Search

The CEA of Cadarache for more than 10 years, has been committed to comply with the ISO-14001 standard. Within this normative framework, the centre has set up an Environmental Management System. In order to support them, each facility at the centre is responsible for carrying out an environmental analysis that must be sent to them. In this thesis, it describes the approach implemented to carry out an environmental analysis within the Cadarache centre. He will know how to present you one of the two analyses I carried out, the one concerning the AGATE Basic Nuclear Facility specialized in the treatment of the centre's effluents.; Le CEA de Cadarache depuis plus de 10 ans, s’est engagé à respecter la norme ISO-14001. Dans ce cadre normatif, le centre a mis en place un Système de Management Environnemental. Afin de les appuyer, chaque installation du centre a pour mission de réaliser une analyse environnementale qui doit leur être transmis. Dans ce mémoire, il est décrit la démarche mise en place afin de réaliser une analyse environnementale au sein du centre de Cadarache. Il vous saura présenté une des deux analyses que j’ai effectuées, celle concernant l’Installation Nucléaire de Base AGATE spécialisée dans le traitement des effluents du centre.

The CEDRA nuclear facility has been created to receive and store radioactive wastes. A unit composed of engineers aims to insure the good operation of the facility with a permanent industrial operator and various companies involved in diverse operations. The facility is checked by many inspections of the Nuclear Safety Authority. Different assignments have been defined in various areas : safety, radiation protection, security, and environment ... Several missions have been introduced in this report, particularly a main one. It concerns a direct request of the ASN, and I worked to 2years on it : we have to prove the resistance to the fall of our different types of packages of radioactive wastes. The other tasks are related to the operation of the facility. For instance, we have to draft documents about commissioning for a radioactive waste package investigation cell, to reinforce the integrity of waste packages, to update the safety documentation.; L’installation nucléaire de base CEDRA, a été conçue pour recevoir et entreposer des déchets radioactifs. Une équipe composée d’ingénieurs est en charge d’assurer le bon fonctionnement de l’installation en s’appuyant sur un opérateur industriel permanent et différentes entreprises intervenant pour des opérations ponctuelles. L’installation se voit confrontée au réexamen de sûreté ainsi qu’aux différentes visites de contrôle de l’Autorité de Sûreté Nucléaire, et nombreuses sont alors les attentes qui en découlent. Afin d’y répondre, différentes missions à plus ou moins long terme ont été définies et ce dans différents domaines : sûreté, radioprotection, sécurité, environnement... Dans ce rapport sont présentées plusieurs missions réalisées sur ces deux années de Master dont essentiellement une principale, étalée sur deux ans, concernant une demande directe de l’ASN : prouver la tenue à la chute de nos différents types de colis de déchets radioactifs. Les autres missions sont en rapport avec l’exploitation de l’installation : rédiger des documents de mise en service d’une cellule d’examen de colis de déchets radioactifs, renforcer l’intégrité des colis de déchets, mettre à jour la documentation sécurité.

14 pages; International audience; — In this paper, early lumping estimation of space-time varying diffusion coefficient and source term for a non-homogeneous linear parabolic partial differential equation (PDE) describing Tokamak plasma heat transport is considered. The analysis of this PDE is achieved in a finite dimensional framework using the cubic b-splines finite element method with the Galerkin formulation. This leads to a finite dimensional linear time-varying state-space model with unknown parameters and inputs. The Extended Kalman Filter with Unknown Inputs Without Direct Feed-through (EKF-UI-WDF) is applied to estimate simultaneously the unknown parameters and inputs and an adaptive fading memory coefficient is introduced in the EKF-UI-WDF, to deal with time varying parameters. Conditions under which the direct problem is well posed and the reduced order model converges to the initial one are established. Insilico and real data simulations are provided to evaluate the performances of the proposed technique.

International audience; We conducted real-time SER measurements on bulk 65 nm SRAMs in the WEST tokamak during long pulse deuterium-deuterium plasma discharges (~60 s), evidencing bursts of SEUs during the most efficient shots and 12% of MCU events.

We use an in-house code based on polyhedral representation of particle shapes and a rigorous contact detection algorithm to investigate the evolution of orthotropic elastic response of regular polyhedral and spherical particle packings. Our DEM algorithm is able to distinguish different types of contact (face-face, face-edge, face-vertex, edge-edge, edge-vertex, vertex-vertex). A linear force law based on normal overlaps at the contact points between particles is implemented and face-face and face-edge contacts are represented by at least three and two contact points, respectively. Dense and isotropic packings of polyhedral particles with different numbers of faces were prepared by isostatic compaction. Then, they were sheared under tri-periodic boundary conditions for different values of interparticle friction coefficient. During shearing, 16 instances corresponding to a wide variety of contact orientation anisotropies were stored and relaxed a static state before applying two distinct strain probes to measure the five independent elastic moduli of the sample. By comparing the simulation data with effective medium theory (EMT), our results clearly show that the elastic moduli are functions of two microstructural parameters: 1) a constraint number that accounts for contact types (face-face and face-edge contacts between polyhedra), and 2) the contact orientation anisotropy. The proposed expression of elastic moduli isolates the direct effect of particle shape, related to the nonaffine particle displacement field from the indirect effect, related to the granular microstructure. The effect of particle shape appears at two levels: on the one hand, through four parameters in the proposed expression, which are independent of friction coefficient, and, on the other hand, through the microstructure reflected by the values of constraint number and fabric anisotropy, which depend on both particle shape and interparticle friction coefficient.

The solution of Poisson-like equations defined on complex geometry is required for gyrokinetic simulations, which are important for the modelling of plasma turbulence in nuclear fusion devices such as the ITER tokamak. In this paper, we compare three solvers capable of solving this problem, in terms of the accuracy of the solution, and their computational efficiency. The first, the Spline FEM solver, uses C 1 polar splines to construct a finite elements method which solves the equation on curvilinear coordinates. The resulting linear system is solved using a conjugate gradient method. The second, the GmgPolar solver, uses a symmetric finite differences method to discretise the differential equation. The resulting linear system is solved using a tailored geometric multigrid scheme, with a combination of zebra circle and radial line smoothers, together with an implicit extrapolation scheme. The third, the Embedded Boundary solver, uses a finite volumes method on Cartesian coordinates with an embedded boundary scheme. The resulting linear system is solved using a multigrid scheme. The Spline FEM solver is shown to be the most accurate. The GmgPolar solver is shown to use the least memory. The Embedded Boundary solver is shown to be the fastest in most cases. All three solvers are shown to be capable of solving the equation on a realistic non-analytical geometry. The Embedded Boundary solver is additionally used to attempt to solve an X-point geometry, highlighting the problems with concave boundaries.

The CEA is an operator of basic nuclear installations (BNIs) and its priority is the control of the risks inherent in its research activities. This company has its own reference system for the management of occupational risks which is created on the basis of regulatory requirements.Many CEA entities have activities that can be described as an annex to the CEA's core business, but they must still apply this safety reference. This is particularly the case for the ITER France Agency, whose main activity is to ensure the implementation of the commitments made by France for the ITER project partners. The ITER France Agency manages different projects requiring forestry or civil works for example.The feedback of this type of intervention is significant because there are a lot of regular and serious accidents so it is really important to manage safety at work. The interest of applying the CEA safety reference is real but it is necessary to adapt it to each of "specifics" interventions.; Le CEA, en tant qu’exploitant d’installations nucléaires de base (INB), a pour priorité la maitrise des risques inhérents à ses activités de recherche. Cet organisme possède ainsi son propre référentiel en matière de gestion des risques professionnels, créé sur la base des prescriptions réglementaires.De nombreuses entités du CEA ont des activités que l’on peut qualifier d’annexe par rapport au cœur de métier du CEA, or elles doivent tout de même appliquer ce référentiel en matière de sécurité. C’est notamment le cas de l’Agence ITER France dont l’activité principale est d’assurer la mise en œuvre des engagements pris par la France, pendant la période de candidature, vis-à-vis des partenaires du projet ITER. Cela se traduit par différents projets nécessitant par exemple des travaux forestiers ou encore des travaux de génie civil.En prenant en compte le retour d’expérience de ce type d’interventions, on s’aperçoit qu’elles ne sont pas si anodines en matière de sécurité au travail et qu’il est nécessaire d’y prêter une attention particulière. L’intérêt d’appliquer le référentiel sécurité du CEA est donc réel mais il est nécessaire de l’adapter à chacune de ses interventions « spécifiques ».

WOS: 000414625700003, In the EXILL campaign a highly efficient array of high purity germanium (HPGe) detectors was operated at the cold neutron beam facility PF1B of the Institut Laue-Langevin (ILL) to carry out nuclear structure studies, via measurements of gamma-rays following neutron-induced capture and fission reactions. The setup consisted of a collimation system producing a pencil beam with a thermal capture equivalent flux of about 10(8) ns(-1)cm(2) at the target position and negligible neutron halo. The targetwas surrounded by an array of eight to ten anti-Compton shielded EXOGAMClover detectors, four to six anti-Compton shielded large coaxial GASP detectors and two standard Clover detectors. For a part of the campaign the array was combined with 16 LaBr3:(Ce) detectors from the FATIMA collaboration. The detectorswere arranged in an array of rhombicuboctahedron geometry, providing the possibility to carry out very precise angular correlation and directional-polarization correlation measurements. The triggerless acquisition system allowed a signal collection rate of up to 6 x 10(5) Hz. The data allowed to set multi-fold coincidences to obtain decay schemes and in combination with the FATIMA array of LaBr3:(Ce) detectors to analyze half-lives of excited levels in the pico-to microsecond range. Precise energy and efficiency calibrations of EXILL were performed using standard calibration sources of Ba-133, Co-60 and Eu-152 as well as data from the reactions Al-27(n, gamma)Al-28 and Cl-35(n,gamma)Cl-36 in the energy range from 30 keV up to 10MeV., U.S. DOE grant [DE-FG02-91ER40609]; NUPNET; German BMBF [O5P12PKNUF, 05P15RDFN1, 05P12RDNUP]; German DFG contract [KR 1796/2-1]; Polish National Science Centre [DEC-2013/09/B/ST2/03485], This work has been partly supported by the Polish National Science Centre under contract DEC-2013/09/B/ST2/03485, the U.S. DOE grant No. DE-FG02-91ER40609, the NUPNET and the German BMBF under contracts O5P12PKNUF, 05P15RDFN1 and 05P12RDNUP and the German DFG contract KR 1796/2-1.

