The fall-cone test is widely used in geotechnical practice to obtain rapid estimates of the undrained shear strength of cohesive soil, and as an index test to determine the liquid limit. This thesis is concerned with numerical modelling of the penetration of solids by conical indenters, and with interpretation of the numerical results in the context of the fall-cone test. Experimental studies of shallow cone penetration in clay are also reported, with the aim of verifying the numerical predictions. The practical significance of the results, in terms of the interpretation of fall-cone test results, is assessed. Results are reported from finite element analyses with the commercial codes ELFEN and Abaqus, in which an explicit dynamic approach was adopted for analysis of continuous cone indentation. Quasi-static analyses using an elastoplastic Tresca material model are used to obtain bearing capacity factors for shallow cone penetration, taking account of the material displaced, for various cone apex angles and adhesion factors. Further analyses are reported in which a simple extension of the Tresca material model, implemented as a user-defined material subroutine for Abaqus, is used to simulate viscous rate effects (known to be important in cohesive soils). Some analyses with the rate-dependent model are displacement-controlled, while others model the effect of rate-dependence on the dynamics of freefall cone indentation tests. Laboratory measurements of the forces required to indent clay samples in the laboratory are reported. Results from displacement-controlled tests with imposed step-changes in cone speed, and from freefall tests, confirm that the numerical rate-dependent strength model represents the observed behaviour well. Some results from experiments to observe plastic flow around conical indenters are also presented. Finally, additional numerical analyses are presented in which a critical state model of clay plasticity is used to study the variation of effective stress, strain and pore pressure around cones in indentation tests at various speeds.

The current thesis focuses on the deformation behaviour and strain distribution in the microstructure of ductile iron during tensile loading. Utilizing Digital Image Correlation (DIC) and in-situ tensile test under optical microscope, a method was developed to measure high resolution strain in microstructural constitutes. In this method, a pit etching procedure was applied to generate a random speckle pattern for DIC measurement. The method was validated by benchmarking the measured properties with the material’s standard properties. Using DIC, strain maps in the microstructure of the ductile iron were measured, which showed a high level of heterogeneity even during elastic deformation. The early micro-cracks were initiated around graphite particles, where the highest amount of local strain was detected. Local strain at the onset of the micro-cracks were measured. It was observed that the micro-cracks were initiated above a threshold strain level, but with a large variation in the overall strain. A continuum Finite Element (FE) model containing a physical length scale was developed to predict strain on the microstructure of ductile iron. The materials parameters for this model were calculated by optimization, utilizing Ramberg-Osgood equation. For benchmarking, the predicted strain maps were compared to the strain maps measured by DIC, both qualitatively and quantitatively. The DIC and simulation strain maps conformed to a large extent resulting in the validation of the model in micro-scale level. Furthermore, the results obtained from the in-situ tensile test were compared to a FE-model which compromised cohesive elements to enable cracking. The stress-strain curve prediction of the FE simulation showed a good agreement with the stress-strain curve that was measured from the experiment. The cohesive model was able to accurately capture the main trends of microscale deformation such as localized elastic and plastic deformation and micro-crack initiation and propagation.

