L'objectif principal est de faire premiers pas vers la conception topologique de structures hétérogènes à comportement non-linéaires. Le deuxième objectif est d’optimiser simultanément la topologie de la structure et du matériau. Il requiert la combinaison des méthodes de conception optimale et des approches de modélisation multi-échelle. En raison des lourdes exigences de calcul, nous avons introduit des techniques de réduction de modèle et de calcul parallèle. Nous avons développé tout d’abord un cadre de conception multi-échelle constitué de l’optimisation topologique et la modélisation multi-échelle. Ce cadre fournit un outil automatique pour des structures dont le modèle de matériau sous-jacent est directement régi par la géométrie de la microstructure réaliste et des lois de comportement microscopiques. Nous avons ensuite étendu le cadre en introduisant des variables supplémentaires à l’échelle microscopique pour effectuer la conception simultanée de la structure et de la microstructure. En ce qui concerne les exigences de calcul et de stockage de données en raison de multiples réalisations de calcul multi-échelle sur les configurations similaires, nous avons introduit: les approches de réduction de modèle. Nous avons développé un substitut d'apprentissage adaptatif pour le cas de l’élasticité non-linéaire. Pour viscoplasticité, nous avons collaboré avec le Professeur Felix Fritzen de l’Université de Stuttgart en utilisant son modèle de réduction avec la programmation parallèle sur GPU. Nous avons également adopté une autre approche basée sur le potentiel de réduction issue de la littérature pour améliorer l’efficacité de la conception simultanée. / High-performance heterogeneous materials have been increasingly used nowadays for their advantageous overall characteristics resulting in superior structural mechanical performance. The pronounced heterogeneities of materials have significant impact on the structural behavior that one needs to account for both material microscopic heterogeneities and constituent behaviors to achieve reliable structural designs. Meanwhile, the fast progress of material science and the latest development of 3D printing techniques make it possible to generate more innovative, lightweight, and structurally efficient designs through controlling the composition and the microstructure of material at the microscopic scale. In this thesis, we have made first attempts towards topology optimization design of multiscale nonlinear structures, including design of highly heterogeneous structures, material microstructural design, and simultaneous design of structure and materials. We have primarily developed a multiscale design framework, constituted of two key ingredients : multiscale modeling for structural performance simulation and topology optimization forstructural design. With regard to the first ingredient, we employ the first-order computational homogenization method FE2 to bridge structural and material scales. With regard to the second ingredient, we apply the method Bi-directional Evolutionary Structural Optimization (BESO) to perform topology optimization. In contrast to the conventional nonlinear design of homogeneous structures, this design framework provides an automatic design tool for nonlinear highly heterogeneous structures of which the underlying material model is governed directly by the realistic microstructural geometry and the microscopic constitutive laws. Note that the FE2 method is extremely expensive in terms of computing time and storage requirement. The dilemma of heavy computational burden is even more pronounced when it comes to topology optimization : not only is it required to solve the time-consuming multiscale problem once, but for many different realizations of the structural topology. Meanwhile we note that the optimization process requires multiple design loops involving similar or even repeated computations at the microscopic scale. For these reasons, we introduce to the design framework a third ingredient : reduced-order modeling (ROM). We develop an adaptive surrogate model using snapshot Proper Orthogonal Decomposition (POD) and Diffuse Approximation to substitute the microscopic solutions. The surrogate model is initially built by the first design iteration and updated adaptively in the subsequent design iterations. This surrogate model has shown promising performance in terms of reducing computing cost and modeling accuracy when applied to the design framework for nonlinear elastic cases. As for more severe material nonlinearity, we employ directly an established method potential based Reduced Basis Model Order Reduction (pRBMOR). The key idea of pRBMOR is to approximate the internal variables of the dissipative material by a precomputed reduced basis computed from snapshot POD. To drastically accelerate the computing procedure, pRBMOR has been implemented by parallelization on modern Graphics Processing Units (GPUs). The implementation of pRBMOR with GPU acceleration enables us to realize the design of multiscale elastoviscoplastic structures using the previously developed design framework inrealistic computing time and with affordable memory requirement. We have so far assumed a fixed material microstructure at the microscopic scale. The remaining part of the thesis is dedicated to simultaneous design of both macroscopic structure and microscopic materials. By the previously established multiscale design framework, we have topology variables and volume constraints defined at both scales.

