Search

International audience

The determination of the hygro-elastic behavior of wood under mechanical loading or environmental changes is of great practical interest, because of its use in many fields such as construction, industry, furniture, etc. Classically, wood can be considered as a hierarchical material, i.e. a material with a multi-scale structure. For many problems involving such materials, the heterogeneous structure at a particular scale can be replaced by a homogeneous structure with similar properties determined by applying homogenization methods. In this thesis, we aim to study the influence of the real cell wall morphology by considering wavy and interconnected cellulosic fibrils on the macroscopic behavior of wood, rather than considering them parallel as usually assumed in the literature. For this purpose, a multi-scale model, which allows to estimate the macroscopic behavior of wood, has been developed. This multi-scale model integrates 3 different scales of wood microstructure: cell wall, tissue and ring. From these scales, we will use both numerical and analytical homogenization methods to determine their effective behavior by defining a periodic elementary cell at each scale. Using this multi-scale model, we show that the influence of the undulations and interconnections of the fibrils is significant for some macroscopic moduli, such as the effective shear moduli, while it is not significant for other moduli at the macroscopic level (growth-ring sscale). Furthermore, although the effect of fibril crosslinks is quite strong for some components of elastic behavior at the cell wall level, it loses its importance at the macroscopic level, especially for low-density wood. This tendency can be explained by the antisymmetric tilting of microfibrils in adjacent cell walls. On the other hand, for denser woods where the interaction between adjacent cell walls is less dominant, the effect of fibrils undulations remains important.; La détermination du comportement hygro-élastique du bois sous l'effet de chargements mécaniques ou de changements environnementaux présente un enjeu pratique important, du fait de son utilisation dans de nombreux domaines tels que la construction, l’industrie, l’ameublement, etc. Classiquement, le bois peut être considéré comme un matériau hiérarchique, c'est-à-dire un matériau possédant une structure multi-échelle. Pour de nombreux problèmes impliquant de tels matériaux, la structure hétérogène à une échelle particulière peut être remplacée par une structure homogène aux propriétés similaires déterminées par l'application de méthodes d'homogénéisation. Dans cette thèse, on vise à étudier l'influence de la morphologie réelle de la paroi cellulaire sur le comportement macroscopique du bois en prenant en compte des fibrilles cellulosiques ondulées et interconnectées entre elles, plutôt que de les considérer parallèles comme cela est supposé habituellement dans la littérature. Pour cela, un modèle multi-échelle, qui permet d’estimer le comportement macroscopique du bois, sera développé. Ce modèle multi-échelle intègre 3 échelles différentes de microstructure du bois : celle de la paroi cellulaire, celle du tissu et celle du cerne, à partir desquelles on s’appuiera sur des méthodes d’homogénéisation à la fois numériques et analytiques pour déterminer leur comportement effectif en définissant à chaque échelle considérée une cellule élémentaire périodique. A l’aide de ce modèle multi-échelle, on montre que l'influence des ondulations et interconnections des fibrilles est significative pour certains modules macroscopiques, tels que les modules de cisaillement effectifs, alors qu'elle ne l'est pas pour d'autres modules au niveau macroscopique (échelle du cerne). De plus, bien que l'effet des réticulations des fibrilles soit assez fort pour certaines composantes du comportement élastique au niveau de la paroi cellulaire, il perd de son importance au niveau macroscopique, surtout pour le bois de faible densité. Cette tendance peut être expliquée par l'inclinaison antisymétrique des microfibrilles dans les parois des cellules adjacentes. En revanche, pour les bois les plus denses où l'interaction entre les parois cellulaires adjacentes est moins dominante, l’effet des ondulations des fibrilles reste important.

Potts shunt (PS) was suggested as palliation for patients with suprasystemic pulmonary arterial hypertension (PAH) and right ventricular (RV) failure. PS, however, can result in poorly understood mortality. Here, a patient-specific geometrical multiscale model of PAH physiology and PS is developed for a paediatric PAH patient with stent-based PS. In the model, 7.6mm-diameter PS produces near-equalisation of the aortic and PA pressures and $$Q_p/Q_s$$ Q p / Q s (oxygenated vs deoxygenated blood flow) ratio of 0.72 associated with a 16% decrease of left ventricular (LV) output and 18% increase of RV output. The flow from LV to aortic arch branches increases by 16%, while LV contribution to the lower body flow decreases by 29%. Total flow in the descending aorta (DAo) increases by 18% due to RV contribution through the PS with flow into the distal PA branches decreasing. PS induces 18% increase of RV work due to its larger stroke volume pumped against lower afterload. Nonetheless, larger RV work does not lead to increased RV end-diastolic volume. Three-dimensional flow assessment demonstrates the PS jet impinging with a high velocity and wall shear stress on the opposite DAo wall with the most of the shunt flow being diverted to the DAo. Increasing the PS diameter from 5mm up to 10mm results in a nearly linear increase in post-operative shunt flow and a nearly linear decrease in shunt pressure-drop. In conclusion, this model reasonably represents patient-specific haemodynamics pre- and post-creation of the PS, providing insights into physiology of this complex condition, and presents a predictive tool that could be useful for clinical decision-making regarding suitability for PS in PAH patients with drug-resistant suprasystemic PAH.

International audience; The magnetic Barkhausen noise energy hysteresis cycles, MBNenergy(H), were evaluated as a nondestructive testing characterization method to identify intern mechanical stress in ferromagnetic parts. Oriented grains electrical steel, and iron-cobalt specimens were tested as model materials of opposite behaviors. Tensile stress tests were run and revealed MBNenergy(H) coercivity as the most correlated indicator. In parallel, a predictive multiscale model was developed to simulate the stress-dependent MBNenergy(H) anhysteretic curves. A hysteresis contribution was added, and the resulting hysteresis predictions were validated by comparison to the tensile-stress experimental tests. 2D simulation predictions reveal the identification of uniaxial tensile stress as more efficient when the magnetic field is applied within an angle between 30° and 75° from the stress direction. The simulation method allows the foresee of the most favorable sensor orientation configurations depending on the material tested and all available a priori knowledge of the stress configuration.

International audience; Germinal centers play a key role in the adaptive immune system since they are able to produce memory B cells and plasma cells that produce high affinity antibodies for an effective immune protection. The mechanisms underlying cell-fate decisions are not well understood but asymmetric division of antigen, B-cell receptor affinity, interactions between B-cells and T follicular helper cells (triggering CD40 signaling), and regulatory interactions of transcription factors have all been proposed to play a role. In addition, a temporal switch from memory B-cell to plasma cell differentiation during the germinal center reaction has been shown. To investigate if antigen affinity-based Tfh cell help recapitulates the temporal switch we implemented a multiscale model that integrates cellular interactions with a core gene regulatory network comprising BCL6, IRF4, and BLIMP1. Using this model we show that affinity-based CD40 signaling in combination with asymmetric division of B-cells result in switch from memory B-cell to plasma cell generation during the course of the germinal center reaction. We also show that cell fate division is unlikely to be (solely) based on asymmetric division of Ag but that BLIMP1 is a more important factor. Altogether, our model enables to test the influence of molecular modulations of the CD40 signaling pathway on the production of germinal center output cells.

This paper is devoted to the investigation of the density sorting of grains using water jigging. Experiments achieved in a laboratory scale water jig for two initial binary bed configurations are studied and compared to numerical results. The vertical composition of the deposit is estimated after different number of water pulses to represent the sorting evolution in time. Simulations are based on a multiscale model in which the fluid is solved at a larger scale than the grain diameter using the finite element method, while the grains are considered as rigid bodies and solved using the nonsmooth contact dynamics. Key parameters are numerically varied and observed to determine the sensitivity of the process.

