Os perfis de aço formados a frio apresentam, em geral, maior esbeltez local (relação largura-espessura dos elementos) em relação aos clássicos perfis laminados, acentuando a instabilidade local. Além disso, em se tratando de seções abertas com paredes muito delgadas, a rigidez à torção resulta muito pequena, o que torna os modos globais de torção e flexo-torção muitas vezes dominantes em relação aos modos de flexão. Outro modo de instabilidade que pode se manifestar é o modo distorcional, característico nos perfis com enrijecedores de borda. Com relação à análise do modo global, as normas para cálculo de perfis formados a frio têm adotado as mesmas curvas de resistência à compressão desenvolvidas para os perfis laminados e soldados, como a curva do SSRC (Structural Stability Research Council), adotada pela NAS (North American Specification), e as curvas européias, adotadas pela norma brasileira. Embora alguns estudos indiquem que as citadas curvas sejam aceitáveis para os perfis formados a frio, há também referências explícitas quanto à necessidade de um maior aprofundamento na investigação sobre o comportamento estrutural destes perfis, uma vez que apresentam particularidades quanto às tensões residuais, imperfeições geométricas e interação entre modos de instabilidade. Nesse trabalho é apresentada uma análise experimental em perfis usualmente empregados no Brasil (perfis U, U enrijecidos e cantoneiras simples e duplas), e uma estratégia de análise numérica não-linear, considerando os efeitos das imperfeições geométricas globais e localizadas (de chapa e distorcional), bem como das tensões residuais, de modo a se obter teoricamente um valor confiável da força normal de compressão resistente da barra. Os resultados permitiram constatar a viabilidade do emprego das atuais curvas de resistência à compressão para os perfis formados a frio. Complementando, foi analisada a aplicação do método da resistência direta (MRD) a todos os perfis estudados, confirmando bons resultados. Especial atenção foi dada ao estudo da estabilidade elástica de cantoneiras, com foco principal na coincidência entre o modo local-chapa e o modo global-torsional, o que tem gerado controvérsias na aplicação dos métodos de cálculo. Além disso, como as cantoneiras não são pré-qualificadas para aplicação do MRD, foram analisadas várias opções para emprego do método, onde pode-se concluir que desconsiderar a torção na análise do modo global conduz a resultados contra a segurança / Cold-formed steel members present, in general, higher local slenderness than classical hot- rolled ones, which make them more prone to local buckling. Besides, thin-walled open sections have small torsional stiffness, and hence global torsional and flexural-torsional instability modes are many times more critical than global flexural ones. Also, distortional mode can happen in sections with lips (edge stiffener). Concerning on global buckling for members under compression, curves used in cold-formed steel design are based on hot-rolled and welded members. For example, the SSRC (Structural Stability Research Council) buckling curve, adopted by NAS (North American Specification), and Eurocode buckling curves, adopted by brazilian codes. Although some papers indicate these curves are acceptable for cold-formed steel members, others claim for a deeper analysis on their unique structural behavior, specially on residual stress, geometric imperfections and coupled buckling modes. It is presented in this thesis an experimental analysis of sections usually used in Brazil (simple and lipped channels, and also single and built-up angles). Moreover, it is developed a strategy for numerical non-linear analysis, considering the effects of global and local (also distortional) geometric imperfections and residual stress as well, in order to obtain a trustable theoretical value for the axial member stength. Results show the viability of the current buckling curves for cold-formed steel members. Finally, direct strength method (DSM) was analysed for all studied members, showing good results. Special attention to angle’s elastic stability, focusing on the coincidence between local-plate and global-torsional mode, which still causes confusion in design methods. Also, due to the fact angles are not pre- qualified sections for using DSM, many options on its application were studied, where it was concluded that negleting torsion in global analysis leeds to unconservative results

Les procédés de traitement de surface sont utilisés à l'échelle industrielle pour améliorer les performances de pièces mécaniques en introduisant des contraintes résiduelles de compression. Cette mise en compression de surface permet de limiter l'amorçage et la propagation de fissures dans le matériau. Ceci permet d'augmenter de façon significative la durée de vie en fatigue des pièces mécaniques ainsi traitées. L'utilisation de ces procédés dans l'industrie a démontré leur efficacité, mais aussi leurs limitations et inconvénients. Les défauts récurrents consistent en une profondeur traitée faible, une dégradation de l'état de surface (rugosité), des difficultés de contrôle, une contamination du matériau traité, etc. Ces défauts ont conduit à l'élaboration de nouveaux procédés innovants qui permettent de meilleures performances en évitant certains des inconvénients succinctement évoqués. Parmi ces procédés innovants, le