due their good proton conductivity and chemical stability. Nevertheless, their proton conductivity decreases for relative humidity below 70% especially at high temperature because of a too low density of proton conducting groups. This characteristic is a limitation for their use in the working conditions of the requirements for the automotive application. With these polymers, an increase of the proton conducting group density leads to a decrease of mechanical and dimensional stability. Yet, this stability is already quite low and decreases the PEMFC durability. The goal of this subject is to develop new membrane structures based on interpenetrating polymer networks that do not present this antagonism between good mechanical stability and proton conductivity. This strategy which has recently been patented by CEA (patent application number 08 06890) is based on the association of two entangled polymer networks, one sulfonated for proton conductivity and one fluorinated for mechanical and chemical stability.
The applicant will make the membranes and then will characterize their mechanical properties, proton conductivity as well as gas permeability. He will also quantify their performance and durability in a running fuel cell.