Composite materials show heterogeneity at different length scales. hence concurrent multiscale analysis is the only reliable method to analyze them. But unfortunately there is no concurrent multi-scale strategy that is efficient, and accurate while addressing all kinds of problems. This lack of reliability is partly because there is no micro-mechanical model which inherently keeps all relevent global information with it. This thesis tries to fill this gap. The presented micro-mechanical model not only homogenizes the micro-structure but also keeps the global information with it. Most of the micro-mechanical models in the literature extract the Representative Volume Element (RVE) from the continuum for analysis which results in loss of information and accuracy. In the present approach also, the RVE has been extracted from the continuum but with the major difference that all the macro/meso-scopic parameters are accounted for. Five macro/meso-scopic one dimensional parameters have been defined which completely define the effect of continuum. 11 for one dimensional stretch, _1 for torsion, __ (_ = 2, 3) for bending and _33 for uniform pressurization due to the presence of the continuum. Further, the above macro/meso-scopic parameters are proven, by the asymptotic, theory to be constant at a cross section but vary, in general, over the length of the fiber. Hence, the analysis is valid for any location and is not restricted to any local domain. Three major problems have been addressed: • Homogenization and analysis of RVE without any defects • Homogenization and analysis of RVE with fiber-matrix de-bonding • Homogenization and analysis of RVE with radial matrix cracking. Variational Asymptotic Method (VAM) has been used to solve the above mentioned problems analytically. The results have been compared against standard results in the literature and against 3D FEA. At the end, results for “Radial deformation due to torsion” problem will be presented which was solved “accidentally.”

Epithelial tissues, one of the four primary tissue structures found in our human body, are known to comprise of tiny cells interconnected in a unique continuous pattern. In most cases, they serve a dual purpose of protecting the internal organs from physical damage, and at the same time, enable in facilitating inter-cellular activities and prevent pathogen break ins. While the tissue mechanics and its proliferation have been scrutinized to great detail, it is their geometric uniqueness, that has remained more or less unexplored. With an intent of doing the same, this thesis identifies and explores those geometric properties/parameters that have an influence on the micro structure’s homogenized and localized response. However, it does so by extracting the microstructures profile and representing its cell edges via three dimensional beam elements - hence the name, bio-inspired structures. The analysis is carried out by first developing a staggered Representative Volume Element (RVE)using finite elements, and identifying its appropriate size. The staggered assembly aids in minimizing boundary effects from creeping in, and at the same time, provides the requisite statistical homogeneity. This is followed by the geometry study. A wide range of epithelial geometries are considered for the study, ranging from completely isotropic skin models, to in plane anisotropic cuboidal structures and out of plane anisotropic stratified geometries. The effects of orientation, relative density and edge length are extracted and studied in great detail. It is observed that cell edges initial orientation has a direct dependence on the particle distribution, whereas the change in orientation is largely dependent on the deformation the microstructure is subjected to. Relative density is documented to show a direct correlation to a materials homogenized response i.e. larger the relative density, greater is the microstructures stiffness and homogenized stress response to the same deformation. Edge length, on the other hand is observed to showcase a downward trend on the cell edge’s axial stress. On average, in any kind of distribution and any kind of deformation, smaller cell edges are known to showcase larger stresses, as compared to the larger cell edges.

Im Spritzguss hergestellte Bauteile mit Kurzfaserverstärkung weisen ein niedriges Gewicht bei hoher Steifigkeit auf und bieten damit beispielsweise in der Automobilbranche eine Alternative zu Bauteilen aus konventionellen Werkstoffen wie Stahl. Die Eigenschaften der Kunststoffbauteile sind das Resultat einer vielschichtigen Prozessgeschichte. Dabei erfährt das Material einen hohen Wärmeaustausch, wechselt seine Phase von flüssig zu fest, kühlt lokal unterschiedlich schnell ab und wird von den Orientierungen der eingebetteten Kurzfasern geprägt. Da die Bauteileigenschaften eine hohe Sensitivität gegenüber Variationen der Prozessparameter besitzen, sollen Simulationen des Fertigungsprozesses kostengünstige Vorhersagen zur Güte des Endproduktes ermöglichen. Den Simulationen liegen mathematische Gleichungen zu Grunde, die das effektive Materialverhalten beschreiben. Die vorliegende Arbeit beschäftigt sich mit der Formulierung eines solchen Materialmodells. Mit Hilfe von Homogenisierungen repräsentativer Volumenelemente wird zunächst der Einfluss der Faserorientierungsverteilung auf die mechanischen und thermischen Eigenschaften analysiert. Die daraus gewonnenen Erkenntnisse fließen anschließend in die Modellierung des Materialverhaltens ein. Der in dieser Arbeit verwendete Modellierungsrahmen ist für große Deformationen ausgelegt, berücksichtigt den Phasenübergang sowie Temperaturabhängigkeiten in den viskoelastischen Steifigkeitsanteilen und stützt sich auf eine effektive Integrationsregel, um die Faserorientierungsverteilung einzubeziehen. Die Identifikation der Materialparameter geschieht mit Hilfe von Experimenten an Proben mit unidirektionaler Faserausrichtung. Das identifizierte Materialmodell wird schließlich in die kommerzielle Finite-Elemente-Umgebung Abaqus implementiert und steht damit Simulationen der Abkühlung und der Beanspruchung eines spritzgegossenen Kettenglieds zur Verfügung.