This dissertation focuses only on the fibre break damage mechanism controlling the failure of unidirectional composite structures. The MPFBM has described this mechanism by including the stochastic nature of fibre break, stress relaxation by viscoelastic matrix and the debonding between fibre and matrix caused by the nearby surrounding breakage. With the simplified FE2 approach, the model has been compared favourably with experimental result in the specimen level, but not for a real scale composite structure. This dissertation has evaluated a new approach based on the concept of Stationary Ergodic Random Function and Integral Range (SERFAIR) to resolve this issue. The concept of SERFAIR essentialy helps to determine a representative volume that would give a representative result in relation to the whole volume of the structure. In this way, the computation runs much faster as the number of degree of freedom has been reduced. Using this concept, the comparison study to the experimental results of racetrack specimen and a type IV PV gives a favourable results, indicating that the concept has succeeded to improve the computational speed without necessarily changing thequality of the strength prediction.; Cette thèse se concentre uniquement sur le mécanisme de rupture des fibres qui contrôle la défaillance des structures composites unidirectionnelles. Le MPTFBM a décrit ce mécanisme en incluant la nature stochastique de la rupture des fibres, la relaxation des contraintes par la matrice viscoélastique et le décollement entre la fibre et la matrice causé par la rupture voisine. Avec l’approche simplifiée FE2, le modèle a été comparé favorablement avec le résultat expérimental au niveau de l’échantillon, mais pas pour une structure composite à échelle réelle. Cette thèse a évalué une nouvelle approche basée sur le concept de Fonction Aléatoire Stationnaire Ergodique et Portée Intégrale (FASTEPI) pour résoudre ce problème. Le concept de SERFAIR aide essentiellement à déterminer un volume représentatif qui donnerait un résultat représentatif par rapport au volume total de la structure. De cette façon, le calcul est beaucoup plus rapide car le nombre de degrés de liberté a été réduit. En utilisant ce concept, l’étude de comparaison avec les résultats expérimentaux d’un spécimen de piste de course et d’un PV de type IV donne un résultat favorable, indiquant que le concept a réussi à améliorer la vitesse de calcul sans nécessairement changer la qualité de la prédiction de la résistance.

The aim of this study is to model the mechanical behavers of two SMC composite materials subjected to fatigue and dynamic loadings. To achieve this goal, a bibliographic study was carried out on the modeling of the dynamic and cyclic behavior of SMC composite. Fiber-matrix interface decohesion is a main local damage mechanism in SMC composites. This phenomenon is introduced in the Mori and Tanaka approach through a local damage criterion which defines step by step the number of micro-cracks to be introduced in the homogenization scheme of this model during loading until final failure.Indeed, this model allows predicting the monotonic behavior of the studied SMCs. The experimental results, extracted from literature, have allowed the extension of the micromechanical model in the cases of cyclic and dynamic loadings. Four complementary approaches, all based on the consideration of interfacial damage, have been developed as a function of the local normal and tangential stress values, the corresponding strain rate and the number of cycles.Thus, the mechanical response under monotonic loading at different strain rates from quasi-static to dynamic was also predicted. For the fatigue loading, the progressive degradation of the mechanical properties and the fatigue lifetime of the two investigated SMC composites were predicted in good agreement with the experimental results obtained at microscopic and macroscopic scale., L’objectif de cette étude est de modéliser les réponses mécaniques de deux matériaux composites SMC soumis à des sollicitations de types fatigue et dynamique. Pour mener à bien cet objectif, une étude bibliographique a été menée sur les propriétés et la modélisation des comportements dynamique et cyclique des matériaux composites à renforts discontinus. L’endommagement à l’interface fibre-matrice apparait comme étant le phénomène moteur dans les composites SMC dont le comportement peut être qualifié d’élastique endommageable. Ainsi, un modèle micromécanique basé sur une technique d’homogénéisation dans lequel un critère local statistique de rupture à l’interface fibre-matrice a été introduit a été développé et a permis de traduire le comportement monotone des SMC étudiés a été développé. Les résultats expérimentaux extraits de la littérature ont permis l’extension du modèle micromécanique aux cas des chargements cycliques et dynamiques. Quatre approches complémentaires tous fondé sur la prise en compte des dommages interfaciaux ont été développées. Notamment, le critère a été formulé en fonction des contraintes limites normales et tangentielles locales, des vitesses de chargement locales correspondantes et du nombre de cycle afin de prédire le comportement dynamique et cyclique. Ainsi, les réponses mécaniques sous chargement monotone, la dégradation progressive des propriétés mécaniques sous chargement dynamique et cyclique et la durée de vie des deux composites SMC étudiés ont pu être prédites avec une bonne concordance avec les résultats expérimentaux obtenus à l’échelle microscopique et macroscopique.

The aim of this study is to model the mechanical behavers of two SMC composite materials subjected to fatigue and dynamic loadings. To achieve this goal, a bibliographic study was carried out on the modeling of the dynamic and cyclic behavior of SMC composite. Fiber-matrix interface decohesion is a main local damage mechanism in SMC composites. This phenomenon is introduced in the Mori and Tanaka approach through a local damage criterion which defines step by step the number of micro-cracks to be introduced in the homogenization scheme of this model during loading until final failure.Indeed, this model allows predicting the monotonic behavior of the studied SMCs. The experimental results, extracted from literature, have allowed the extension of the micromechanical model in the cases of cyclic and dynamic loadings. Four complementary approaches, all based on the consideration of interfacial damage, have been developed as a function of the local normal and tangential stress values, the corresponding strain rate and the number of cycles.Thus, the mechanical response under monotonic loading at different strain rates from quasi-static to dynamic was also predicted. For the fatigue loading, the progressive degradation of the mechanical properties and the fatigue lifetime of the two investigated SMC composites were predicted in good agreement with the experimental results obtained at microscopic and macroscopic scale.; L’objectif de cette étude est de modéliser les réponses mécaniques de deux matériaux composites SMC soumis à des sollicitations de types fatigue et dynamique. Pour mener à bien cet objectif, une étude bibliographique a été menée sur les propriétés et la modélisation des comportements dynamique et cyclique des matériaux composites à renforts discontinus. L’endommagement à l’interface fibre-matrice apparait comme étant le phénomène moteur dans les composites SMC dont le comportement peut être qualifié d’élastique endommageable. Ainsi, un modèle micromécanique basé sur une technique d’homogénéisation dans lequel un critère local statistique de rupture à l’interface fibre-matrice a été introduit a été développé et a permis de traduire le comportement monotone des SMC étudiés a été développé. Les résultats expérimentaux extraits de la littérature ont permis l’extension du modèle micromécanique aux cas des chargements cycliques et dynamiques. Quatre approches complémentaires tous fondé sur la prise en compte des dommages interfaciaux ont été développées. Notamment, le critère a été formulé en fonction des contraintes limites normales et tangentielles locales, des vitesses de chargement locales correspondantes et du nombre de cycle afin de prédire le comportement dynamique et cyclique. Ainsi, les réponses mécaniques sous chargement monotone, la dégradation progressive des propriétés mécaniques sous chargement dynamique et cyclique et la durée de vie des deux composites SMC étudiés ont pu être prédites avec une bonne concordance avec les résultats expérimentaux obtenus à l’échelle microscopique et macroscopique.

International audience; The current requirements for composite cylinders are still based on an arbitrary approach derived from the behaviour of metal structures, that the designed burst pressure should be at least 2.5 times the maximum in-service pressure [1]. This could lead to an over-designed composite cylinder for which the weight saving would be less than optimum. Moreover, predicting the lifetime of composite cylinders is a challenging task due to their anisotropic characteristics. A federal research institute in Germany (BAM) has proposed a minimum load-cycle requirement that mitigates this issue by using a Monte-Carlo analysis of the burst test results [2-3]. To enrich this study, more experiments are required however they are normally limited by the necessity of long duration testing times (loading rate and number of cylinders) and the design (stacking sequence of the composite layer). A multi-scale model incorporating the micromechanical behaviour of composite structures has been developed at Mines ParisTech. The model has shown similar behaviour to that of composite cylinders under different loading rates [4]. This indicates that the model could assist the Monte-Carlo analysis study. An evaluation of the multi-scale model therefore has been carried out to determine its limitations in predicting lifetimes of composite cylinders. The evaluation starts with the comparison of burst pressures with type IV composite cylinders under different loading rates. A μCT-Scan of a type IV cylinder has been carried out at the University of Southampton. The produced images were analysed using the Fast-Fourier Transform (FFT) technique to determine the configuration of the composite layers, which is required by the model. Finally, the time dependent effect studied, by using the multi scale model, has been described. In the long-term, this study can be used to conduct a parametric study for creating more efficient design of type IV cylinders.

