京都大学 / 0048 / 新制・課程博士 / 博士(農学) / 甲第21374号 / 農博第2298号 / 新制||農||1066(附属図書館) / 学位論文||H30||N5147(農学部図書室) / 京都大学大学院農学研究科地域環境科学専攻 / (主査)教授 村上 章, 教授 藤原 正幸, 教授 渦岡 良介 / 学位規則第4条第1項該当 / Doctor of Agricultural Science / Kyoto University / DFAM

This dissertation mainly focuses on the development of new numerical models to simulate transport phenomena and predict the occurrence of macrosegregation defects known as freckles in directional solidification processes. Macrosegregation models that include double diffusive convection are very complex and require the simultaneous solution of the conservation equations of mass, momentum, energy and solute concentration. The penalty method and Galerkin Least Squares (GLS) method are the most commonly employed methods for predicting the interdendritic flow of the liquid melt during the solidification processes. The solidification models employing these methods are computationally inefficient since they are based on the formulations that require the coupled solution to velocity components in the momentum equation Motivated by the inefficiency of the previous solidification models, this work presents three different numerical algorithms for the solution of the volume averaged conservation equations. First, a semi explicit formulation of the projection method that allows the decoupled solution of the velocity components while maintaining the coupling between body force and pressure gradient is presented. This method has been implemented with a standard Galerkin finite element formulation based on bi-linear elements in two dimensions and tri-linear elements in three dimensions. This formulation is shown to be robust and very efficient in terms of both the memory and the computational time required for the macrosegregation computations. The second area addressed in this work is the use of adaptive meshing with linear triangular elements together with the Galerkin finite element method and the projection formulation. An unstructured triangular mesh generator is integrated with the solidification model to produce the solution adapted meshes. Strategies to tackle the different length scales involved in macrosegregation modeling are presented. Meshless element free Galerkin method has been investigated to simulate the solidification processes to alleviate the difficulties associated with the dependence on the mesh. This method is combined with the fractional step method to predict macrosegregation. The performance of these three numerical algorithms has been analyzed and two and three dimensional simulations showing the directional solidification of binary Pb-Sn and multicomponent Ni base alloys are presented.

[pt] Neste trabalho a Análise de Vibrações é tratada no contexto da formulação fraca. Um sistema contínuo é formulado abstratamente em um espaço de Hilbert e uma base de projeção é escolhida para a dinâmica. Um esquema de convergência para a aproximação é garantido à medida em que se aumenta o número de funções da base usada para representar a resposta do problema. Esta é a idéia por traz de métodos como o Método dos Elementos Finitos e o Método dos Modos Supostos, que derivam do Método de Galerkin. Esta estratégia é diferente do que comumente é ensinado nos cursos de vibrações, onde um sistema massa-mola é analisado, e sistemas discretos formados por massas, molas e amortecedores são discutidos. Nestes casos não se sabe qual é o erro cometido na análise numérica. A Análise de Vibrações é muito usada na manutenção preditiva de máquinas rotativas. Alguns fenômenos observados nesses equipamentos motivaram o desenvolvimento de um modelo numérico que pudesse reproduzir tais fenômenos para melhor entendê-los. Um sistema rotor-mancal é modelado e sua resposta dinâmica comparada qualitativamente com a resposta dinâmica captada através de acelerômetros fixados nos mancais de um exaustor da Companhia Siderúrgica de Tubarão (CST). Durante o trabalho diversos programas foram desenvolvidos através da plataforma MATLAB. / [en] Vibration Analysis is treated in the context of weak formulation. A continuous system is formulated in the Hilbert space and one base is selected to project the dynamics. An approximation scheme is guaranteed by increasing the number of functions in the base used to represent the response. This is the idea behind methods like the Finite Element Method and Assumed Modes Method, which derive from Galerkin Method. This strategy is different from what is commonly taught in vibration courses, where a mass-spring system is analyzed and discrete systems composed by masses, springs and dashpots are discussed. In those cases the error of the numerical analysis is not known. Vibration Analysis is very used in predictive maintenance of rotating machines. Some phenomenons observed in those machines motivated the development of a numerical model that could reproduce such phenomenons to better understand them. A rotor-bearing system is modelled and its dynamic response is qualitative compared to the dynamic response captured by accelerometers fixed on the bearings of a blower of the steel company Companhia Siderúrgica de Tubarão (CST). During this work several programs were developed using MATLAB software.

