This thesis contributes towards understanding of mechanical response of 3D composites and ceramics. Composite materials have widespread applications ranging from aerospace, civil sectors to sports and drones. One important application is in composite armours where composites and ceramic layers are used together. Therefore, it is important to study the mechanical response of these components to develop better armour systems. The first part of this thesis concerns with dynamic penetration response of confined ceramic targets. In the second part, mechanics of a novel 3D composite consisting of orthogonal carbon fibre tows is studied. The dynamic penetration of ceramic target by a long-rod projectile is studied using a mechanism based ceramic constitutive model. This is to capture and explain the essential physics observed during penetration of a ceramic target such as dwell and structural size effect. Dwell is captured using the constitutive model and the related physics is studied along with identification of causes of dwell. Origins of structural size effect in ceramics are identified and their influences are studied. In the second part of the thesis a novel 3D composite consisting of three mutually perpendicular orthogonal tows is studied under compression, indentation and three-point bending. Under compression along low fibre volume fraction direction (Z), the 3D composite forms stable and multiple kinks in the Z tows resulting in 10% ductility. This contrasts with traditional UD or 2D composites which fail catastrophically at 2% strain. The stability in the case of the 3D composite is due to the constraint imposed by the surrounding material. Under indentation, the 3D composite has a near isotropic and ductile response. In contrast, traditional cross-ply composites show highly anisotropic response where indentation results in brittle failure along in-plane direction. Under three-point bending, the response was ductile in Z-direction and brittle in other two directions. Overall, the 3D composite studied in this thesis shows improvement over traditional CFRPs in ductility and energy absorption capability. The 3D composite has been demonstrated to have smooth load-displacement curves reminiscent to indentation of metal in all three directions achieved at densities significantly lower than structural metals that display equivalent ductility. Thus, these 3D composites are strong candidates for applications where loading direction is unknown a-priori, and where high energy absorption is required along with reusability of the material.

No trabalho estuda-se o crescimento de trincas em carga monotônica e cíclica nos casos de materiais quase-frágeis, introduzindo uma lei de dano cíclico. Revisam-se conceitos sobre modelos coesivos, leis de carga-descarga, leis de evolução de dano e efeito de escala. É seguido o modelo coesivo irreversível proposto por Wang e Siegmund (2006). Em particular se dá ênfase aos efeitos de escala não estatísticos. O modelo de zona coesiva irreversível apresenta uma formulação de dano e considera carregamento em fadiga. Quando o tamanho estrutural é reduzido (ou as trinca se extendem), a fratura por fadiga não mais ocorre por propagação de trinca, mas sim por uma decoesão uniforme. O objetivo desde trabalho é implementar este modelo e verificar sua potencialidade na captura de efeitos de escala, comparando com experimentos e dados disponíveis na literatura. / At present work is intended to study crack growth in cyclic and monotonic loading in the case of quasi-brittle materials, introducing a damage mechanism, is reviewed concepts of cohesive models, loading-unloading laws, damage evolution laws and effect of scale. The irreversible cohesive zone model proposed by Wang e Siegmund (2006) is followed. In particular emphasizes in the not statistical size effects. The irreversible cohesive zone model, presents a damage formulation and considers fatigue loading. It is demonstrated in this study that, when the structure size is reduced (or extend cracks), the fatigue fracture no longer occurs by crack propagation, then occurs by uniform decohesion . The objetive of this work is implementing this model and verify its capability to capture the scale effect compared with experiments and data available in literature.