International audience; Thermal-hydraulics safety requirements for the second and third generation of nuclear reactors led to the development of innovative passive safety systems. In particular, new devices must be developed involving numerical simulations for turbulent two-phase flows around complex geometries. To reduce the time-consuming mesh generation phase when testing various geometries, we use a fictitious domain approach. More specifically, we choose the Penalized Direct Forcing Method to take into account inflow obstacles. Following a recent work, involving the resolution of the one-phase incompressible Navier-Stokes equations using a projection scheme and the Finite Element Method, this paper focuses on different techniques to recover data from the discrete immersed boundary and different ways to achieve order 2 in space via linear interpolation. Indeed, we investigate two data reconstruction approaches (one based on various weighted averaging, the other based on optimization) and compare their results for cylindrical and NACA0012 airfoil shapes: they provide similar accuracy but the weighting is much faster in terms of execution time. We also investigate three different interpolation types: unidirectional, multi-directional and a new hybrid between the two. The Taylor-Couette flow and the flow around a circular cylinder are used to carry out mesh convergence studies. Globally, order 2 in space is numerically assessed in both  2 and  ∞ norms for all the interpolation types, which is consistent with theoretical expectations-even if the space convergence order is a bit higher for the multi-directional approach. For the flow around a circular cylinder, the values of aerodynamic coefficients and Strouhal number are in good agreement with the literature, especially when using directional interpolation. Finally, an industrial case, representative of passive safety systems, is presented to assess the robustness and capability of the method. The simulations tend to show that, here again, the directional interpolation offers the best behavior when dealing with complex geometries and relatively coarse meshes.

International audience; The Management and Uncertainties of Severe Accidents (MUSA) project, funded in HORIZON 2020 and coordinated by CIEMAT (Spain), aims at consolidating a harmonized approach for the analysis of uncertainties and sensitivities associated with Severe Accidents (SAs) focusing on Source Term (ST). In this framework, the objectives of the Innovative Management of Spent Fuel Pool Accidents (IMSFP – WP6), led by IRSN (France), are to quantify and rank the uncertainties affecting accident analyses in a Spent Fuel Pool (SFP), to review existing and contemplated SA management measures and systems and to assess their possible benefits in terms of reduction of radiological onsequences.To quantify the propagation of the uncertainties of the input parameters to the output uncertainties of severe accident codes (ASTEC, MELCOR, RELAP/SCDAP), a diverse set of uncertainty quantification (UQ) tools (DAKOTA, RAVEN, SUNSET, SUSA) are used. The statistical framework used by the different UQ-tools is similar e.g. pure random (Monte Carlo) and Latin hypercube sampling (LHS).Fourteen partners from three different world regions are involved in the WP6 activities. The target of this paper is to describe the achievements during the first three years ofthe project. In a first part, a description is given of the SFP accidental scenario, of the key target variables and radionuclides chosen as ST Figure of Merits (FoM) and of theidentified uncertainty sources in models and input parameters. A key element when defining the SFP scenario has been the consideration (or not) of the reactor building,as it is expected to significantly affect analyses. In a second part, the first insights coming out from the calculation phase of the project are presented. The review of existing SA management measures is also exposed, as well as systems whose benefits will be assessed in the second phase of the project. Finally, challenges that arise from such an exercise are discussed, as well as major difficulties found when applying UQ methodologies to SFP scenarios and solutions adopted.

International audience; In future thermonuclear fusion reactors, like ITER and DEMO, plasma-facing components with tungsten armor material will have to sustain high thermal fluxes (10 MW/m² in steady state and 20 MW/m² in quasi-steady-state). For such extreme conditions, tungsten may reach temperatures higher than 1000 °C and, consequently can be prone to recrystallize, which can limit the lifetime of the divertor targets under cycling thermal loadings. Characterization of recrystallization kinetics involves the use of different methods and devices (furnace, laser heating) depending on the annealing temperature regime to be analyzed. For these methods, the heating and the subsequent characterization of around 10 samples (5 × 4 × 4 mm) per annealing temperature is needed. In this article, a method is proposed to quantify the recrystallization kinetics between 1300 °C and 1600 °C using a limited number of samples and heating conditions. The basic idea is to induce a steady-state temperature gradient in tungsten rods by heating one side of the rod using a laser heating system. Tungsten rods (50 × 4 × 5 mm) were annealed with this method. The resulting microstructures are characterized using optical microscopy and hardness measurements. The different stages of tungsten softening are clearly evidenced. These first results can be considered as a proof of concept for the use of such a methodology to assess tungsten recrystallization at high temperature.

The effect of divertor closure and nitrogen on the detachment process has been studied by performing 2D numerical simulations of TCV H-mode divertor scenarios with SOLEDGE3X-EIRENE edge plasma transport code. The outcomes reveal that: In the cases with only deuterium gas fuelling, there exists a divertor neutral pressure threshold about 0.76 Pa for detachment, despite the difference in divertor closure. Nitrogen can cool the target temperature with little effect on the upstream density and momentum loss but drops the upstream pressure, leading to a decrease in the target particle flux and divertor neutral pressure. Furthermore, when the radiation front starts to move up from the outer target, the peak parallel heat flux level at the outer target remains approximately the same (about 2.3 MW/m²), despite the difference in nitrogen seeding and divertor closure. An empirical partial detachment qualifier calibrated on AUG experimental data was compared with the TCV simulations. The results show good agreement in detachment state prediction, indicating the potential of this detachment qualifier to be applied in devices of different sizes, emphasizing the combined influence of input power entering the divertor, neutral pressure, and the concentration and species of impurity in achieving divertor detachment.

International audience; A numerical FEM (Finite Element Method) model was implemented to represent the mechanical state of the wooden panel of the Mona Lisa, as it is conserved in its exhibition case, and constrained in its auxiliary frame. The model is based on the integration of advanced numerical analysis and various experimental examinations carried out non-invasively on the artwork by the authors during over 15 years. This includes visual, microscopic and X-ray observations together with mechanical measurements and monitoring of panel deformations and constraining external forces. In addition to the development of non-invasive techniques to characterise the mechanical properties of the panel, the FEM model reliably evaluated the strains and stresses generated in the panel by the various actions it experiences. The paper consists of the following parts: (i) a short summary of the experimental measurements and other observations, (ii) a detailed description of the FEM numerical model, of the hypotheses it is based on, and of its advantages and limits, (iii) the main results obtained by running the model. This includes the identification of local strains and stresses, the location of most critical areas, an evaluation of the risk that the existing ancient crack may propagate, and an evaluation of safe ranges for the forces acting on the wooden panel, (iv) the validation criteria for such results, and (v) a discussion about the significance of the mechanical model.

International audience; The emission of $\gamma$-rays after a neutron capture in a cryogenic detector can generate mono-energetic nuclear recoils in the sub-keV regime, of direct interest for the calibration of Dark Matter and Coherent Elastic Neutrino Nucleus Scattering experiments. Here we show that accurate predictions of the nuclear recoil spectra induced by neutron captures require taking into account the interplay between the development in time of the de-excitation $\gamma$-cascade of the target nucleus and that of the associated atomic collisions in matter. We present detailed simulations coupling the FIFRELIN code for the description of the $\gamma$-cascades and the IRADINA code for the modelling of the fast atomic movements in matter. Nuclear recoil spectra are predicted, and made available to the community, for concrete cases of Al$_2$O$_3$, Si, Ge and CaWO$_4$ crystals exposed to a low intensity beam of thermal neutrons. We find that timing effects cause new calibration peaks to emerge in the recoil spectra and also impact the shape of the continuous recoil distribution. We discuss how they could give access to a rich physics program, spanning the accurate study of the response of cryogenic detectors in the sub-keV range, tests of solid state physics simulations and tests of nuclear models.

International audience

Boiling is a phenomenon that can occur in reactors under normal or incidental operating conditions. In the latter case, it can lead to the appearance of a boiling crisis, the mechanisms of which are still not fully understood.The proposed subject therefore concerns the characterisation of nucleate boiling from the appearance of the first bubble (Onset of Nucleate Boiling - ONB) to the appearance of the boiling crisis over a wide pressure range from atmospheric pressure to the pressure of the primary circuit of a Pressurised Water Reactor (155 bar). Indeed, the literature indicates that there is a change in flow structure around 70 bar, with the flow becoming emulsive. The study of the coupling between the thermal phenomena inside the wall and the flow in the near-wall zone (bubble size at detachment, detachment frequency, wall void rate...) is of particular interest. The final aim of this work is to provide data on the various boiling models that can be used in CFD-type numerical simulation tools.The methodology followed in this work is based on the design of an experimental device allowing a direct visualisation of the boiling two-phase flow coupled with a characterisation of the temperature field of the heating element.As the thermal characterisation is to be carried out by Infrared thermography coupled with an inverse method (allowing the temperature measured outside the test section to be compared with the wetted temperature), the determination of the geometric characteristics of the heating element (in particular its thickness) is of prime importance. A detailed study based on both one- and two-dimensional approaches to wall thermics is therefore carried out. The originality of this approach lies in the consideration of inhomogeneous boundary conditions in time and space in order to simulate boiling at the plate. This work enabled the heating element to be dimensioned as representative of the industrial device (in terms of thermal inertia and surface state) and allowing the desired measurements to be made with a target accuracy.Visualisation in the visible range is a second important aspect of the work. This is done using a high-speed camera (25,000 fps), coupled with high-precision lenses and high-performance lighting, by observing the flow through transparent sapphire windows. Focus is placed on determining and optimising the instrumentation to be used to carry out precision measurements. It is shown that the use of a bi-telecentric lens associated with an ombroscopic setup will allow obtaining precise results. The potential distortion effects of near-wall images due to the optical index gradient resulting from the temperature gradient in the liquid is specifically studied. This effort shows that on the one hand, this effect remains weak, given that in our conditions of under-saturated boiling the temperature gradient at the wall is of small magnitude, and that on the other hand, this effect could be corrected via the use of specially designed test patterns (with size comparable to the measured object).All these elements lead to the design of a thermal-hydraulic installation and the development of a test model on which preliminary tests can be carried out.; L'ébullition est un phénomène qui peut survenir en réacteur en conditions de fonctionnement normal ou incidentel. Dans ce dernier cas, elle peut conduire à l'apparition de la crise d'ébullition dont les mécanismes sont aujourd'hui encore non déterminés.Le sujet proposé concerne donc la caractérisation de l'ébullition nucléée depuis l'apparition de la première bulle (Onset of Nucleate Boiling - ONB) jusqu'à l'apparition de la crise d'ébullition dans un spectre de pression large allant de la pression atmosphérique jusqu'à la pression du circuit primaire (155 bar). En effet, il semble d'après la littérature qu'il existe un changement de structure d'écoulement autour de 70 bar, l'écoulement devenant émulsif. On s'intéressera en particulier à l'étude du couplage entre les phénomènes thermiques dans la paroi et l'écoulement dans la zone de proche paroi (taille de bulles au détachement, fréquence de détachement, taux de vide à la paroi...). Ce travail vise au final à fournir des données relatives aux différents modèles d'ébullition susceptibles d'être utilisés dans les outils de calculs numériques de type CFD.La méthodologie qu’on se propose d’utiliser dans ce travail repose sur la conception d’un dispositif expérimental qui devra permettre une visualisation directe de l’écoulement diphasique bouillant couplée à une caractérisation du champ de température de l’élément chauffant de ce dispositif.La caractérisation thermique devant se faire par thermographie Infra-Rouge couplée à une méthode inverse (permettant de remonter de la température mesurée à l’extérieur de la veine d’essais à la température mouillée), la détermination des caractéristiques géométriques de l’élément chauffant (en particulier son épaisseur) est donc primordiale. Une étude poussée s’appuyant sur une approche mono et bidimensionnelle de la thermique de paroi a donc été réalisée. L’originalité de celle-ci repose sur la prise en compte de conditions aux limites inhomogènes en temps et en espace afin de pouvoir simuler l’ébullition au droit de la plaque. Ce travail a permis de dimensionner l’élément chauffant afin qu’il soit d’un part représentatif du dispositif industriel (en terme d’inertie thermique et d’état de surface) et d’autre part qu’il permette les mesures souhaitées respectant une précision cible.La visualisation dans le domaine visible a constitué un second pan important ce de travail. Elle se fera grâce à une caméra rapide (25000 i/s) couplés à des objectifs de haute précision et un éclairage performant en observant l’écoulement à travers des hublots transparents en Saphir. On s’est en particulier attaché à déterminer et optimiser l’instrumentation devant être mise en œuvre afin de réaliser des mesures de précision. On a ainsi montré que l’utilisation d’un objectif bi-télécentrique associé à un montage en ombroscopie permettrait d’obtenir des résultats précis. Un focus particulier a été réalisé sur les effets de distorsion potentiel des images réalisées en proche paroi du fait du gradient d’indice optique résultant du gradient de température dans le liquide. Cette étude a montré que d’une part cet effet restait faible compte-tenu que dans nos conditions d’ébullition sous saturées le gradient de température à la paroi reste faible et que d’autre part cet effet pouvait se corriger via l’utilisation de mires spécialement conçues (taille comparable à l’objet mesuré).Tous ces éléments ont conduit au dimensionnement d’une installation thermohydraulique et à la mise au point d’une maquette d’essais sur laquelle des tests préliminaires ont pu être réalisés.