Cette étude tente de modéliser les taux de transport de sédiments et l'évolution du lit dans un milieu estuarien complexe : l'estuaire de la Gironde, caractérisé par une grande hétérogénéité dans la composition des sédiments de lit, avec la présence de sédiments cohésifs et non-cohésifs ainsi que des mélanges sablo-vaseux. Notre objectif principal est d'étendre un modèle morphodynamique 2D développé par Huybrechts et al. (2012b) pour les sédiments non-cohésifs, afin de tenir compte de la présence de la vase et d'établir une étape préliminaire pour un modèle morphodynamique de sédiment mixte. Notre cadre est le système Telemac (version 6.1) où l'approche bi-dimensionnelle a été sélectionnée pour des simulations à grandes échelles spatiales (150 km) et moyen terme (5 ans).La première partie de ce travail est consacrée à la compréhension des processus de sédimentation-consolidation de la vase pure, en combinant expériences et modèles 1D verticaux. Les processus du sédiment cohésif sont ensuite intégrés dans le modèle morphodynamique de l'estuaire de la Gironde. Des expériences d'érosion et de dépôt ont été réalisées au laboratoire RWTH (Université d'Aachen, Allemagne) pour calibrer les paramètres des lois d'érosion et de dépôt. En outre, l'effet de la consolidation est pris en compte à travers la mise en œuvre d'un modèle 1DV de sédimentation et consolidation basé sur la théorie de Gibson (Thiébot et al., 2011) en utilisant des équations de fermeture analytique pour la perméabilité et la contrainte effective. Une attention particulière est accordée à l'initialisation de la structure verticale du lit sédimentaire. Les mesures et les résultats du modèle sont comparés pour les concentrations des sédiments en suspension et pour l'évolution de fond en moyen terme (5 ans).Dans la deuxième partie, un nouveau modèle 1DV pour la sédimentation entravée des mélanges sablo-vaseux a été développé sur la base des modèles de bi-disperse non-cohésifs. La solution numérique a été réalisée en prenant en considération un schéma de haute précision dans l'espace par la technique de reconstruction WENO et en temps par un Discontinue de Galerkin local (DG). Le modèle est ensuite validé par rapport à une large gamme de données expérimentales (mono-disperse sable, vase, non-cohésif bi-disperse et le mélange non-cohésif/cohesif) / This study attempts to model sediment transport rates and the resulting bed evolution in a complex estuarine environment: the Gironde estuary, characterized by a high hetereogeneity in the sediment bed composition, with the presence of both cohesive and non-cohesive sediments and sand/mud mixtures. Our main objective is to extend an existing 2D morphodynamic model developped by Huybrechts et al (2012b) for non-cohesive sediments, to account for the presence of mud and to draw some preliminary step for a fully mixte sediment morphodynamic model. Our framework is the finite element Telemac system (release 6.1), where the two-dimensional (depth averaged) approach has been selected for large scale and medium term simulations.The first part of this work is devoted to the understanding of sedimentation-consolidation processes for pure mud, combining laboratory experiments and 1D vertical models. Cohesive processes are then integrated in the 2D (depth-averaged) large scale morphodynamic model of the Gironde estuary developed by Huybrechts et al. (2012b). Erosion/deposition experiments were performed at the RWTH laboratory (University of Aachen, Germany) to calibrate the erosion and deposition law parameters. Moreover, the effect of consolidation is taken into account through the implementation of a 1DV Gibson-based sedimentation-consolidation model (Thiebot et al., 2011) using analytical closure equations for permeability and effective stress. Special attention is paid to the initialisation of the bed structure. Comparisons between measurements and model results are achieved on both suspended sediment concentration records and on medium term (5-year) bed evolutions.In the second part, a new 1DV model for the hindered settling of sand-mud mixtures has been developed based on the background of non-cohesive bi-disperse models. The numerical solution has been constructed by considering a high-order of accuracy in space via a Weighted Essentially Non Oscillatory (WENO) reconstruction technique and in time via a local space-time Discontinuous Galerkin (DG).The model is then validated against a large range of experimental data (mono-disperse sand, mud, non-cohesive bi-disperse and non-cohesive/cohesive mixture)

Nous avons donc proposé dans cette thèse une stratégie pour suivre l’évolution des fissures dans une structure en béton précontraint. L'objectif principal est de définir des briques de modélisation permettant d'étudier le comportement des fissures dans les enceintes de confinement des centrales nucléaires. Chaque brique de modélisation permet ainsi de représenter un phénomène non-linéaire présent dans ces structures : la fissuration du béton, les grilles d'armatures, les cables de précontrainte, la décohésion de l'acier dans le béton ainsi que le vieillissement du béton. Elles ont été validées grâce à plusieurs essais expérimentaux et appliquées à différentes zones. de l'enceinte de confinement pour simuler l'amorçage et la propagation des fissures. Enfin, une stratégie pour définir un état initialement fissuré a été développé à l'aide de ces différentes briques ainsi qu'une méthode de zoom pour reproduire les conditions limites de la précontrainte et la montée en pression sur une partie de l'enceinte. / The difficulty of predicting cracks on reinforced concrete structures is due to the heterogeneity of the concrete and the dependence on environmental boundary conditions. To overcome this problem, we introduce directly a potential crack at the initial time, using cohesive zone models. A major difficulty will be to juggle the cohesive zone models describing cracking and the phenomena of creep and drying, and thus to consider three types of nonlinearities appearing in the calculation : cracking, decohesion of steelconcrete and creep. We also propose a strategy to represent an initially cracked structure. In this thesis, the crack behavior in different parts of the confinement building are modeled using the above described method.