traitement de surface par impulsion électromagnétique semble particulièrement intéressant. Ce procédé met en œuvre un puissant champ magnétique transitoire pour engendrer des forces de Laplace dans une pièce métallique et induire des contraintes résiduelles. Il n'existe que peu d'informations dans la littérature et il n'existe aucun dispositif expérimental de ce procédé. Cette thèse est dédiée à la conception et la réalisation d'un prototype de mise en compression électromagnétique. Le premier chapitre de cette thèse est un état de l'art des technologies de mise en compression et du procédé de mise en compression par impulsion électromagnétique. Ainsi, les besoins de ce procédé sont identifiés et les technologies pouvant répondre à ces besoins sont explorées. Le deuxième chapitre, après une sélection de la structure globale du dispositif, va consister aux dimensionnements des éléments du prototype EMP. Cette étude commencera avec une étude sur l'inducteur qui va être utilisé avant de continuer sur le dimensionnement du stockage d'énergie et de l'interrupteur de décharge. Afin de valider le dimensionnement des composants précédents, une simulation électromagnétique 3D du système est réalisée. L'assemblage du prototype est présenté dans le troisième chapitre ainsi qu'une première campagne d'essai sur un alliage d'aluminium. Deux types d'éprouvettes sont testées : une éprouvette fine pour vérifier visuellement la mise en compression (essai Almen) et une éprouvette massive afin d'évaluer la profondeur traitée. Une modélisation multiphysique 3D du procédé est réalisée afin de corréler ces résultats avec l'expérience. Dans un dernier chapitre, une étude exploratoire est menée sur un matériau ferromagnétique, le mumétal, pour visualiser l'influence des contraintes résiduelles sur les propriétés magnétiques de ce dernier. / Penning processes are widely used in industries to apply compressive residual stresses into the most solicited part of mechanical pieces. In that way, the compressive residual stresses limit the priming and the propagation of micro-cracks in the material. This increases significantly the lifespan of the treated mechanical piece under fatigue stresses. These existing peening processes have proved their efficiency and also their limitations and weaknesses. The main recurrent defaults are a shallow depth of treatment, a degradation of the surface condition, a random control of the treatment, a material contamination, etc. These problems have led towards the development of news innovative peening processes which allow better performance avoiding some previous defaults briefly evoked. Among these news processes, the electromagnetic peening process seems especially interesting. This process uses high energy electromagnetic fields to induce Lorentz forces into a metallic piece and thus residual stresses. Actually, there is not much information about this process in the literature and no prototype was ever built. The work of this thesis is dedicated to development and realization of an electromagnetic peening prototype. The first chapter of this thesis adresses the state of the art of major peening processes actually in industrial use. Next, the electromagnetic peening process, or EMP process, is described and the electrical needs are exposed. A second state of the art is made about the technological solutions to respond to the EMP needs. The second chapter is about the conception of the EMP prototype with the electrical structure adopted in the previous chapter. The first step is about the inductor sizing to generate an electromagnetic field sufficient enough for a peening application. Next, the storage system is designed depending on the inductor parameters and finally the closing switch is created considering the electrical parameters used for the EMP process. To validate the previous results, a 3D electromagnetic simulation is done. The prototype assembly is presented in the third chapter and also the first experimental test on the EMP prototype. To begin with, an aluminium alloy with low yield strength is selected to be treated. Two different samples forms are used, a thin one, to realize a similar test to the Almen test and thick one to check the EMP depth of treatment. A 3D multiphysics simulation of these experiments is made and these numeric results are next correlated to the experimental ones. In the fourth chapter, an exploratory study is realized on the effects of the residual stresses on magnetic properties of ferromagnetic material, the mumetal

This thesis describes advancements in the modelling, optimisation, process control and mechanical performance of novel high deposition rate gas metal arc welding processes for large scale additive manufacturing applications. One of the main objectives of this study was to develop fundamental understanding of the mechanisms involved during processing with particular focus on single layer welds made of carbon steel using both pulsed-current gas metal arc welding and cold metal transfer processes. The effects of interactions between critical welding process variables and weld bead and plate fusion characteristics are studied for single and multi-layers. It was shown that several bead and plate fusion characteristics are strongly affected by the contact tip to work distance, TRIM, wire feed speed, wire feed speed to travel speed ratio, and wire diameter in pulsed-current