Le risque individuel de rupture des anévrismes cérébraux est un enjeu majeur dans la prise en charge clinique des anévrismes asymptomatiques. La rupture anévrismale se produit lorsque la contrainte intra-pariétale dépasse la contrainte à rupture du matériau composant la paroi. Notre étude a pour objectif d'être un pas vers une nouvelle mesure biomécanique du risque individuel de rupture des anévrismes cérébraux. Dans un premier temps, une étude expéri- mentale fut menée pour caractériser le comportement biomécanique de la paroi anévrismale sur 16 échantillons d'anévrismes prélevés chirurgicalement. L'expérimentation sur les échan tillons de poche anévrismale a permis de dégager trois grandes classes de tissus pour chaque sexe (homme et femme) : souple, rigide et intermédiaire. Il apparaît que tous les anévrismes non rompus appartiennent à la catégorie rigide ou intermédiaire et que tous les anévrismes rompus correspondent à la catégorie souple. Ceci permet de mettre en évidence une corrélation entre le risque de rupture et les propriétés du matériau composant la paroi anévrismale. Dans un deuxième temps, des simulations d'interaction fluide/structure (FSI) ont été réalisées pour comparer les déformations d'un anévrisme " patient spécifique " constitué d'un matériau dégradé et non dégradé. Les résultats montrent que les propriétés du matériaux ont un impact majeur sur l'ampleur de la variation de volume anévrismale diastolosystolique. Les changements en terme de variations de volume en fonction des caractéristiques du tissu sont potentiellement visualisables à l'aide de l'imagerie médicale. Une analyse des incertitudes des paramètres est aussi présentée et montre la robustesse des résultats aux incertitudes des données d'entrée. Il a ensuite été démontré sur 12 cas " patient-spécifique " d'anévrismes différents (forme, taille, localisation et conditions aux limites différentes) qu'il existe toujours une différence significative en terme de variation de volume au cours du cycle cardiaque entre un anévrisme dont la paroi est composé d'un matériau rigide et d'un matériau souple. Cette étude suggère donc que la variation de volume anévrismale pourrait être utilisée comme une base pour une évaluation individuelle du risque de rupture des anévrismes cérébraux. / The individual risk of rupture of cerebral aneurysm is a major stake in the clinical treatment. The aneurismal rupture occurs when the intra-parietal stress exceeds the rupture stress of the material of the aneurismal wall. The goal of our study is to be a step toward a new biomechanical measure of an individual risk of rupture of cerebral aneurysm. First, an experimental study was performed to characterize the biomechanical behavior of the aneurismal wall on 16 samples of aneurysms removed by neurosurgery. The experimentation on the samples allowed us to reach three main categories of tissues for each sex (female and male): soft, intermediate and stiff. All the unruptured aneurysms belong to the stiff category or the intermediate category and all the ruptured aneurysms belong to the soft category. This is allowed us to give prominence to the correlation between the risk of rupture and the properties of the material of the aneurismal wall. Then, Fluid/Structure interaction computations (FSI) were performed to compare the strain of a “patient-specific” aneurysm composed of a degraded and undegraded material. The results show that the properties of the material have a major impact on the scope of the aneurismal volume variation over the cardiac cycle. The volume variation changes depending on the properties of the tissue are potentially viewable by medical imaging. A study of the uncertainties of the parameters is also proposed and shows the robustness of the results. We also demonstrated on 12 cases of “patient-specific” aneurysms that a significant difference stiff exists in terms of volume variation over the cardiac cycle between an aneurysm composed of a stiff and a soft material. This study suggests that the aneurismal volume variation could be used as a basis for an evaluation of the individual risk of rupture of cerebral aneurysms.

This work aims at dimensional reduction of nonlinear material models in an asymptotically accurate manner. The three-dimensional(3-D) nonlinear material models considered include isotropic, orthotropic and dielectric compressible hyperelastic material models. Hyperelastic materials have potential applications in space-based inflatable structures, pneumatic membranes, replacements for soft biological tissues, prosthetic devices, compliant robots, high-altitude airships and artificial blood pumps, to name a few. Such structures have special engineering properties like high strength-to-mass ratio, low deflated volume and low inflated density. The majority of these applications imply a thin shell form-factor, rendering the problem geometrically nonlinear as well. Despite their superior engineering properties and potential uses, there are no proper analysis tools available to analyze these structures accurately yet efficiently. The development of a unified analytical model for both material and geometric nonlinearities encounters mathematical difficulties in the theory but its results have considerable scope. Therefore, a novel tool is needed to dimensionally reduce these nonlinear material models. In this thesis, Prof. Berdichevsky’s Variational Asymptotic Method(VAM) has been applied rigorously to alleviate the difficulties faced in modeling thin shell structures(made of such nonlinear materials for the first time in the history of VAM) which inherently exhibit geometric small parameters(such as the ratio of thickness to shortest wavelength of the deformation along the shell reference surface) and physical small parameters(such as moderate strains in certain applications). Saint Venant-Kirchhoff and neo-Hookean 3-D strain energy functions are considered for isotropic hyperelastic material modeling. Further, these two material models are augmented with electromechanical coupling term through Maxwell stress tensor for dielectric hyperelastic material modeling. A polyconvex 3-D strain energy function is used for the orthotropic hyperelastic model. Upon the application of VAM, in each of the above cases, the original 3-D nonlinear electroelastic problem splits into a nonlinear one-dimensional (1-D) through-the-thickness analysis and a nonlinear two-dimensional(2-D) shell analysis. This greatly reduces the computational cost compared to a full 3-D analysis. Through-the-thickness analysis provides a 2-D nonlinear constitutive law for the shell equations and a set of recovery relations that expresses the 3-D field variables (displacements, strains and stresses) through thethicknessintermsof2-D shell variables calculated in the shell analysis (2-D). Analytical expressions (asymptotically accurate) are derived for stiffness, strains, stresses and 3-D warping field for all three material types. Consistent with the three types of 2-D nonlinear constitutive laws,2-D shell theories and corresponding finite element programs have been developed. Validation of present theory is carried out with a few standard test cases for isotropic hyperelastic material model. For two additional test cases, 3-Dfinite element analysis results for isotropic hyperelastic material model are provided as further proofs of the simultaneous accuracy and computational efficiency of the current asymptotically-correct dimensionally-reduced approach. Application of the dimensionally-reduced dielectric hyperelastic material model is demonstrated through the actuation of a clamped membrane subjected to an electric field. Finally, the through-the-thickness and shell analysis procedures are outlined for the orthotropic nonlinear material model.