This manuscript assembles my contributions in developing new mathematical and computational methods for analyzing the dynamics of charged particles, like electrons or ions, as a multiscale phenomenon. The underlying mechanisms of this general physical problem are described byVlasov--Poisson systems. The objective of this work is to study these equations and to implement various efficient numerical methods to approximate their solutions. In Chapter 1, two strategies are proposed in order to cope with the multiscale issue. First, we obtain reduced models by means of asymptotic analysis in the frame of the two-scale convergence. Second, we treat the full model with a numerical method based on exponential integrator, by which the high frequency oscillations are exactly solved whereas the slower process is treated in an approximate way.In Chapter 2, we focus on the performance of Particle-in-Cell simulations for solving Vlasov--Poisson systems in six dimensional phase space. Mainly, we address specific data structures in order to optimize the memory accesses. In addition, we exploit efficiently parallelism patterns, like vectorization,multithreading, and multiprocessing.In Chapter 3, we develop a computational framework for modelling and simulating complex problems in plasma physics. More precisely, we study the problems of the diocotron instability in a non-neutral plasma and of the dynamics of two species of charged particles following an edge-localized mode event in a tokamak's scrape-off layer. In this direction, we propose and solve different kinetic and fluid equations to treat the modelling questions, while for the numerical aspect, an asymptotic preserving strategy turns out to be fruitful to deal with the multiscale issue., Dans ce mémoire, nous présentons différentes stratégies pour approcher les solutions de modèles cinétiques faisant intervenir plusieurs échelles en temps. Plus précisément, nous nous intéressons à des systèmes de type Vlasov-Poisson pour décrire l'évolution dans l’espace des phases de particules chargées sous l’influence d’un champ électromagnétique. L’objectif de ce travail est d’étudier le comportement multi-échelle de ces équations à l’aide de l’analyse asymptotique et de proposer des méthodes numériques adaptées pour approcher leurs solutions.Dans la première partie, nous proposons deux méthodes pour traiter la difficulté liée au caractère multi-échelle. La première est une méthode d’homogénéisation en temps, basée sur la notion de convergence à deux échelles. Dans cette approche, nous obtenons un modèle réduit approchant l’équation originelle de Vlasov. La deuxième méthode est un nouveau schéma en temps pour l’équation de Vlasov. Basé sur un intégrateur exponentiel, le schéma résout la petite échelle tout en utilisant des pas de temps macroscopiques. Des cas-tests issus de la physique des plasmas illustrent la précision et l’efficacité de la méthode.Dans la deuxième partie, nous analysons la performance d’une implémentation “Particle-in-Cell” pour résoudre numériquement le système de Vlasov–Poisson. Le problème de la performance se pose lors de l’utilisation de schémas de discrétisation explicites, avec des paramètres numériques résolvant la petite échelle, ce qui entraîne un coût de calcul important. Nous développons des structures de données spécifiques pour optimiser les accès mémoire. Ce travail est complété par le développement d’une approche spécifique de répartition des calculs par parallélisme hybride, utilisant à la fois les mécanismes de mémoire partagée et de mémoire distribuée, que nous implémentons sur processeurs multi-coeurs. Des mesures classiques pour évaluer la performance des codes sont analysées.Dans la troisième partie, nous développons un cadre de modélisation et de simulation pour résoudre des problèmes pertinents en physique des plasmas. Plus précisément, nous étudions (i) la dynamique d’un plasma d’ions et d’électrons suivant un “edge-localized mode”, qui est une instabilité se produisant au bord d’un tokamak et (ii) un problème d’instabilité de diocotron dans un plasma. Pour modéliser ces phénomènes, nous proposons différents couplages faisant intervenir des équations de type Poisson, des équations fluides et des équations cinétiques de type Vlasov avec termes source et de collision. Leur résolution est effectuée avec des schémas numériques de type“asymptotic-preserving” permettant de traiter efficacement le caractère multi-échelle.

Bone is a complex material showing a hierarchical and porous structure but also a natural ability to remodel thanks to cells sensitive to fluid flows. Based on these characteristics, a multiscale numerical model has been developed in order to represent the bone response under mechanical solicitation. It relies on the homogenization technique, simulating bone as a homogeneous structure having a porous microstructure saturated with bone fluid. The numerical modeling of the loading of a finite volume of bone enables the determination of an equivalent poroelastic stiffness. Focusing on two extreme fluid boundary conditions, the study of the corresponding structural response provides an overview of the fluid contribution to the poroelastic behavior, impacting the stiffness of the considered material. This parameter is either reduced (when the fluid can flow out of the structure) or increased (when the fluid is kept inside the structure) and quantified through this model. The presented poroelastic numerical model is here developed in the perspective of providing a bio-reliable model of bones, to determine the critical parameters that might impact bone remodeling.

The composite materials are strongly conditioned by the ability of the company to design the automotive structures under various complexes loadings such as fatigue. The aim of this thesis is to develop a multi-scale modeling coupled to a phenomenological approach in order to provide a response to the dimensioning need of structural parts subjected to cyclic loading at different temperatures of 23°C, 80°C and -30°C. By this way, the work was conducted along two main lines; firstly, an experimental investigation under monotonic and fatigue loadings. The results of this experimental study provide the necessary data for the construction of a micromechanical model which constitute the basis of the second part of this work; the predictive approaches of the fatigue lifetime for SMC composite. Thus, two hybrid, phenomenological/micromechanical, modeling approaches have been proposed. Both are based on a micromechanical modeling that allows describing the mechanical behavior of our material under monotonic loading, taking into account the microstructure and the damage phenomenon. These two predictive models require only a limited number of experimental tests and investigations but remain very reliable in their capacity to predict the lifetime of an SMC composite material under fatigue loading. The approach is validated in the case of thermomechanical sequential loadings at variable temperatures or amplitudes., L’utilisation des matières composites est fortement conditionnée par la capacité du constructeur ou du sous-traitant à dimensionner les structures automobiles sous divers types de chargements complexes tel que la fatigue. Le présent travail de thèse a pour objectif de développer un outil de modélisation par transition d’échelles couplée à une approche phénoménologique afin d’apporter une réponse à un besoin de maîtrise du dimensionnement de pièce de structure en composite SMC (polyester chargé en fibres de verre) soumis à des sollicitations de type fatigue sous différents niveaux de température 23°C, 80°C et -30°C. Pour ce faire, le travail a été mené selon deux axes principaux. En premier lieu, une investigation expérimentale sous chargement monotone et fatigue. Les résultats de l’analyse expérimentale de l’endommagement du matériau a permis d’accéder aux données nécessaires à la construction d’un modèle micromécanique constituant la base des approches prédictives de la durée de vie des SMC sous chargements cycliques constituant la seconde phase de travail. Ainsi, deux approches de modélisations hybrides phénoménologiques/micromécaniques ont été proposées. Elles sont toutes les deux basées sur une modélisation micromécanique qui permet de traduire le comportement mécanique du matériau d’étude sous chargement monotone avec prise en compte de la microstructure et de l’endommagement. Ces deux approches prédictives ne nécessitent qu’un nombre limité d’essais et d’investigations expérimentales mais restent bien fiables et pertinentes dans leurs capacités de prédire la durée de vie d’un matériau composite SMC sous chargement en fatigue. L’approche est validée dans le cas de chargements thermomécaniques séquentiels à température ou amplitude variable.

The composite materials are strongly conditioned by the ability of the company to design the automotive structures under various complexes loadings such as fatigue. The aim of this thesis is to develop a multi-scale modeling coupled to a phenomenological approach in order to provide a response to the dimensioning need of structural parts subjected to cyclic loading at different temperatures of 23°C, 80°C and -30°C. By this way, the work was conducted along two main lines; firstly, an experimental investigation under monotonic and fatigue loadings. The results of this experimental study provide the necessary data for the construction of a micromechanical model which constitute the basis of the second part of this work; the predictive approaches of the fatigue lifetime for SMC composite. Thus, two hybrid, phenomenological/micromechanical, modeling approaches have been proposed. Both are based on a micromechanical modeling that allows describing the mechanical behavior of our material under monotonic loading, taking into account the microstructure and the damage phenomenon. These two predictive models require only a limited number of experimental tests and investigations but remain very reliable in their capacity to predict the lifetime of an SMC composite material under fatigue loading. The approach is validated in the case of thermomechanical sequential loadings at variable temperatures or amplitudes.; L’utilisation des matières composites est fortement conditionnée par la capacité du constructeur ou du sous-traitant à dimensionner les structures automobiles sous divers types de chargements complexes tel que la fatigue. Le présent travail de thèse a pour objectif de développer un outil de modélisation par transition d’échelles couplée à une approche phénoménologique afin d’apporter une réponse à un besoin de maîtrise du dimensionnement de pièce de structure en composite SMC (polyester chargé en fibres de verre) soumis à des sollicitations de type fatigue sous différents niveaux de température 23°C, 80°C et -30°C. Pour ce faire, le travail a été mené selon deux axes principaux. En premier lieu, une investigation expérimentale sous chargement monotone et fatigue. Les résultats de l’analyse expérimentale de l’endommagement du matériau a permis d’accéder aux données nécessaires à la construction d’un modèle micromécanique constituant la base des approches prédictives de la durée de vie des SMC sous chargements cycliques constituant la seconde phase de travail. Ainsi, deux approches de modélisations hybrides phénoménologiques/micromécaniques ont été proposées. Elles sont toutes les deux basées sur une modélisation micromécanique qui permet de traduire le comportement mécanique du matériau d’étude sous chargement monotone avec prise en compte de la microstructure et de l’endommagement. Ces deux approches prédictives ne nécessitent qu’un nombre limité d’essais et d’investigations expérimentales mais restent bien fiables et pertinentes dans leurs capacités de prédire la durée de vie d’un matériau composite SMC sous chargement en fatigue. L’approche est validée dans le cas de chargements thermomécaniques séquentiels à température ou amplitude variable.