Wood is a complex organic orthotropic viscoelastic material with acellular structure. When stressed, wood will deform over timethrough a process called creep. Creep affects all wooden structureand can be difficult, time-consuming and expensive to measure. For this thesis, a simple computer model of the woodenmicrostructure was developed. The hypothesis was that the modelledmicrostructure would display similar elastic and viscoelasticproperties as the macroscopic material. The model was designed by finding research with cell geometries ofconiferous trees measured. The model considered late- and earlywoodgeometries as well as growth rings. Rays were ignored as they onlycomposed 5-10% of the material. By applying a finite element method, the heterogeneous late- andearlywood cells could be homogenized by sequentially loading thestrain vector and calculating the average stress. The computer model produced stiff but acceptable values for theelastic properties. Using the standard linear solid method to modelviscoelasticity, the computer model assembled creep curvescomparable to experimental results. With the model sufficiently validated, parametric studies on thecell geometry showed that the elastic and viscoelastic propertieschanged greatly with cell shape. An unconventional RVE was alsotested and shown to give identical result to the standard RVE. Although not perfect, the model can to a certain degree predict theelastic and viscoelastic characteristics for wood given itscellular geometry. Inaccuracies were thought to be caused byassumptions and approximations when building the model.

<p>Injection to the subcutaneous (SC) tissue is one of the preferred methods for drug delivery of pharmaceuticals, from small molecules to monoclonal antibodies. Delivery to SC has become widely popular in part thanks to the low cost, ease of use, and effectiveness of drug delivery through the use of auto-injector devices. However, injection physiology, from initial plume formation to the eventual uptake of the drug in the lymphatics, is highly dependent on SC mechanics, poroelastic properties in particular. Yet, the poroelastic properties of SC have been understudied. In this thesis, I present a two-pronged approach to understanding the poroelastic properties of SC. Experimentally, mechanical and fluid transport properties of SC were measured with confined compression experiments and compared against gelatin hydrogels used as SC-phantoms. It was found that SC tissue is a highly non-linear material that has viscoelastic and porohyperelastic dissipation mechanisms. Gelatin hydrogels showed a similar, albeit more linear response, suggesting a micromechanical mechanism may underline the nonlinear behavior. The second part of the thesis focuses on the multiscale modeling of SC to gain a fundamental understanding of how geometry and material properties of the microstructure drive the macroscale response. SC is composed of adipocytes (fat cells) embedded in a collagen network. The geometry can be characterized with Voroni-like tessellations. Adipocytes are fluid-packed, highly deformable and capable of volume change through fluid transport. Collagen is highly nonlinear and nearly incompressible. Representative volume element (RVE) simulations with different Voroni tesselations shows that the different materials, coupled with the geometry of the packing, can contribute to different material response under the different kinds of loading. Further investigation of the effect of geometry showed that cell packing density nonlinearly contributes to the macroscale response. The RVE models can be homogenized to obtain macroscale models useful in large scale finite element simulations of injection physiology. Two types of homogenization were explored: fitting to analytical constitutive models, namely the Blatz-Ko material model, or use of Gaussian process surrogates, a data-driven non-parametric approach to interpolate the macroscale response.</p>