No description available.

Underwater explosions (UNDEX) and a ship's vulnerability to them are problems of interest in early-stage ship design. A series of events occur sequentially in an UNDEX scenario in both the fluid and structural domains and these events happen over a wide range of time and spatial scales. Because of the complexity of the physics involved, it is a common practice to separate the description of UNDEX into early-time and late-time, and far-field and near-field. The research described in this dissertation is focused on the simulation of near-field and early-time UNDEX. It assembles a hybrid framework of algorithms to provide results while maintaining computational efficiency. These algorithms include Runge-Kutta, Discontinuous Galerkin, Level Set, Direct Ghost Fluid and Embedded Boundary methods. Computational fluid dynamics (CFD) solvers are developed using this framework of algorithms to demonstrate the computational methods and their ability to effectively and efficiently solve UNDEX problems. Contributions, made in the process of satisfying the objective of this research include: the derivation of eigenvectors of flux Jacobians and their application to the implementation of the slope limiter in the fluid discretization; the three-dimensional extension of Direct Ghost Fluid Method and its application to the multi-fluid treatment in UNDEX flows; the enforcement of an improved non-reflecting boundary condition and its application to UNDEX simulations; and an improvement to the projection-based embedded boundary method and its application to fluid-structure interaction simulations of UNDEX problems. / Doctor of Philosophy / Underwater explosions (UNDEX) and a ship's vulnerability to them are problems of interest in early-stage ship design. A series of events occur sequentially in an UNDEX scenario in both the fluid and structural domains and these events happen over a wide range of time and spatial scales. Because of the complexity of the physics involved, it is a common practice to separate the description of UNDEX into early-time and late-time, and far-field and near-field. The research described in this dissertation is focused on the simulation of near-field and early-time UNDEX. It assembles a hybrid framework of algorithms to provide results while maintaining computational efficiency. These algorithms include Runge-Kutta, Discontinuous Galerkin, Level Set, Direct Ghost Fluid and Embedded Boundary methods. Computational fluid dynamics (CFD) solvers are developed using this framework of algorithms to demonstrate these computational methods and their ability to effectively and efficiently solve UNDEX problems.

Composite materials and hybrid materials such as fiber-metal laminates, and functionally graded materials are increasingly common in application in aerospace structures. However, adhesive bonding of dissimilar materials makes these materials susceptible to delamination. The use of integrated Sinc methods for predicting interlaminar failure in laminated composites and hybrid material systems was examined. Because the Sinc methods first approximate the highest-order derivative in the governing equation, the in-plane derivatives of in-plane strain needed to obtain interlaminar stresses by integration of the equilibrium equations of 3D elasticity are known without post-processing. Interlaminar stresses obtained with the Sinc method based on Interpolation of Highest derivative were compared for the first-order and third-order shear deformable theories, the refined zigzag beam theory and the higher-order shear and normal deformable beam theory. The results indicate that the interlaminar stresses by the zigzag theory compare well with those obtained by a 3D finite element analysis, while the traditional equivalent single layer theories perform well for some laminates. The philosophy of the Sinc method based on Interpolation of Highest Derivative was extended to create a novel weak form based approach called the Integrated Local Petrov-Galerkin Sinc Method. The Integrated Local Petrov-Galerkin Sinc Method is easily utilized for boundary-value problem on non-rectangular domains as demonstrated for analysis of elastic and elastic-plastic plane-stress panels with elliptical notches. The numerical results showed excellent accuracy compared to similar results obtained with the finite element method. The Integrated Local Petrov-Galerkin Sinc Method was used to analyze interlaminar debonding of composite and fiber-metal laminated beams. A double-cantilever beam and a fixed-ratio mixed mode beam were analyzed using the Integrated Local Petrov-Galerkin Sinc Method and the results were shown to correlate well with those by the finite element method. An adaptive Sinc point distribution technique was implemented for the delamination analysis which significantly improved the methods accuracy for the present problem. Delamination of a GLARE, plane-strain specimen was also analyzed using the Integrated Local Petrov-Galerkin Sinc Method. The results correlate well with 2D, plane-strain analysis by the finite element method, including interlaminar stresses obtained by through-the-thickness integration of the equilibrium equations of 3D elasticity. / Ph. D.

