Les céramiques à base d’Al2O3 et de 3Y-ZrO2 à structures submicroniques présentent, au vu de leurs propriétés physiques et mécaniques, un fort potentiel pour l’usinage de composites en bois. Neuf nuances ont été développées par coulage-gélification (CG) et par pressage isostatique à froid (CIP). Elles ont ensuite été testées en usinage dans des conditions industrielles de coupe. L’objectif de la thèse a été la compréhension, à partir de ces résultats, des mécanismes d’usure, afin d’identifier les propriétés physiques et mécaniques nécessaires pour l’usage d’un matériau nouveau pour cette application. Des outils prototypes de fraisage et de tournage ont été conçus. L’état d’endommagement et les contraintes résiduelles dans la céramique, dus aux étapes de fabrication des prototypes, ont été déterminés. Les résultats des expérimentations de coupe montrent que différents mécanismes opèrent sur l’usure des céramiques. Pour les nuances d’alumine, l’usure de l’arête se fait par arrachement cyclique de grains, tandis que les nuances à base de zircone se déforment plastiquement. Il apparaît que les propriétés physiques, telles que la taille de grain et la densité, sont prédominantes sur les propriétés mécaniques. L’utilisation de céramiques à structure submicronique avec des densités supérieures à 99 % améliore les performances en usinage. C’est la nuance Al2O3-10 % ZrO2 produite par CG qui a montré la meilleure résistance à l’usure. Lors de l'affûtage et de la coupe, l'oxyde de zirconium non stabilisé opère un changement de phase qui induit des contraintes de compression au sein de la microstructure et augmente ainsi la résistance micromécanique des joints de grain. / High mechanical properties of sub-μm grain Al2O3 and 3Y-ZrO2 present high potential for the cutting of wood-based materials. Nine grades were produced by gelcasting (GC) and cold isostatic pressing (CIP). They were tested in real industrial conditions. From those results, this thesis’ aim was to highlight the wear mechanisms in cutting, and then deduce major physical and mechanical properties that are task of material development. Consequently, prototypes for cutting trials have been developed. As a result, the manufacturing of prototype tools, X-Ray diffraction, analytical and numerical investigations were performed to quantify residual stress that occurred in ceramic. Observation of the results of experimentations in machining showed different wear modes. For alumina’s grade, single grain pull-out was the main wear mechanism of the cutting edge, while microplastic deformation occurred in zirconia’s composition. Physical properties of ceramic structure, like grain size and density, are more important than mechanical properties. To improve cutting lifetime, density must be higher than 99%, with a structure with sub-μm grain size. It appeared that there was a predominant role of micromechanical stability while cutting abrasive materials. Fracture toughness or thermal shock resistance seem to have marginal influence when ceramic were produced by low-defect process like GC. Regarding all ceramic grades the Al2O3-10ZrO2 made by GC showed the highest wear resistance in machining. Unstabilized zirconia brought micromechanical stability between grains of the cutting edge while working.

L'évolution du contexte industriel pousse l'industrie du transport, et plus particulièrement le secteur automobile, à réaliser des gains de de masse. Ceci passe, pour partie, par le développement de nouvelles solutions en matériaux composites. Le présent travail de thèse est consacré à la caractérisation mécanique et à la modélisation micromécanique d'un nouveau matériau composite SMC renforcé de mèches de fibres de carbone. L'objectif est de constituer une première base de connaissances sur le comportement de ce SMC en fatigue. Les investigations expérimentales passent notamment par l'analyse de la microstructure, la caractérisation du comportement mécanique sous sollicitation quasi-statique et de fatigue ainsi que par l'analyse des modes de dégradations. L'approche multi-échelle développée prend en compte la microstructure du matériau aux deux échelles mises en évidence à travers deux homogénéisations successives par une méthode Mori-Tanaka. Cette stratégie de modélisation permet de relier le comportement des fibres et le comportement élasto-plastique de la matrice à travers une loi de comportement dédiée à celui du matériau composite, et enfin d'intégrer la distribution d'orientation des mèches induites par le procédé de thermocompression.Le modèle multi-échelle a été identifié par une méthode inverse à partir des bases de données expérimentales constituées lors des travaux. La loi constitutive globale, à l'échelle d'un volume élémentaire représentatif, a été implémentée dans la bibliothèque scientifique SMART+ en langage C++ et a été conçue pour être compatible dans le cadre d'analyse de structures par éléments finis. En régime non-linéaire intégrant l’endommagement. / The evolution of the industrial context is pushing the transport industry, and more specifically the automotive sector, toward better energy efficiency. This objective is partly achieved by the development of new composite material solutions. The current work is devoted to the mechanical characterization and the multiscale modeling of a new SMC composite material reinforced with short carbon fibers bundles. The objective is to build a first knowledge base on the behavior of this SMC under fatigue loading. Experimental investigations include the analysis of the microstructure, the characterization of the mechanical behavior under quasi-static and fatigue loading, as well as the analysis of the damage mechanisms. The proposed multiscale approach takes into account the microstructure of the composite at the two scales highlighted through two successive homogenizations by the means of a Mori-Tanaka based method. Such modelling strategy makes it possible to relate elasto-plastic behavior of the matrix through dedicated local constitutive laws and the behavior of the fibers to that of the composite material, and to integrate the orientation distribution of the bundles induced by the manufacturing process... The multiscale model was identified by the mean of an inverse computation method applied on the experimental results taken from the databases created during the works.The global constitutive law, computed at the scale of a representative elementary volume, has been implemented in the SMART+ scientific library in C++ language in such a manner to be compatible in a finite element analysis (FEA) framework.