International audience; This paper presents an unresolved computational fluid dynamic–discrete element method (CFD–DEM) model for the simulation of flows mixing fluid and grains. The grains trajectories are solved at a fine scale using a discrete element method. It provides the velocities and the trajectories of the grains with an accuracy that is needed to describe microscopic phenomena like clogging in pipe happening in these flows. Solved at a coarse scale using the finite element method, the fluid motion is deduced from a mean continuous representation of the fluid phase giving computational performance and keeping variables evolutions that are of interest for a lot of simulation processes. The key point of this method lays in the coupling of the two different representation scales. An empirical drag formula for monodisperse granular media parametrises the transfer of momentum between the phases. Applying this model to the well-known problem of suspension drops provides validation and insight in this kind of methodology. Simulations in which inertia is non-negligible are achieved to prove the generality and adaptability of the unresolved CFD–DEM model compared to other models.

International audience; We are developing a Mathematical Virtual Simulator (MVS)[1] that combines multiscale mathematical modeling, advanced numerical techniques and effcient computational tools motivated by biomedical applications in the context of ophthalmology. The design of MVS is constituted by a three-dimensional poro-elastic system, representing the biomechanics and hemodynamics of the ocular posterior tissues, and a lumped-parameter circuit that describes the blood circulation in the retina and the blood flow in the central retinal vessels. Driven by the physics of the problem, which requires a blood systemic view, we do not impose artificial boundary condition for the hemodynamics; rather, we couple the description of the flow by means of a system of partial differential equations with a system of Ordinary Differential Equations (ODEs).We have elaborated a novel splitting approach to numerically solve the multiscale problem in an efficient, accurate and affordable manner, which involves the coupling between Darcy equations and ODE systems.The algorithm is inspired by the work of Carichino et al.[2] and Glowinski[3] and it is based on a semi-discretization in time based on operator splitting in order to guarantee the physical energy balance at the discrete level. As a result, unconditional stability with respect to the time step choice is ensured by the implicit treatment of interface conditions within the Darcy sub-steps, whereas the coupling between the 3d and 0d sub-steps is enforced via appropriate initial conditions for each sub-step. Notably, unconditional stability is attained without the need of sub-iterating between Darcy and ODE sub-steps. In our case, the Darcy problem is solved using HDG (Hybridizable Discontinuous Galerkin) method to achieve high accuracy on fluxes, which have a fundamental role from a biological viewpoint.In conclusion, we have formulated and implemented a numerical algorithm combining HDG discretization with an unconditionally stable operator splitting method for multiscale problems in order to simulate the physiological behavior of blood flow in the ocular posterior segment. More generally, the algorithm might be utilized for other problems arising in engineering and life sciences, where a robust and efficient couplingbetween a 3d poro-elastic model and a 0d fluid flow circuit described by a system of ODEs is required.References[1] L. Sala, C. Prud'homme, G. Guidoboni, M. Szopos., Towards a full model for ocular biomechanics, fluid dynamics and hemodynamics.. J. for Modeling in Ophthalmology, to appear.[2] L. Carichino, G. Guidoboni, M. Szopos., Energy-based operator splitting approach for the time discretization of coupled systems of partial and ordinary differential equations for fluid flows: the Stokes case.Journal of Computational Physics, 364, 235-256.[3] R. Glowinski, Finite Element methods for incompressible viscous flow. In Handbook of Numerical Analysis, Vol. IX, North-Holland, Amsterdam, 2003, pp. 3-1176.

The current work focuses on the development of a micromechanical modeling tool in the form of a virtual test machine which, used with the structural calculation codes, allows to determine the complex anisotropic behavior of polypropylene matrix composites reinforced with short glass fibers under dynamic loading. The core-skin microstructure induced by the material injection process is investigated experimentally by μCT. The dynamic behavior is characterized for strain rates of up to 200s-1 using an experimental methodology based on the use of a damping joint and specimen optimization. The mechanisms of damage are analyzed experimentally by in situ SEM testing. They highlight the importance of the debonding phenomenon in the damage scenario. Based on these experimental results, the multiscale approach developed consists of an incremental Mori Tanaka method applied to an elastoviscoplastic matrix and coated reinforcements integrating the evolution of damage at the mesoscopic scale. The damage introduced into the coatings disturbs the load transfer between the matrix and the reinforcements. In addition, the strain rate, orientation, and fiber rate dependence of the model are correlated by testing. The virtual testing machine is validated by modeling structures. The developed predictive tool thus takes into account the minimum necessary to describe the microstructure while being reliable and relevant in the modeling of composites under moderate dynamic stress.; Le présent travail de thèse a pour objectif de développer un outil de modélisation par transition d’échelles sous forme de machine d’essais virtuels. Celle-ci, utilisée conjointement aux codes de calculs de structures, permet de déterminer le comportement anisotrope complexe de composites à matrice polypropylène chargés en fibres de verre courtes sous sollicitations dynamiques. La microstructure en cœur-peau induite par le procédé d’injection du matériau est investiguée expérimentalement par μCT. Le comportement dynamique est caractérisé pour des vitesses de déformation allant jusqu’à 200s-1 au moyen d’une une méthodologie expérimentale basée sur l’utilisation d’un joint d’amortissement et d’une optimisation des éprouvettes. Les mécanismes d’endommagement sont analysés expérimentalement par essai in situ. Ils mettent en évidence le phénomène d’endommagent prépondérant qui est la décohésion de l’interface fibre matrice. Basé sur ces résultats expérimentaux, l’approche multi échelles développée consiste en une méthode de Mori Tanaka incrémentale appliquée à une matrice élastoviscoplastique et des renforts enrobés intégrant l’évolution de l’endommagement à l’échelle mésoscopique. L’endommagement introduit dans les enrobages perturbe le transfert de charge entre la matrice et les renforts. De plus, la dépendance à la vitesse de déformation, aux orientations et aux taux de fibre du modèle sont corrélés par des essais. La machine d’essais virtuels est validée par modélisation de structures. L’outil prédictif ainsi développé prend en compte le minimum nécessaire à la description de la microstructure tout en étant fiable et pertinent dans la modélisation de composites sous sollicitations dynamiques modérées.

International audience; A sodium fast reactor (SFR) fuel bundle is usually a strong hexagonal tube containing around 200 fuel pins made of steel. These pins should not touch each other to avoid overheating and damage of the bundle, which constitutes the first confinement barrier. For that purpose a steel wire is often wrapped around each pin from the bottom to the top with a helical shape. This technology maintains a distance between pins and ensures a proper mixing of the sodium, two necessary conditions so as to avoid hot points and damage. Nevertheless the intense fast neutron irradiation induces a significant isotropic swelling and creeping of the pins. If continued, this progressively generates contact closures (phase 1), then mainly a helical bow of the pin (phase 2), and in the end, a strong ovalisation of the cladding sections (phase 3). At such an interaction level, there is a possibility of cladding damage due to thermal creep accumulation. Several 3D finite elements models of the fuel bundle have been developed including the tube, all the 217 pins and their wrapped wires, at least 7.000 contact, with thermal and irradiation creeping laws. If conventional solid or even shell elements are used, then the huge memory cost and CPU time become completely inconsistent with the project expectations. The first analyses have shown that it was necessary to reduce the number of degree of freedom, but also, to keep the representation of all the pins and all the contact areas. Thus, a new approach is used here. First, the helical deformation of each pin is captured by a simple viscoplastic vertical beam; second, a new connecting element is developed and positioned horizontally on every potential contact area, between the neutral fibers of the two adjacent pins concerned. This connecting element, based on a bar finite element with only two nodes, takes into account:-the gap and the contact between a wire and the neighbor pin, or between a wire and the tube,-the thermal expansion and irradiation swelling of the pin and wire cross sections,-the radius increase by creeping under internal pressure loading,-the elastic rigidity of the pin crushing (locally pinching a short section of a tube),-the contact force, altered by the two creeping behaviors, when pinching this 3D tube section,-the counter ovalisation force due to internal pressure,-the hot point stress and strain tensors on the internal skin (strain concentration),-the hot point damage accumulation due to thermal creeping only. The two main difficulties are the following ones:-the damage evaluation supposes a really high precision in the stress tensor calculation, around 1%,-several nonlinear structural effects should be considered in only one integration point. These bar connecting element have been implemented in the Cast3M finite element code (http://www-cast3m.cea.fr). In the end, the bar element is validated by detailed 3D solid calculation. Then, a few relevant bundle experiments, which were conducted in the PHENIX French SFR, are fully simulated. They show good agreement to the measurements and a CPU performance gain between 100 and 1000.