Cette thèse porte sur l’étude d’une méthode de type Galerkin discontinu en domaine temporel (GDDT), afin de résoudre numériquement les équations de Maxwell instationnaires sur des maillages hybrides tétraédriques/hexaédriques en 3D (triangulaires/quadrangulaires en 2D) et non-conformes, que l’on note méthode GDDT-PpQk. Comme dans différents travaux déjà réalisés sur plusieurs méthodes hybrides (par exemple des combinaisons entre des méthodes Volumes Finis et Différences Finies, Éléments Finis et Différences Finies, etc.), notre objectif principal est de mailler des objets ayant une géométrie complexe à l’aide de tétraèdres, pour obtenir une précision optimale, et de mailler le reste du domaine (le vide environnant) à l’aide d’hexaèdres impliquant un gain en terme de mémoire et de temps de calcul. Dans la méthode GDDT considérée, nous utilisons des schémas de discrétisation spatiale basés sur une interpolation polynomiale nodale, d’ordre arbitraire, pour approximer le champ électromagnétique. Nous utilisons un flux centré pour approcher les intégrales de surface et un schéma d’intégration en temps de type saute-mouton d’ordre deux ou d’ordre quatre. Après avoir introduit le contexte historique et physique des équations de Maxwell, nous présentons les étapes détaillées de la méthode GDDT-PpQk. Nous réalisons ensuite une analyse de stabilité L2 théorique, en montrant que cette méthode conserve une énergie discrète et en exhibant une condition suffisante de stabilité de type CFL sur le pas de temps, ainsi que l’analyse de convergence en h (théorique également), conduisant à un estimateur d’erreur a-priori. Ensuite, nous menons une étude numérique complète en 2D (ondes TMz), pour différents cas tests, des maillages hybrides et non-conformes, et pour des milieux de propagation homogènes ou hétérogènes. Nous faisons enfin de même pour la mise en oeuvre en 3D, avec des simulations réalistes, comme par exemple la propagation d’une onde électromagnétique dans un modèle hétérogène de tête humaine. Nous montrons alors la cohérence entre les résultats mathématiques et numériques de cette méthode GDDT-PpQk, ainsi que ses apports en termes de précision et de temps de calcul. / This thesis is concerned with the study of a Discontinuous Galerkin Time-Domain method (DGTD), for the numerical resolution of the unsteady Maxwell equations on hybrid tetrahedral/hexahedral in 3D (triangular/quadrangular in 2D) and non-conforming meshes, denoted by DGTD-PpQk method. Like in several studies on various hybrid time domain methods (such as a combination of Finite Volume with Finite Difference methods, or Finite Element with Finite Difference, etc.), our general objective is to mesh objects with complex geometry by tetrahedra for high precision and mesh the surrounding space by square elements for simplicity and speed. In the discretization scheme of the DGTD method considered here, the electromagnetic field components are approximated by a high order nodal polynomial, using a centered approximation for the surface integrals. Time integration of the associated semi-discrete equations is achieved by a second or fourth order Leap-Frog scheme. After introducing the historical and physical context of Maxwell equations, we present the details of the DGTD-PpQk method. We prove the L2 stability of this method by establishing the conservation of a discrete analog of the electromagnetic energy and a sufficient CFL-like stability condition is exhibited. The theoritical convergence of the scheme is also studied, this leads to a-priori error estimate that takes into account the hybrid nature of the mesh. Afterward, we perform a complete numerical study in 2D (TMz waves), for several test problems, on hybrid and non-conforming meshes, and for homogeneous or heterogeneous media. We do the same for the 3D implementation, with more realistic simulations, for example the propagation in a heterogeneous human head model. We show the consistency between the mathematical and numerical results of this DGTD-PpQk method, and its contribution in terms of accuracy and CPU time.