A large fraction of greenhouse gases (about 60%) released into the atmosphere are due to CO2 emissions from industrial processes and the burning of fossil fuels [1]. One of the strategies employed to reduce the emissions is rapping them securely in the subsurface [2-4]. Dissolution trapping, in particular, involves injection of CO2 into the subsurface where the supercritical CO2 (sCO2) dissolves in the aquifer brine and forms a CO2 enriched layer within solution. The interface between the high density CO2 rich brine on the top and the ambient low density aquifer water below results in destabilization of the aforementioned layer [2-4]. This leads to a gravitational instability which then causes a natural convection of CO2 rich brine to lower layers, thereby accelerating further dissolution of the sCO2 into the fresh brine.The study of Brouzet et al. shows that traditional continuume scale, Darcy law-governed, models underestimate the timescales of the convective dissolution’s dynamics, owing to local heterogeneity in the pore-scale flow, and that it may thus be necessary to take pore-scale fluctuations into account [5]. We present here a 2D experimental study using miscible analog fluids with a contrast in densities to understand the convective transport of the dissolved sCO2. The fluids and the granular media are refractive index matched, which renders the medium transparent and helps in accurate quantification of experimental findings at various Rayleigh (Ra) and Darcy numbers (Da). Darcy scale simulations are used to complement the two-dimensional experimental measurements and it was found that Darcy scale simulations underpredict the experimental findings by several orders of magnitude, which is consistent with the findings by Brouzet et al. We investigate convective dynamics for various values of the number by changing the density of fluids, the properties of the granular medium (permeability, size of the granular medium) which determines the size of the instability with respect to pore size. When that number is much smaller than 1, obvious causes for the failure of the continuum scale description can be excluded, yet discrepancies remain between the experimental results and the simulations.References:[1] Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, IPCC 2014.[2] Emami-Meybodi, H., Hassanzadeh, H., Green, C. P., & Ennis-King, J. (2015). Convective dissolution of CO2 in saline aquifers: Progress in modeling and experiments. International Journal of Greenhouse Gas Control, 40, 238-266.[3] Pau, G. S., Bell, J. B., Pruess, K., Almgren, A. S., Lijewski, M. J., & Zhang, K. (2010). High-resolution simulation and characterization of density-driven flow in CO2 storage in saline aquifers. Advances in Water Resources, 33(4), 443-455.[4] Meunier, P., & Nadal, F. (2018). From a steady plume to periodic puffs during confined carbon dioxide dissolution. Journal of Fluid Mechanics, 855, 1-27.[5] Brouzet, C., Méheust, Y., & Meunier, P. (2022). CO2 convective dissolution in a three-dimensional granular porous medium: An experimental study. Physical Review Fluids, 7(3), 033802.

Convective dissolution is a perennial trapping mechanism of carbon dioxide in geological formations saturated with an aqueous phase. This process, which couples dissolution of supercritical CO2, convection of the liquid containing the dissolvedCO2, and mixing of the latter within the liquid, has so far not been studied in two-dimensional porous media. In order to do so, two-dimensional (2D) porous micromodels (patterned Hele-Shaw cells) have been fabricated from UV-curable NOA63 glue. NOA63 is used instead of PDMS, which is permeable to CO2 and does not allow for a controlled no flux boundary condition at the walls. The novel fabrication protocol proposed here, based on the bonding of a patterned photo-lithographed NOA63 layer on a flat NOA63 base, shows good reproducibility regardless of the pattern’s typical size, and allows for easy filling of the cell despite the small value of the gap. A pressure chamber allows pressurizing the CO2 and outside of the flow cell up to 10 bars. Experiments were performed in 11 different porous media geometries. As expected, a gravitational fingering instability is observed upon injection of gaseous carbon dioxide in the cell, resulting in the downwards migration of dissolved CO2 plumes through the 2D porous structure. The initial wavelength of the fingers is larger in the presence of a hexagonal lattice of pillars. This effect can be correctly predicted from the theory for the gravitational instability in a Hele-Shaw cell devoid of pillars, provided that the permeability of the hexagonal porous medium is considered in the theory instead of that of the Hele-Shaw cell. Fluctuations around the theoretical prediction observed in the data are mostly attributed to a hitherto unknown weak locking of the wavelength on the distance between closest pillars.

In the next step magnetic fusion devices like DEMO, a large fraction of power, more than 95%, needs to be dissipated before the plasma impacts the dedicated regions on the wall, the so-called divertor targets. The divertor detachment is one of the solutions envisaged to protect target tiles by keeping the radiation front detached from the target, thus radiating most of the heating power on a larger surface, keeping the heat flux deposit on the target at a manageable level, and reducing the sputtering and accumulation of impurities. The present work is devoted to investigating the strategies to control detachment in the tokamak single null divertor configuration, which is planned for the International Thermonuclear Experimental Reactor (ITER) experiment presently in construction at Cadarache. The detachment must be feedback controlled in real-time to guarantee the continuous production of energy by the reactor and a long life of the plasma-facing components. To develop reliable feedback control methods, the first step is to use good control variables and actuators. With the help of edge numerical simulations and reduced models, this thesis aims to find the more appropriate control variable that will be used in a feedback scheme to adjust the actuators available in fusion reactors. 2D numerical simulations are obtained using SOLEDGE3X-EIRENE transport code with input parameters specially defined to reproduce the experimental scenarios in tokamaks, and more precisely in the TCV tokamak located in Switzerland and the WEST tokamak located in Cadarache, France. Profile feedback control has been developed as an interpretative tool to extract edge plasma transport coefficients from experimental data. Simulation results are compared with available experimental measurements and show good agreement. Modeling density scan is performed via adjusting deuterium gas puffing or impurity seeding to achieve different levels of detachment.The impacts of wall geometry (like divertor closure), impurities, and input power on detachment have been discussed. Simulation results show that the case with a more closed divertor has better performance in trapping the neutral particles and has higher neutral pressure near the target. Increasing the neutral pressure or impurity concentration in the divertor leads to greater power dissipation, thus facilitating the access to the detached regime. Both considering more closed divertor configurations and increasing nitrogen seeding will induce the divertor detachment at lower upstream separatrix density, up to 24% in the cases we have considered. The threshold values of variables like target heat flux, and radiation front height in the divertor for detachment are found to be not significantly affected by the change of divertor closure and impurities concentration. Higher input power can significantly increase the detachment threshold in upstream plasma density, divertor neutral pressure, target particle flux, etc. The dependence of numerical results from few physical quantities is also analyzed with the help of reduced models to find the observables that mostly affect the detachment process. Finally, the relative advantage of the different observables regarding detachment controllability is evaluated. A new criterion for detachment is proposed, which provides a good indicator of the optimal detachment condition in all the analyzed plasma scenarios (the situations with different levels of divertor closure, impurities concentration, and input power). Several strategies for controlling the divertor detachment have been designed based on good control variables and compared with existing control methods. The robustness of a control strategy for WEST tokamak has been investigated by numerical simulations, and it will be tested in future experiments.; Dans les machines de fusion magnétique de prochaine génération, comme DEMO, une grande partie de la puissance injectée et produite par les réactions de fusion doit être dissipée avant que le plasma n'entre en contact avec la paroi au niveau de composants appelés « cibles » dans le divertor. Le détachement du plasma est l'une des solutions envisagées pour protéger les tuiles des cibles en maintenant le front de rayonnement à une certaine distance de la cible. Ceci permet d’évacuer sous forme de rayonnement la majeure partie de la puissance de chauffage sur une plus grande surface et de maintenir le flux de chaleur à des niveaux compatibles avec leur design. Ce travail de thèse se concentre sur l'étude des caractéristiques du plasma détaché dans différentes configurations divertor de deux tokamaks : « TCV », situé en Suisse, et « WEST » situé à Cadarache, en France. Le détachement doit être contrôlé en temps réel pour garantir à la fois une production d’énergie continue par le réacteur, et une durée de vie longue pour composants face au plasma. Pour développer des méthodes de contrôle de rétroaction fiables, la première étape consiste à sélectionner les bonnes variables de contrôle et les bons actionneurs (injection de gaz, d’impureté, etc.). A l’aide de simulations numériques de bord et de modèles réduits, cette thèse vise à trouver la variable de contrôle la plus appropriée qui sera utilisée dans un système de rétroaction pour contrôler les actionneurs disponibles dans les réacteurs de fusion. Des simulations numériques 2D sont obtenues en utilisant le code de transport SOLEDGE3X-EIRENE avec des paramètres d'entrée spécialement définis pour reproduire les scénarios expérimentaux. Pour obtenir ces paramètres, une méthode de contrôle par rétroaction sur le profil du plasma a été développée comme un outil permettant de déterminer les coefficients de transport radial du plasma de bord à partir des données expérimentales. Les résultats des simulations sont comparés aux mesures expérimentales disponibles et montrent un bon accord. Les impacts de la forme géométrique de la paroi (comme le degré de fermeture du divertor), de la quantité d’impuretés injectées, et de la puissance d'entrée sont discutés. Les résultats de la simulation montrent que le cas avec un divertor plus fermé montre de meilleures performances pour piéger les particules neutres, avec une pression neutre plus élevée près de la cible. L’augmentation à la fois de la pression de neutres ou de la concentration en impuretés dans le divertor entraîne une plus grande dissipation de puissance, et facilite l'accès au régime détaché. Un divertor plus fermé, ainsi que l’injection d’azote peuvent permettre au plasma de se détacher avec une densité plasma dans la région centrale plus faible (jusqu'à 24%) dans les cas que nous avons examinés. Les valeurs seuils à partir desquelles le plasma détache pour les paramètres tels que le flux de chaleur sur la paroi et la hauteur du front de rayonnement ne sont pas significativement affectés par la modification de la fermeture du baffle ou la concentration d'impuretés. Une puissance d'entrée plus élevée peut augmenter de manière significative le seuil de détachement en termes de densité du plasma en amont, de la pression neutre du divertor, du flux de particules cibles, etc. La dépendance des résultats numériques pour quelques quantités physiques est également analysée à l’aide de modèles réduits pour trouver les observables qui affectent principalement le processus de détachement.Enfin, les avantages relatifs des différentes observables au regard la contrôlabilité du détachement sont analysés. Un nouveau critère pour le détachement est proposé, fournissant un bon indicateur de l’état de détachement optimal dans tous les scénarios de plasma analysés (les situations avec différents degré de fermeture du divertor, concentration d’impuretés, et puissance d’entrée). Plusieurs stratégies de contrôle du détachement du divertor sont conçues en fonction des variables de contrôle choisies et comparées aux méthodes de contrôle existantes. La robustesse d’une stratégie de contrôle du tokamak WEST est étudiée par des simulations numériques, et sera testée dans de futures expériences.

