Les composites SiC/SiC pr´esentent d’excellentes propri´et´es thermom´ecaniques `a hautes temp´eratures. Ils apparaissent donc comme des candidats cr´edibles pour remplacer les alliages m´etalliques dans les zones chaudes de moteurs a´eronautiques civils afin d’en r´eduire l’impact environnemental. Comprendre et pr´evoir l’apparition des premiers endommagements constitue donc un enjeu industriel majeur. La d´emarche multi-´echelle propos´ee permet d’int´egrer dans des mod`eles num´eriques les sp´ecificit´es du mat´eriau. Elle s’articule autour d’une phase exp´erimentale de caract´erisation des endommagements et d’une phase de mod´elisation par ´el´ements finis aux ´echelles microscopique et m´esoscopique. Des essais in situ sous microscopes et sous micro-tomographie X (μCT) sont e↵ectu´es pour visualiser et quantifier les m´ecanismes d’endommagement `a des ´echelles compatibles avec les mod`eles num´eriques. Sur la base des observations exp´erimentales, des calculs d’endommagement sont r´ealis´es `a l’´echelle microscopique afin de simuler la fissuration transverse des torons. Des essais virtuels permettent alors d’identifier des lois d’endommagement `a l’´echelle sup´erieure et de mod´eliser l’apparition des premi`eres fissures dans des textures tiss´ees 3D `a l’´echelle m´esoscopique. Cela permet de mettre en ´evidence les liens entre l’organisation du mat´eriau aux di↵´erentes ´echelles et l’initiation des premiers endommagements. Des confrontations essais/calculs sont finalement propos´ees, en comparant notamment les sites d’amor¸cage des endommagements observ´es exp´erimentalement lors des essais in situ sous μCT avec ceux pr´edits par les simulations. / SiC/SiC composites display excellent thermomechanical properties at high temperatures. They appear as promising candidates to replace metallic alloys in hot parts of aircraft engines to reduce their environmental impact. Thus, to understand and to predict the onset of damage in such materials is critical. An integrated multi-scale approach is developed to construct numerical models that integrate the specificities of the material at the di↵erent relevant scales. This work is twofold : an experimental characterization of the damage, and finite element modeling at the microscopic and mesoscopic scales. In situ tensile tests are carried out under microscopes and X-ray micro-tomography (μCT). Images are analyzed to visualize and quantify the damage mechanisms at scales consistent with the numerical models. Based on these observations, damage calculations are performed at the microscopic scale to simulate the transverse yarns cracking. Virtual tests are then used to identify damage laws at the upper scale and to simulate the first cracks in 3D woven composites at the mesoscopic scale. Through these simulations, we highlight the links between the organization of the material at di↵erent scales and the initiation of the damages. Comparisons between experiments and calculations are finally performed. In particular, the predicted damage events are compared to those obtained experimentally on the same specimen during in situ μCT tensile tests.

<div>In contrast to traditional steel and aluminum, short fiber reinforced polymer composites (SFRCs) provide promising alternatives in material selection for automotive and aerospace applications due to their potential to decrease weight while maintaining excellent mechanical properties. However, uncertainties about the influence of complex microstructures and defects on mechanical response have prevented widespread adoption of material models for</div><div>SFRCs. In order to build confidence in models’ predictions requires deepened insight into the heterogenous damage mechanisms. Therefore, this research takes a micro-mechanics standpoint of assessing the damage behavior of SFRCs, particularly micro-void nucleation at the fiber tips, by passing information of microstructural attributes within neighborhoods of incipient damage and non-damage sites, into a framework that establishes correlations between the microstructural information and damage. To achieve this, in-situ x-ray tomography of the gauge sections of two cylindrical injection molded dog-bone specimens, composed of E-glass fibers in a polypropylene matrix, was conducted while the specimens were monotonically loaded until failure. This was followed by (i) the development of microstructural characterization frameworks for segmenting fiber and porosity features in 3D images, (ii) the development of a digital volume correlation informed damage detection framework that confines search spaces of potential damage sites, and (iii) the use of a Gaussian process classification framework to explore the dependency of micro-void nucleation on neighboring microstructural defects by ranking each of their contributions. Specifically, the analysis considered microstructural metrics related to the closest fiber, the closest pore, and the local stiffness, and the results demonstrated that less stiff resin rich areas were more relevant for micro-void nucleation than clustered fiber tips, T-intersections of fibers, or varying porosity volumes. This analysis provides a ranking of microstructural metrics that induce microvoid nucleation, which can be helpful for modelers to validate their predictions on proclivity of damage initiation in the presence of wide distributions of microstructural features and</div><div>manufacturing defects. </div>

[pt] entendimento do macro comportamento dos materiais, este trabalho apresenta soluções inovadoras para a análise de imagens 3D obtidas por microtomografia computadorizada de raios-X (microCT). O material estudado conhecido pelo termo em inglês “strain-hardening cement-based composites” ou pela abreviação SHCC é um compósito cimentício reforçado com fibras que atinge deformações significativas através da formação de múltiplas fissuras, estabelecendo um material cimentício com característica pseudo-dúctil. O primeiro desafio deste trabalho foi reconhecer e quantificar as fases constituintes nas imagens 3D de SHCC obtidas por microCT. Materiais com estruturas complexas podem apresentar imagens em que as fases internas não podem ser distinguidas pela técnica de limiarização clássica, exigindo o uso de outra técnica como a segmentação por Deep Learning (DL). Portanto, este trabalho utilizou DL como solução para esta tarefa. Desta forma, as características de cada fases puderam ser correlacionadas ao comportamento mecânico macro do material em ensaios de microCT in-situ. Outro método moderno de análise de imagens 3D utilizado foi a correlação digital de volume (em inglês, digital volume correlation - DVC). O DVC é uma técnica que estima o campo de deformação sobre todo o volume da amostra, correlacionando as imagens 3D nos estados descarregado e carregado. Assim, as imagens obtidas nos ensaios de tração e compressão in-situ puderam ter seus deslocamentos internos medidos e deformações calculadas. Em síntese, este trabalho trouxe avanços ao campo do processamento digital e análise de imagens 3D, aplicadas a materiais cimentícios, mas que também podem se adaptar à análise de diversos materiais. / [en] Considering the importance of micro and mesoscale analyses to understand the macro behavior of materials, this work brings innovative solutions for analyzing 3D images obtained by X-ray micro-computed tomography (microCT). The studied material was the strain-hardening cement-based composites (SHCC), a fiber reinforced cementitious composite that achieves significant deformations through multiple cracks formation, resulting in a cementitious material with pseudo ductile features. The first challenge of this work was to recognize and quantify the constituent phases in the 3D images of SHCC obtained by microCT. Materials with complex structures may present images in which the internal phases cannot be distinguished by the classical thresholding technique, requiring the use of another technique such as segmentation by Deep Learning (DL). Therefore, this work used DL as a solution for this task. Then, the features of each phase could be correlated to the macro mechanical behavior of the material in in-situ microCT tests. Another modern method for analyzing 3D images used was the digital volume correlation (DVC). DVC is a technique that estimates full-field strain in 3D over the entire volume of the specimen by correlating imaging volumes of the specimen in unloaded and loaded states. Thus, the images obtained from tensile and compression in-situ tests could have their internal displacements measured and strain calculated. In summary, this work brought advances to the 3D image processing and analysis field, applied to cementitious materials, but which could also adapt for the analysis of various materials.