Le travail présenté a pour objectif la prédiction numérique de la résistance résiduellede grandes structures vis-à-vis d’évènements accidentels, tels que ceux rencontrés p.ex. dans le cas de la collision de navires ou d’impact d’oiseaux en aéronautique. Cesévènements peuvent dans certain cas conduire à la rupture, qui est ici considéréeductile. La difficulté de cette étude, consiste à reproduire dans une méthodologieunifiée basée sur la méthode des éléments finis les étapes successives menant àla ruine ultime de la structure. Ces étapes sont : l’endommagement ductile, lalocalisation de la déformation et la propagation de la fissure. Un élément essentiel pour la conception d’un modèle de fissuration ductile prédictif est le traitement numérique de la phase transitoire critique de localisation de la déformation induite par l’endommagement dans une bande de matière étroite.A cet effet, trois points de vue différents en termes de champ de déplacement àtravers la bande de localisation sont proposés. Ces trois approches se distinguentpar le type de discontinuité considérée : forte, faible et régularisée (expression nonlinéaire). Un cadre variationnel consistant est élaboré pour chacune des trois approches.Ainsi la cinématique enrichie est incorporée dans la formulation de l’élément fini enutilisant la méthode des éléments finis enrichis (X-FEM). Puis, la performance deces méthodes est évaluée vis-à-vis de leur capacité à modéliser la phase transitoireentre endommagement diffus (mécanique des milieux continus) et propagation defissure (mécanique de la rupture). Ces travaux sont réalisés dans le contexte dematériaux ductiles. D’après les analyses réalisées, la combinaison du modèle de ’discontinuité fortecohésive’ et la X-FEM semble être la plus prometteuse des trois approches étudiéespour allier physique et numérique. Le développement d’un tel modèle est discutéen détail. Enfin, deux critères supplémentaires sont définis : le premier pour lepassage de l’endommagement diffus au modèle de bande cohésive et un deuxièmepour le passage du modèle de bande cohésive à la rupture. / The present work aims at numerically predicting the current residual strengthof large engineering structures made of ductile metals regarding accidental events,e.g. ships collision or bird strike in aviation, which may potentially lead to failure.With this aim in view, the challenge consists in reproducing within a unified finiteelement (FE)-based methodology the successive steps of micro-voiding-induceddamage, strain localization and crack propagation, if any.A key ingredient for a predictive ductile fracture model is the proper numericaltreatment of the critical transition phase of damage-induced strain localizationinside a narrow band. For this purpose, three different viewpoints in terms ofdisplacement field across the localization band are proposed involving a strong,weak and (non-linearly) regularized discontinuity, respectively.A consistent variational framework is elaborated for each of the three methods,whereby the enriched kinematics is embedded into the FE formulation using theeXtended FEM. Then, within a comparative procedure, the performance of thesemethods is assessed regarding their ability of modeling the transition phase betweendiffuse damage (continuum mechanics framework) and crack propagation (fracturemechanics framework), always in the context of ductile materials.According to the aforementioned analyses, the combination of the strong discontinuitycohesive model and the X-FEM appears to be the most promising of thethree studied approaches to bring together physics and numerics. The developmentof such a model is discussed in detail. Finally, two supplementary criteria aredefined: the first one for the passage from diffuse damage to the cohesive bandmodel and the second one for the passage from the cohesive band model to thecrack.