Le domaine de la fusion par confinement magnétique et la technologie des tokamaks privilégient l'utilisation de câbles supraconducteurs pour la circulation de haute densité de courant (typiquement 70kA). Ces conducteurs se composent de milliers brins câblés sur plusieurs étages, insérés en conduite et refroidis à température cryogénique. Les chargements extrêmes d’origine thermique et électromagnétique appliqués sont susceptibles d’engendrer des déformations locales dégradant les qualités conductrices de ces câbles. Ces dégradations d'origine mécanique sont observées mais peu de modèles en décrivent le comportement. Dans le cadre de la modélisation des conducteurs de type câble-en-conduit (CICCs), nous proposons l’utilisation et l’adaptation d’un code élément fini de calcul de structures dédié aux milieux entremêlés: Multifil. Le problème du comportement mécanique global et local des câbles est posé sous la forme de la recherche d’équilibre d’un assemblage de poutres en grande déformation et en interaction de contact frottement sous différents chargements. Le développement de cette méthode numérique s’appuie sur un modèle de poutre à cinématique enrichie, intégrant notamment des déformations de sections et sur une résolution algorithmique implicite de type Newton-Raphson. Cette modélisation est largement basée sur les travaux déjà réalisés quant à l’étude de la mécanique interne de câble métallique et de textile où le traitement des contacts et primordiale. Pour ce qui est de l’application de Multifil aux modèles de CICCs, les adaptations et nouveaux développements du code de calcul ont été introduits lors de cette thèse.En premier lieu, la géométrie des câbles formés étant a priori inconnue, nous proposons une simulation du processus de mise en forme des conducteurs. Celle-ci consiste en une compaction d’une configuration théorique de câblage initiale au moyen d’outils rigides. Une partie importante de cette thèse a été dédiée au développement de conditions aux limites, dites pseudopériodiques, pertinentes face aux problématiques liées à la modélisation des câbles. Dans un deuxième temps et dans un effort de validation des modèles de câbles formés, nous présentons les résultats obtenus lors d’essais de traction/compression longitudinale et de compression transverse de câbles formés. Les comportements mécaniques des câbles en sollicitation axiale et transversale prédits par les modèles ont montré un bon accord avec les données expérimentales disponibles. Cela a été rendu possible grâce à l’identification des lois de comportements des brins sur des essais de traction cycliques et de pincement effectués au cours de cette thèse. L’observation au Microscope Electronique à Balayage de coupe de brins sollicités a permis de caractériser un critère de ruptures pour les micro-filaments Nb3Sn des brins. Des simulations complètes des conducteurs ITER en condition opératoire ont été réalisées avec succès pour différentes configurations de câbles. L’analyse des données fournies par le code de calculs a permis de mettre en lumière l’importante hétérogénéité des déformations axiales dans les conducteurs ainsi que la présence de déformations, dites critiques, susceptibles de pouvoir expliquer en partie les dégradations des propriétés supraconductrices des conducteurs d’ITER. Enfin les données fournies par Multifil ont été utilisées par deux codes de calculs électromagnétiques, CARMEN et JackPot, dans le but de décrire les propriétés électriques des conducteurs en fonction des déformations prédites par les modèles mécaniques. / The ITER Cable-In-Conduit Conductors (CICC) are composed of an assembly of pure copper wires and composite superconducting strands (with embedded brittle Nb3Sn microfilaments) cabled together and inserted in a stainless steel jacket. If the current carrying capacities of individual ITER strand are clearly identified, by a dependence of the critical current on the applied strain and by a statistical quantification of possible microfilaments breakage, the characterization of cable-in-conduit is not yet fully achieved. What are the local strain values of the strands inside CICCs under operating conditions is still an open question. A deeper understanding of how local strains develop and where critical strains appear in complex cabled structures could help to optimize CICCs designs in term of the losses of conductivity.The present work aims at providing for a finite element model of conductors, able to predict local strains, especially the bending strain, at the scale of individual strands. The finite element software, Multifil, initially developed to model various kinds of entangled media, has been adapted to consider the specific issues related to the conductors. The Multifil’s main feature is basically to handle the evolution of contact-friction interactions between wires. In this study, the initial conductors’ geometry (trajectories of all individual wires), a priori unknown, is determined by a simulation of the shaping process by means of moving rigid tools. Starting from formed cables, both the thermal restraint and the transverse Lorentz loads are simulated through successive applications of proper loading. An important part of this thesis is dedicated to the implementation in the code of proper transverse boundary conditions that are relevant to the cable modelling. Moreover, the numerical work is supported by experiments performed at ECP regarding the characterization of the axial and transverse material constitutive law of the strands of the cable. The, experimental and numerical “Force/Displacements” curves, obtained on cables under standard axial and transverse loading, show good agreement. At last, the results of the full conductor simulations (from initial shaping to magnetic loading) are presented for various conductors design. Relevant information at the scale of strands (axial strains across and along the strands of the cable) can be retrieved from these simulations. The careful analysis of these data have led to highlight the high non-uniformity of the axial strains in loaded conductor with occurrence of localized critical strains that could explain the conductivity loss observed in ITER conductors. At last, the mechanical information provided by the Multifil have been put to good use by two different electromagnetic codes, CARMEN and JackPot in order to predict the superconducting properties of the conductor according to the axial strains measured along and across the strand.