L'absence de structure cristalline dans les verres métalliques tendrait à suggérer qu'aucun effet d'échelle sur leur comportement mécanique ne pourrait exister. Cependant, la diminution de la taille des éprouvettes de verre métallique, principalement solliciter en micro-compression sur des micro-pilliers usiné au FIB à partir de BMG révèlent une transition entre des mécanismes de déformation localisés dans des bandes de cisaillement pour des échelles supérieures à 400 nm et des mécanismes qui présentent des déformations homogènes en dessous de 400 nm. Des interrogations demeurent cependant sur l'impact des procédés d'élaboration des échantillons. Dans cette thèse, on s'intéresse à une autre voie de caractérisation des effets d'échelle sur le comportement mécanique des verres métalliques par l'élaboration et la caractérisation d'éprouvettes obtenues par des procédés de microélectronique dans des films minces de verre métallique de différentes épaisseurs, sans recours à de l'usinage FIB. La structure amorphe et la composition des dépôts réalisés par MS-PVD ont été confirmées par des analyses DRX et MET. L'homogénéité des dépôts entre les différentes épaisseurs a été confirmée par l'invariance du module de Young et du module de cisaillement déterminé par diffusion Brillouin sur chaque épaisseur. Des mesures de nano indentation ont cependant révélé une diminution de la dureté, une augmentation du module de compressibilité et du coefficient de Poisson avec l'augmentation de l'épaisseur des films. L'observation des déformations d'éprouvettes de flexion et des empreintes en nano indentation confirme l'existence d'une taille critique d'éprouvette. Les relations entre les différentes propriétés mécaniques et les observations sur les effets d'échelle sont discutées à partir du modèle des STZ et d'une loi de comportement élasto-plastique parfaite pour toutes les épaisseurs. / The lack of cristalline structure in metallic glasses suggests that no mechanical size effect could be expected. However, if the sample size decreases, a transition around 400 nm in size between deformations localized in shear band and homogeneous deformations was observed mainly on micropillars machined by FIB. But there are still a few questions which remained unanswered on elaboration process and their influence on amorphous structure with the size decreasing. This thesis deals with on a new way to characterize the size effect on the mechanical behaviour of metallic glasses. Metallic glasses thin films were elaborated by MS-PVD and mechanical samples were designed thanks to microelectronic process. Homogeneity of the amorphous structure and of the composition were determined by XRD and TEM. Elastic constant such as Young modulus and shear modulus obtained by Brillouin scattering analysis are also steady whatever the thicknesses. Nevertheless, nano indentation measures showed that Poisson ratio and Bulk modulus increase when the thin films thicknesses increase. Deformations observed on nano bending sample and imprints in nano indentation show also a transition between homogeneous deformation and deformations with shear bands. The origins of these transitions between the mechanical properties are discussed from a elastic perfectly plastic model and the Shear Transition Zone model.

This paper shows the existence of the size effect in the cryptocurrency market. The size effect is a market phenomenon observed in the stock market in which smaller assets outperform larger assets. Recent literature has revealed the size effect in other financial markets as well. In order to explain the size effect, this paper proposes a general quantitative theory that supports its existence in any financial markets under specific conditions. Furthermore, the paper tests for the size effect in the cryptocurrency market using daily price data from April 2013 to April 2018. The paper finds a statistically significant size effect across the cryptocurrency market during the sample period. In the process, we test a profitable pair-trading strategy that involves opening a short position on the higher rank (larger assets) and opening a long position on the lower rank (smaller assets) of the cryptocurrency market. Based on our findings, we discuss the implications on modern finance, specifically on the subjects of Efficient Market Hypothesis and asset pricing models.

No trabalho estuda-se o crescimento de trincas em carga monotônica e cíclica nos casos de materiais quase-frágeis, introduzindo uma lei de dano cíclico. Revisam-se conceitos sobre modelos coesivos, leis de carga-descarga, leis de evolução de dano e efeito de escala. É seguido o modelo coesivo irreversível proposto por Wang e Siegmund (2006). Em particular se dá ênfase aos efeitos de escala não estatísticos. O modelo de zona coesiva irreversível apresenta uma formulação de dano e considera carregamento em fadiga. Quando o tamanho estrutural é reduzido (ou as trinca se extendem), a fratura por fadiga não mais ocorre por propagação de trinca, mas sim por uma decoesão uniforme. O objetivo desde trabalho é implementar este modelo e verificar sua potencialidade na captura de efeitos de escala, comparando com experimentos e dados disponíveis na literatura. / At present work is intended to study crack growth in cyclic and monotonic loading in the case of quasi-brittle materials, introducing a damage mechanism, is reviewed concepts of cohesive models, loading-unloading laws, damage evolution laws and effect of scale. The irreversible cohesive zone model proposed by Wang e Siegmund (2006) is followed. In particular emphasizes in the not statistical size effects. The irreversible cohesive zone model, presents a damage formulation and considers fatigue loading. It is demonstrated in this study that, when the structure size is reduced (or extend cracks), the fatigue fracture no longer occurs by crack propagation, then occurs by uniform decohesion . The objetive of this work is implementing this model and verify its capability to capture the scale effect compared with experiments and data available in literature.