O objetivo deste estudo é compreender a dinâmica das partículas no estuário do rio Araranguá (SC) em termos de transporte em suspensão, deposição e exportação, em diferentes condições físico-químicas da água, com pH baixo (<4,5) e normal (>=7), e sob condições de regime transiente da descarga fluvial. O estuário do rio Araranguá é classificado como altamente estratificado e dominado pela descarga fluvial. Um aspecto fundamental nos processos que controlam a dinâmica de materiais em suspensão no estuário é a atividade de lavra e beneficiamento do carvão, que gera uma drenagem ácida no rio Mãe Luzia. O seu pH baixo (&#732;3) indica que o processo de floculação é intensificado no interior do estuário. Com o objetivo de melhor entender a hidrodinâmica do estuário, um modelo tridimensional foi implementado para a região. O modelo foi calibrado e validado com dados observados. Simulações com diferentes picos de descarga fluvial (100, 400, 700 e 1000 m3.s-1) e baixa descarga (30 m3.s-1) foram realizadas. Para calcular o transporte do material particulado em suspensão foi desenvolvido um modelo Lagrangeano da trajetória da partícula. O modelo da trajetória utiliza os resultados do modelo hidrodinâmico e a velocidade de decantação das partículas. O estuário respondeu rapidamente às variações da descarga fluvial. A cunha salina foi transportada para fora do estuário durante os picos de 700 e 1000 m3.s-1. A descarga fluvial é o fator dominante no controle da intrusão da salinidade no estuário e, em condições de baixa descarga, o transporte advectivo de sal é dominado pelo termo dispersivo da circulação gravitacional. O tempo de residência médio das partículas no estuário foi menor em condições de pH baixo. A exportação das partículas aumentou com o aumento da descarga fluvial (pico de 1000 m3.s-1), sendo que 27% e 30% foram exportadas para a plataforma interna durante os experimentos com pH baixo e pH normal, respectivamente. A região de deposição das partículas indica que com a diminuição do pH, as partículas passam a depositar mais à jusante no estuário. O evento de descarga de 1000 m3.s-1 não foi capaz de remover a grande maioria das partículas do estuário, mas foi responsável por redistribuir as partículas ao longo do médio estuário, onde se depositaram em condições favoráveis com baixas velocidades e ausência de salinidade. O trabalho demonstra que apesar da alta estratificação do estuário, a acidez de suas águas, seu padrão de circulação, com baixas velocidades e a presença da cunha salina, possuem um papel fundamental no transporte das partículas, resultando em maior potencial de retenção do material em suspensão no estuário do rio Araranguá. / The present study assessed the dynamics of particles of the Araranguá (SC) river estuary in terns of suspended transport, deposition and exportation in different conditions of physical-chemical of water, with low (<4,5) and normal pH (>=7), and transient of river discharge conditions. The Araranguá river estuary is a highly stratified flash-flood dominated. The river basin of the Mãe Luzia river, one of the main tributaries of the Araranguá river, is performed the activity of mining and coal processing, wich generates an acidic drainage. Its low pH indicates that flocculation process is intensified in estuary inner. With the objective of better understanding the estuarine processes, a three-dimensional hydrodynamics model was implemented. The model was calibrated and verified using observational data. Simulations were conducted with freshet peak ranged over 100, 400, 700 and 1000 m3.s-1. Lagrangian particle tracking algorithms was developed to represent the suspended particle transport at estuary. The particle tracking used the montion resulting from hydrodynamics model and the particle settling velocity. The estuary responded quickly the river discharge variations. During the pulse event the velocities at water column were unidirectional down-estuary is shown the barotropic dominates over baroclinic component. During the highest flash flood event (700 and 1000 m3.s-1) the salt water was pushed out of the estuary. The river discharge is a dominating factor affecting the salinity intrusion in the estuarine system and under low discharge conditions the advective salt transport is dominated by dispersive term of gravitational circulation. The mean residence time of the particles was lowest in pH low conditions. The particles export is great as the river discharge is rising (1000 m3.s-1) with 27% and 30% exported to inner shelf during the experiments with low and normal pH, respectively. The particles deposition local indicates that with the pH decrease the particles become to be deposited further downstream. The discharge event of 1000 m3.s-1 was not able to remove all particles out, but was response to redistribute the particles along of the meddle estuary, where settling in favorable conditions at low velocities and absence of salinity. The study shows that although high stratification of estuary, the acidity of water, the pattern of circulation, at low velocities, and the salt wedge presence play a fundamental role in the transport of particle matter resulting in increased the potential retention of suspended matter in the Araranguá river estuary.