Orientadores: Rubens Caram Junior, Itamar Ferreira / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecanica / Made available in DSpace on 2018-08-11T18:56:13Z (GMT). No. of bitstreams: 1 Cremasco_Alessandra_M.pdf: 37516014 bytes, checksum: edfb74db9675b3aa63db74dea7998021 (MD5) Previous issue date: 2008 / Resumo: As ligas de titânio correspondem à classe mais versátil de materiais metálicos empregados em implantes ortopédicos. Dentre tais ligas, encontram-se as do tipo ß, que devem se tornar em breve o padrão nesse tipo de aplicação.O presente trabalho tem como objetivo a preparação e a caracterização de amostras da liga de titânio tipo ß ?de composição Ti-35Nb (% em peso) processadas sob duas condições distintas de resfriamento, rápida e lentamente, a partir do campo de temperaturas ß. Foram realizados estudos ligados à estabilidade microestrutural, ao efeito de tratamentos térmicos na microestrutura, ao comportamento mecânico, à resistência à corrosão e finalmente, à resistência à fadiga, em meio neutro e em meio agressivo. Observou-se que sob resfriamento lento, a microestrutura da liga Ti-35Nb é formada pela fase ß ?com precipitados da ß e de martensita a". Tratamentos térmicos de envelhecimento resultaram na transformação da martensita a" em fase ß, seguida da nucleação das fases ? ?e a. Ensaios mecânicos mostraram que o módulo de elasticidade e a dureza de ligas resfriadas lentamente foram significativamente superiores aos valores obtidos com as amostras resfriadas rapidamente. Da mesma forma, observou-se que as amostras resfriadas lentamente apresentam maior limite de resistência à tração e inferior nível de ductilidade. Ensaios de resistência à corrosão mostraram comportamento semelhante para ambas as amostras resfriadas rápida e lentamente, o que mostra que a taxa de resfriamento e conseqüentemente, as fases formadas não alteram significativamente a resistência à corrosão. Os ensaios de fadiga mostraram que as amostras processadas sob as duas condições de resfriamento apresentaram limite de resistência à fadiga inferior ao da liga Ti-6Al-4V. A liga resfriada rapidamente apresentou valor maior que a mesma liga resfriada em forno. Os valores de resistência à fadiga em solução de 0,9% NaCl de amostras resfriadas rápida e lentamente exibiram leve tendência de redução em relação ao meio neutro. / Abstract: Titanium alloys form the most versatile class of metallic materials used as biomaterials. Among them, one finds the ß ?type alloy, which must soon become the standard metallic materials for orthopedic applications. The aim of this work is to prepare and characterize a type ß ?titanium alloy containing 35 wt.% Nb. Samples of such alloy were rapidly and slowly cooled from the ß ?field temperatures. This work includes studies of microstructure stability, effects of heat treatment on the microstructure, evaluation of corrosion resistance and finally, analysis of fatigue resistance in non-aggressive and aggressive media. The results obtained allowed one to observe that microstructure of slowly cooled samples is formed by precipitates of ß ?and a phases in a ? ?matrix. Under rapid cooling, the microstructure consists of ß ?phase and a" martensite. Aging heat treatments resulted in the transformation of a" martensite into ß ?phase, followed by ? ?phase precipitation and hence, a ?phase nucleation. Mechanical testing showed that the elastic modulus and Vickers hardness of slowly cooled samples were significantly higher than that obtained in the case of the rapidly cooled samples. On the other hand, it was observed that slowly cooled samples showed higher tensile strength and lower ductility. Corrosion resistance testing led to the conclusion that both rapidly and slowly cooled samples have similar corrosion behavior, which allows one to conclude that cooling rate and consequently, the phases in the microstructure do not have a relevant influence on the corrosion resistance. Fatigue testing showed that both types of sample (rapidly and slowly cooled samples) have fatigue resistance lower than that of the Ti-6Al- 4V alloy. The rapidly cooled sample showed fatigue resistance higher than that of slowly cooled samples. The fatigue resistance measured in a NaCl (0,9% vol.) solution of both types of samples showed a slight tendency of reduction when compared to the same samples tested in nonaggressive medium. / Mestrado / Materiais e Processos de Fabricação / Mestre em Engenharia Mecânica