gas metal arc welding. The arc-length control, dynamic correction and the contact tip to work distance are shown to strongly influence the weld bead geometry in the cold metal transfer process. This fundamental knowledge was essential to ensure the successful development of predictive interaction models capable of determining the weld bead geometry from the welding process parameters. The models were developed using the least-squares analysis and multiple linear regression method. The gas tungsten constricted arc welding process was utilised for the first time for out-of-chamber fabrication of a large scale and high-quality Ti-6Al-4V component. The main focus was, however, in the use of the cold metal transfer process for improving out-of-chamber deposition of Ti-6Al-4V at much higher deposition rates. The effect of the cold metal transfer process on the grain refinement features in the fusion zone of single layer welds under different torch gas shielding conditions was investigated. It was shown that significant grain refinement occurs with increasing helium content. The morphological features and static mechanical performance of the resulting multi-layered Ti-6Al-4V walls were also examined and compared with those in gas tungsten constricted arc welding. The results show that a considerable improvement in static tensile properties is obtained in both testing directions with cold metal transfer over gas tungsten constricted arc welding. It was suggested that this improvement in the mechanical behaviour could be due to the formation of more fine-grained structures,which are therefore more isotropic. The average ultimate tensile strength and yield strength of the as-deposited Ti-6Al-4V material processed via cold metal transfer meet the minima specification values recommended for most Ti-6Al-4V products. Neutron diffraction technique was used to establish the effect of repeated thermo-mechanical cycling on the generation, evolution and distribution of residual stresses during wire and arc additive manufacturing. The results show a significant redistribution of longitudinal residual stresses along both the substrate and multi-bead with repeated deposition. However, a nearly complete relaxation occurs along the built, once the base plate constraint is removed.

Fracture Mechanics process of Welded Joint is a very vast research area and has many possibilities for solution and prediction. Although the fatigue strength (FAT) and stress intensity factor (SIF) solutions are reported in several handbooks and recommendations, these values are available only for a small number of specimens, components, loading and welding geometries. The available solutions are not always adequate for particular engineering applications. Moreover, the reliable solutions of SIF are still difficult to find in spite of several SIF handbooks have been published regarding the nominal applied SIF. The effect of residual stresses is still the most challenge in fatigue life estimation. The reason is that the stress distributions and SIF modified by the residual stresses have to be estimated. The stress distribution is governed by many parameters such as the materials type, joint geometry and welding processes. In this work, the linear elastic fracture mechanics (LEFM), which used crack tip SIFs for cases involving the effect of weld geometry, is used to calculate the crack growth life for some different notch cases. The variety of crack configurations and the complexity of stress fields occurring in engineering components require more versatile tools for calculating SIFs than available in handbook’s solutions that were obtained for a range of specific geometries and load combinations. Therefore, the finite element method (FEM) has been used to calculate SIFs of cracks subjected to stress fields. LEFM is encoded in the FEM software, FRANC, which stands for fracture analysis code. The SIFs due to residual stress are calculated in this work using the weight function method. The fatigue strength (FAT) of load-carrying and non-load carrying welded joints with lack of penetration (LOP) and toe crack, respectively, are determined using the LEFM. In some studied cases, the geometry, material properties and loading conditions of the joints are identical to those of specimens for which experimental results of fatigue life and SIF were available in literature so that the FEM model could be validated. For a given welded material and set of test conditions, the crack growth behavior is described by the relationship between cyclic crack growth rate, da/dN, and range of the stress intensity factor ( K) , i.e., by Paris’ law. Numerical integration of the Paris’ equation is carried out by a FORTRAN computer routine. The obtained results can be used for calculating FAT values. The computed SIFs along with the Paris’ law are used to predict the crack propagation. The typical crack lengths for each joint geometry are determined using the built language program by backward calculations. To incorporate the effect of residual stresses, the fatigue crack growth equations which are sensitive to stress ratio R are recommended to be used. The Forman, Newman and de Konig (FNK) solution is considered to be the most suitable one for the present purpose. In spite of the recent considerable progress in fracture mechanics theories and applications, there seems to be no, at least