Nas últimas décadas, considerável evidência tem sido direcionada no desenvolvimento de métodos computacionais que analisam os comportamentos não lineares das estruturas. O método da corda é considerado o tipo de método de seguimento de trajetórias que possui a técnica mais comum e versátil para analisar tais comportamentos não lineares. No entanto, testes realizados pelos autores concluem que tais métodos ainda apresentam algumas dificuldades quando as estruturas possuem formas complexas e comportamentos fortemente não lineares. Devido a isso, o objetivo do trabalho em questão é acrescentar duas modificações ao método da corda, a fim de desenvolver um modelo mais seguro e estável. A esses métodos, inclui-se (i) uma equação alternativa para o passo previsor, que tornará o método mais estável ao ultrapassar pontos críticos; e (ii) um parâmetro de controle, que através do conceito de soma dos comprimentos de corda, é chamado de método da corda acumulado. Ao final, por consequência destas alterações, um procedimento mais robusto para o método da corda é obtido. Sua validação é apresentada, pelos autores, através de exemplos utilizando estruturas treliçadas (bidimensionais e tridimensionais) em um programa de simulação numérica desenvolvida. / In the last decades, a considerable effort has been directed to develop of computational methods to analyze the nonlinear behavior of structures. The arc-length method is considered the type of path following method which have the most common and versatile techniques for analyzing such nonlinear behaviors. Nevertheless, tests performed by the authors conclude that such methods present some difficulties when the structures possess complex shapes and strongly nonlinear behaviors. Due to this, the purpose of the present work is to add two amendments to the arc-length method, in order to develop a model safer and more stable. In such methods, it is included (i) an alternative equation for the predictor step, that will make the method most stable to overcoming critical points, and (ii) a control parameter, through the concept of sum of the arc length up to now, referred to as accumulated arc-length method. By the end, as a result of these modifications, a more robust procedure for the method of the string is obtained. And its validation is presented by the authors through examples using truss structures in a two-dimensional and three-dimensional numerical simulation program developed.

Nas últimas décadas, considerável evidência tem sido direcionada no desenvolvimento de métodos computacionais que analisam os comportamentos não lineares das estruturas. O método da corda é considerado o tipo de método de seguimento de trajetórias que possui a técnica mais comum e versátil para analisar tais comportamentos não lineares. No entanto, testes realizados pelos autores concluem que tais métodos ainda apresentam algumas dificuldades quando as estruturas possuem formas complexas e comportamentos fortemente não lineares. Devido a isso, o objetivo do trabalho em questão é acrescentar duas modificações ao método da corda, a fim de desenvolver um modelo mais seguro e estável. A esses métodos, inclui-se (i) uma equação alternativa para o passo previsor, que tornará o método mais estável ao ultrapassar pontos críticos; e (ii) um parâmetro de controle, que através do conceito de soma dos comprimentos de corda, é chamado de método da corda acumulado. Ao final, por consequência destas alterações, um procedimento mais robusto para o método da corda é obtido. Sua validação é apresentada, pelos autores, através de exemplos utilizando estruturas treliçadas (bidimensionais e tridimensionais) em um programa de simulação numérica desenvolvida. / In the last decades, a considerable effort has been directed to develop of computational methods to analyze the nonlinear behavior of structures. The arc-length method is considered the type of path following method which have the most common and versatile techniques for analyzing such nonlinear behaviors. Nevertheless, tests performed by the authors conclude that such methods present some difficulties when the structures possess complex shapes and strongly nonlinear behaviors. Due to this, the purpose of the present work is to add two amendments to the arc-length method, in order to develop a model safer and more stable. In such methods, it is included (i) an alternative equation for the predictor step, that will make the method most stable to overcoming critical points, and (ii) a control parameter, through the concept of sum of the arc length up to now, referred to as accumulated arc-length method. By the end, as a result of these modifications, a more robust procedure for the method of the string is obtained. And its validation is presented by the authors through examples using truss structures in a two-dimensional and three-dimensional numerical simulation program developed.

Finite element analysis using Cohesive Zone Method (CZM) is a commonly used method to investigate delamination in laminated composites. In this study, two plane strain, zero-thickness six-node quadratic (6-NQ) and four-node linear (4-NL) interface elements are developed to implement CZM. Two main approaches for CZM formulation are categorized as Unified Mode Approach (UMA) and Separated Mode Approach (SMA), and implemented into 6-NQ interface elements to model a double cantilever beam (DCB) test of a unidirectional laminated composite. The results of the approaches are nearly identical. However, it is theoretically shown that SMA spawns non-symmetric tangent stiffness matrices, which may lower convergence and/or overall performance, for mixed-mode loading cases. Next, a UMA constitutive relationship is rederived. The artificial modifications for improving convergence rates such as lowering penalty stiffness, weakening interfacial strength and using 6-NQ instead of 4-NL interface elements are investigated by using the derived UMA and the DCB test model. The modifications in interfacial strength and penalty stiffness indicate that the convergence may be improved by lowering either parameter. However, over-softening is found to occur if lowering is performed excessively. The morphological differences between the meshes of the models using 6-NQ and 4-NL interface elements are shown. As a consequence, it is highlighted that the impact to convergence performance and overall performance might be in opposite. Additionally, benefits of selecting CZM over other methods are discussed, in particular by theoretical comparisons with the popular Virtual Crack Closure Technique. Finally, the numerical solution scheme and the Arc-Length Method are discussed.