International audience; Safety issues pertaining to Li-ion batteries justify intensive testing all along their value chain. However, progress in scientific knowledge regarding lithium based battery failure modes, as well as remarkable technologic breakthroughs in computing science, now allow for development and use of prediction tools to assist designers in developing safer batteries. Subsequently, this paper offers a review of significant modeling works performed in the area with a focus on the characterization of the thermal runaway hazard and their relating triggering events. Progress made in models aiming at integrating battery ageing effect and related physics is also discussed, as well as the strong interaction with modeling-focused use of testing, and the main achievements obtained towards marketing safer systems. Current limitations and new challenges or opportunities that are expected to shape future modeling activity are also put in perspective. According to market trends, it is anticipated that safety may still act as a restraint in the search for acceptable compromise with overall performance and cost of lithium-ion based and post lithium-ion rechargeable batteries of the future. In that context, high-throughput prediction tools capable of screening adequate new components properties allowing access to both functional and safety related aspects are highly desirable.

The anisotropic magnetoresistance (AMR) of ferromagnetic materials is widely used as the basic phenomenon for measuring or detecting magnetic field. Owing to the relationship between magnetic domain configuration and macroscopic resistivity, the application of an external magnetic field changes the resistivity of ferromagnetic materials. Although this effect is widely used in industrial applications, some basic aspects of AMR behavior are still unsufficiently understood. For example, the role of crystallographic texture is not accurately described by conventional modeling tools. As a consequence of the direct relationship between microstructure and AMR, models for AMR effect are generally based on micromagnetic calculations. For these calculations, the number of degrees of freedom and interactions can grow exponentially when investigating macroscopic behavior (case of polycrystals for example).The thesis deals with the numerical modeling of AMR effect in ferromagnetic materials. This new 3D modeling tool can overcome this major drawback of micromagnetic approaches. A model to describe the effects of magneto-elastic coupling using a micro-macro approach is available at the laboratory GeePs. Based on the same principles of micro-macro modeling, an AMR effect simulation tool has been developed including the effect of mechanical stress and the role of crystallographic texture of materials.The modeling strategy is as follows:Three scales of description of the behavior are introduced: the Representative Volume Element (RVE) of polycrystals (macro scale), the single crystal or grain, and finally the magnetic domain (micro scale).A first step, named localization, determines the magneto-mechanical loading (magnetic field and mechanical stress) within a grain depending on the external applied load. The introduction of internal variables and corresponding evolution laws allow describing in a statistical way the evolution of the magnetic domain microstructure under the influence of the local load. Also at this scale, the use of the phenomenological Doring model allows for each area, to calculate the resistivity as a function of the relative orientation between local magnetization and electric current. Once this local resistivity is known, a so-called homogenization step based on the Bruggeman model is used to determine the macroscopic resistivity of the RVE. It is thus possible to predict the variation in resistivity between an initial demagnetized state and a state under any magneto-mechanical loading.The results obtained by this approach were successfully compared to experimental results from literature on polycrystalline nickel, pure iron or Permalloy.Then simulations reproducing AMR sensors operating conditions were carried out. These simulations lead to the conclusion that it is possible to improve the sensitivity of AMR sensors by introducing an appropriate biaxial residual stress., La magnétorésistance anisotrope (AMR) des matériaux ferromagnétiques est largement utilisée comme le phénomène de base pour la mesure ou la détection de champ magnétique. En raison de la relation entre la configuration en domaines magnétiques et la résistivité macroscopique, l'application d'un champ magnétique externe modifie la résistivité des matériaux ferromagnétiques. Bien que cet effet soit largement utilisé dans des applications industrielles, certains aspects fondamentaux du comportement AMR sont encore assez mal compris. Par exemple, le rôle de la texture cristallographique dans le comportement effectif n'est pas décrit avec précision par les outils classiques de modélisation. En raison de ce lien direct entre la microstructure en domaines et l'effet AMR, les modèles de description de l'effet AMR reposent généralement sur des calculs micromagnétiques. Pour ces calculs, le nombre de degrés de liberté et d'interactions peuvent se multiplier rapidement si on recherche à décrire un comportement macroscopique (cas des polycristaux par exemple).La thèse porte sur la modélisation numérique de l'effet de magnétorésistance anisotrope des matériaux ferromagnétiques. Ce nouvel outil de modélisation 3D peut remédier à cet inconvénient majeur des approches micromagnétiques. Un modèle permettant de décrire les effets de couplage magnéto-élastique en utilisant une approche micro-macro est disponible au laboratoire GeePs. Sur la base des mêmes principes de la modélisation micro-macro, un outil de simulation de l'effet AMR en fonction de la contrainte mécanique et de la texture cristallographique des matériaux a été développé.La stratégie de modélisation est la suivante:Trois échelles de description du comportement sont introduites: le Volume Elémentaire Représentatif (VER) polycristallin (échelle macro), le monocristal ou grain, et enfin le domaine magnétique (échelle micro).Une première étape dite de localisation permet de déterminer le chargement magnéto-mécanique (champ magnétique et contrainte mécanique) à l'échelle d'un grain en fonction du chargement extérieur appliqué. L'introduction de variables internes et des lois d'évolution correspondantes permet de décrire de façon statistique l'évolution de la microstructure en domaines magnétiques sous l'influence de ce chargement local. Toujours à cette échelle, l'utilisation du modèle phénoménologique de Doring permet, pour chaque domaine, de calculer la résistivité en fonction de l'orientation relative entre aimantation locale et courant électrique. Une fois cette résistivité locale connue, une étape dite d'homogénéisation s'appuyant sur le modèle de Bruggeman permet de déterminer la résistivité macroscopique du VER polycristallin. Il est ainsi possible de prédire la variation de la résistivité entre un état initial désaimanté et un état sous chargement magnéto-mécanique quelconque.Les résultats obtenus par cette démarche ont été comparés avec succès à des résultats expérimentaux extraits de la littérature portant sur des polycristaux de Nickel, de Fer pur ou encore de Permalloy.Ensuite des simulations reproduisant les conditions de fonctionnement des capteurs AMR ont été effectuées. Ces simulations permettent de conclure qu'il est possible d'améliorer la sensibilité des capteurs AMR en générant une contrainte résiduelle biaxiale.

The thesis focuses on the numerical simulation of a biomathematical, multiscale model explaining the phenomenon of selection within the population of ovarian follicles, and grounded on a cellular basis. The PDE model consists of a large dimension hyperbolic quasilinear system governing the evolution of cell density functions for a cohort of follicles (around twenty in practice).The equations are coupled in a nonlocal way by control terms involving moments of the solution, defined on either the mesoscopic or macroscopic scale.Three chapters of the thesis, presented in the form of articles, develop the method used to simulate the model numerically. The numerical code is implemented on a parallel architecture. PDEs are discretized with a Finite Volume scheme on an adaptive mesh driven by a multiresolution analysis. Flux discontinuities, at the interfaces between different cellular states, require a specific treatment to be compatible with the high order numerical scheme and mesh refinement.A chapter of the thesis is devoted to the calibration method, which translates the biological knowledge into constraints on the parameters and model outputs. The multiscale character is crucial, since parameters are used at the microscopic level in the equations governing the evolution of the density of cells within each follicle, whereas quantitative biological data are rather available at the mesoscopic and macroscopic levels.The last chapter of the thesis focuses on the analysis of computational performances of the parallel code, based on statistical methods inspired from the field of uncertainty quantification., La thèse porte sur la calibration d'un modèle biomathématique multi-échelle expliquant le phénomène de sélection des follicules ovariens à partir du niveau cellulaire. Le modèle EDP consiste en un système hyperbolique quasi linéaire de grande taille gouvernant l'évolution des fonctions de densité cellulaire pour une cohorte de follicules (en pratique, une vingtaine).Les équations sont couplées de manière non locale par l'intermédiaire de termes de contrôle faisant intervenir les moments de la solution, intégrée à l'échelle mésoscopique et macroscopique. Trois chapitres de la thèse présentent, sous forme d'articles publiés, la méthode développée pour simuler numériquement ce modèle. Elle est conçue pour être implémentée sur une architecture parallèle. Les EDP sont discrétisées avec un schéma Volumes Finis sur un maillage adaptatif piloté par une analyse multirésolution. Le modèle présente des discontinuités de flux aux interfaces entre les différents états cellulaires, qui nécessitent la mise en ½uvre d'un couplage spécifique, compatible avec le schéma d'ordre élevé et le raffinement de maillage.Un chapitre de la thèse est dévolu à la méthode de calibration, qui consiste à traduire les connaissances biologiques en contraintes sur les paramètres et sur les sorties du modèle. Le caractère multi-échelle est là encore crucial. Les paramètres interviennent au niveau microscopique dans les équations gouvernant l'évolution des densités de cellules au sein de chaque follicule, alors que les données biologiques quantitatives sont disponibles aux niveaux mésoscopique et macroscopique.