to the author’s knowledge, systematic study of the effect of welding geometries and residual stresses upon fatigue crack propagation based completely on an analytical approach where the SIF due to external applied load (Kapp) is calculated using FEM. In contrast, the SIF due to residual stresses (Kres) is calculated using the analytical weight function method and residual stress distribution. To assess the influence of the residual stresses on the failure of a weldment, their distribution must be known. Although residual stresses in welded structures and components have long been known to have an effect on the components fatigue performance, access to reliable, spatially accurate residual stress field data are limited. This work constitutes a systematic research program regarding the concept for the safety analysis of welded components with fracture mechanics methods, to clarify the effect of welding residual stresses upon fatigue crack propagation. / Die Bewertung einer Schweißnaht ist ein großes Forschungsgebiet und hat viele Möglichkeiten für Lösungskonzepte und Vorhersagen. Obwohl für die Schwingfestigkeit und die Spannungsintensitätsfaktor (SIF)-Lösungen in verschiedenen Handbüchern Empfehlungen ausgewiesen sind, sind diese Werte nur für eine geringe Anzahl von Proben, Komponenten, Belastungsfälle und Schweißgeometrien verfügbar. Die vorhandenen Lösungsansätze sind nicht immer für spezielle technische Anwendungen geeignet. Darüber hinaus sind zuverlässige bewährte Lösungen von Spannungsintensitätsfaktoren immer noch schwierig zu finden, obwohl verschiedene SIF-Handbücher mit Hinweis auf den anliegenden nominalen SIF veröffentlicht sind. Der Einfluss von Eigenspannungen ist eine der größten Herausforderungen bei der Lebensdauerabschätzung. Aufgrund der Tatsache, dass infolge der Eigenspannungen sowohl die Spannungsverteilung als auch der SIF verändert werden, muss eine Abschätzung erfolgen. Die Spannungsverteilung wird durch viele Parameter beeinflusst, wie zum Beispiel den Werkstoff, die Nahtgeometrie und den Schweißprozess. In der vorliegenden Arbeit wurde für die Berechnung des Ermüdungsrisswachstums unter verschiedenen Kerbfällen das Konzept der linear-elastischen Bruchmechanik (LEBM) verwendet, welches K-Lösungen für die Rissspitze bei unterschiedlichen Fällen der Schweißgeometrie berücksichtigt. Aufgrund der Komplexität der Risskonfigurationen und der Spannungsfelder in praxisrelevanten Komponenten werden weitere Hilfsmittel zur Berechnung von Spannungsintensitätsfaktoren benötigt, welche die herkömmlichen Lösungen in Handbüchern erweitern. Deshalb wurde die Finite Elemente Methode (FEM) zur Berechnung von Spannungsintensitätsfaktoren an Rissen verwendet. Die LEBM wird in der FEMSoftware FRANC berücksichtigt. Die aus Eigenspannungen resultierenden Spannungsintensitätsfaktoren wurden mit Hilfe der Gewichtsfunktionsmethode berechnet. Die Ermüdungslebensdauer (Schwingfestigkeit) von tragenden und nichttragenden Schweißnähten mit ungenügender Durchschweißung beziehungsweise Kerbriss wurden mit Hilfe der LEBM durch Integration der Zyklischen Risswachstumskurve ermittelt. Zur Validierung des FEM-Modells konnte in einigen untersuchten Fällen auf experimentelle Ergebnisse zur Lebensdauer und zum SIF aus der Literatur zurückgegriffen werden, wo identische Geometrien, Materialeigenschaften und Belastungsverhältnisse der Naht vorlagen. Unter Vorgabe des Werkstoffes und der Prüfbedingungen wurde das Risswachstumsverhalten mit dem Zusammenhang von Risswachstumsgeschwindigkeit da/dN und zyklischem Spannungsintensitätsfaktor K mit dem Paris-Gesetz beschrieben. Eine numerische Integration der Paris-Gleichung erfolgte über ein FORTRAN-Programm. Die damit erhaltenen Ergebnisse sind als Ermüdungslebensdauer (Schwingfestigkeit) verwendbar. Die berechneten SIF‘en entlang der Paris-Geraden werden zur Vorhersage des Risswachstums benutzt. Die typischen Risslängen für jede Nahtgeometrie wurden mit Hilfe des eigens integrierten Programmes ermittelt. Zur Berücksichtigung des Einflusses von Eigenspannungen wird empfohlen, Risswachstumsgleichungen zu nutzen, die empfindlich auf das Spannungsverhältnis R reagieren. Für die vorliegende Zielsetzung gilt der Lösungsansatz nach Forman, Newman und de Konig (FNK) als der am besten geeignete. Trotz der jüngsten, beträchtlichen Fortschritte in den bruchmechanischen Theorien und Anwendungen sind systematische Studien zum Einfluss der Schweißgeometrie und der Eigenspannungen auf das Ermüdungsrisswachstum, in welchen der SIF aufgrund extern anliegender Beanspruchungen (Kapp) mit der FEM berechnet wurde, in der Literatur kaum vorhanden. Im Gegensatz dazu wurde der SIF infolge von Eigenspannungen (Kres) mit Hilfe der analytischen Gewichtsfunktionsmethode und der Eigenspannungsverteilung berechnet. Um den Einfluss von Eigenspannungen auf das Versagen einer Schweißverbindung abzuschätzen, muss deren Verteilung bekannt sein. Obwohl die Wirkung von Eigenspannungen auf das Ermüdungsverhalten in geschweißten Strukturen und Komponenten schon lange bekannt ist, ist der Zugriff auf verlässliche und präzise Daten von räumlichen Eigenspannungsfeldern begrenzt. Bezüglich einer konzeptionellen Sicherheitsanalyse von geschweißten Komponenten mit bruchmechanischen Methoden begründet diese Arbeit einen systematischen Ansatz, um den Einfluss von Schweißeigenspannungen auf das Ermüdungsrisswachstum zu verdeutlichen.