Le principe du contrôle par flux de fuite magnétique (Magnetic Flux Leakage MFL) consiste à aimanter la pièce à contrôler par un champ magnétique et à détecter à l'aide d'un capteur magnétique les fuites des lignes du champ qui résultent de la présence d'un défaut dans la pièce. Dans le but d'améliorer les performances d'un dispositif de détection, le CEA et la société Vallourec collaborent pour développer un modèle numérique dédié au contrôle virtuel des défauts longitudinaux dans les tubes ferromagnétiques. Le dispositif expérimental comprend un circuit magnétique tournant à une vitesse constante autour du tube qui défile. Dans le cadre de cette thèse, on débute le problème de la modélisation sans tenir compte des effets de la vitesse de rotation, il s'agit donc de résoudre un problème d'électromagnétisme en régime magnétostatique.Pour résoudre ce problème, on propose de comparer une approche semi-analytique basée sur le formalisme des équations intégrales (EI) et une approche purement numérique utilisant les éléments finis (EF).Dans la première partie de cette thèse, après avoir établi le formalisme théorique par EI, un premier modèle considérant des matériaux ferromagnétiques à perméabilité magnétique constante (régime linéaire) a été mis en œuvre en 2D. Ce modèle a été appliqué pour un exemple de système extrait de la littérature et validé numériquement par une comparaison des résultats EI/EF. Pour une meilleure détection, il est opportun de saturer magnétiquement la pièce. Le matériau ferromagnétique est alors caractérisé par une courbe B(H) non-linéaire. Par conséquent, la deuxième partie de la thèse a été consacrée à la mise en œuvre du modèle en régime non linéaire qui tient compte de cette caractéristique.Différentes méthodes de discrétisation ont été étudiées afin de réduire le nombre d'inconnues et le temps de calcul. L'originalité de la thèse réside dans l'utilisation des fonctions d'interpolation d'ordre élevé (polynôme de Legendre) pour une discrétisation des équations intégrales par une approche de type Galerkin. Les premiers essais de validation numérique de ce modèle ont été effectués sur un système MFL simplifié. Des premiers essais de validation expérimentale pour des données obtenues par EF ont été effectués en deux phases : La première a consisté à vérifier le distribution du champ magnétique pour un tube sain et en régime magnétostatique. La deuxième phase a consisté à calculer la réponse d'un défaut dans le tube ferromagnétique en tenant en compte les effets éventuels de la rotation du circuit magnétique par rapport au tube. / The principle of the non destructive testing by magnetic flux leakage (MFL) is to magnetize the part to be inspected by a magnetic field and to detect a flaw thanks to magnetic leakage field lines due to the strong decreasing of the magnetic permeability in the flawed region. In order to improve the performance of detection, the CEA and the Vallourec society collaborate to develop a numerical model dedicated to the virtual NDT of longitudinal defects in ferromagnetic tubes. The experimental system includes a magnetic circuit rotating at a constant speed around the tube to be inspected. The modeling task is started without considering the effects of the rotational speed, so the magnetostatic regime is considered to solve the modeling problem. In the framework of this thesis, we propose to compare a semi-analytical approach based on the formalism of integral equations method (IEM) and a purely numerical approach using finite element method (FEM).In the first part of this thesis, the theoretical formalism was established. A first simple discretization scheme is been implemented in the linear regime considering a constant magnetic permeability. This first numerical model has been validated for a simplified MFL configuration extracted and modified from the literature.For better detection, it is wishable to magnetically saturate the piece under-test. The ferromagnetic material is then characterized by a B(H) curve. Therefore, the second part of the thesis was devoted to the implementation of the model in the non-linear regime that takes into account this non-linear characteristic. Different discretization schemes have been studied in order to reduce the number of unknowns and the computational time. The originality of the thesis lies in the use of basis function of high order (Legendre polynomials) associated to a Galerkin approach for the discretization of integral equations. The first numerical result has been validated on a simplified MFL system. The first results of the experimental validation based on simulated data obtain by FEM have been performed in two steps. The first one consists to verify the distribution of the magnetic field for a ferromagnetic tube without any defect and in the magnetostatic regime. The objective of the second one was to compute the response of the flaw and to evaluate the effects of the rotational speed of the magnetic circuit around the tube.