International audience; Recent results, using different experimental methods, evidenced a shortening of UO distance in UO2 PDF, while the average unit cell is expanding with increasing temperature. This shortened UO distance induces a change in the uranium coordination polyhedron, going from cubic in the regular Fm-3m to octahedral with shorter UO distances. Different interpretations were given for this change of coordination polyhedron, amongst which static multi-domains, dynamic domains or coexisting phases. Whatever its origin, this change in uranium coordination polyhedron might have consequences on UO2 behavior at low and high temperature. At high temperature, the existence of a local octahedral environment could be a reason why UO2 can incorporate so many fission products, because this environment is more favorable for fission products. At low temperature, the PDF uranium coordination polyhedron is very close to the one in Pa-3 symmetry, which is the symmetry of UO2 under Neel transition.

Within the framework of Best-Estimate-Plus-Uncertainty approaches, the assessment of model parameter uncertainties, associated with numerical simulators, is a key element in safety analysis. The results (or outputs) of the simulation should be compared and validated against experimental values, when such data is available. This validation step is required to ensure a reliable use of the simulator for modeling and prediction. In addition, it must take into account both model and experimental uncertainties (measurement uncertainties). This work aims to define quantitative criteria to carry out this validation for multivariate outputs, while taking into account the different sources of uncertainty. For this purpose, different statistical indicators, based on likelihood or statistical depths, are investigated and extended to the multidimensional case. First, the properties of the criteria are studied, either analytically or by simulation, for some specific cases (Gaussian distribution for experimental uncertainties, identical distributions of experiments and simulations, particular discrepancies). Then, some natural extensions to multivariate outputs are proposed, with guidelines for practical use depending on the objectives of the validation (strict/hard or average validation). From this, transformed criteria are proposed to make them more comparable and less sensitive to the dimension of the output. It is shown that these transformations allow for a fairer and more relevant comparison and interpretation of the different criteria. Finally, these criteria are applied to a code dedicated to nuclear material behaviour simulation. The need to reduce the uncertainty of the model parameters is thus highlighted, as well as the outputs on which to focus.

International audience; ABSTRACT. Implementing an instrumentation that allows to record real seismic events is an essential complement to numerical simulation approaches of site effect assessment. Such " empirical " approaches allow for example to apply the Standard Spectral Ratio (SSR) and to evaluate the high frequencies parameter " kappa " (κ). While these approaches seem readily available in areas of high seismicity (Japan, Greece ...), they are more difficult to implement in areas of low seismicity, as in metropolitan France. The first objective of this study is to provide some elements to evaluate the usefulness of such site instrumentation, especially in the aim of site effect evaluation. A second objective is to give some recommendations for the implementation of instrumentation. This study presents the databases acquired over a period of about two years at two sites in SouthEastern France. A very large number of seismic recordings was acquired. To obtain such results, we recommend the use of velocimeters recording in continuous mode. SSR method is robust even when a limited number of events are available. It is also possible to assess the κ 0 parameter, even though the conventional method requires a good signal to noise ratio up to high frequencies. The displacement method is very encouraging to evaluate κ 0 in less seismic zones. MOTS-CLÉS : Instrumentation, sismicité faible à modérée, effets de site; La mise en place d'une instrumentation in situ permettant l'enregistrement d'événements sismiques est un complément essentiel aux approches d'évaluation des effets de site par simulation numérique. Une telle approche « empirique » permet, par exemple, l'emploi de la méthode site/référence (SSR) pour évaluer la fonction de transfert spécifique au site, ainsi que la mesure à haute fréquence du paramètre « kappa » (κ). Alors que ces approches semblent déjà utilisables dans les régions fortement sismiques (Japon, Grèce…), elles se heurtent souvent à un a priori défavorable pour les régions de sismicité faible à modérée (p. ex. la France métropolitaine) en raison du temps supposé qu'il faudrait pour obtenir un nombre suffisant d'enregistrements de qualité. Le premier objectif de cette étude est d'apporter quelques éléments permettant d'évaluer l'utilité de l'instrumentation des sites, en particulier pour ce qui concerne l'évaluation des effets de site. Un deuxième objectif est de fournir quelques recommandations pour leur mise en place. Cette étude présente les bases de données acquises sur une durée d'environ deux ans sur deux sites du Sud-Est de la France. Un nombre très important d'enregistrements sismiques a été acquis. Pour obtenir de tels résultats, nous recommandons l'emploi de vélocimètres enregistrant en continu. La méthode SSR se montre assez robuste même lorsque l'on sélectionne qu'un nombre limité d'événements. Il est également possible d'évaluer le paramètre κ 0 , alors même que la méthode classique nécessite un bon rapport signal sur bruit jusqu'à haute fréquence. La méthode en déplacement est très encourageante pour approcher la mesure de κ 0 dans les zones les moins sismiques.

The acoustic detection of defects or leaks inside a cylindrical shell containing a fluid is of prime importance in the industry, particularly in the nuclear field. This paper examines the beamforming technique which is used to detect and locate the presence of an acoustic monopole inside a cylindrical elastic shell by measuring the external shell vibrations. In order to study the effect of fluid–structure interactions and the distance of the source from the array of sensors, a vibro-acoustic model of the fluid-loaded shell is first considered for numerical experiments. The beamforming technique is then applied to radial velocities of the shell calculated with the model. Different parameters such as the distance between sensors, the radial position of the source, the damping loss factor of the shell, or of the fluid, and modifications of fluid properties can be considered without difficulty. Analysis of these different results highlight how the behaviour of the fluid-loaded shell influences the detection. Finally, a test in a water-filled steel pipe is achieved for confirming experimentally the interest of the presented approach.

International audience; This work addresses the use of plastic scintillators as an alternative to 3He neutron detectors for radioactive waste drum characterization. The time response of scintillators is three orders of magnitude faster than that of gas proportional counters, and they offer similar neutron detection efficiency at lower cost. However, they are sensitive to gamma rays and the neutron-gamma pulse shape discrimination (PSD) is not possible with standard PVT scintillators. The proposed approach uses the neutron and gamma times of flight in triple coincidences recorded with 252Cf, AmBe, and 60Co sources. A 2-D histogram of time delays between the second and first detected pulses, on the x -axis, and between the third and second pulses, on the y -axis, evidences a specific region of interest for spontaneous fission coincidences. MCNPX-PoliMi simulations are performed, which are in good agreement with previous experiments and allow for investigating the types of coincidences ( γγγ , nnn, γ nn, γγn ) and optimizing the rejection of neutron and gamma scattering crosstalk. The method was also experimentally tested with a 118-L mock-up drum filled with a metallic or an organic matrix, showing a correct estimation of the net fission signal up to an “alpha ratio” of 12 between fission and ( α ,n) neutron emissions. Matrix and localization effects were also measured, showing a sensitivity (useful signal per gram of 240Pueq) for a homogeneous distribution of plutonium in the iron matrix about three times larger than in the wood matrix, due to neutron slowing-down in the latter. This difference can be taken into account by the prior knowledge of the matrix characteristics or by neutron and gamma attenuation measurements. In contrast, in case of a point source with an unknown position in the matrix, the relative localization uncertainty is 26% for the metallic drum, and 41% for the organic one.

International audience; Spray drying produces spherical agglomerates to obtain powders with good flowability. Three alumina powders of different size and/or specific surface area were spray-dried without any binder using either two-fluid or ultrasonic nozzle technology. Compared to the two-fluid nozzle, the ultrasonic one shows a better ability to form agglomerates from a coarse powder and increases the size of the agglomerates when spray drying is carried out under the same operating conditions. The redispersion of spray-dried agglomerates in solution was analysed by applying variable ultrasound durations during size measurement by wet laser diffraction. The initial state of the powder plays a predominant role in relation to the spray drying operating conditions both in terms of size characteristics and redispersion properties. Regardless of the alumina powders and spray drying conditions, a similar trend in deagglomeration behaviour was observed with first-order kinetics and a characteristic kinetic time depending on the initial powder properties.