Os perfis de aço formados a frio apresentam, em geral, maior esbeltez local (relação largura-espessura dos elementos) em relação aos clássicos perfis laminados, acentuando a instabilidade local. Além disso, em se tratando de seções abertas com paredes muito delgadas, a rigidez à torção resulta muito pequena, o que torna os modos globais de torção e flexo-torção muitas vezes dominantes em relação aos modos de flexão. Outro modo de instabilidade que pode se manifestar é o modo distorcional, característico nos perfis com enrijecedores de borda. Com relação à análise do modo global, as normas para cálculo de perfis formados a frio têm adotado as mesmas curvas de resistência à compressão desenvolvidas para os perfis laminados e soldados, como a curva do SSRC (Structural Stability Research Council), adotada pela NAS (North American Specification), e as curvas européias, adotadas pela norma brasileira. Embora alguns estudos indiquem que as citadas curvas sejam aceitáveis para os perfis formados a frio, há também referências explícitas quanto à necessidade de um maior aprofundamento na investigação sobre o comportamento estrutural destes perfis, uma vez que apresentam particularidades quanto às tensões residuais, imperfeições geométricas e interação entre modos de instabilidade. Nesse trabalho é apresentada uma análise experimental em perfis usualmente empregados no Brasil (perfis U, U enrijecidos e cantoneiras simples e duplas), e uma estratégia de análise numérica não-linear, considerando os efeitos das imperfeições geométricas globais e localizadas (de chapa e distorcional), bem como das tensões residuais, de modo a se obter teoricamente um valor confiável da força normal de compressão resistente da barra. Os resultados permitiram constatar a viabilidade do emprego das atuais curvas de resistência à compressão para os perfis formados a frio. Complementando, foi analisada a aplicação do método da resistência direta (MRD) a todos os perfis estudados, confirmando bons resultados. Especial atenção foi dada ao estudo da estabilidade elástica de cantoneiras, com foco principal na coincidência entre o modo local-chapa e o modo global-torsional, o que tem gerado controvérsias na aplicação dos métodos de cálculo. Além disso, como as cantoneiras não são pré-qualificadas para aplicação do MRD, foram analisadas várias opções para emprego do método, onde pode-se concluir que desconsiderar a torção na análise do modo global conduz a resultados contra a segurança / Cold-formed steel members present, in general, higher local slenderness than classical hot- rolled ones, which make them more prone to local buckling. Besides, thin-walled open sections have small torsional stiffness, and hence global torsional and flexural-torsional instability modes are many times more critical than global flexural ones. Also, distortional mode can happen in sections with lips (edge stiffener). Concerning on global buckling for members under compression, curves used in cold-formed steel design are based on hot-rolled and welded members. For example, the SSRC (Structural Stability Research Council) buckling curve, adopted by NAS (North American Specification), and Eurocode buckling curves, adopted by brazilian codes. Although some papers indicate these curves are acceptable for cold-formed steel members, others claim for a deeper analysis on their unique structural behavior, specially on residual stress, geometric imperfections and coupled buckling modes. It is presented in this thesis an experimental analysis of sections usually used in Brazil (simple and lipped channels, and also single and built-up angles). Moreover, it is developed a strategy for numerical non-linear analysis, considering the effects of global and local (also distortional) geometric imperfections and residual stress as well, in order to obtain a trustable theoretical value for the axial member stength. Results show the viability of the current buckling curves for cold-formed steel members. Finally, direct strength method (DSM) was analysed for all studied members, showing good results. Special attention to angle’s elastic stability, focusing on the coincidence between local-plate and global-torsional mode, which still causes confusion in design methods. Also, due to the fact angles are not pre- qualified sections for using DSM, many options on its application were studied, where it was concluded that negleting torsion in global analysis leeds to unconservative results