No contexto do método dos elementos de contorno, este trabalho apresenta comparativamente três formulações em distintos aspectos. Visando a análise de sólidos bidimensionais no campo da mecânica da fratura, primeiramente é estudada a equação singular ou em deslocamentos. Em seguida, a formulação hiper-singular ou em forças de superfície é avaliada. Por último, a formulação dual, que emprega ambas equações é analisada. Para esta análise, elementos contínuos e descontínuos são empregados, equações numéricas e analíticas com ponto fonte dentro e fora do contorno são testadas, usando aproximação linear. A formulação é inicialmente empregada a problemas da mecânica da fratura elástica linear e em seguida extendida a problemas não-lineares, especialmente o modelo coesivo. Exemplos numéricos diversos averiguam as formulações, comparando com resultados analíticos ou disponíveis na literatura. / In this work three boundary elment formulations applied to fracture mechanics are studied. Aiming the analysis of two-dimensional solids with emphasis on the crack problem, the first considered method is the one based on using displacement equations only (singular formulation). The second scheme discussed in this work is a formulation based on the use of traction equations (hyper-singular formulation). Finally the dual boundary element method that uses singular and hyper-singular equations is considered. The numerical schemes have been implemented using continuous and discontinuous linear boundary and crack elements. The boundary and crack integral were all carried out by using analytical expressions, therefore increasing the accuracy of the algebraic system obtained for each one of the studied schemes. The developed numerical programs were applied initially to elastic fracture mechanics and then extended to analyze cohesive cracks. Several numerical examples were solved to verify the accuracy of each one of the studied models, comparing the results with the analytical solutions avaliable in the literature.

Made available in DSpace on 2016-04-28T19:33:54Z (GMT). No. of bitstreams: 1 Bianca Rodrigues de Oliveira dos Santos Rezende.pdf: 562748 bytes, checksum: cf1d68dcdfdb97a954d5252574179ab1 (MD5) Previous issue date: 2015-02-19 / Secretaria da Educação do Estado de São Paulo / This dissertation is part of the line of research "Reading, Writing and English Language Teaching" of the Postgraduate Program in Portuguese Language at PUC-SP. The overall aim of this study is to analyze the argumentative and textual sequences and the cohesive markers in essays of students of the 3rd year of high school (Ensino Médio in Brazil). The specific objectives are to identify, describe, analyze and interpret these argumentative sequences and these cohesive markers, as well as reflect on their uses in argumentative texts produced by students of the 3rd year of high school. The theoretical basis is the Textual Linguistics and more specifically the Textual Analysis of Discourse. In order to achieve the goal, we seek the study of textual sequences in Adam (2011), Silva (2012) and Marquesi (2014), supplemented by Charaudeau (2014); the analysis of the argumentative text sequences is supported by Adam (2011) and supplemented by Charaudeau (2014), Cabral (2012, 2013 and 2014) and Sayeg (1995); and the study of the cohesive markers is supported by Adam (2011), Koch (2009), Koch and Elias (2010), Koch and Travaglia (1993). The analysis of the essays pointed the need of teaching of argumentative textual sequences and cohesive markers in high school, so those students can build arguments and establish a link between the constituent parts of argumentative text and consequently make progress in the writing skill of Portuguese language / Esta dissertação insere-se na linha de pesquisa Leitura, Escrita e Ensino de Língua Portuguesa , do Programa de Pós-Graduados em Língua Portuguesa da PUC-SP. O objetivo geral deste estudo é analisar as sequências textuais argumentativas e os marcadores coesivos em redações de alunos da 3ª série do Ensino Médio. Os objetivos específicos são identificar, descrever, analisar e interpretar essas sequências argumentativas e esses marcadores coesivos, bem como refletir sobre seus usos em