Viele systematische Überblicksarbeiten und Meta-Analysen haben sich mit den einzelnen Faktoren befasst, die zu Schulerfolg beitragen. Diese Doktorarbeit erweitert den bisherigen Forschungsstand. Es wird untersucht, wie Gewissenhaftigkeit, Neurotizismus und Selbstwirksamkeit zu Schulerfolg in der Schule beitragen. In Anlehnung an das Big Five Narrow Trait Modell (B5NT) wird angenommen, dass Selbstwirksamkeit den Zusammenhang zwischen Persönlichkeit und Schulerfolg mediiert. Außerdem wird der Einfluss verschiedener Moderatoren untersucht. Bivariate Meta-Analysen werden genutzt, um Forschungsergebnisse bezüglich der bivariaten Zusammenhänge zwischen den Variablen zu aggregieren. MASEM werden genutzt, um die Mediationshypothese zu testen. Um die Abhängigkeitsstruktur der Effektstärken zu berücksichtigen, werden Drei-Ebenen-Modelle und robuste Varianzschätzmethoden (RVE) angewandt. Insgesamt konnte eine der Hauptannahmen des Big Five Narrow Trait Modells bestätigt werden. Der Einfluss von Persönlichkeit auf Schulerfolg wird (partiell) durch Selbstwirksamkeit mediiert. Während Selbstwirksamkeit den gesamten Einfluss von Neurotizismus auf Schulerfolg mediiert, wirkt Selbstwirksamkeit als partieller Mediator in dem Zusammenhang von Gewissenhaftigkeit und Schulerfolg. Diese Ergebnisse haben praktische Relevanz für den Bildungssektor. Leistungsschwache Schüler sollten demnach vor allem Selbstwirksamkeitstrainings erhalten, um ihren Schulerfolg zu verbessern. Dies hätte den Größten Einfluss auf Schulerfolg, mit einer mittleren bis großen Effektstärke. Persönlichkeitsinventare hingegen könnten diagnostisch eingesetzt werden. So könnten Schüler identifiziert werden, die eine Unterstützung durch Selbstwirksamkeitstrainings am meisten benötigen. Gewissenhaftigkeit hängt am stärksten mit Selbstwirksamkeit zusammen. Demnach würden Schüler mit geringen Werten in Gewissenhaftigkeit am meisten von den Trainings profitieren. / The factors that individually contribute to academic achievement have been the topic of many systematic reviews. This doctoral thesis contributes to the current state of research by meta-analytically testing a comprehensive theory regarding the underlying processes through which conscientiousness, neuroticism, and self-efficacy are influencing achievement in primary and secondary school. Following the Big Five Narrow Trait (B5NT) model, it is examined whether self-efficacy mediates the relationship between personality and achievement. Also, the influence of different potential moderator variables is analyzed. Methodologically, this doctoral thesis is applying a combination of separate bivariate meta-analyses and meta-analytical structural equation modeling (MASEM). Bivariate meta-analyses are used to aggregate research on the bivariate relationships between all variables. MASEM techniques are used to test the mediation hypothesis. Three-level random-effects models and robust variance estimation methods are used to account for the dependency among effect sizes. Overall, the main idea of the B5NT model was supported by showing that self-efficacy (partially) mediates the relationship between personality and scholastic achievement. Self-efficacy fully mediated the relationship between neuroticism and achievement. Whereas the relationship between conscientiousness and achievement was only partially mediated. These findings have practical implications for the educational sector. Most important, self-efficacy trainings are recommended for low-achieving students. This would lead to the biggest improvement in scholastic achievement – with a medium-to-large effect size. Personality inventories could be used for identifying students that need support through self-efficacy trainings. Conscientiousness was having the strongest effect on self-efficacy beliefs. Thus, students with low levels of conscientiousness should be primarily targeted.

First, the theory of MSG is extended to aperiodic heterogeneous solid structures. Integral constraints are introduced to decompose the displacements and strains of the heterogeneous material into a fluctuating part and a macroscopic part, of which the macroscopic part represents the responses of the homogenized material. One advantage of this theory is that boundary conditions are not required. Consequently, it is capable of handling micro-structures of arbitrary shapes. In addition, periodic constraints can be incorporated into this theory as needed to model periodic or partially periodic materials such as textile composites. In this study, the newly developed method is employed to investigate the finite thickness effect of textile composites.<div><br></div><div>Second, MSG is enabled to deal with Timoshenko beam-like structures with spanwise heterogeneity, which provide higher accuracy than the previous available Euler–Bernoulli beam model. Its reduced form, the MSG beam cross sectional analysis, is found to be able to analyze generalized free-edge problems with arbitrary layups and subjected to general loads. In this method, the only assumption applied is that the laminate is long enough so that the Saint-Venant principle can be adopted. There is no limitation on the cross section of the laminate since no ad hoc assumption is involved with the microstructure geometry. This method solve the free-edge problem from a multiscale simulation point of view.<br></div><div><br></div>