Na indústria automobilística há uma grande preocupação em relação ao custo e ao peso das peças utilizadas em projetos e os adesivos vem sendo utilizados para substituir a solda ponto com esta finalidade. Neste contexto o adesivo epóxi estrutural Betamate 1496 vem sendo empregado pela Volkswagen do Brasil S.A para substituição de solda a ponto na fabricação da porta dianteira do veículo Pólo Hatch. O adesivo Betamate 1496 é utilizado na fabricação da carroçaria do veículo Pólo Hatch da Volkswagen do Brasil SA com essa finalidade. O objetivo do presente trabalho é propor e testar uma metodologia para avaliar o estado de cura do adesivo epóxi Betamate 1496 por meio da determinação do seu módulo de elasticidade e de sua dureza, quando este é empregado na fabricação das portas dianteiras do veículo Pólo Hatch. Estas propriedades foram determinadas pela medição das propriedades mecânicas do adesivo em amostras retiradas em seis posições da porta em duas fases de processamento: estado pré-curado, durante a montagem da porta e estado curado, após a cura final na estufa de pintura. O estudo foi realizado através do ensaio de indentação instrumentada, utilizando os procedimentos experimentais propostos por Franco Jr. e colaboradores e Pintaúde e colaboradores na avaliação das propriedades do polímero. O estado de cura do polímero foi avaliado de modo qualitativo também pela técnica de calorimetria diferencial exploratória (Differential Scanning Calormetry). Os resultados obtidos pela metodologia proposta por Franco Jr. e colaboradores apresentaram grande dispersão de valores, provavelmente pelo fato do método propagar erros de arredondamento na aplicação do algoritmo. Conforme observado nos resultados experimentais obtidos pela metodologia proposta por Pintaúde e colaboradores, no estado pré-curado todas as amostras apresentam valores de módulo de elasticidade muito próximos, em torno de 1,0 + 0,1. GPa, independentemente da posição na porta ou do tempo de descarregamento, já no estado curado observa-se uma diferença significativa dos valores de módulo de elasticidade, E, e dureza H dependendo da posição original da amostra na porta, com as amostras retiradas das posições inferiores apresentando E=0,8 + 0,1 GPa, enquanto que as demais apresentam E=1,0+0,1 GPa, os valores de dureza, H, para todas as amostras curadas foi H=0,03 GPa. A analise qualitativa do ensaio térmico pelas as curvas obtidas no ensaio térmico de DSC mostra que as curvas da amostra no estado pré-curado e curado são idênticas, indicando o ciclo de cura do polímero já se completa na fase de montagem da porta, antes mesmo da cura final na estufa de pintura. / Cost and weight reduction are major concerns of the automobilist industry, this leads to the increasing substitution of spot weld by adhesives in the assembling of automobile parts. In this context Volkswagen do Brazil S.A uses the Betamate 1496 structural epoxy adhesive in the manufacture of the chassis of the Polo v. Hatch. The aim of the present work is to suggest and to test a methodology to evaluate the state of cure of this adhesive by means of the determination of its hardness and elastic modulus when used in the manufacture of the front doors of the Polo v. Hatch vehicle. These properties have been determined by the measurement in six samples removed from different positions at the two phases of processing: after assembling of the chassis and cured state, after final cure at painting. The study was carried through instrumented indentation testing, which used the experimental procedures suggested by Franco Jr. and collaborators and Pintaúde and collaborators. The state of cure of polymer was evaluated also in qualitative way by Differential Scanning Calorimetry. The results obtained by the Franco Jr. methodology show great dispersion of values, probable due to the fact that the method propagates rounding errors in the application of the algorithm. The experimental results obtained by the Pintaúde methodology show that, in the state after assembling of the chassis all samples present elastic modulus, E, around 1,0 + 0.1 GPa, independently of the position in the door. In the cured state is observed depending on the sample’s position at the door: samples taken from the bottom of the door shaved E=0,8 + 0,1 GPa, while the remaining showed E= 1,0 + 0,8 GPa. The hardness, H, of the adhesive is constant (H=0,03 GPa) for all samples. The comparison of DSC curves obtained from the incurred and cured samples, showed that they are very similar, suggesting that the cure cycle of the polymer is already completed at assembling stage of the door.