International audience; 1- Occupancy estimates can inform biodiversity managers about the distribution of elusive species, such as the Pyrenean desman Galemys pyrenaicus, a small semi-aquatic mammal that lives along streams. Occupancy models rely on replication within a sampling site and provide estimates of the probability of detection. However, we still do not know how occupancy and detection estimates obtained from spatial vs. temporal replications differ or the appropriateness of using one or the other when cost and logistics make one approach prohibitive. Recently, the Markovian occupancy model has been developed to analyse adjacent spatial replicates and to test for spatial dependence between them. This model has already been applied to large and highly mobile mammals using trails, but never tested for any species with linear home ranges. 2- We compared detection and occupancy estimates obtained from both temporal and spatial sampling designs that were subsequently organized into four data configurations (sites with both spatial and temporal replicates, adjacent spatial replicates only, temporal replicates only at the segment and site scales). From that, five occupancy models with different assumptions (the standard occupancy model, the standard multiscale model, the multiscale model with Markovian process for detection, the Markovian detection model and the Markovian occupancy model) were used. We also assessed which occupancy model was the most appropriate for each data configuration to determine whether it is necessary to incorporate correlation into models. 3 - We found that the estimated detection probabilities were relatively high (≥0·58) and similar when the same model was applied to each data configuration. 4 - Spatial replication weakly underestimated occupancy. But when using this design, the Markovian occupancy model was the most supported and minimized the underestimation of occupancy, highlighting a spatial dependence between adjacent replicates. 5 - Synthesis and applications. We show that a survey based on adjacent spatial replicates for a mammal living along linear features of the landscape is a good compromise between cost and occupancy estimates, while using the Markovian occupancy model to estimate detection and occupancy. Our finding may have wider applications for the monitoring of species especially when temporal replicates are difficult or unrealistic. Spatial design, for surveys based on sign detection, could thus be applied for species with linear home ranges or when surveys are constrained by linear habitats.

International audience; Solving boundary value problems requires sufficiently robust constitutive models to be implemented. In this context, this presentation focuses on slope stability. For this purpose, an advanced micromechanical model for granular soils is presented.This model introduces an intermediate scale made up of elementary hexagonal patterns of adjoining particles. This is a breakthrough with respect to current micromechanical models that generally describe the material by a single distribution of contacts. It is shown how specific degradation laws influence the existence of a bifurcation domain, detected by the vanishing of the determinant of the symmetric part of the constitutive tensor. The bifurcation domain, located inside the plastic limit surface, gathers all mechanical states at which a failure may occur, depending on the incremental loading applied. This basic notion is proved to be fundamental in order to predict the occurrence of large, catastrophic landslides.

International audience; Because irregular bearing generates major agronomic issues in fruit-tree species, particularly in apple, the selection of regular cultivars is desirable. Here, we aimed to define methods and descriptors allowing a diagnostic for bearing behaviour during the first years of tree maturity, when tree production is increasing. Flowering occurrences were collected at whole-tree and (annual) shoot scales on a segregating apple population. At both scales, the number of inflorescences over the years was modelled. Two descriptors were derived from model residuals: a new biennial bearing index, based on deviation around yield trend over years and an autoregressive coefficient, which represents dependency between consecutive yields. At the shoot scale, entropy was also considered to represent the within-tree flowering synchronicity. Clusters of genotypes with similar bearing behaviours were built. Both descriptors at the whole-tree and shoot scales were consistent for most genotypes and were used to discriminate regular from biennial and irregular genotypes. Quantitative trait loci were detected for the new biennial bearing index at both scales. Combining descriptors at a local scale with entropy showed that regular bearing at the tree scale may result from different strategies of synchronization in flowering at the local scale. The proposed methods and indices open an avenue to quantify bearing behaviour during the first years of tree maturity and to capture genetic variations. Their extension to other progenies and species, possible variants of descriptors, and their use in breeding programmes considering a limited number of years or fruit yields are discussed.

International audience; An advanced elastic-plastic self-consistent polycrystalline model, accounting for intragranular microstructure development and evolution, is coupled with a bifurcation-based localization criterion and applied to the numerical investigation of the impact of microstructural patterns on ductility of single-phase steels. The proposed multiscale model, taking into account essential microstructural aspects, such as initial and induced textures, dislocation densities, and softening mechanisms, allows us to emphasize the relationship between intragranular microstructure of B.C.C. steels and their ductility. A qualitative study in terms of forming limit diagrams for various dislocation networks, during monotonic loading tests, is conducted in order to analyze the impact of intragranular substructure parameters on the formability of single-phase B.C.C. steels.

In design process, uncertainties have to be taken into account. Stochastic methods have therefore been proposed. Furthermore, in many cases, local defects affect strongly the behavior of a structure in a localized region while the rest of the structure is only slightly affected. In these cases, it is not reasonable to model the structure entirely at a fine scale, and multiscale methods are thus appealing. In this framework, we focused on the evaluation of a local specific quantity of interest when the Arlequin method is used to couple a deterministic model with a stochastic one. First, we give ingredients needed for the use of the method in this particular context. Second, to control the quality of the approximate solution obtained with such an approach, a goal-oriented method is introduced. Using residual-types estimates and adjoint-based techniques, a strategy for goal-oriented error estimation is presented for this coupling. Contributions of various error sources (modeling, space discretization, and discretization along the random dimension) are assessed. From information on error sources, an adaptive procedure is proposed to guaranty a given error tolerance. Finally, the described method is applied to study the infiltration of resin inside collagen network in the dentine.; La prise en compte de l’aléa dans le calcul des structures est souvent nécessaire pour le dimensionnement de celle-ci. Des méthodes stochastiques sont alors proposées. De plus, dans de nombreux cas, des altérations ou défauts affectent localement le comportement de la structure, alors que le reste n’est que faiblement impacté. Il n’est alors pas raisonnable d’utiliser une échelle d’analyse fine sur l’ensemble de la structure. On fait alors appel aux méthodes dites multi-échelles. Dans ce contexte, nous nous intéressons à l’estimation d’une quantité d’intérêt spécifique locale lorsque la méthode Arlequin est utilisée pour coupler un modèle déterministe à un modèle stochastique. Dans un premier temps, nous donnons les éléments nécessaires à l’utilisation de la méthode dans ce cadre de couplage stochastique. Pour contrôler ensuite la qualité de l’approximation obtenue par une telle approche, une méthode d’estimation d’erreur de type Goal-Oriented est proposée. En introduisant le résidu du problème de référence et un problème adjoint, une stratégie d’estimation de l’erreur est décrite. Nous étudions aussi les contributions des différentes sources de l’erreur à l’erreur totale (erreur de modèle, erreur de discrétisation, erreur stochastique). Nous proposons une technique pour estimer ces différentes erreurs et piloter un processus d’adaptation afin de contrôler l’erreur totale commise. Finalement, la méthode décrite est utilisée pour l’étude de l’infiltration de résine médicale dans le cas du traitement de la carie.

International audience; A multiscale model was developed to simulate the rate of landing of a winged insect, the grain aphid, Sitobion avenae F. At a large scale (kilometric scale), it is convenient to use a deterministic model of their dispersion, based upon diffusion-advection-reaction partial derivative equations. At a small scale (hectometric scale), the process of site selection ('landing') is only partially understood, but is known to include the perception of field colour and landscape characteristics. Several hypotheses on aphids' behaviour were tested to simulate the rate of landing: the simulation was done by a cellular automata submodel under five different hypotheses on the precise organisation of the landing behaviour. We found a strong interaction between the effect of the proportion of wheat crops in the landscape and their spatial organisation. The spatial correlation between the places occupied by wheat crops appears crucial to determine the global rate of landing of the aerial insect stock. The shape of the response surface of landing rate against the proportion of surface occupied by wheat, and the spatial autocorrelation of wheat plots, appears very regular and relatively simple to model by ad hoc mathematical functions. Large scale simulations using the results of the small scale model in a diffusion-reaction equation solved numerically, showed that, on a real landscape extracted from a GIS on the whole Brittany region (western France), the spatial pattern of the aphid landing is sensitive to the hypothesis tested on their landing behaviour. This hierarchical modelling combining two different approaches at two different scales (mathematical deterministic equations on a large multi-kilometric scale, and partly stochastic cellular automata on a small hectometric scale), requires methods to validate its results in the field, in the framework of a decision support system. Such amultiscalemodel has a wide field of application including not only plant protection but also management and conservancy of animal species dispersing by flight.