textos dissertativos produzidos por alunos da 3ª série do Ensino Médio. A fundamentação teórica situa-se na Linguística Textual e, mais especificamente, na Análise Textual dos Discursos. Para atingirmos nossos objetivos, buscamos os estudos das sequências textuais em Adam (2011), Silva (2012) e Marquesi (2014), complementados por Charaudeau (2014); para as sequências textuais argumentativas, em Adam (2011), complementados por Charaudeau (2014), Cabral (2012, 2013 e 2014) e Sayeg (1995); e para os marcadores coesivos, em Adam (2011), Koch (2009), Koch e Elias (2010), Koch e Travaglia (1993). Os resultados obtidos na análise das redações apontam para a necessidade da abordagem das sequências textuais argumentativas e dos marcadores coesivos no ensino médio, a fim de que os alunos possam construir a argumentação e estabelecer a articulação entre as partes constitutivas do texto dissertativo, e, consequentemente, avançar na produção escrita

CoordenaÃÃo de AperfeÃoamento de Pessoal de NÃvel Superior / Materiais compÃsitos vem sendo amplamente estudados devido aos seus inÃmeros benefÃcios em relaÃÃo aos materiais metÃlicos, principalmente a elevada razÃo resistÃncia/peso, bom iso-lamento tÃrmico e boa resistÃncia à fadiga. CompÃsitos laminados, foco do presente trabalho, sÃo produzidos pelo empilhamento de um conjunto delÃminas, cada uma composta de fibras unidirecionais ou bidirecionais imersas em uma matriz polimÃrica. As estruturas de materiais compÃsitos apresentam comportamento nÃo linear, tanto fÃsico quanto geomÃtrico. Devido à elevada resistÃncia, estruturas de material compÃsito tendem a ser bastante esbeltas, podendo apresentar grandes deslocamentos e problemas de estabilidade. Adicionalmente, a consideraÃÃo da nÃo linearidade fÃsica tambÃm à importante para a simulaÃÃo de falha de estruturas laminadas. Um dos modos de falha mais importantes destas estruturas à a delaminaÃÃo, que consiste no descolamento de duas lÃminas adjacentes. No projeto de estruturas laminadas, o MÃtodo dos Elementos Finitos à a ferramenta de anÃlise mais utilizada devido a sua robustez, precisÃo e relativa simplicidade. Afim de permitir a anÃlise nÃo linear de estruturas laminadas submetidas a grandes deslocamentos, foi desenvolvida neste trabalho uma formulaÃÃo de elementos finitos sÃlidos laminados baseados na abordagem Lagrangiana Total. A simulaÃÃo do inÃcio e propagaÃÃo da delaminaÃÃo foi realizada neste trabalho utilizando Modelos de Zona Coesiva. Para este fim, foi desenvolvida uma formulaÃÃo de elementos isoparamÃtricos de interface com espessura nula e utilizados diferentes modelos constitutivos para representar a relaÃÃo entre as tensÃes e os deslocamentos relativos das faces da trinca coesiva, incluindo tanto o caso de modo I puro quanto de modo misto. As formulaÃÃes desenvolvidas neste trabalho foram implementadas no software de cÃdigo aberto FAST utilizando afilosofiade ProgramaÃÃo Orientada a Objetos. Estas implementaÃÃes sÃo apresentadas utilizando as convenÃÃes da UML. VÃrios exemplos foram utilizados para verificar e validar as implementaÃÃes realizadas. Excelentes resultados foram obtidos utilizando elementos sÃlidos laminados na anÃlise de estruturas de casca, mesmo empregando malhas com apenas um elemento sÃlido na espessura. No que diz respeito à delaminaÃÃo, verificou-se que o uso de Modelos de Zona Coesiva requer muito cuidado na escolha dos parÃmetros utilizados na anÃlise, principalmente no que diz respeito à relaÃÃo tensÃo-deslocamento relativo, tamanho dos elementos e mÃtodo de integraÃÃo numÃrica. Contudo, utilizando-se a integraÃÃo de Newton-Cotes e elementos de interface de tamanho adequado, obteve-se uma concordÃncia muito boa com resultados teÃricos e experimentais disponÃveis na literatura. De forma geral,verificou-se que o modelo coesivo exponencial apresenta maior robustez e eficiÃncia computacional que o modelo bilinear. / Composite materials has been widely studied thought the years because of it benefits compared to metals (elevated resistance/weight ratio, good thermal isolation and good fatigue resistance). Laminate composites are the focus of this work. Produced by stacked layers of fibers embed- ded on polymeric matrices, structures of composite materials