An important part of the electricity produced in France is supplied by nuclearreactors. Research in this field is therefore an important industrial challenge. Thetopic of this PhD thesis fits into this context and is the result of a collaborationbetween the CEA and the Paul Sabatier University (Toulouse, France).The subject of this thesis is the study of an industrial code modeling a severenuclear accident, namely the fuel-coolant interaction, and more precisely thetreatment of uncertainties on the modeling parameters of this code. To achievethis goal, various statistical tools of the uncertainty quantification framework wereused into a global methodology consisting in five main steps:1. Exploration of the input space using sampling methods. Knowingthe variation domain of the input variables, the input space was randomlysampled. A code run was then performed on each sampled point of theexperimental design. The obtained sample of inputs/outputs constitutedthe learning sample. The goal is here to explore with this sample the inputspace and get as much information as possible about the behavior of thecode output. Only having a limited budget for simulations, we used a nearrandom sampling method, named Latin hypercube sampling (LHS) insteadof a classic random sampling method. This method ensures a better spacecoverage of the marginal distributions while maintaining some properties ofclassical Monte-Carlo sampling. Asymptotic convergence of statistics used inour methodology under LHS were also discussed. In particular, a theoreticalwork on the asymptotic convergence of Z-estimators under LHS has beencarried out.2. Sensitivity analysis of code failures. Running the simulations of theexperimental design defined in step 1 revealed that a significant portion ofthe code executions did not converge. We therefore analyzed these codefailures to understand which of the inputs have the most influence on them.3. Selection of the inputs from sensitivity analysis of the outputs.From the learning sample, the main influential inputs were identified usingtwo sensitivity analysis tools. The first one is based on the estimation offirst-order Sobol’ indices. They quantify the contribution of each input,alone (independently from the other inputs), to the output variance. Inpractice, a rank based estimator is used to compute the indices. These indicesidentified the impact of the main effects of the variables. In addition, to detecta broader dependence between input and output (including interactions),another sensitivity tool relying on a kernel based dependency measure, namelythe Hilbert Schmidt Independence Criterion, was used. From the analysis ofSobol’ inidices and HSIC measures, a pool of significantly influential inputswere deduced.4. Emulation of the outputs by regression models (called metamodels).From the learning sample, regression models were built to fit the simulatoroutputs. These models are intended to replace the initial code. For this,Gaussian Process (GP) regression method was used. This method hasseveral advantages compared to other regression methods notably its abilityto propose both a prediction and a quantification of the uncertainty on itspredictions.5. Use of metamodels (built at Step 4) to perform a more completesensitivity analysis. A quantitative sensitivity analysis were thenperformed using the regression models instead of computer code. Thesemetamodels, by their negligible evaluation time, allow an intensive explorationof the input space. This is made possible by the negligible evaluationtime of the metamodels. Hence, they were used to estimate both first-orderand total Sobol’ indices. The latter quantify the effect of variables alone andin interaction with the other inputs. The information provided is thereforemuch more complete than that resulting from the previous estimation of thefirst-order indices alone performed in Step 3.; Une grande part de l’électricité consommée en France étant produite par desréacteurs nucléaires, la recherche dans ce domaine constitue un enjeu industrielimportant. Ce travail de thèse s’inscrit dans ce contexte et est le résultat d’unecollaboration entre le CEA et l’Université Paul Sabatier (Toulouse, France).Cette thèse a pour objet l’étude d’un code industriel modélisant un accidentnucléaire grave, l’interaction corium-eau, et plus précisément le traitement desincertitudes sur les paramètres de modélisation de ce code. Pour étudier ce code,une méthodologie basée sur différents outils statistiques avancés a été proposée.Elle se décompose en cinq grandes étapes :1. Échantillonnage de l’espace des entrées. Connaissant le domaine devariation des paramètres d’entrée, l’espace de ces paramètres a été échantillonné.Une exécution du code a été ensuite effectuée sur chaque pointdu plan d’expérience ainsi construit. Le jeu de données entrées/sortiesobtenu constitue l’échantillon d’apprentissage. Le but ici est d’exploreravec cet échantillon l’espace des entrées et d’obtenir autant d’informationsque possible sur le comportement des sorties du code. Ne disposant qued’un budget limité pour les simulations, nous avons utilisé une méthoded’échantillonnage quasi aléatoire, appelée Latin Hypercube Sampling (LHS),au lieu d’une méthode Monte Carlo classique. Cette méthode assure unemeilleure couverture spatiale des distributions marginales tout en conservantcertaines propriétés de l’échantillonnage purement aléatoire. La convergenceasymptotique des statistiques utilisées dans notre méthodologie sousLHS a également été discutée, avec notamment un travail théorique sur laconvergence asymptotique de la classe des Z-estimateurs avec un LHS.2. Analyse de sensibilité des échecs de code. Au cours de l’explorationde l’espace d’entrée, nous avons constaté qu’une proportion importante desexécutions du code ne convergeait pas. Nous avons donc analysé ces échecsde code pour voir quelles entrées ont le plus d’influence sur eux, ce qui nousa permis de mieux comprendre le fonctionnement du code.3. Sélection des entrées à partir de l’analyse de sensibilité des sorties.À partir du jeu de données d’apprentissage, un tri a été effectué en utilisantdeux outils d’analyse de sensibilité. Le premier outil de sélection est basé surl’estimation des indices de Sobol’ de premier ordre. Ces indices quantifientla contribution de chaque entrée, seule (indépendamment des autres entrées),à la variance de la sortie. En pratique, un estimateur basé sur le rang a étéutilisé pour les estimer. Pour détecter plus de formes de dépendance entreles entrées et la sortie (notamment les interactions entre entrées), un autreoutil d’analyse sensibilité reposant sur une mesure de dépendance basée surun noyau, à savoir le critère d’indépendance de Hilbert Schmidt (HSIC), aété utilisé. Ces deux critères de sélection ont permit d’identifier un ensembled’entrées ayant une influence considérée comme significative sur les sortiesdu code.4. Approximation des sorties par des modèles de régression (appelésmétamodèles). A partir de l’échantillon d’apprentissage, des métamodèlesont été construits pour ajuster les sorties du simulateur. Ces modèles sontdestinés à remplacer le code initial dans les études de sensibilité. Pourcela, des modèles de type processus gaussien ont été utilisés. Ces modèlesprésentent plusieurs avantages par rapport aux autres modèles de régression,comme la possibilité d’évaluer l’incertitude des prédictions effectuées.5. Utilisation des modèles de régression (construits à l’étape 4) poureffectuer une analyse de sensibilité plus précise. Une analyse de sensibilitéa ensuite été réalisée en utilisant les modèles de régression construitsprécédemment. Ces métamodèles, par leur temps d’évaluation négligeable,permettent une exploration approfondie de l’espace d’entrée. Par conséquent,ils ont pu être utilisés pour estimer les indices de Sobol’ de premier ordre etles indices totaux. Ces derniers quantifient l’effet total (seul et en interaction)de chaque entrée sur chaque sortie. Les informations obtenues sont doncbeaucoup plus complètes que celles fournies par l’estimation des seuls indicesde premier ordre effectuée à l’étape 3.

Previous experiments on cubic UO2 have suggested that the temperature dependences of the nearest-neighbour U–O and U–U distances are different. We have acquired total-scattering neutron diffraction patterns out to Q = 23.5 Å−1 for 50 < T < 1023 K and produced via Fourier transform a pair-distribution function P D F ( r ) . The P D F ( r ) shows quite clearly that r ( U–O ) , defined by the maximum of the U–O peak in the P D F ( r ) , does in fact decrease with increasing temperature, whereas r ( U–U ) follows the lattice expansion as expected. We also observe that the r ( U–O ) contraction accelerates continuously above T ≈ 400 K, consistent with earlier experiments by others. Furthermore, by analysing the eigenvectors of the phonon modes, we show that the Δ 5 ( T O 1 ) phonon tends to separate the eight equivalent U–O distances into six shorter and two longer distances, where the longer pair contribute to a high-r tail observed in the U–O distance distribution becoming increasingly anisotropic at higher T. These results have significance for a wide range of materials in which heavy and light atoms are combined in a simple atomic structure.

International audience; Emission of colloids from treatment wetlands solid deposit (SD) may affect the speciation, toxicity and transfer of organic and inorganic pollutants, thereby controlling the associated environmental risks. This study investigated the nature of the colloids and their influence on the leaching of major and trace elements over a wide range of pH between 2 and 12. Results evidenced the mobilization of high quantity of organic carbon leached from the SD and Al-Fe colloidal minerals embedded by large organic molecules. Based on their varying affinity for the different colloidal phases, three groups of chemical species were established: (i) elements with low affinity for colloidal phases, mainly released in truly dissolved fraction (30 kDa), preferentially associated to organic colloids (Li, Ca, Mg, Sr, Ba, Mn, Zn) or to mineral colloids (Pb, Cr, Sc, Y); and (iii) element with high affinity for large and truly dissolved organic molecules (Cu, Co, Cd, Ni). Differences observed in the partitioning of elements can be attributed to variations in the pH/Eh of the solutions, the nature of the colloidal phases, and the ionic properties of the considered elements. Small pH variations were found to modify drastically the speciation of the elements solubilized with potential effect on their (bio)availability. Results obtained on this study can be used to enhance wastewater treatment through constructed wetland process and improve recovery technics applied on solid deposits (agriculture, compounds extraction…).

International audience; Tungsten is the privileged option as plasma-facing material for the divertor region of ITER and DEMO tokamaks. Under repeated high thermal fluxes (10 MW/m² in steady state and 20 MW/m² in quasi-steady state), Plasma-Facing Units (PFU) can be damaged through thermal fatigue phenomena, which are enhanced by in-service tungsten recrystallization. Slowing down tungsten recrystallization is thus a way to optimize PFU lifetime. Tungsten-recrystallization kinetics is known to depend upon the initial microstructure. In the present paper, a mean-field model is developed to highlight microstructural effects upon tungsten restoration mechanisms at high temperature. The effects of grain size, initial recrystallized fraction, initial dislocation density, i.e. stored energy, and dislocation density distribution have been investigated. It is clearly quantified how a higher stored energy can lead to a drastic increase of the recrystallization rate. It is also shown how a prior recovery stage helps to slow down recrystallization. Finally, a conclusion is brought upon the ability of such a mean field model to optimize the tungsten initial microstructure for fusion engineering.

International audience; Background: The N14(n,p)C14 reaction is of interest in neutron capture therapy, where nitrogen-related dose is the main component due to low-energy neutrons, and in astrophysics, where N14 acts as a neutron poison in the s process. Several discrepancies remain between the existing data obtained in partial energy ranges: thermal energy, keV region, and resonance region.Purpose: We aim to measure the N14(n,p)C14 cross section from thermal to the resonance region in a single measurement for the first time, including characterization of the first resonances, and provide calculations of Maxwellian averaged cross sections (MACS).Method: We apply the time-of-flight technique at Experimental Area 2 (EAR-2) of the neutron time-of-flight (n_TOF) facility at CERN. B10(n,α)Li7 and U235(n,f) reactions are used as references. Two detection systems are run simultaneously, one on beam and another off beam. Resonances are described with the R-matrix code sammy.Results: The cross section was measured from subthermal energy to 800 keV, resolving the first two resonances (at 492.7 and 644 keV). A thermal cross section was obtained (1.809±0.045 b) that is lower than the two most recent measurements by slightly more than one standard deviation, but in line with the ENDF/B-VIII.0 and JEFF-3.3 evaluations. A 1/v energy dependence of the cross section was confirmed up to tens of keV neutron energy. The low energy tail of the first resonance at 492.7 keV is lower than suggested by evaluated values, while the overall resonance strength agrees with evaluations.Conclusions: Our measurement has allowed determination of the N14(n,p) cross section over a wide energy range for the first time. We have obtained cross sections with high accuracy (2.5%) from subthermal energy to 800 keV and used these data to calculate the MACS for kT=5 to kT=100 keV.