In design process, uncertainties have to be taken into account. Stochastic methods have therefore been proposed. Furthermore, in many cases, local defects affect strongly the behavior of a structure in a localized region while the rest of the structure is only slightly affected. In these cases, it is not reasonable to model the structure entirely at a fine scale, and multiscale methods are thus appealing. In this framework, we focused on the evaluation of a local specific quantity of interest when the Arlequin method is used to couple a deterministic model with a stochastic one. First, we give ingredients needed for the use of the method in this particular context. Second, to control the quality of the approximate solution obtained with such an approach, a goal-oriented method is introduced. Using residual-types estimates and adjoint-based techniques, a strategy for goal-oriented error estimation is presented for this coupling. Contributions of various error sources (modeling, space discretization, and discretization along the random dimension) are assessed. From information on error sources, an adaptive procedure is proposed to guaranty a given error tolerance. Finally, the described method is applied to study the infiltration of resin inside collagen network in the dentine., La prise en compte de l’aléa dans le calcul des structures est souvent nécessaire pour le dimensionnement de celle-ci. Des méthodes stochastiques sont alors proposées. De plus, dans de nombreux cas, des altérations ou défauts affectent localement le comportement de la structure, alors que le reste n’est que faiblement impacté. Il n’est alors pas raisonnable d’utiliser une échelle d’analyse fine sur l’ensemble de la structure. On fait alors appel aux méthodes dites multi-échelles. Dans ce contexte, nous nous intéressons à l’estimation d’une quantité d’intérêt spécifique locale lorsque la méthode Arlequin est utilisée pour coupler un modèle déterministe à un modèle stochastique. Dans un premier temps, nous donnons les éléments nécessaires à l’utilisation de la méthode dans ce cadre de couplage stochastique. Pour contrôler ensuite la qualité de l’approximation obtenue par une telle approche, une méthode d’estimation d’erreur de type Goal-Oriented est proposée. En introduisant le résidu du problème de référence et un problème adjoint, une stratégie d’estimation de l’erreur est décrite. Nous étudions aussi les contributions des différentes sources de l’erreur à l’erreur totale (erreur de modèle, erreur de discrétisation, erreur stochastique). Nous proposons une technique pour estimer ces différentes erreurs et piloter un processus d’adaptation afin de contrôler l’erreur totale commise. Finalement, la méthode décrite est utilisée pour l’étude de l’infiltration de résine médicale dans le cas du traitement de la carie.

MOCHA Investigators: Andrew Taylor, MD, Alessandro Giardini, MD, Sachin Khambadkone, MD, Silvia Schievano, PhD, Marc de Leval, MD, and T.-Y. Hsia, MD (Institute of Child Health, London, United Kingdom); Edward Bove, MD, and Adam Dorfman, MD (University of Michigan, Ann Arbor, MI); G. Hamilton Baker, MD, and Anthony Hlavacek (Medical University of South Carolina, Charleston, SC); Francesco Migliavacca, PhD, Giancarlo Pennati, PhD, and Gabriele Dubini, PhD (Politecnico di Milano, Milan, Italy); Alison Marsden, PhD (University of California, San Diego, CA); Jeffrey Feinstein, MD (Stanford University, Stanford, CA) Irene Vignon-Clementel (National Institute of Research in Informatics and Automation, Paris, France); Richard Figliola, PhD, and John McGregor, PhD (Clemson University, Clemson, SC).; International audience; Single ventricle malformations are complex congenital heart defects which require a three-stage surgical treatment, starting from the very first days of life, to separate the systemic and pulmonary circulations, and restore the serial circuit occurring in normal patients. The final surgery results in a total cavopulmonary connection (TCPC), where both the superior and the inferior vena cava are connected to the right pulmonary artery. Several clinical and computational studies have been done to optimize the geometry of the TCPC, with the aim of minimizing energy losses and improving surgical outcomes. To date, only few modeling studies have taken into account respiration and exercise as important factors to quantify the performance of a Fontan geometry. The objective of this work is to test the dependence of fluid dynamic variables and energy efficiency on respiration in patient-specific models of Fontan circulation, when subjected to exercise tests.|A closed-loop multiscale approach was used, including a simple respiration model that modulates the extravascular pressures in the thoracic and abdominal cavities, to generate physiologic time-varying flow conditions. A lumped parameter network (LPN) representing the whole circulation was coupled to a patient-specific 3D finite volume model of the preoperative bidirectional cavo-pulmonary anastomosis (BCPA) with detailed pulmonary anatomy. Subsequently, three virtual TCPC alternatives were coupled to the LPN and investigated in terms of both local and global hemodynamics. In particular, a T-junction of the venae cavae to the pulmonary arteries, a design with an offset between the venae cavae and a Y-graft design were compared under exercise conditions.|Results showed that the BCPA model is able to realistically capture oscillations due to both cardiac and respiratory effects, when compared to the venous Doppler velocity tracings acquired preoperatively on the patient.|The differences in hemodynamics between the three investigated TCPC options were minimal and similar to those obtained without inclusion of respiratory effects. Hence, the three surgical options result to be equivalent according to the analyzed parameters. Moreover, although the simulation of the Fontan circulation with a respiratory model requires a longer computational time, the developed framework allows for a more physiologic method to incorporate respiratory effects that was not possible using other methods.

International audience; A computational model is potentially a powerful tool to apprehend complex phenomena like solid tumour growth and to predict the outcome of therapies. To that end, the confrontation of the model with experiments is essential to validate this tool. In this study, we develop a computational model specifically dedicated to the interpretation of tumour growth as observed in a mouse model with a dorsal skinfold chamber. Observation of the skin vasculature at the dorsal window scale shows a sparse network of a few main vessels of several hundreds micrometers in diameter. However observation at a smaller scale reveals the presence of a dense and regular interconnected network of capillaries about ten times smaller. We conveniently designate this structure as the submicrovascular network (SMVN).(1) The question that we wish to answer concerns the necessity of explicitly taking into account the SMVN in the computational model to describe the tumour evolution observed in the dorsal chamber. For that, simulations of tumour growth realised with and without the SMVN are compared and lead to two distinct scenarios. Parameters that are known to strongly influence the tumour evolution are then tested in the two cases to determine to which extent those parameters can be used to compensate the observed differences between these scenarios. Explicit modelling of the smallest vessels appears mandatory although not necessarily under the form of a regular grid. A compromise between the two investigated cases can thus be reached.

The effect of stress on the magnetic behaviour of ferrites has been investigated by the use of experiments (parallelepipedic samples submitted to compression and tensile stress, and toric samples submitted to compressive stress) and a multiscale model. The measurements allow us to estimate the magnetic and magnetostrictive behaviours of the ferrite. The numerical model is based upon three scales: the magnetic domain, the single crystal and the polycrystal. At the scale of the magnetic domain, a minimisation of the potential energy gives the direction of magnetisation under the combined effect of magnetic field, stress, and crystal orientation. At the scale of the single crystal, the volumetric fractions of domains of different orientations are computed from their respective energies through a constitutive equation (Boltzmann function). The polycrystalline behaviour is deduced from the behaviour of a sufficient number of grains thanks to an averaging operation. The model allows us to simulate the magnetic and magnetostrictive behaviours. Experimental and numerical results are in good accordance. The modelling of the magnetic behaviour under high frequency magnetic field conditions constitutes the final point to the work. It allows us to describe the effect of stress on the domain walls motion and gyromagnetism.; L'effet des contraintes sur le comportement magnétique des ferrites est étudié de façon expérimentale (essai en traction et compression sur des éprouvettes parallélépipédiques et de compression sur des tores) et par une modélisation multiéchelle. Les mesures réalisées permettent de caractériser à la fois les comportements magnétique et magnétostrictif du ferrite étudié. Le modèle multiéchelle développé repose sur trois échelles : le domaine magnétique, le grain et le polycristal. A l'échelle du domaine magnétique, une minimisation de l'énergie totale permet de calculer l'orientation de l'aimantation sous les effets couplés de la contrainte mécanique, du champ magnétique appliqué et de l'orientation cristalline. A partir de ce calcul énergétique, une fonction de Boltzmann fournit une estimation de la fraction volumique des différents domaines du grain. Le comportement du polycristal est calculé par homogénéisation du comportement d'un nombre suffisant de grains. Le modèle ainsi élaboré permet de simuler le comportement magnétique et magnétostrictif. Les résultats sont en bonne adéquation avec les mesures expérimentales. Une modélisation du comportement sous champ magnétique haute fréquence décrivant l'effet des contraintes sur les phénomènes de relaxation de parois et de gyromagnétisme complète le modèle.