presents material and geometrical non-linear behavior. Because of it elevated resistance, composite materials allow designers to create very slender structures which might present large displacements and stability problems. Additionally, considering material non-linearity is also important for collapse simulation of la- minated structures. One of the most important failure modes on laminated structures is delami- nation. Delamination is the detachment of adjacent layers. On laminated structures simulation, the Finite Element Method is one of the most used analysis tool. It is a robust, precise and relative simple operating tool. Intending analyzing non-linear behavior of laminated structures subjected to large displacements, was developed on this work a laminated solid finite element formulation based on Full Lagrangian formulation. Simulation of delamination beginning and propagation was developed on this work using Cohesive Zone models. To achieve this goal, an isoparametric formulation of interface finite elements without thickness and many constitutive models to represent the relation tension à displacement jump (relative displacement between crack faces) were developed. These models consider pure mode I and mixed mode. The formu- lations developed on this work were implemented on the open source finite element code FAST using Oriented Object Programing philosophy. These implementations are presented on UML conventions. Many examples were tested for verifying and validating all the implementations. Excellent results were obtained using laminated solid elements on the analysis of a shell struc- ture, even using meshes with only one element though thickness. On the delamination analysis, was verified that Cohesive Zone Models are very sensible related to the parameters used on the analysis, mainly tension à displacement jump model, size of elements and numerical integra- tion. Spite of it, using Newton-Cotes integration and interface elements of appropriate size, good agreements were obtained compared with theoretical results obtained on literature. In general, was observed that cohesive exponential model presents greater robustness and compu- tational efficiency than bilinear model.

No contexto do método dos elementos de contorno, este trabalho apresenta comparativamente três formulações em distintos aspectos. Visando a análise de sólidos bidimensionais no campo da mecânica da fratura, primeiramente é estudada a equação singular ou em deslocamentos. Em seguida, a formulação hiper-singular ou em forças de superfície é avaliada. Por último, a formulação dual, que emprega ambas equações é analisada. Para esta análise, elementos contínuos e descontínuos são empregados, equações numéricas e analíticas com ponto fonte dentro e fora do contorno são testadas, usando aproximação linear. A formulação é inicialmente empregada a problemas da mecânica da fratura elástica linear e em seguida extendida a problemas não-lineares, especialmente o modelo coesivo. Exemplos numéricos diversos averiguam as formulações, comparando com resultados analíticos ou disponíveis na literatura. / In this work three boundary elment formulations applied to fracture mechanics are studied. Aiming the analysis of two-dimensional solids with emphasis on the crack problem, the first considered method is the one based on using displacement equations only (singular formulation). The second scheme discussed in this work is a formulation based on the use of traction equations (hyper-singular formulation). Finally the dual boundary element method that uses singular and hyper-singular equations is considered. The numerical schemes have been implemented using continuous and discontinuous linear boundary and crack elements. The boundary and crack integral were all carried out by using analytical expressions, therefore increasing the accuracy of the algebraic system obtained for each one of the studied schemes. The developed numerical programs were applied initially to elastic fracture mechanics and then extended to analyze cohesive cracks. Several numerical examples were solved to verify the accuracy of each one of the studied models, comparing the results with the analytical solutions avaliable in the literature.