A presente tese tem por objetivo contribuir para o conhecimento do comportamento dependente do tempo das rochas de anidrita e gipso. O maior interesse é fornecer argumentos convincentes sobre os mecanismos físicos responsáveis pelas deformações lentas observadas em algumas escavações subterrâneas realizadas em maciços dessas rochas na Europa, aspecto este que hoje está bastante confuso e escassamente estudado. A hipótese central desta pesquisa é que o comportamento dependente do tempo das rochas de anidrita e gipso deve-se parcial e/ou totalmente às propriedades reológicas (fluência) e não exclusivamente a expansão decorrente da transformação anidrita – gipso como a maioria dos autores sustenta. Este problema será analisado a partir de uma perspectiva experimental através de ensaios específicos. Os ensaios de expansão axial livre em rochas de anidrita permitiram concluir que quando submersas em água destilada, experimentam diminuição de volume por dissolução. Já as amostras de anidrita em contato com solução saturada em 'CA'SO IND.4' mostram uma relação não linear entre deformação axial e o tempo indicando uma taxa de expansão que decresce com o tempo. As taxas de expansão axial variam entre 0,3 e 2,4%/ano. A novidade é que esta expansão se deve à deposição de uma camada contínua formada por neocristais de gipso e não a expansão por hidratação da anidrita como acontece com os argilominerais. O mecanismo de transformação consiste na dissolução da anidrita e posterior precipitação do gipso em condições de sobre-saturação da solução. A transformação mineralógica acontece, na natureza, em condições propícias de temperatura e saturação da solução com íons 'CA POT.2+' e 'SO IND.4'POT.2-'. Propõe-se uma teoria alternativa sobre deformações lentas de maciços sulfáticos baseada na força de cristalização decorrente da deposição de cristais de gipso nas juntas do maciço rochoso. Propõe-se utilizar as formulações termodinâmicas propostas para expansão de concretos por crescimento de etringita e gipso. Os ensaios de fluência uniaxial sob compressão axial constante mostram que ambos os tipos de rochas sulfáticas apresentam comportamentos elasto-visco-plásticos e taxas de deformação axial e lateral da ordem de '10 POT.-12' a '10 POT.-10' 'S POT.-1'. A deflagração da fluência secundária acontece sob tensões de 4 – 6 MPa para gipsitas e de 25 – 40 MPa para anidritas. As análises sobre os possíveis mecanismos de deformação por fluência foram inconsistentes e nenhuma conclusão definitiva foi atingida. Aparentemente a baixas tensões atuariam mecanismos de difusão e dissolução por pressão. A tensões intermediárias dominariam mecanismos de deslocamentos intra e intercristalinos e a altas tensões as rochas deformariam por propagação de microfissuramento. Ensaios de fluência acoplados a sensores de emissão acústica mostram a manifestação de eventos microssísmicos que usualmente se atribuem à propagação de microfraturas mas que podem dever-se a outros mecanismos. Tanto as taxas de expansão axial como de fluência são compatíveis com as taxas de convergência medidas em algumas obras subterrâneas escavadas em maciços sulfáticos, verificando a hipótese central da pesquisa / The purpose of this thesis is to contribute for the knowledge of time-dependent behavior of sulfatic rocks of anhydrite and gypsum. Emphasis is given to provide straightforward arguments about physical mechanisms responsible for creep observed in some underground excavations in those rocks in Europe. Explanation for this phenomena is not clear yet and studies about them are still scarce. The backbone idea behind this research is that the time-dependent behavior of anhydrite and gypsum is totally or at least partially due to rheological properties, and not exclusively due to swelling resulting from the anhydritegypsum transformation. This problem was investigated experimentally with specific tests. Free swelling tests on anhydrite led to the conclusion that when immersed into distilled water, volume decrease due to dissolution is observed. Anhydrite samples in contact with a saturated solution of 'CA'SO IND.4' present a nonlinear relationship between axial strain and time, indicating swelling rate decreasing with time. Axial swelling rates vary between 0,3% and 2,4%/year. The new concept is that this swelling is due to the deposition of a continuous layer of gypsum composed by gypsum neo-crystals, and not due to hydration swelling of anhydrite, a usual with clay minerals. The transformation mechanism consists of anhydrite dissolution and later precipitation of gypsum under over-saturation condition of the solution. The mineralogical transformation takes place in nature under favorable conditions of temperature and solution saturation with 'CA POT.2+' and 'SO IND.4'POT.2-' ions. An alternative theory is proposed about creep of sulfatic rock masses based on crystallization forces resulting from the deposition of gypsum crystals in the rock masses joints. The use of thermodynamic formulations is proposed for concrete swelling due to the growth of ettringite and gypsum. Uniaxial compression creep tests show that both types of sulfatic rocks present elasto-visco-plastic behavior and axial and lateral strain rates of the order of '10 POT.