Fornos petroquímicos constituem a parte mais importante da indústria de derivados de petróleo. São nesses fornos que se processam as reações químicas imprescindíveis ao processamento dos produtos que constituem nosso cotidiano, como os fertilizantes, polímeros, produtos farmacêuticos e alimentícios. Devido ao caráter fortemente endotérmico dos processos, associado a reações catalíticas provocada numa mistura reacional de vapor de água com hidrocarbonetos (caso da reforma para obtenção de hidrogênio) ou craqueamento por pirólise (processo de pirólise), estes fornos são continuamente aquecidos. Com o objetivo de aumento de produção e produtividade, os fornos têm sido submetidos a condições de temperatura e pressão cada vez mais extremas. Nos fornos de pirólise, adicionalmente tem-se aumentado consideravelmente a velocidade do fluido e para tanto têm tido as seções transversais dos condutos diminuídas levando a um aumento das tensões de trabalho. Além disso, paradas e partidas desses fornos têm se mostrado um elemento crucial na operação. É comum que durante a parada desses fornos se detecte fragilização desse material após certo tempo de serviço à temperatura de trabalho. Quedas significativas nos valores de alongamento têm sido reportadas sendo que em algumas situações, componentes apresentam valores nulos de alongamento à temperatura ambiente quando ensaios mecânicos são efetuados em corpos de prova retirados dos mesmos. Devido a essa fragilização, com conseqüente diminuição nos valores de alongamento, operações de reparo por soldagem têm se mostrado freqüentemente impraticáveis resultando no sucateamento do componente. Nesse trabalho caracterizou-se a fragilização ocorrida no material 20Cr32Ni+Nb que é usado nos coletores de saída de fornos que embora operem em temperaturas mais baixas, estão sujeitos também aos mesmos requisitos de pressão do forno. Para melhor caracterizar o material em estudo, um tubo centrifugado foi analisado tomando-se o cuidado de extrair corpos de prova em regiões que propiciassem microestrutura reprodutível. Para garantir que os corpos de prova fossem submetidos a temperaturas e tempos similares com boa precisão, estes foram envelhecidos em equipamento de ensaios de fluência, porém sem aplicação de tensão. Os corpos de prova envelhecidos foram então ensaiados com tração à temperatura ambiente e o alongamento da fratura foi registrado, evidenciando a fragilização. A região da cabeça do corpo de prova foi examinada por metalografia óptica e por microscópio eletrônico de varredura (MEV). As fraturas foram igualmente examinadas no microscópio eletrônico de varredura. Os resultados obtidos na microscopia ótica, com e sem ataque, bem como no MEV não evidenciaram quaisquer alterações microestruturais significativas que justificassem a alteração do comportamento mecânico com o envelhecimento. Análises fratográficas também não evidenciaram alteração significativa no mecanismo de fratura que ocorre pela nucleação de micro-cavidades (dimples). Adicionalmente procurou-se comprovar ou rejeitar a hipótese de que a fragilização seria causada ela precipitação de um intermetálico frágil em contorno de grão: a chamada fase G (um silicieto misto de nióbio e níquel com estequiometria Ni16Nb6Si7, estrutura cúbica de faces centradas e parâmetro de rede a0 = 1,13nm). A análise dos resultados não evidencia fase nova, nem uma mudança de modo de fratura de trans para intergranular, que poderiam comprovar uma influência da fase G no processo de fragilização (ao menos até o limite de resolução das técnicas experimentais aqui empregadas). / Petrochemical furnaces play an important role in the oil industry. In these furnaces, the essential chemical reactions take place in order to get everyday products like fertilizers, polymers, products for pharmaceutic and food industry. Due to the highly endothermic characteristics of these processes, these furnaces are heated up. It is common during shutdowns and startups of these furnaces, to have fragilization of this material after an aging time at working temperatures. This fragilization impairs repair when needed. In order to have an increase in production and productivity, the furnaces are continuously submitted to higher temperature and pressure working conditions. Pyrolisis furnaces have been also subjected to an increase in fluid speed and in order to reach this, the transverse area of tubes has been significantly decreased resulting in an increase in working stresses. Besides, shutdowns and startups have been a crucial part of operation. Fragilization in this material in service at high temperatures is frequently detected during shutdowns. Significant losses in elongation have been reported and sometimes components show no elongation at room temperature when samples are subjected to tensile testing. Due to this fragilization and low elongation values, repair by welding is sometimes useless resulting in scrapping the component. In this work the characterization of the fragilization that occurs in a 20Cr32Ni+Nb centrifugally cast tube, whose application is the outlet collectors that, though operate in lower temperature, are subjected to the pressure requirements of the furnace. In order to better characterize the material, one centrifugally cast tube was analysed taking care of having the test samples from regions that could provide similar microstructures. In order to assure that the samples were submitted to temperature and time with good accuracy, they were aged in a creep testing machine, but without application of load. The aged samples were then tensile tested at room temperature and the fracture elongation was calculated, evidencing the fragilization. The head of the test sample was examined by optical metalography and Scanning Electron Microscopy (SEM). The fractured areas were also examined by SEM. The results from optical microscopy, with or without etching, as well as SEM did not show any microstructural changes that could explain the changes in mechanical behavior with during aging. Additionally one aims to confirm or reject the hypothesis of the fragilization being caused by the precipitation of an fragile intermetallics in the grain boundary: the socalled G-phase (a silicide of Niobium and Nickel with stoichiometry Ni16Nb6Si7 , face centered cubic and lattice parameter of a0 = 1,13nm). The results did not either evidence this new phase nor a change in the fracture mode from trans to intergranular which could confirm the influence of this phase in the fragilization process (at least in the resolution limits of the experimental procedures here employed).