Dans cette thèse on s'intéresse aux jets en écoulement transverse dans une configuration générique de celle du refroidissement par effusion de chambres de combustion aéronautiques. L'amélioration des modèles de paroi avec transfert de masse passe par une meilleure connaissance de l'interaction entre les jets et l’écoulement principal. Nous avons donc réalisé la simulation numérique directe d'un jet issu d'un perçage incliné avec ou sans giration, pour des écoulements isothermes, turbulents et à bas nombre de Mach, dans un contexte compressible. Pour cela nous avons travaillé avec la bibliothèque AeroSol d'éléments finis continus et discontinus sur maillage hybride. En particulier nous nous sommes intéressés à la stabilité des flux numériques pour le compressible instationnaire associés à la méthode de Galerkin discontinue lorsque le nombre de Mach tend vers zéro. Nous avons pu mettre en évidence des comportements instables lors de l'utilisation de discrétisation temporelle explicite que nous avons corrigés en proposant un nouveau flux. Dans un deuxième temps, nous avons effectué les développements nécessaires à la réalisation des calculs. Nous nous sommes en particulier intéressés à la génération d'un champ de vitesse turbulent synthétique par la méthode SEM (Synthetic Eddy Method) que nous avons implantée dans AeroSol et validée. Grâce aux outils de post-traitement développés, nous avons conduit l'analyse de nos résultats. Dans le cas sans giration, les comparaisons avec les résultats expérimentaux et les résultats de simulations RANS que nous avons obtenus en parallèle sur la configuration du banc d'essai MAVERIC sont encourageants. La structure moyenne d'ensemble du jet est notamment correctement reproduite. En ce qui concerne la cas avec giration, le comportement attendu de déflexion successive du jet dans les deux plans caractéristiques (plan d'injection et plan de l'écoulement transverse) est bien reproduit et illustre tout le potentiel prévisionnel de la librairie de calcul que nous avons contribué à développer. / In this work we are interested in jet in crossflow in a generic configuration to the one used in effusion cooling for combustion chambers. Improved wall models with mass transfer requires a better knowledge of the interaction between the jets and the main flow. We therefore carried out the direct numerical simulation of a jet issuing from an inclined hole with or without gyration, for isothermal turbulent flow at low Mach number, in a compressible context. To achieved this, we worked with the continuous and discontinuous finite element library : AeroSol on hybrid grid. In particular we studied the stability of numerical flux for the unsteady compressible flow associated with discontinuous Galerkin method when the Mach number tends to zero. We were able to demonstrate unstable behavior when using explicit time discretization and we corrected them by providing a new flux. In a second time, we have performed the necessary development to achieve the calculations. We have been especially interested in the generation of a synthetic turbulent velocity field using the SEM method (Synthetic Eddy Method) that we have implemented in aerosol and validate. Thanks to the developed post-processing tools, we have conducted an analysis of our results. In the case without gyration, comparisons with experimental results and the results of RANS simulations we obtained on the Maveric test-bench configuration are encouraging. The mean flow of the jet is correctly reproduced. In the case with gyration, the expected behavior of successive deflection of the jet in both planes (injection plane and transverse plane of the flow) is reproduced and shows all the potential of the AeroSol library we helped to develop.