-12' to '10 POT.-10' 'S POT.-1'. Triggering of secondary creep takes places under stresses of the order of 4 – 6 MPa for gypsum and 25 – 40 MPa for anhydrite. The analyses about the possible creep mechanisms were not consistent and no definite conclusion has been reached get. Apparently under low stresses diffusion and pressure dissolution mechanism dominate. Under intermediate stresses, intra- and inter crystalline displacement mechanisms seem to dominate, and under high stresses rock deform due to microcrack propagation. Creep tests monitored with acoustic emission devices show the occurrence of microseismic events attributed to microcrack propagation. Both axial swelling and creep rock are compatible with convergence rocks measured in some underground works excavated in sulfatic rock masses, following the central idea of this thesis

Le grenaillage est souvent utilisé sur les pieds d’aube de turbine haute pression afin de retarder l’apparition des fissures dans les zones de concentration de contraintes. Ce traitement de surface génère des contraintes résiduelles et de l’écrouissage en surface de la pièce, ce qui influe sur sa durée de vie. L’enjeu de cette thèse est de mettre en place une méthodologie permettant de prendre en compte cet état mécanique initial, ainsi que son évolution en service, dans l’analyse de durée de vie d’une aube élaborée en superalliage monocristallin à base de nickel (AM1). Tout d’abord, cet état mécanique (contraintes résiduelles et écrouissage) est déterminé expérimentalement. Les contraintes résiduelles sont notamment évaluées par diffraction des rayons X en utilisant la méthode d’Ortner. Cet état mécanique est ensuite introduit dans les calculs de structure. Pour cela, on s’inspire de la méthode connue de l’introduction directe du champ d’eigenstrains qui est, ensuite, étendue afin d’introduire également les variables d’écrouissage ainsi que l’état mécanique anisotrope complet dans toute la structure. L’étape suivante a visé à suivre expérimentalement et à modéliser l’évolution de ces quantités sous sollicitations thermique, d’une part et cycliques à température constante (650°C), d’autre part. Enfin, la chaîne complète de calcul de durée de vie de l’AM1 est appliquée afin d’analyser la durée de vie des éprouvettes grenaillées. Les résultats obtenus sont discutés en regard des essais de fatigue effectués sur éprouvettes représentatives / Shot-peening is widely used on roots of high pressure turbine blade to postpone crack initiation in stress concentration area. This pre-stressing introduces compressive residual stress and strain hardening in a surface layer which will influence lifetime. The aim of this thesis is to propose a methodology which allows taking into account the impact of such pre-stressing and their evolution on the fatigue behaviour of a single crystal nickel-based superalloy (AM1) used for high pressure turbine blades. Firstly, the experimental work is devoted to the determination of the initial mechanical state (residual stresses and strain hardening). Residual stresses are namely determined using X-ray diffraction involving the use of a specific method developed by Ortner. This experimental initial mechanical state is then introduced as an input in structure calculation. The well-known method involving the direct introduction of eigenstrain profiles is used and improved to also introduce strain hardening variables and the complete anisotropic mechanical state in all the integration points of the structure. The next step consist in following experimentally and modeling the evolution of these quantities under thermal and mechanical loads under an uniform temperature (650°C). Finally, the complete lifetime assessment is carried out on representative shot-peened samples. Results are discussed with respect to experimental fatigue tests