Este estudo analisa a influência do porte empresarial no retorno das ações do mercado brasileiro, lidando com a sobreposição aparentemente existente entre porte da empresa e liquidez das ações e buscando elucidar qual variável é dominante na determinação de retornos acionários. Para tanto, utiliza-se uma análise de dados em painel como ferramenta econométrica para a obtenção de resultados significativos. Os modelos tradicionais de precificação de ativos, notadamente o CAPM, partem de algumas premissas que possuem pouca aderência à realidade dos mercados acionários. Há evidências de que, no Brasil, o CAPM tem sido ineficaz em explicar os retornos do mercado acionário. Por essa razão, algumas variáveis são comumente adicionadas ao modelo original, sendo que o porte empresarial se encontra no topo da lista. Contudo, a legitimidade do porte empresarial reduzido como fator de risco a ser remunerado ao acionista não é consensual em mercados mais maduros como o americano tampouco no mercado brasileiro. O estudo leva em conta um intervalo de tempo de 16 anos (de 1995 até 2011) e analisa a influência dessas variáveis no retorno das ações com diferentes defasagens de tempo entre as variáveis (de 1 a 5 anos). O trabalho conclui que há uma significativa influência do porte empresarial nos retornos das ações. A relação estatística é negativa, isto é, empresas de menor porte tendem a ter custo de capital próprio mais alto. Esse resultado está alinhado com a pesquisa internacional sobre o assunto. O modelo quantitativo sugere que a cada variação de uma unidade logarítmica no porte há uma variação de 4,03% no retorno esperado da ação. Em relação à liquidez não foi encontrada, de forma consistente, relação estatística de que essa variável afeta o retorno das ações quando controlada pelo porte empresarial no mercado brasileiro. / This study analyses the influence of company size in stock returns in Brazilian market, dealing with the apparent superposition between company size and stock liquidity and searching to elucidate which variable is dominant in determining stock returns. With that purpose, panel data analysis has been used as econometric technic, searching to obtain significant results. Traditional pricing models, mainly CAPM, are based on certain premises which have little in common with the stock markets. There is evidence that, in Brazil, CAPM has been unable to explain stock returns properly. For this reason, some variable are commonly added to the original model and the company size is at the top of the list. However, the legitimacy of the size effect as a risk factor to be paid off to the stockholder is neither consensual in more mature markets, as the American one, nor in the Brazilian market. The study takes into account a time span of 16 years (from 1995 to 2011) and analyses the influence of these variables in stock returns with different delays among variables (from 1 to 5 years). The work concludes that there is a significant size effect in stock returns. The statistical relation is negative, that is, smaller companies tend to have higher cost of equity. This result is aligned with the international research on the subject. The quantitative model suggest that for a logarithmic unit variation in company size there is a 4,03% additional cost of equity. Concerning liquidity, it has not been found, consistently, statistical relation that this variable affects stock returns when controlled by company in Brazilian market.

Nowadays technology development places increasing demands on miniature and micro components and products, and micro-milling is one of the most flexible machining processes in manufacturing 3D structures and complex structured surfaces. A thorough and scientific understanding on fundamentals of the micro-milling process is essential for applying it in an industrial scale. Therefore, in-depth scientific understanding of the micro-cutting mechanics is critical, particularly on size effect, minimum chip thickness, chip formation, tool wear and cutting temperature, etc. so as to fulfil the gap between fundamentals and industrial scale applications. Therefore, three key fundamental research topics are determined for this research, and a comprehensive study on those topics is conducted by means of modeling, simulation, experiments. The topics include chip formation process in micro-milling, novel cutting force modeling in multiscale and study on the tool wear and process monitoring. The investigation into chip formation process in micro-milling consists of three stages; the micro-cutting process is firstly simulated by means of FEA with a primary focus on finding the minimum chip thickness for different tool/material pair and explaining the size effect; the simulation results are then validated by conducting micro-cutting experiment on the ultra-precision lathe. Experiments are carried out on aluminium 6082-T6 with both natural diamond and tungsten carbide tool. By knowing the minimum chip thickness for different tool/material pair, the chip formation process is investigated by performing comparative study by using the diamond and tungsten carbide micro-milling tools. As the minimum chip thickness for diamond micro-milling tool is smaller than that for tungsten carbide tool compared to nominal chip thickness, MCT is ignored in diamond micro-milling. Thus the comparative study is conducted by utilizing both tools with perfectly sharpened cutting edge and tools with the rounded cutting edge in micro-milling. The chips are inspected and associated with cutting force variations in the micro-milling process. The findings are further consolidated by comparing with research results by other researchers. The cutting force modeling is developed in three different aspects, e.g. cutting force on the unit length or area and cutting force on the unit volume in order to better understand the micro-cutting mechanics in aspects of size effect, tool wear mechanism and the cutting energy consumption. The mathematical modeling firstly starts with a novel instantaneous chip thickness algorithm, in which the instantaneous chip thickness is computed by taking account of the change of tool geometry brought about by the tool runout; then the collected cutting forces are utilized to calibrate the model coefficients. For accurate measurement on cutting forces, the Kalman Filter technique is employed to compensate the distortion of the measured cutting force. Model calibration is implemented using least-square method. The proposed cutting force model is then applied in micro-milling to represent the conditions of tool wear and the cutting energy consumption. Further study on the surface generation simulation is based on force model and its comparison with the machined surface is also performed. Cutting experiments using the new tungsten carbide tool are carried out and the tool wear is monitored offline at different machining stages. The dominant tool wear types are characterised. Tool wear is investigated by mainly analysing cutting force at different tool wear status. Frequency analysis by Fourier Transform and Wavelet Transform are carried out on the force signals, and features closely related to the tool wear status are identified and extracted. The potential of applying these features to monitoring the tool wear process is then discussed. Experimental studies to machine the structured surface and nano-metric level surface roughness are presented, the machining efficiency, dimensional accuracy and tool-path strategies are optimised so as to achieve the desired outcomes. Moreover, investigation on cutting temperature in micro-cutting is also studied to some extent by means of simulation; the influence of cutting edge radius on cutting temperature is particularly investigated. Investigation on above aspects provides systematic exploration into the micro-milling process and can contribute substantially to future micro-milling applications.