Les projets d'Ingénierie Structurale doivent s’adapter aux critères de performance, de sécurité, de fonctionnalité, de durabilité et autres, établis dans la phase d’avant-projet. Traditionnellement, les projets utilisent des informations de nature déterministe comme les dimensions, les propriétés des matériaux et les charges externes. Toutefois, la modélisation des systèmes structuraux complexes implique le traitement des différents types et niveaux d'incertitudes. Dans ce sens, la prévision du comportement doit être préférablement faite en termes de probabilités puisque l'estimation de la probabilité de succès d'un certain critère est une nécessité primaire dans l’Ingénierie Structurale. Ainsi, la fiabilité est la probabilité rapportée à la parfaite opération d'un système structural donné durant un certain temps en des conditions normales d'opération pour trouver le meilleur compromis entre coût et sécurité pour l’élaboration des projets. Visant à pallier les désavantagés des méthodes traditionnelles FORM et SORM (First and Second Order Reliability Method), cette thèse propose une méthode d’analyse de fiabilité basée sur des techniques d’optimisation heuristiques (HBRM, Heuristic-based Reliability Method). Les méthodes heuristiques d’optimisation utilisées par cette méthode sont : Algorithmes Génétiques (Genetic Algorithms), Optimisation par Essaims Particulaires (Particle Swarm Optimisation) et Optimisation par Colonie de Fourmis (Ant Colony Optimization). La méthode HBRM ne requiert aucune estimation initiale de la solution et opère selon le principe de la recherche multi-directionnelle, sans besoin de calculer les dérivées partielles de la fonction d’état limite par rapport aux variables aléatoires. L’évaluation des fonctions d’état limite est réalisée en utilisant modèles analytiques, semi analytiques et numériques. Dans ce but, la mise en oeuvre de la méthode de Ritz (via MATLAB®), la méthode des éléments finis (via MATLAB® et ANSYS®) et la méthode sans maillage de Galerkin (Element-free Galerkin sous MATLAB®) a été nécessaire. La combinaison d’analyse de fiabilité, des méthodes d’optimisation et méthodes de modélisation, ci-dessus mentionnées, configure la méthodologie de conception fiabiliste proposée dans ce mémoire. L’utilisation de différentes méthodes de modélisation et d’optimisation a eu pour objectif de mettre en évidence leurs avantages et désavantages pour des applications spécifiques, ainsi pour démontrer l’applicabilité et la robustesse de la méthodologie de conception fiabiliste en utilisant ces techniques numériques. Ce qui a été possible grâce aux bons résultats trouvés dans la plupart des applications. Dans ce sens, des applications uni, bi et tridimensionnelles en statique, stabilité et dynamique des structures explorent l’évaluation explicite et implicite des fonctions d’état limite de plusieurs variables aléatoires. Procédures de validation déterministe et analyses stochastiques, et la méthode de perturbation de Muscolino, donnent les bases de l’analyse de fiabilité des applications en problèmes d’interaction fluide-structure bi et tridimensionnelles. La méthodologie est particulièrement appliquée à une structure industrielle. Résultats de applications uni et bidimensionnelles aux matériaux composites stratifiés, modélisés par la méthode EFG sont comparés avec les obtenus par éléments finis. A la fin de la thèse, une extension de la méthodologie à l’optimisation fiabiliste est proposée à travers la méthode des facteurs optimaux de sûreté. Pour cela, sont présentes des applications pour la minimisation du poids, en exigent un indice de fiabilité cible, aux systèmes modélisés par la méthode de EF et par la méthode EFG. / Structural Engineering designs must be adapted to satisfy performance criteria such as safety, functionality, durability and so on, generally established in pre-design phase. Traditionally, engineering designs use deterministic information about dimensions, material properties and external loads. However, the structural behaviour of the complex models needs to take into account different kinds and levels of uncertainties. In this sense, this analysis has to be made preferably in terms of probabilities since the estimate the probability of failure is crucial in Structural Engineering. Hence, reliability is the probability related to the perfect operation of a structural system throughout its functional lifetime; considering normal operation conditions. A major interest of reliability analysis is to find the best compromise between cost and safety. Aiming to eliminate main difficulties of traditional reliability methods such as First and Second Order Reliability Method (FORM and SORM, respectively) this work proposes the so-called Heuristic-based Reliability Method (HBRM). The heuristic optimization techniques used in this method are: Genetic Algorithms, Particle Swarm Optimization and Ant Colony Optimization. The HBRM does not require initial guess of design solution because it’s based on multidirectional research. Moreover, HBRM doesn’t need to compute the partial derivatives of the limit