International audience

International audience; We present a new Gogny-type finite-range effective interaction including a third Gaussian in the central term. Based on simple arguments valid for an arbitrary radial form factor, the three ranges are fixed from physical grounds, relating them to one-boson exchange interactions. Moreover, some parameters of the longest range are fixed through the G-matrix elements of the One Pion Exchange Potential. On top of giving a fairly good description of atomic nuclei properties, the resulting interaction leads to a remarkable improvement of some infinite matter properties that are relevant for astrophysical calculations.

International audience; IntroductionLes lomboradiculalgies postopératoires (LRPO), comme d’autres types de douleur, sont considérées comme une expérience sensorielle et émotionnelle désagréable influencée par des facteurs biologiques, psychologiques et sociaux [1, 2, 3]. Malgré l’innovation constante de la technologie numérique et de l’intelligence artificielle, la douleur est toujours évaluée par des « outils de référence » tels que l’échelle visuelle analogique de la douleur (EVA) [4]. La principale limite de cet outil est qu’il ne tient pas compte des nombreuses dimensions de la douleur, comme les incapacités fonctionnelles, la détresse psychologique ou de la qualité de vie [3]. En pratique, les évaluations de ces dimensions sont considérées indépendamment, même si l’expérience clinique souligne l’imbrication de ces dimensions les unes avec les autres. Notre objectif était de développer un nouvel indice multidimensionnel de réponse clinique (MCRI), incluant l’intensité de la douleur, la capacité fonctionnelle, l’anxiété/dépression, la qualité de vie et la cartographie de la douleur, l’objectif étant une évaluation multidimensionnelle des douleurs.Matériel et méthodesC’est une étude observationnelle, prospective, multicentrique (PREDIBACK) incluant 200 patients avec des LRPO avec un suivi de 12 mois. L’intensité de la douleur, la surface de la douleur corporelle (en cm2 via une interface tactile), l’incapacité fonctionnelle (Oswestry Disability Index), l’anxiété/dépression (Hospital Anxiety and Depression scale), la qualité de vie (EuroQol-5 dimensions). Ces résultats ont été combinés à l’aide d’une analyse factorielle pour créer l’indice composite MCRI. Le MCRI a été comparé aux autres mesures par des tests de corrélations. La capacité du MCRI à détecter la satisfaction du patient après une thérapie a également été analysée.RésultatsLe MCRI représentait de façon plus pertinente toutes les dimensions de la douleur comparée aux autres scores classiques (corrélation avec les autres critères>0,603). De plus, comparé aux autres critères, le MCRI a montré la plus grande sensibilité (77,4 %) et spécificité (79,8 %) pour détecter de la satisfaction du patient par rapport aux autres scores.Discussion/conclusionCe nouvel indice composite MCRI pourrait contribuer à affiner l’évaluation de l’état de santé général du patient douloureux chronique sur la base de paramètres objectifs. Ce nouvel indice MCRI doit être éprouvé en milieu clinique afin de déterminer l’efficacité des thérapies.

International audience; Background. The mechanical response and damage mechanisms of heterogeneous cementitious materials result from the mechanical properties of their constituents at the mesoscale. Objective. In this study, a finite element based Digital Volume Correlation (FE-DVC) method with mechanical regularization tailored for heterogeneous materials was applied to study crack propagation in mortar. Method. A realistic 3D mesh was built consistently with the actual microstructure based on high resolution X-ray tomography. A notched mesobeam was subjected to in situ three-point flexure. Results. The introduction of mechanical contrast in DVC analyses allowed the strain heterogeneities related to the underlying microstructure to be described and inelastic strain localization due to cracks to be captured. Furthermore, the relaxation of mechanical regularization in damaged elements improved the trustworthiness of displacement measurements. Conclusion. Estimations of crack opening displacements and characterization of the crack morphology were carried out, providing insights into the damage processes at the mesoscale.

In the context of its R&D actions by non-destructive testing, the CEA's nuclear measurement laboratory uses a high energy tomograph. The photon source of this system is an electron linear accelerator that can reach energies of 21 MeV and provide high dose rates (60 Gy/min at 1 m). The detector is composed of a scintillator imaged by a low noise camera. Until now, this tomograph allows the laboratory to make radiographies and tomographies of massive waste packages (up to 5 tons and 140 cm in diameter) with millimeter spatial resolution. For new needs, such as the control of the cracking of the concrete coating of waste packages or for the inspection of metal parts resulting from additive manufacturing, the laboratory wanted to have a high resolution version of this device. The objective of this thesis is therefore to update the tomograph in order to obtain a spatial resolution of a hundred micrometers for objects with equivalent steel thicknesses between 50 and 200 mm.In this perspective, a first part of the thesis consists of a hardware update of the detector taking into account the different sources of blur involved in the measurement chain. To do this, two technological solutions, inspired by the tomographs studied in the state of the art, have been explored. First, the characterization and purchase of a flat panel detector providing the best resolution on the market (about 200 µm for a 100 µm pixel). The second solution explored was was the optimization of the detector initially used by the laboratory in order to adapt its elements (camera, lens, and scintillator) to the high resolutionThe performances of these detectors have been evaluated on 9 MeV tomographies of real objects. The flat panel detector allows to obtain images with a spatial resolution of about 310 µm for metallic objects of 50 mm diameter, but provides very low contrast images for more absorbent objects. The updated detector provides a slightly lower spatial resolution, of the order of 350 µm, but is better adapted to the tomography of absorbent parts, up to 200 mm of steel.In order to improve the spatial resolution of the measurement chain, the rest of the thesis is focused on the test and development of numerical methods to post-process the acquired projections to improve their quality. To do this, different deconvolution algorithms classically used in imaging and described in the literature have been tested. At the end of this work, three algorithms were selected. The different tests show good performances of deblurring at the expense of the creation of artifacts and a certain amplification of the noise.To surpass the performances of these three algorithms, the last part of this thesis focuses on a new approach of post-processing: blur deconvolution by convolutional neural network. For this purpose, a training image dataset has been generated by simulation in order to be as representative as possible of real data. A convolutional neural network was then adapted, trained and evaluated. Each parameter of the network was then specially optimized. Finally, this network was validated on 9 MeV tomographies in order to evaluate the final performances obtained, but also to understand the limitations of this type of approach.Finally, spatial resolutions between 150 and 250 µm were obtained for metallic objects of 50 mm diameter, while allowing to distinguish holes of 1 mm diameter and cracks of 0.3 mm width in steel objects of 200 mm diameter. These performances correspond to the needs of the laboratory and represent a significant improvement compared to existing industrial high-energy tomographs in Europe or in the United States.; Dans le cadre de ses actions de R&D par contrôle non destructif, le laboratoire de mesures nucléaires du CEA dispose d'un tomographe haute énergie. La source photonique de ce système est un accélérateur linéaire d'électrons pouvant atteindre des énergies de 21 MeV et fournir des débits de dose élevés (60 Gy/min à 1 m). Le détecteur est quant à lui composé d'un scintillateur imagé par une caméra bas-bruit. Jusqu’à présent, ce tomographe permet au laboratoire de réaliser des radiographies et des tomographies sur des colis de déchets massifs (jusqu’à 5 tonnes et 140 cm de diamètre) avec une résolution spatiale millimétrique. Pour de nouveaux besoins, comme le contrôle de la fissuration de l’enrobage béton des colis de déchets ou pour l'inspection de pièces métalliques issues de la fabrication additive, le laboratoire a souhaité disposer d'une version haute résolution de ce dispositif. L'objectif de cette thèse est donc de mettre à jour le tomographe afin d’obtenir une résolution spatiale d'une centaine de micromètres pour des objets ayant des épaisseurs équivalentes de 50 à 200 mm d'acier.Dans cette optique, une première partie de la thèse consiste en une mise à jour matérielle du détecteur en tenant compte des différentes sources de flous intervenant dans la chaîne de mesure. Pour ce faire, deux solutions technologiques, inspirées des tomographes étudiés dans l'état de l'art, ont été explorées. Tout d'abord, la caractérisation puis l’approvisionnement d’un détecteur à panneau plat (Flat Panel) offrant la meilleure résolution sur le marché (environ 200 µm pour un pixel de 100 µm). La deuxième solution explorée a été l’optimisation du détecteur à reprise d'image initialement utilisé par le laboratoire afin d'adapter ses éléments (caméra, objectif, et scintillateur) à la haute résolution.Les performances de ces détecteurs ont été évaluées sur des tomographies à 9 MeV d’objets réels. Le détecteur à panneau plat permet d'obtenir des images avec une résolution spatiale d'environ 310 µm pour des objets métalliques de 50 mm de diamètre, mais fournis des images très peu contrastées pour des objets plus absorbants. Le détecteur à reprise d'image mis à jour fournis une résolution spatiale légèrement inférieure, de l'ordre de 350 µm, mais est mieux adapté à la tomographie de pièces absorbantes, allant jusqu'à 200 mm d'acier.Dans le but d’améliorer la résolution spatiale de la chaîne de mesure, la suite de la thèse s'oriente sur le test et le développement de méthodes numériques afin de post-traiter les porjections acquises pour en améliorer la qualité. Pour ce faire, différents algorithmes de déconvolution classiquement utilisés en imagerie et décrits dans la littérature ont été testés. À l’issue de ce travail, trois algorithmes ont été sélectionnés. Les différents tests montrent de bonnes performances de défloutage au détriment de la création d’artefacts et d'une certaine amplification du bruit.Afin de surpasser les performances de ces trois algorithmes, la dernière partie de cette thèse s'intéresse à une approche de post-traitement novatrice : la déconvolution de flou par réseau de neurones convolutifs. Pour ce faire, un jeu de données d’images d’apprentissage a été généré par simulation dans le but d’être le plus représentatif possible de données réelles. Un réseau de neurones convolutif a ensuite été adapté, entraîné et évalué. Chaque paramètre du réseau a alors été spécialement optimisé. Enfin, ce réseau a pu être validé sur des tomographies à 9 MeV afin d’évaluer les performances finales obtenues, mais aussi comprendre les limitations de ce type d’approche.In fine, des résolutions spatiales comprise entre 150 et 250 µm ont été obtenues pour des objets métalliques de 50 mm de diamètre, tout en permettant de distinguer des perçages de 1 mm de diamètre et des fentes de 0,3 mm de large dans des objets en acier de 200 mm de diamètre.