Orientador: Gabriela Rezende Fernandes / Resumo: Neste trabalho é apresentada uma formulação do MEC (Método dos Elementos de Contorno) considerando-se não-linearidade física para analisar microestruturas de materiais heterogêneos no contexto da análise em multi-escala. A microestrutura, também denominada como EVR (Elemento de Volume Representativo), é modelada como uma chapa em sub-regiões onde vazios ou inclusões podem ser considerados dentro da matriz, sendo diferentes propriedades elásticas e modelos constitutivos definidos para cada sub-região. A equação integral para o deslocamento é obtida a partir do Teorema de Betti, onde para considerar o fenômeno dissipativo, um campo de esforços iniciais é considerado. A equação algébrica da chapa é obtida após a discretização do contorno externo e interface em elementos e do domínio das subregiões em células. Na análise multi-escala cada ponto da estrutura (macrocontínuo) é representado por um EVR, onde o comportamento do material não é definido por um modelo constitutivo, mas através da solução do problema de equilíbrio do EVR quando sujeito à deformação referente ao ponto do macrocontínuo. O problema de equilíbrio do EVR é definido em termos da flutuação dos deslocamentos, sendo o mesmo satisfeito quando seu campo de forças se encontra em equilíbrio. Após a solução do EVR, os deslocamentos no contorno e as forças dissipativas são atualizados e as forças de superfície sobre o contorno recalculadas para se obter a tensão homogeneizada. O custo computacional obtido com a presente... (Resumo completo, clicar acesso eletrônico abaixo) / Mestre

Submitted by LUIS HENRIQUE DE REZENDE CROZARIOL null (luiscrozariol@gmail.com) on 2017-09-26T22:06:47Z No. of bitstreams: 1 Dissertação Luis 26-09.pdf: 3642072 bytes, checksum: 6bd5d174cfe596f9dc2def41f1270185 (MD5) / Approved for entry into archive by Monique Sasaki (sayumi_sasaki@hotmail.com) on 2017-09-28T13:33:18Z (GMT) No. of bitstreams: 1 crozariol_lhr_me_ilha.pdf: 3642072 bytes, checksum: 6bd5d174cfe596f9dc2def41f1270185 (MD5) / Made available in DSpace on 2017-09-28T13:33:18Z (GMT). No. of bitstreams: 1 crozariol_lhr_me_ilha.pdf: 3642072 bytes, checksum: 6bd5d174cfe596f9dc2def41f1270185 (MD5) Previous issue date: 2017-07-31 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Neste trabalho é apresentada uma formulação do MEC (Método dos Elementos de Contorno) considerando-se não-linearidade física para analisar microestruturas de materiais heterogêneos no contexto da análise em multi-escala. A microestrutura, também denominada como EVR (Elemento de Volume Representativo), é modelada como uma chapa em sub-regiões onde vazios ou inclusões podem ser considerados dentro da matriz, sendo diferentes propriedades elásticas e modelos constitutivos definidos para cada sub-região. A equação integral para o deslocamento é obtida a partir do Teorema de Betti, onde para considerar o fenômeno dissipativo, um campo de esforços iniciais é considerado. A equação algébrica da chapa é obtida após a discretização do contorno externo e interface em elementos e do domínio das subregiões em células. Na análise multi-escala cada ponto da estrutura (macrocontínuo) é representado por um EVR, onde o comportamento do material não é definido por um modelo constitutivo, mas através da solução do problema de equilíbrio do EVR quando sujeito à deformação referente ao ponto do macrocontínuo. O problema de equilíbrio do EVR é definido em termos da flutuação dos deslocamentos, sendo o mesmo satisfeito quando seu campo de forças se encontra em equilíbrio. Após a solução do EVR, os deslocamentos no contorno e as forças dissipativas são atualizados e as forças de superfície sobre o contorno recalculadas para se obter a tensão homogeneizada. O custo computacional obtido com a presente formulação é menor que aquele referente ao modelo desenvolvido pelo Método dos Elementos Finitos, sendo a resposta homogeneizada do EVR comparada ao modelo de elementos finitos a fim de validar a formulação apresentada nesse trabalho. / A BEM formulation, considering dissipative phenomena, to analyze microstructures of heterogeneous materials in the context of multi-scale analysis is presented. The microstructure, also denoted as RVE (Representative Volume Element), is modelled as a zoned plate where voids or inclusions can be considered inside a matrix, being different elastic properties and constitutive models defined for each sub-region. The integral representation for displacement is obtained from Betti’s Theorem, where to consider the dissipative phenomena, an initial forces field is considered. The plate algebraic equation is obtained after discretizing the external boundary and interfaces into elements and the sub-regions domain into cells. In the multi-scale analysis, each macro-continuum point is represented by a RVE, being the material behaviour not governed by a phenomenological constitutive model, but defined after the solution of the RVE equilibrium problem due to the macro strain. The RVE equilibrium problem is defined in terms of displacement fluctuations, being satisfied when the forces field is in equilibrium. After the RVE solution, the boundary displacements and dissipative forces are updated and the boundary tractions recalculated to obtain the homogenized stress. The computational cost obtained with the proposed formulation is smaller than the formulation developed by the Finite Element Method. Besides, the homogenized response is compared to the finite element model to show its accuracy.