Le dimensionnement au crash des structures automobiles en matériaux composites à renforts discontinus tels que les SMC est généralement réalisé sur la base de données expérimentales recueillies sur des matériaux vierges n’ayant subi aucun chargement depuis leur mise en œuvre. Or, les accidents proviennent après quelques années de mise en service durant lesquelles la structure composite est soumise généralement à des sollicitations de type fatigue à plus ou moins grande amplitude. La prise en compte d’un pré-endommagement éventuel en fatigue devient alors essentielle si l’on veut rester réaliste. Par ailleurs, de nouvelles formulations de matériaux composites SMC ont été récemment développées. Leur comportement en fatigue et sous sollicitation rapide étaient jusqu’à lors inexplorés. Cette étude à caractère fortement expérimentale a donc pour but d’apporter la connaissance nécessaire au dimensionnement de structures en matériaux SMC de diverses formulations, notamment le A-SMC et le LD-SMC. Le premier correspond à une matrice vinylester fortement renforcée de mèches de fibres de verre (50%). Le second correspond à une formulation proche de celle d’un SMC standard dans laquelle on a rajouté un fort taux de billes de verre creuse afin de réduire la densité. Une analyse multi-échelle permet de mettre en évidence l’influence de la microstructure sur les phénomènes d’endommagement sous sollicitation quasi-statique, dynamique et fatigue. Des méthodes originales d’analyse expérimentale sont développées afin de corréler ces mécanismes aux comportements macroscopiques observés. Les essais dynamiques rapides optimisés sont réalisés jusqu’à des vitesses de déformation de l’ordre de 80 s-1 et mettent en évidence un comportement visco-endommageable pour les deux matériaux SMC étudiés. Le décalage du seuil d’endommagement et la baisse de la cinétique d’endommagement observés à l’échelle macroscopique sont directement corrélés au seuil et cinétique des mécanismes d’endommagement observés à l’échelle locale telles que la rupture à l’interface fibre-matrice ou bille-matrice, la microfissuration de la matrice et le pseudo-délaminage entre les mèches de fibres de verre. Par ailleurs, une analyse multi-échelle du même type est également réalisée sous sollicitation de type fatigue dans laquelle les fréquences varient de 10 à 100 Hz. Une étude de l’influence des phénomènes d’auto-échauffement sur l’endommagement et le comportement du A-SMC est proposée. Enfin, une analyse originale des propriétés résiduelles sous sollicitation rapide d’échantillons préalablement fatigués à différents niveaux de fraction de durée de vie met en évidence une forte influence de l’histoire du chargement sur la sensibilité du A-SMC à la vitesse de sollicitation. L’ensemble des résultats de cette étude, de par son apport de compréhension des phénomènes mis en jeu, constitue la base expérimentale nécessaire à la construction d’outils de dimensionnement adaptés aux structures SMC sous sollicitations cycliques et dynamiques. / Classically, crash design of automotive composite structures made of discontinuous reinforcements such as SMC is usually based on experimental data obtained from virgin materials without any history of loading. However, crash events always occur after a few years of use during which composite structure is generally subjected to more or less important fatigue loading. Taking into account of a potential pre-damage in fatigue becomes essential if we want to stay realistic. Moreover, new formulations of SMC composites have been developed recently. Up to now, their fatigue and dynamic behavior were never being explored. This study, of a mostly experimental nature, aims to provide the knowledge necessary for design of structures made of various SMC formulations including A-SMC and LD-SMC. The first one is a vinylester matrix reinforced with a high content of glass fibers (50%). The second corresponds to a standard SMC formulation in which we added a high content of hollow glass micro-sphere in order to reduce density. A systematic multi-scale analysis allows to highlight the influence of the microstructure on damage phenomena under quasi-static, dynamic and fatigue loading. Original experimental methods are developed to correlate these mechanisms to the observed macroscopic behavior. Optimized dynamic tests are performed at strain rate up to 80 s-1 and highlight a visco-damageable behavior for the two SMC materials in study. Damage threshold delay and reduction of damage kinetics observed at the macroscopic scale are directly correlated to the threshold and kinetic of damage development observed at the local scale such as fiber-matrix or matrix-microsphere interfaces, matrix micro-cracking and pseudo-delamination between the bundles of glass fiber. Moreover, a similar multi-scale analysis is also performed under fatigue loading in which the frequency ranges from 10 to 100 Hz. A study of the influence of the self-heating phenomenon on the fatigue damage behavior of A-SMC is proposed. Finally, an original analysis of the remaining dynamic properties of samples which were previously submitted to fatigue loading until different levels of lifetime highlights a strong influence of the history of the loading on the strain rate sensitivity of the A-SMC. Considering the high contribution of this study in term of comprehension of the phenomena involved in the mechanical response of SMC materials, all of the experimental findings of this study provide the physical background crucial to build structural design tools for SMC subjected to cyclic and dynamic loading.