The non-destructive characterization of large, concrete radioactive waste packages is a major challenge for criticality safety aspects and long-term management. On the one hand, the fissile mass present inside must be estimated precisely for their transport and storage in the safest possible way. On the other hand, a precise knowledge of their alpha activity makes it possible to send them towards the most appropriate final repository, which potentially implies significant cost savings. The 870 L concrete packages produced by the CEA and stored at the Cadarache center cannot be characterized by gamma spectrometry or passive and active neutron measurements because of insufficient sensitivity and very penalizing uncertainties associated with the high attenuation of neutrons and gamma rays in the waste or in the surrounding concrete layer. Therefore, Active Photon Interrogation (API) is being investigated to locate and quantify fissile material in these bulky concrete waste drums. This technique is based on the photofission phenomenon when actinides are interrogated with photons of energy higher than 6 MeV. Since all actinides have a similar photofission cross section, the method does not allow a direct estimate of the mass of fissile material (e.g. 235U, 239Pu) or the alpha activity (mainly due to plutonium isotopes). It is therefore necessary to use a differentiation method to evaluate the proportions of the actinides of interest. To do this, differences in photofission product yields between actinides are exploited via the measurement of delayed gamma-ray ratios with a high purity germanium detector. New experimental data concerning the production rates of photofission products of 235U, 238U and 239Pu have been measured by irradiating uranium and plutonium samples with a Bremsstrahlung photon beam produced with a linear electron accelerator located in the CINPHONIE irradiation cell of the IRESNE institute, at CEA Cadarache. A preliminary characterization of the Bremsstrahlung photon beam and of the photoneutron flux produced by the accelerator was carried out via photon and neutron activation measurements of reference materials. From the cumulative photofission yields measurements, the most effective photofission product pairs for actinide differentiation were then identified for different mixtures of 235U, 238U and 239Pu isotopes. A method for localizing actinides inside a mock-up of a concrete waste package was also developed and tested, based on both experimental data (differential attenuation of delayed gamma rays emitted by 138Cs) and Monte-Carlo simulations with the MCNP6 code. Localization is an essential step to correct for matrix effects, since the differential attenuation of delayed gamma rays affects actinide discrimination, but also to estimate the uncertainty on the fissile mass evaluated by API. Finally, detection limits ranging from a few grams to tens of grams depending on the waste characteristics and the position of actinides in an 870 L waste package were estimated by simulation with the experimentally validated MCNP model. Ways of improving these detection limits are being studied, in particular by reducing the active background or by measuring high-energy delayed gamma rays with NaI(Tl) scintillation detectors.; La caractérisation non-destructive des colis de déchets radioactifs bétonnés et volumineux est un défi important pour des questions de sûreté-criticité et de gestion sur le long terme. D’une part, la masse fissile présente à l’intérieur doit être estimée précisément pour leur transport et leur entreposage de la façon la plus sûre possible. D’autre part, une connaissance précise de leur activité alpha permet de les envoyer vers le mode de stockage définitif le plus approprié, ce qui implique potentiellement des économies conséquentes. Les colis bétonnés de 870 L produits par le CEA et entreposés au centre de Cadarache ne peuvent être caractérisés par spectrométrie gamma ou mesures neutroniques passives et actives en raison d’une sensibilité insuffisante et d’incertitudes très pénalisantes associées à une forte atténuation des neutrons et des rayonnements gamma dans les déchets ou le béton. Par conséquent, l’Interrogation Photonique Active (IPA) est étudiée pour localiser et quantifier la matière fissile dans ces fûts de déchets bétonnés volumineux. Cette technique est basée sur le phénomène de photofission lorsque des actinides sont interrogés avec des photons d’énergie supérieure à 6 MeV. Puisque tous les actinides ont une section efficace de photofission similaire, la méthode ne permet pas d’estimer directement la masse de matière fissile (par exemple 235U, 239Pu) ou l’activité alpha (principalement due aux isotopes du plutonium). Il faut donc recourir à une méthode de différenciation pour évaluer les proportions des actinides d’intérêt. Pour ce faire, les différences dans les taux de production des produits de photofission selon l’actinide sont exploitées via la mesure de ratios de pics de gamma retardés avec un détecteur germanium haute pureté. De nouvelles données expérimentales concernant les taux de production des produits de photofission de l’235U, l’238U et du 239Pu ont été mesurées en irradiant des échantillons d’uranium et de plutonium avec un faisceau de photons de Bremsstrahlung issu d’un accélérateur linéaire d’électrons situé dans la cellule d’irradiation CINPHONIE de l’institut IRESNE, au CEA de Cadarache. Une caractérisation préalable du faisceau de photons de Bremsstrahlung et du flux de photoneutrons produits par l’accélérateur a été menée via des mesures d’activation photonique et neutronique de matériaux de référence. À partir des mesures des taux de production cumulatifs, les couples de produits de photofission les plus efficaces pour différencier les actinides ont ensuite été identifiés pour différents mélanges d’isotopes 235U, 238U et 239Pu. Une méthode de localisation des actinides dans une maquette de colis de déchets en béton a aussi été développée et testée, basée à la fois sur des données expérimentales (atténuation différentielle des rayonnements gamma retardés émis par le 138Cs) et des simulations Monte-Carlo avec le code MCNP6. La localisation est une étape essentielle tout d’abord pour corriger les effets de matrice, puisque l’atténuation différentielle des rayonnements gamma retardés affecte la discrimination des actinides, mais également pour estimer l’incertitude sur la masse fissile évaluée par IPA. Enfin, des limites de détection de l’ordre de quelques grammes à dizaines de grammes selon les caractéristiques du colis et la position de la matière nucléaire dans le fût de 870 L ont été estimées par simulation avec le modèle MCNP validé expérimentalement. Des pistes d’amélioration de ces limites de détection sont étudiées, notamment via la réduction du bruit de fond actif ou la mesure des rayonnements gamma retardés de haute énergie avec des détecteurs à scintillation de type NaI(Tl).

The ICFDT6 Proceedings will be published in Journal of Instrumentation (JINST), a multidisciplinary, peer-reviewed and online-only journal created by the International School of Advanced Studies (SISSA) and IOP Publishing.; International audience

International audience; This paper aims to develop two criteria for the lowest and the highest (when exists) multipactor thresholds, that are independent of the amount of reflected power. Two geometries are analyzed herein: a coaxial transmission line of characteristic impedance 50 Ω, and a T-junction type component (called bridge hereafter) used in the IC antennas of WEST. The rest of the paper is organized as follows; first, the previous findings on the determination of the multipactor thresholds under standing wave regimes are discussed and then the geometries’ models, analyzed here, are presented. Later the methodology is explained, to finally detail the results.

Nuclear fission is a complex phenomenon during which a heavy nucleus splits into two lighter nuclei, referred as fission fragments. During this process, as fission leaves the fragments in an excited state, prompt particles (neutrons, photons and electrons) will be emitted shortly after scission. The characteristics of these prompt particles, e.g., spectral distribution, multiplicity, as well as the fragment yields and the isomeric ratios, are called fission observables. For nuclear-energy applications, these data are of primary interest as they characterize the number and the energy of particles produced by the fission reaction.First, multi-parameter measurements of such fission observables will be presented. Correlated data were acquired with the VESPA setup (VErsatile γ SPectrometer Array) at the European Commission Joint Research Center (EC-JRC) in Geel, using the spontaneous fission of Cf-252. VESPA is made of several LaBr3(Ce) scintillators, which are used to measure the prompt fission gamma-ray spectra (PFGS). A double Frisch-grid ionization chamber is mounted in the center of the setup to measure simultaneously the kinetic energy of both fission fragments (2E method). Thus, the PFGS have been obtained as a function of fission fragments characteristics, i.e., mass, total kinetic energy (TKE) and total excitation energy (TXE). These spectra are then unfolded to extract the mass-, TKE- and TXE- dependent average prompt fission γ-ray multiplicity and total energy. The study of short-lived isomers, with half-lives ranging from nanosecond up to a couple of microseconds, can also be performed with this setup and will be discussed.Second, improvements of the Monte-Carlo code FIFRELIN, developed by the CEA Cadarache, will be presented. FIFRELIN simulates the prompt-particles emission from fission fragments to predict fission observables. For this purpose, it includes several nuclear models defining the initial state of the fragments on the one hand, as e.g. the total excitation energy sharing between the fragments and their total angular momentum, and on the other hand the prompt de-excitation mechanism through an Hauser-Feshbach algorithm, such as nuclear level densities, gamma strength function, etc. These models, previously highly phenomenological, are now including microscopic ones, e.g., based on the Hartree-Fock-Bogoliubov approach. The implementation of such models into FIFRELIN was one of the main developments carried out in this work. Finally, the comparison of the results from FIFRELIN calculations with the experimental data from VESPA helped constraining several nuclear models and their parameters. The presented interplay of experiment and theory may help to understand better the fission process, eventually improving FIFRELIN’s predictability of fission observables and corresponding nuclear data.; La fission nucléaire est un phénomène complexe au cours duquel un noyau lourd se divise en deux noyaux plus légers, aussi appelés fragments de fission. Après la scission du noyau composé, les fragments sont laissés dans un état fortement excité et vont alors émettre un certain nombre de particules promptes (neutrons, photons et électrons) lors de leur désexcitation. Les caractéristiques de ces particules promptes (spectres, multiplicités…) ainsi que les rendements de fission et les rapports isomériques constituent des observables de fission. Dans le domaine de l’énergie nucléaire, ces données sont, au même titre que les sections efficaces de réaction, de première importance puisqu’elles caractérisent le nombre et l’énergie des particules issues de la réaction de fission.Dans un premier temps, des mesures multi-paramètres de ces observables de fission seront présentées. Ces mesures ont été effectuées à l’aide du dispositif expérimental VESPA (VErsatile γ SPectrometer Array) du European Commission Joint Research Center (EC-JRC) de Geel, pour la fission spontanée du Cf-252. Ce dispositif est formé d’un ensemble de scintillateurs LaBr3(Ce), permettant la mesure des spectres de γ prompts de fission (PFGS). Ces spectres sont ensuite déconvolués afin de déterminer la multiplicité moyenne et l’énergie totale des γ prompts de fission. Une double chambre d’ionisation à grilles de Frisch est également placée au centre du dispositif, et mesure simultanément l’énergie cinétique des deux fragments de fission (méthode 2E). Ainsi, les caractéristiques des γ prompts de fission ont pu être obtenues en fonction de la masse, de l’énergie cinétique totale, et de l’énergie d’excitation totale des fragments de fission. Le dispositif rend également possible l’analyse d’isomères de demi-vie courte (entre la nanoseconde et la microseconde), ce qui sera également discuté.Dans un second temps, les développements effectués sur le code Monte Carlo FIFRELIN, développé par le CEA Cadarache, seront présentés. Ce code modélise l'émission des particules promptes par les fragments, pour prédire les observables de fission. Pour cela, il intègre plusieurs modèles nucléaires qui définissent d’une part l’état initial des fragments de fission, comme par exemple la répartition de l’énergie d’excitation totale entre les deux fragments, et leur moment angulaire total, et d’autre part le mécanisme d’émission des particules promptes via les densités de niveaux nucléaires ou les fonctions de force gamma, effectué au travers d’un algorithme de type Hauser-Feshbach. Récemment, de nouveaux modèles basés sur une description microscopique des noyaux (via la méthode Hartree-Fock-Bogoliubov par exemple), concurrencent les modèles phénoménologiques établis. Une part importante des développements réalisés sur FIFRELIN a ainsi été l’implémentation de tels modèles au sein de ce code. La confrontation des résultats obtenus à l’issue des calculs de FIFRELIN aux données expérimentales de VESPA permet alors de contraindre divers modèles nucléaires ainsi que leurs paramètres. Cette interaction entre approche expérimentale et approche théorique peut nous permettre de mieux comprendre le processus de fission, et d’améliorer la prédiction par FIFRELIN des observables de fission, et des données nucléaires associées.