state function with respect to the random variables. The evaluation of these functions is carried out using analytical, semi analytical and numerical models. To this purpose were carried out the following approaches: Ritz method (using MATLAB®), finite element method (through MATLAB® and ANSYS®) and Element-free Galerkin method (via MATLAB®). The combination of these reliability analyses, optimization procedures and modelling methods configures the design based reliability methodology proposed in this work. The previously cited numerical tools were used to evaluate its advantages and disadvantages for specific applications and to demonstrate the applicability and robustness of this alternative approach. Good agreement was observed between the results of bi and three-dimensional applications in statics, stability and dynamics. These numerical examples explore explicit and implicit multi limit state functions for several random variables. Deterministic validation and stochastic analyses lied to Muscolino perturbation method give the bases for reliability analysis in 2-D and 3-D fluidstructure interaction problems. This methodology is applied to an industrial structure lied to a modal synthesis. The results of laminated composite plates modelled by the EFG method are compared with their counterparts obtained by finite elements. Finally, an extension in reliability based design optimization is proposed using the optimal safety factors method. Therefore, numerical applications that perform weight minimization while taking into account a target reliability index using mesh-based and meshless models are proposed. / Os projectos de Engenharia Estrutural devem se adaptar a critérios de desempenho, segurança, funcionalidade, durabilidade e outros, estabelecidos na fase de anteprojeto. Tradicionalmente, os projectos utilizam informações de natureza deterministica nas dimensões, propriedades dos materiais e carregamentos externos. No entanto, a modelagem de sistemas complexos implica o tratamento de diferentes tipos e níveis de incertezas. Neste sentido, a previsão do comportamento deve preferivelmente ser realizada em termos de probabilidades dado que a estimativa da probabilidade de sucesso de um critério é uma necessidade primária na Engenharia Estrutural. Assim, a confiabilidade é a probabilidade relacionada à perfeita operação de um sistema estrutural durante um determinado tempo em condições normais de operação. O principal objetivo desta análise é encontrar o melhor compromisso entre custo e segurança. Visando a paliar as principais desvantagens dos métodos tradicionais FORM e SORM (First and Second Order Reliability Method), esta tese propõe um método de análise de confiabilidade baseado em técnicas de optimização heurísticas denominado HBRM (Heuristic-based Reliability Method). Os métodos heurísticos de otimização utilizados por este método são: Algoritmos Genéticos (Genetic Algorithms), Optimização por Bandos Particulares (Particle Swarm Optimisation) e Optimização por Colónia de Formigas (Ant Colony Optimization). O método HBRM não requer de uma estimativa inicial da solução e opera de acordo com o princípio de busca multidirecional, sem efetuar o cálculo de derivadas parciais da função de estado limite em relação às variáveis aleatórias. A avaliação das funções de estado limite é realizada utilizando modelos analíticos, semi analíticos e numéricos. Com este fim, a implementação do método de Ritz (via MATLAB®), o método dos elementos terminados (via MATLAB® e ANSYS®) e o método sem malha de Galerkin (Element-free Galerkin via MATLAB®) foi necessária. A combinação da análise de confiabilidade, os métodos de optimização e métodos de modelagem, acima mencionados, configura a metodologia de projeto proposta nesta tese. A utilização de diferentes métodos de modelagem e de otimização teve por objetivo destacar as suas vantagens e desvantagens em aplicações específicas, assim como demonstrar a aplicabilidade e a robustez da metodologia de análise de confiabilidade utilizando estas técnicas numéricas. Isto foi possível graças aos bons resultados encontrados na maior parte das aplicações. As aplicações foram uni, bi e tridimensionais em estática, estabilidade e dinâmica de estruturas, as quais exploram a avaliação explícita e implícita de funções de estado limite de várias variáveis aleatórias. Procedimentos de validação déterministica e de análises estocásticas, aplicando o método de perturbação de Muscolino, fornecem as bases da análise de confiabilidade nas aplicações de problemas de iteração fluído-estrutura bi e tridimensionais. A metodologia é testada com uma estrutura industrial. Resultados de aplicações bidimensionais em estratificados compostos, modelados pelo método EFG são comparados com os obtidos por elementos finitos. No fim da tese, uma extensão da metodologia à optimização baseada em confiabilidade é proposta aplicando o método dos factores óptimos de segurança. Finalmente são apresentadas as aplicações para a minimização do peso em sistemas modelados pelo método de EF e o método EFG que exigem um índice de confiabilidade alvo.