Ce travail se situe dans le cadre du projet ANR VBD ECOREB qui œuvre avec le projet national PN RECYBETON à la levée des verrous technologiques propres à l’utilisation des granulats recyclés en vue de la formulation des bétons de structure.Pour cette étude six bétons de granulats recyclés correspondant à deux classes de résistance C25/30 et C35/40 et à une classe de consistance S4 sont élaborés. Ces formulations, définies dans le cadre du PN RECYBETON, sont conçues à partir de deux formulations de référence en substituant partiellement ou complètement les granulats naturels par des matériaux issus du recyclage des bétons. Un programme expérimental a été réalisé pour déterminer les propriétés du béton à l’état frais et à l’état durci. Les résultats des essais mécaniques montrent que pour la même classe de résistance à la compression, l’incorporation des granulats recyclés dans le béton induit une diminution de la résistance à la traction et du module d'élasticité .aussi bien qu’une augmentation de la déformation au pic et de la déformation ultime. Sous compression cyclique du type charge-décharge, l’endommagement au pic de contrainte augmente lorsque le taux de substitution est important alors que la vitesse d’endommagement devient plus faible. La campagne d’essais du fluage menée sur des poutres chargées en flexion trois points indique que la cinétique du fluage est influencée par la présence des granulats recyclés et elle est d’autant plus accélérée que le taux de remplacement est important.Les résultats expérimentaux de cette étude ainsi qu’un nombre important des résultats rapportés dans les références bibliographiques ont permis d’évaluer la validité des relations analytiques développées pour prédire les propriétés mécaniques du béton des granulats naturels en vue de leur utilisation pour le béton des granulats recyclés. On montre que ces expressions nécessitent la connaissance de la résistance moyenne à la compression, fcm aussi bien qu’un coefficient qui tient compte de l’effet des granulats recyclés. La comparaison entre les courbes contrainte-déformation obtenues dans le cadre de cette thèse et certains modèles montre, qu’avec les modifications proposées, ils décrivent d’une manière satisfaisante le comportement jusqu’à la rupture. / This work is part of the ANR VBD ECOREB which aims with the project PN RECYBETON to remove the technological locks for the use of recycled aggregates for the formulation of structural concretes.For this study, six concretes of recycled aggregates corresponding to two classes of compressive strength C25 / 30 and C35 / 40 and S4 class of workability are developed. These mixtures, defined in the framework of the PN RECYBETON, are derived from two reference formulations by substituting partially or completely the natural aggregates with materials resulting from the recycling of concretes. An experimental program was carried out to determine the properties of concrete in fresh state and hardened state. The results of mechanical tests show that for the same class of compressive strength, the incorporation of recycled aggregates into the concrete induces a decrease in the tensile strength and the elastic modulus as well as an increase in the peak and ultimate strains. Under loading-unloading compression, the damage at the peak stress increases when the substitution rate is high while the damage rate becomes lower The creep tests carried out on beams loaded under three-point bending indicate that creep kinetics are influenced by the presence of recycled aggregates and is accelerated as the replacement rate is important.The experimental results of the present study together with an extensive number of results reported in the literature have allowed evaluating the current relationships used for predicting mechanical properties of recycled concrete aggregates (RAC). The validity of many analytical expressions of the stress-strain relationship has been also studied and the effect of replacement ratio was taken into account. It was pointed out that design codes relationships dedicated to assess the mechanical properties and the stress-strain compressive curve of natural aggregates concretes (NAC) are not adequate to predict the behavior of recycled aggregates concrete (RAC). It is established that these properties are related to fcm and to a parameter, which takes into account the effect of the recycled aggregates. The comparison between the stress-strain curves obtained in this thesis and some models shows that, with the proposed modifications, these models satisfactorily describe the behavior up to failure.