International audience

International audience

International audience

Commercial pressurized water reactors (PWR) typically rely on high electrical output power, greater than or around 1 GWe. With the emergence of newly identified uses such as power supply in remote area or heat production, small modular reactors (SMR) concepts have been designed with an electrical output power below 300 MWe. Although soluble boron is commonly used in PWR to control the reactivity, some SMR are designed without soluble boron in the primary circuit. The reactivity control in soluble boron-free (SBF) reactors relies mainly on the insertion or withdrawal of control rods. The introduction of burnable absorbers also contributes by design to the reactivity control. Soluble boron-free SMR can be justified regarding safety, economic and environmental advantages.Two-level deterministic computational schemes are generally used for PWR modelling. The first step consists in 2D detailed assembly calculations where refined energy and spatial meshes are considered to solve the neutron transport equation. In the standard computational scheme, a first depletion calculation is performed over a full range of burnup steps in an All-Rods-Out (ARO) configuration to compute isotopic concentrations. Branch calculations are then carried out at each burnup step by varying different parameters: moderator density, fuel temperature and rod insertion. The fine energy mesh cross sections are then collapsed to two-group cross sections and homogenized over the pin cell. This cross section library is then used at the core level to compute the 3D power distribution by solving the two-group diffusion equations. With SBF cores, the assumption of an ARO configuration for computing the fuel depletion can be questioned since the depletion history includes the presence of inserted control rods that modifies the neutron spectrum.The first objective of this thesis is to quantify the numerical biases introduced by the standard neutronic computational scheme in the modelling of SBF fuel assemblies, cluster and SMR core including a large quantity of burnable absorber (in the form of gadolinium oxide). The second and main objective is to develop a method able to simulate spectrum history effects over the entire core at reasonable computing cost.This work is carried out using the APOLLO3® deterministic code. The developed method called COHLISEE is inspired from previous work found in the scientific literature about spectrum history effects. COHLISEE is a hybrid method that consists in using both microscopic depletion and spectrum history indicator in 3D diffusion core calculation. The multigroup cross section libraries are generated by merging All Rods Out (ARO) and All Rods In (ARI) depleted libraries. The method has been verified and validated on fuel assemblies and 2x2 cluster 2D calculations. Reference results have been obtained with 2D transport calculations performed with controls rods inserted following specific sequences. COHLISEE results (reactivity, power, material balance) are compared to standard schemes using either macroscopic or microscopic depletion. The V&V process shows that macroscopic standard calculations are leading to massive numerical biases on reactivity for all control rods sequences studied. With the microscopic standard scheme, numerical biases are relatively small for UO2 fuel assembly without burnable absorbers but can be comparable to macroscopic standard scheme for poisoned fuel assemblies. Core calculations performed with the COHLISEE method show a reduction of the numerical biases, especially on pin-by-pin power distributions, at a computing cost comparable to standard microscopic scheme.In order to compare the COHLISEE method to standard schemes, 3D whole-core calculations are performed for a SBF-SMR benchmark. This study evaluates the magnitude of numerical differences between COHLISEE and standard schemes.; La puissance d’un réacteur à eau pressurisée (REP) industriel est généralement de l’ordre de 1000 MWe. Avec l’émergence des nouveaux marchés tels que l’alimentation de sites isolés ou la production de chaleur, des concepts de petits réacteurs modulaires (small modular reactor ou SMR) sont proposés avec des puissances inférieures à 300 MWe. Certains concepts de SMR n’utilisent pas de bore soluble dans le circuit primaire, malgré son utilisation quasi-systématique dans les REP. Le contrôle de la réactivité de ces réacteurs repose donc principalement sur l’insertion et le retrait des grappes de contrôle. L’utilisation de poisons consommables à poste fixe permet également un contrôle de la réactivité à la conception et limite la quantité d’absorbants mobiles insérés dans le cœur.Un schéma de calcul déterministe en deux étapes est généralement utilisé pour la modélisation des REP. La première étape est un calcul de transport neutronique sur une géométrie 2D détaillée d’un assemblage combustible en réseau infini. Dans un schéma de calcul standard, une évolution toutes grappes extraites (TGE) est réalisée sur une plage de burnup. Des calculs de reprise sont effectués à chaque pas de burnup en faisant varier les paramètres principaux de fonctionnement tels que la densité du modérateur, la température du combustible et l’insertion de la grappe de contrôle. Les sections efficaces multigroupes sont ensuite condensées à deux groupes et homogénéisées à l’échelle de la cellule combustible. Les librairies de sections efficaces ainsi obtenues sont utilisées à l’étape cœur pour le calcul des distributions 3D de puissance par résolution de l’équation de la diffusion à deux groupes d’énergie. L’hypothèse d’évolution du combustible en configuration TGE nécessite d’être remise en question au regard de l’insertion prolongée des grappes de contrôle pour un cœur de REP sans bore soluble, qui modifie le spectre neutronique d’évolution.Le premier objectif de ce travail de thèse est de quantifier les écarts numériques introduits par le schéma de calcul neutronique standard pour la modélisation d’assemblage, de cluster et de cœur d’un REP sans bore soluble fortement empoisonné. Le second objectif est de développer une méthode qui permette de simuler des effets d’historique de spectre à l’échelle d’un cœur complet de SMR sans pénalité excessive sur le temps de calcul.Ce travail a été réalisé avec la plateforme déterministe multi-filières APOLLO3®. La méthode développée, baptisée COHLISEE, est une méthode hybride qui consiste à utiliser à la fois une évolution microscopique et un indicateur d’historique de spectre lors d’un calcul de diffusion. La méthode est vérifiée et validée grâce à des calculs d’assemblages et de cluster 2x2 en 2D. Les résultats de référence sont obtenus par des calculs en transport réalisés avec des séquences particulières d’insertion de grappe de contrôle. Les résultats de COHLISEE (réactivité, distributions de puissance, bilan matière) sont comparés à ceux obtenus par des schémas standards utilisant des modèles d’évolution macroscopique ou microscopique au niveau cœur. Le processus de vérification et de validation (V&V) montre que le schéma macroscopique présente des biais considérables sur la réactivité pour toutes les séquences étudiées. Les écarts numériques avec le schéma microscopique sont relativement faibles pour un assemblage UO2 sans poisons consommables mais peuvent être comparables au schéma macroscopique en présence d’absorbants gadoliniés. Les calculs de cœur réalisés avec la méthode COHLISEE réduisent les biais numériques, notamment sur les distributions de puissance crayon par crayon, avec un coût de calcul comparable au schéma standard microscopique. Afin de comparer COHLISEE avec les schémas standards, des calculs critiques en 3D sur un cœur de benchmark de SMR sans bore soluble sont réalisés. Cette étude permet d’évaluer les écarts numériques entre COHLISEE et les schémas standards pour un cœur complet.

IRSN has been involved with CEA for some years in projects led by ONET Technologies (OT) and funded by Japanese government (METI). These projects aim at evaluating a strategy for the retrieval of fuel debris fell over the floor of Fukushima Dai-Ichi (1F) reactor pedestal, by a laser cutting technology developed by CEA and adapted by OT for a potential use in the 1F site.However, an important issue regarding aerosols dispersion and potential release into the environment is present around this well-proven technology and must be under control. To do that, IRSN has contributed in this project to provide quantitative information to assess the risk induced by aerosols generation and dispersion. Concomitant experimental and numerical simulations (CFD) studies were led in order to characterize the source term of aerosols from laser cutting of fuel debris simulants (Porcheron et al. 2020), to characterize their fate inside the reactor pedestal by the contribution of airflows coming from the laser cutting head (Gelain et al., 2019) (Gélain et al., 2018) and to propose solutions to help mitigating dispersed aerosols (Porcheron et al. 2019) (Sun et al., 2020) (Porcheron et al., 2021a) (Porcheron et al., 2021b).The present article is dedicated to summarizing a global CFD study led to evaluate the impact of fuel debris laser cutting in a representative design of 1F2 reactor pedestal in terms of aerosols dispersion and mitigation means. Given that laser cutting (as cutting in general) produces a large amount of aerosols, laser cutting head was first equipped with a local particle collection system close to the emission source. This system includes an extraction device (a flexible duct and filtration systems) allowing to catch a large part of emitted aerosols with a particle size distribution characterized by a mass median diameter of around 300 nm. However, even if this particle collection means is very efficient, a part of emitted aerosols is not collected and still dispersed in the pedestal. Hence, to collect them, IRSN proposed to implement a spray system, commonly used in French PWR to washout fission products in the event of a severe accident.This article presents different results of CFD calculations showing the necessity of mitigation means to avoid high aerosols concentration in the pedestal which could lead to a release into the environment. These calculations give also quantitative information about the performance of intended collection systems (extraction and spray) implemented in a representative geometry of 1F2 reactor pedestal of Fukushima and their complementarity. Finally, they highlight the evolution of the aerosols emitted during a cutting scenario as well for the phases during which the laser cutting (cutting phase) is activated as the phases during it is stopped (non-cutting phase). This last part will allow to discuss about the options to help decreasing faster the aerosol concentration during the non-cutting phase.

International audience

International audience

International audience; We validated a method for predicting the Soft Error Rate (SER) in the WEST tokamak operated with deuterium plasmas, and we applied it to predict the SER in the ITER tokamak operated with deuterium-tritium plasmas.