In Nonlinear Finite Element Analysis (FEA) applied to structures, displacements at which the tangent stiffness matrix KT becomes singular are called critical points, and correspond to instabilities such as buckling or elastoplastic softening (e.g., necking). Prior to the introduction of Arc Length Methods (ALMs), critical points posed severe computational challenges, which was unfortunate since behavior at instabilities is of great interest as a precursor to structural failure. The original ALM was shown to be capable in some circumstances of continued computation at critical points, but limited success and unattractive features of the formulation were noted and addressed in extensive subsequent research. The widely used Crisfield Cylindrical and Spherical ALMs may be viewed as representing the 'state-of-the-art'. The more recent Stiff Arc Length method, which is attractive on fundamental grounds, was introduced in 2004, but without implementation, benchmarking or performance assessment. The present thesis addresses (a) implementation and (b) performance comparisons for the Crisfield and Stiff methods, using simple benchmarks formulated to incorporate elastoplastic softening. It is seen that, in contrast to the Crisfield methods, the Stiff ALM consistently continues accurate computation at, near and beyond critical points.

Made available in DSpace on 2014-12-17T14:48:18Z (GMT). No. of bitstreams: 1 EstefaneGML_DISSERT.pdf: 1743429 bytes, checksum: 278db485c2938a47ce6dc5f0b6962c1f (MD5) Previous issue date: 2014-02-28 / This work presents the positional nonlinear geometric formulation for trusses using different strain measures. The positional formulation presents an alternative approach for nonlinear problems. This formulation considers nodal positions as variables of the nonlinear system instead of displacements (widely found in literature). The work also describes the arc-length method used for tracing equilibrium paths with snap-through and snap-back. Numerical applications for trusses already established in the literature and comparisons with other studies are provided to prove the accuracy of the proposed formulation / Este trabalho apresenta a formula??o posicional n?o linear geom?trica para treli?as usando diferentes medidas de deforma??o. A formula??o posicional ? uma abordagem alternativa para problemas n?o lineares. Essa formula??o considera as posi??es nodais como vari?veis do sistema n?o linear em vez dos deslocamentos (que ? largamente utilizado na literatura). O trabalho tamb?m descreve o m?todo do comprimento de arco, usado para tra?ar caminhos de equil?brio com snap-through e snap-back. Aplica??es num?ricas com treli?as j? consagradas na literatura e compara??es com outros trabalhos s?o fornecidos para provar a acur?cia da formula??o proposta

Le plissement est généralement observé dans les structures minces ayant un comportement de type membrane. Ces structures minces ne supportent pas d'effort de flexion et sont donc sollicitées en traction. Dans cette thèse, nous avons développé une technique de réduction de modèle pour la modélisation du plissement des membranes minces. Cette technique, basée sur les séries de Fourier à double échelle, permet de déduire d'un modèle complet de membrane, un modèle réduit capable de prendre en compte les instabilités globales et locales. Les valeurs critiques de charge et longueur d'onde sont déterminées analytiquement puis numériquement. Des exemples numériques nous ont permis de valider le modèle numérique par rapport au modèle analytique. Les modèles numériques étudiés prennent en compte le modèle complet et le modèle réduit de la membrane. Le modèle complet est simulé dans Abaqus et résolu numériquement à l'aide de la méthode de la longueur d'arc et le modèle réduit est implémenté dans Matlab et résolu numériquement à l'aide de la méthode asymptotique numérique. Nous avons étudié le comportement de la membrane sous sollicitation mécanique, thermique et thermo-mécanique. Les résultats obtenus montrent que le modèle réduit est capable de se substituer au modèle complet dans la détermination des contraintes critiques et longueurs d'onde correspondantes. Le gain en temps de calcul obtenu est important, ceci grâce à la très faible densité de maillage requis par le modèle réduit. Le modèle réduit est très sensible aux conditions aux bords de la membrane et requiert d'avoir une longueur d'onde des plis quasiment constante dans la largeur de la membrane / Wrinkling is an instability phenomenon generally observed in thin structures with membrane's behavior. Those thin structures have no rigidity to flexion and are therefore used in traction. In this thesis, we developed a reduction model's technique for the modeling of wrinkling phenomenon in thin membranes. This technique, based on the double scale Fourier series, allow us to deduce from a full membrane model, a reduced membrane model that is able to take into account the global and local instability of the structure. The critical load and critical wavelength are determined analytically on one side, then numerically on the other side. Numerical exemples are conducted to validate the numerical model towards the analytical one. Numerical models studied take into account both full and reduce membrane models. The full model is simulated in Abaqus and solved numerically using the arc length method and the reduced model is implemented in Matlab and solved numerically using the asymptotic numerical method. We studied the membrane behavior under mechanical, thermal and thermo-mechanical loading. The results obtained show that the full membrane model can be replaced by the reduced one in determining critical loads and corresponding wavelengths. The gain in computation time obtained is important, due to the coarse mesh required by the reduced model. The reduced model is very sensitive to membrane's boundaries conditions and requires to have a quasi constant wavelength along the membrane width