De par leur diversité et leur large gamme d’applications, les polymères se sont imposés dans notre environnement. Dans le cas d’applications techniques ces matériaux peuvent être exposés à des environnements agressifs conduisant à une altération de leurs propriétés. Les effets de cette dégradation sont reliés à la notion de durée de vie, c'est-à-dire au temps nécessaire pour qu’une propriété atteigne un seuil en dessous duquel le matériau devient inutilisable. Le suivi du vieillissement des matériaux polymères présente donc des enjeux importants. La spectroscopie de fluorescence est une technique qui permet d’apporter certaines réponses à ce problème. Dans le cadre de cette étude, l’accent a été porté sur l’utilisation de la spectroscopie de fluorescence pour l’étude des phénomènes intervenant lors de l’irradiation UV ou de la sollicitation mécanique d’un polymère. Dans le cas du polyéthylène haute densité, l’absence de signal fluorescent intrinsèque impose l’ajout d’un colorant. Ce colorant donnant une réponse en fluorescence dépendant de son microenvironnement, toutes modifications des chaînes du polymère engendrent un déplacement du pic de fluorescence de la sonde. Ce travail peut être séparé en deux grandes parties indépendantes, d’un coté l’influence du vieillissement UV sur la réponse fluorescente et de l’autre l’influence d’une sollicitation mécanique. Dans la première partie, l’utilisation de techniques complémentaires telles que l’IRTF ou l’AED a permis de corréler les différents résultats avec les mécanismes de vieillissement connus du polyéthylène. Les résultats obtenus dans cette partie montrent la grande sensibilité de la spectroscopie de fluorescence aux réarrangements microstructuraux intervenant dans le matériau. Dans la seconde partie, la dépendance entre la contrainte appliquée au matériau et la longueur d’onde de fluorescence a permis a partir de modèles simples d’évaluer les contraintes internes qui se développent au cours d’une sollicitation cyclique. / Due to their diversity and their wide range of applications, polymers have emerged in our environment. For technical applications, these materials can be exposed to aggressive environment leading to an alteration of their properties. The effects of this degradation are linked to the concept of life duration, corresponding to the time required for a property to reach a threshold below which the material becomes unusable. Monitoring the ageing of polymer materials constitute a major challenge. Fluorescence spectroscopy is a technique able to provide accurate information concerning this issue. In this study, emphasis was placed on the use of fluorescence spectroscopy to study the phenomena involved in either the UV radiation or mechanical stresses of a polymer. In the case of high density polyethylene, the lack of intrinsic fluorescent signal leads to the use of a dye. This dye gives a fluorescent response depending on its microenvironment. All modifications in the macromolecular chain generate a shift of the fluorescent peak. This work can be dissociated in two major parts, on one hand the influence of UV aging on the fluorescent response and in another hand the influence of mechanical stresses. In the first part, complementary analyses like FTIR or DSC are used to correlate fluorescent results with known photo degradation mechanisms. The results show the great sensibility of the technique to the microstructural rearrangement in the polymer. In the second part, the dependence between the stress and the fluorescence emission gives opportunity to evaluate internal stresses in the material during cyclic solicitations.

La recherche en biomécanique des chocs est une nécessité pour améliorer la sécurité dans les transports. Pour une meilleure évaluation des critères lésionnels lors des simulations de crash, le manque de représentativité des modèles EF du thorax humain pourrait être comblé par une démarche de personnalisation aussi bien au niveau géométrique que mécanique. Cette thèse se base sur l’étude de 18 sujets humains post-mortem. A partir des données d’imagerie, les différentes dimensions des côtes sont analysées. La corrélation de ces paramètres aboutit à la prédiction de 192 dimensions à partir d’un unique paramètre d’entrée. A une échelle inférieure, un protocole innovant a permis de coupler des informations microstructurales issues d’un μCT avec la forme extérieure des côtes. 2 hémi-thorax ont été micro-scannés afin de générer une cartographie complète des épaisseurs d’os cortical. Une stratégie a été mise en place pour proposer un algorithme prédisant l’intégralité de cette géométrie locale d’après un seul tronçon de côte. La pertinence de cette personnalisation a été évaluée par une étude de sensibilité sur des modèles EF. Les résultats d’essais de traction sur os cortical montrent un comportement différent entre les éprouvettes prélevées sur la table interne ou externe des côtes. Une caractérisation précise de la structure interne de l’os cortical, couplé à des essais de micro-traction in-situ, a pu apporter des éléments de réponse sur cette différence. Unalgorithme de personnalisation a été aussi proposé pour les propriétés mécaniques, complétant ainsi la démarche d’adapter les modèles EF du thorax à chaque individu afin d’améliorer leur biofidélité. / For a better assessment of injury criteria on the human thorax, realistic numerical simulations need accurate geometrical characterization and an understanding of the mechanical behavior of the rib. Thelack of representation of the FE models of the human thorax could be filled by a personalization of these two aspects. This thesis is based on the study of 18 post-mortem human subjects. From medical data (CT-scans), the different dimensions of the ribs were analyzed. The correlation of the measurements led to the prediction of 192 dimensions from a single input parameter. At a lower scale, an innovative protocol enabled us to combine microstructural information obtained from a μCT with the external shape of the ribs. 2 hemi-thoraxes were scanned to generate a complete map of the thickness of cortical bone and cross-section area evolution. A strategy was implemented to provide an algorithm, predicting this entire local geometry from a single rib’s sample. The relevance of this customization was evaluated by a sensitivity analysis on FE models. The results of tensile tests on cortical bone showed different behaviors between the samples harvested from the inner or outer side of the rib. A precise characterization of the internal structure of the cortical bone, coupled with in-situ micro-tensile device, revealed certain answers about this difference. An algorithm is also proposed topersonalize the mechanical properties, completing the approach of adapting the FE models of the thorax of each individual to improve their biofidelity.