The effect of stress on the magnetic behaviour of ferrites has been investigated by the use of experiments (parallelepipedic samples submitted to compression and tensile stress, and toric samples submitted to compressive stress) and a multiscale model. The measurements allow us to estimate the magnetic and magnetostrictive behaviours of the ferrite. The numerical model is based upon three scales: the magnetic domain, the single crystal and the polycrystal. At the scale of the magnetic domain, a minimisation of the potential energy gives the direction of magnetisation under the combined effect of magnetic field, stress, and crystal orientation. At the scale of the single crystal, the volumetric fractions of domains of different orientations are computed from their respective energies through a constitutive equation (Boltzmann function). The polycrystalline behaviour is deduced from the behaviour of a sufficient number of grains thanks to an averaging operation. The model allows us to simulate the magnetic and magnetostrictive behaviours. Experimental and numerical results are in good accordance. The modelling of the magnetic behaviour under high frequency magnetic field conditions constitutes the final point to the work. It allows us to describe the effect of stress on the domain walls motion and gyromagnetism.; L'effet des contraintes sur le comportement magnétique des ferrites est étudié de façon expérimentale (essai en traction et compression sur des éprouvettes parallélépipédiques et de compression sur des tores) et par une modélisation multiéchelle. Les mesures réalisées permettent de caractériser à la fois les comportements magnétique et magnétostrictif du ferrite étudié. Le modèle multiéchelle développé repose sur trois échelles : le domaine magnétique, le grain et le polycristal. A l'échelle du domaine magnétique, une minimisation de l'énergie totale permet de calculer l'orientation de l'aimantation sous les effets couplés de la contrainte mécanique, du champ magnétique appliqué et de l'orientation cristalline. A partir de ce calcul énergétique, une fonction de Boltzmann fournit une estimation de la fraction volumique des différents domaines du grain. Le comportement du polycristal est calculé par homogénéisation du comportement d'un nombre suffisant de grains. Le modèle ainsi élaboré permet de simuler le comportement magnétique et magnétostrictif. Les résultats sont en bonne adéquation avec les mesures expérimentales. Une modélisation du comportement sous champ magnétique haute fréquence décrivant l'effet des contraintes sur les phénomènes de relaxation de parois et de gyromagnétisme complète le modèle.

This work is an original contribution to the study of the thermostructural composite materials thermal behaviour. It aims to develop a methodology with a new experimental device for thermal characterization adapted to this type of material and to model the heat transfer by conduction within these heterogeneous media. The first part deals with prediction of the thermal effective conductivity of stratified composite materials in the three space directions. For that, a multiscale model using a rigorous morphology analysis of the structure and the elementary properties is proposed and implemented. The second part deals with the thermal characterization at high temperature. It shows how to estimate simultaneously the thermal effusivity and the thermal conductivity. The present method is based on the observation of the heating from a plane sample submitted to a continuous excitation generated by Joule effect. Heat transfer is modelled with the quadripole formalism, temperature is here measured on two sides of the sample. The development of both resistive probes for excitation and linear probes for temperature measurements enables the thermal properties measured up to 1000°C. Finally, some experimental and numerical application examples lead to review the obtained results.; Ce travail apporte une contribution originale à l'étude du comportement thermique des composites thermostructuraux à haute température. Il se consacre à la fois au développement d'une méthodologie et d'un nouveau dispositif expérimental de caractérisation thermique par contact adapté pour ce type de matériau ainsi qu'à la modélisation des transferts de chaleur par conduction au sein de ces milieux hétérogènes. La première partie du document traite de la prédiction de la conductivité thermique effective de composites stratifiés 2D dans les trois directions de l'espace. Pour cela, une stratégie de modélisation multiéchelle, s'appuyant sur une analyse morphologique précise des matériaux étudiés et sur la connaissance de leurs propriétés élémentaires, est proposée puis appliquée. La seconde partie du document aborde la caractérisation thermique en température et démontre comment identifier simultanément l'effusivité et la conductivité thermique. La méthode développée est fondée sur l'observation de l'échauffement d'un échantillon plan soumis à un créneau de flux généré par Effet Joule. L'écriture du modèle direct (1-D) fait appel au formalisme des quadripôles thermiques, les observables sont les températures en face avant et en face arrière de l'échantillon. La mise au point de sondes résistives et de capteurs de température linéiques, revêtus de faible épaisseur de céramique isolante permet la réalisation de mesures jusqu'à 1000°C. Enfin, divers exemples d'applications expérimentales et numériques conduisent à une analyse critique de l'ensemble des résultats

This work is an original contribution to the study of the thermostructural composite materials thermal behaviour. It aims to develop a methodology with a new experimental device for thermal characterization adapted to this type of material and to model the heat transfer by conduction within these heterogeneous media. The first part deals with prediction of the thermal effective conductivity of stratified composite materials in the three space directions. For that, a multiscale model using a rigorous morphology analysis of the structure and the elementary properties is proposed and implemented. The second part deals with the thermal characterization at high temperature. It shows how to estimate simultaneously the thermal effusivity and the thermal conductivity. The present method is based on the observation of the heating from a plane sample submitted to a continuous excitation generated by Joule effect. Heat transfer is modelled with the quadripole formalism, temperature is here measured on two sides of the sample. The development of both resistive probes for excitation and linear probes for temperature measurements enables the thermal properties measured up to 1000°C. Finally, some experimental and numerical application examples lead to review the obtained results.; Ce travail apporte une contribution originale à l'étude du comportement thermique des composites thermostructuraux à haute température. Il se consacre à la fois au développement d'une méthodologie et d'un nouveau dispositif expérimental de caractérisation thermique par contact adapté pour ce type de matériau ainsi qu'à la modélisation des transferts de chaleur par conduction au sein de ces milieux hétérogènes. La première partie du document traite de la prédiction de la conductivité thermique effective de composites stratifiés 2D dans les trois directions de l'espace. Pour cela, une stratégie de modélisation multiéchelle, s'appuyant sur une analyse morphologique précise des matériaux étudiés et sur la connaissance de leurs propriétés élémentaires, est proposée puis appliquée. La seconde partie du document aborde la caractérisation thermique en température et démontre comment identifier simultanément l'effusivité et la conductivité thermique. La méthode développée est fondée sur l'observation de l'échauffement d'un échantillon plan soumis à un créneau de flux généré par Effet Joule. L'écriture du modèle direct (1-D) fait appel au formalisme des quadripôles thermiques, les observables sont les températures en face avant et en face arrière de l'échantillon. La mise au point de sondes résistives et de capteurs de température linéiques, revêtus de faible épaisseur de céramique isolante permet la réalisation de mesures jusqu'à 1000°C. Enfin, divers exemples d'applications expérimentales et numériques conduisent à une analyse critique de l'ensemble des résultats

International audience; In this paper, we present a two-level physiological model that is able to reflect morphology and function of vascular networks, in clinical images. Our approach results from the combination of a macroscopic model, providing simulation of the growth and pathological modifications of vascular network, and a microvascular model, based on compartmental approach, which simulates blood and contrast medium transfer through capillary walls. The two-level model is applied to generate biphasic computed tomography of hepatocellular carcinoma. A contrast-enhanced sequence of simulated images is acquired, and enhancement curves extracted from normal and tumoral regions are compared to curves obtained from in vivo images. The model offers the potential of finding early indicators of disease in clinical vascular images.

International audience; In this work a three-scale model of the mechanical behaviour of titanium aluminides is presented. The first scale transition between the macroscopic length scale and the scale of an individual $α_2$, $γ$ or lamellar $α_2 + γ$ grain is made using elastically self-consistent transformation field analysis (TFA) with anisotropic elasticity. The constitutive equations of the $α_2$ and $γ$-phases are obtained through the framework of crystal plasticity. The effective behaviour of the lamellar $α_2–γ$ phase, however, necessitates a second scale transition. This transition, derived from a system of linear equations for a multilayer, is obtained and shown to be equivalent to the formalism of transformation field analysis. Some simple simulations are presented to show the viability of the method.