During machining play the size off component deciding role from the viewpoint of their behaviour. This is result of „size effect”, which turns common characteristic cutting process. The aim of diploma thesis was contribute piece of knowledge verification of this effect and the further exploit during machining. The main interest is directed to the relation between the cutting edge and depth of cut.

The paper proposes a method for characterizing the in-situ interaction between filaments in a multifilament yarn. The stress transfer between neighboring filaments causes the reactivation of a broken filament at some distance from the break. The utilized statistical bundle models predict a change in the slope of the mean size effect curve once the specimen length becomes longer than the stress transfer length. This fact can be exploited in order to determine the stress transfer length indirectly using the yarn tensile test with appropriately chosen test lengths. The identification procedure is demonstrated using two test series of tensile tests with AR-glass and carbon yarns.

Des simulations de dynamique de dislocation (DD) sont utilisées pour l’étude de l'effet Hall-Petch (HP) et des contraintes internes à long-portée induites par les hétérogénéités de déformation dans les matériaux polycristallins.L'effet HP est reproduit avec succès grâce à des simulations de DD réalisées sur de simples agrégats polycristallins périodiques composés de 1 ou de 4 grains. De plus, l'influence de la forme des grains a été explorée en simulant des grains avec différents rapports d'aspect. Une loi généralisée de HP est proposée pour quantifier l'influence de la morphologie du grain en définissant une taille de grain effective. La valeur moyenne de la constante HP $K$ calculée avec différentes orientations cristallines à faible déformation est proche des valeurs expérimentales.Les dislocations stockées pendant la déformation sont principalement localisées à proximité des joints de grain et peuvent être traitées comme une distribution surfacique de dislocations. Nous avons utilisé des simulations DD pour calculer les contraintes associées aux parois de dislocations de différentes hauteurs, longueurs densités et caractères. Dans tous les cas, la contrainte est proportionnelle à la densité surfacique de dislocations géométriquement nécessaires (GNDs) et sa variation est capturée par un ensemble d'équations empiriques simples. Une prévision de contraintes à long-portée dans les grains est réalisée en sommant les contributions des GNDs accumulées de part et d’autre des joints de grains.L'augmentation de la contrainte interne liée au stockage de GNDs est linéaire avec la déformation plastique et est indépendante de la taille des grains. L'effet de taille observé dans les simulations de DD est attribué au seuil de déformation plastique, contrôlé par deux mécanismes concurrents : la contrainte critique de multiplication des sources et la contrainte critique de franchissement de la forêt. En raison de la localisation de la déformation dans les matériaux à gros grains, le modèle d’empilement des dislocations doit être utilisé pour prédire la contrainte critique dans ce cas. En superposant cette propriété aux analyses que nous avons fait à partir de simulations de DD dans le cas d'une déformation homogène, l'effet HP est justifié pour une large gamme de tailles de grains. / Dislocation Dynamics (DD) simulations are used to investigate the Hall-Petch (HP) effect and back stresses induced by grain boundaries (GB) in polycrystalline materials.The HP effect is successfully reproduced with DD simulations in simple periodic polycrystalline aggregates composed of 1 or 4 grains. In addition, the influence of grain shape was explored by simulating grains with different aspect ratios. A generalized HP law is proposed to quantify the influence of the grain morphology by defining an effective grain size. The average value of the HP constant K calculated with different crystal orientations at low strain is close to the experimental values.The dislocations stored during deformation are mainly located at GB and can be dealt with as a surface distribution of Geometrically Necessary Dislocations (GNDs). We used DD simulations to compute the back stresses induced by finite dislocation walls of different height, width, density and character. In all cases, back stresses are found proportional to the surface density and their spatial variations can be captured using a set of simple empirical equations. The back stress calculation inside grains is achieved by adding the contributions of GNDs accumulated at each GB facet.These back stresses are found to increase linearly with plastic strain and are independent of the grain size. The observed size effect in DD simulations is attributed to the threshold of plastic deformation, controlled by two competing mechanisms: the activation of dislocation sources and forest strengthening. Due to strain localization in coarse-grained materials, the pile-up model is used to predict the critical stress. By superposing such property to the analysis we made from DD simulations in the case of homogeneous deformation, the HP effect is justified for a wide range of grain sizes.