L'intégrité des portées d'étanchéité des robinets des centrales REP nécessite l'emploi de revêtements durs comme par exemple le Norem02, base Fe, déposé par soudage PTA (Plasma Transferred Arc). La prédiction de leur tenue en service nécessite la connaissance de leur état mécanique en fin de fabrication. Pour cela, la simulation numérique du procédé de dépôt doit être effectuée. La microstructure du Norem02 est essentiellement constituée d'une matrice austénitique parsemée d'ilôts de ferrite et de carbures. Son évolution et ses éventuelles transformations allotropiques sont étudiées durant des cycles thermiques représentatifs de ceux subis par le matériau au cours du dépôt, amenant à la conclusion que l'évolution métallurgique de l'alliage n'a que très peu d'influence sur ses propriétés mécaniques. En s'appuyant sur des essais jusqu'à 1000°C de dilatométrie libre, de traction isotherme, ainsi que de traction-compression, l'évolution des caractéristiques thermomécaniques est déterminée en fonction de la température et une loi de comportement élasto-plastique à écrouissage cinématique non linéaire est identifiée. Des mesures expérimentales de températures, déplacements, déformations et contraintes résiduelles lors d'essais de dépôt PTA sur des maquettes représentatives de la géométrie réelle d'un robinet permettent l'identification d'un chargement thermique équivalent. Enfin, des simulations numériques sont conduites en utilisant le code éléments finis développé au sein d'EDF R et D Code_Aster. Elles permettent la validation du caractère prédictif des modèles développés par comparaison entre les résultats mécaniques simulés et expérimentaux. Une excellente corrélation est ainsi observée entre les températures mesurées et simulées tout au long du procédé. De la même façon, à l'issue du refroidissement, les distorsions, déformations et contraintes résiduelles expérimentales et simulées sont en adéquation, validant ainsi la stratégie de modélisation proposée dans ces travaux. / Hardfacing coatings are widely used on the surfaces of parts subjected to drastic loadings. Norem02 alloy, Fe-based, is used in PWR nuclear power plants on valves seating surfaces. Its microstructure consists of a dendritic austenite structure with ferrite islets and carbides. This work tends to demonstrate that for this alloy, metallurgical evolution during the welding process has very little influence on mechanical properties. Tensile behaviour was characterised and completed by dilatometry tests in welding process temperature range until 1000°C, in order to identify an elastoplastic model with non linear kinematic hardening rule. Temperature, displacements, distorsions and residual stresses were measured during the PTAW process and used to identify an equivalent thermal loading by solving an inverse problem. Finally, the numerical simulation of the whole process using the EDF R and D FEM software Code_Aster is presented. Predicted temperatures are consistants with experimental ones. In the same way, predicted displacements, residual distorsions and residual stresses at the end of the cooling phase are close to experimental measures, validating the modelisation strategy presented in this work.

La défaillance des pièces mécaniques est très souvent initiée par un défaut de surface. Par conséquent, la génération de contraintes résiduelles compressives sur des pièces mécaniques via une déformation plastique hétérogène améliore la tenue en fatigue et augmente la durée de vie des pièces. Parmi les procédés permettant d'introduire des contraintes résiduelles dans les pièces, le traitement par choc laser est plus intéressant à plusieurs titres. D'une part, il permet de produire des pressions en surface du matériau de l'ordre de 1 à 6 GPa sur de courtes durées d'impulsion allant de 3 à 30 nanosecondes. D'autre part, il offre la possibilité d'introduire des contraintes résiduelles de compression sur une certaine profondeur tout en conservant l'état initial de la pièce traitée. Ainsi, les simulations numériques par réalisation de modèles simples permettent de cerner les physiques mises en jeux. Dans cette perspective, la méthode semi-analytique offre d'énormes avantages, notamment la simplicité des modèles et la réduction des temps de calcul. Cependant, cette méthode n’a jamais été étendue aux problème dynamiques. Dans cette thèse la méthode semi-analytique a été étendue aux problèmes dynamiques et le modèle mis en place été appliqué pour la simulation du procédé de choc Laser / The failure of the mechanical parts is very often initiated by a surface defects. Consequently, the generation of compressive residual stresses on mechanical parts by introducing a heterogeneous plastic strain improves the resistance to fatigue and increases the lifetime of the parts. Among the processes making it possible to introduce residual stresses into the parts, the laser shock peening is more interesting for several reasons. On the one hand, it makes it possible to produce pressures on the surface of material of about 1 to 6 going GPa over short pulse times from 3 to 30 nanoseconds. In addition, he gives the opportunity of introducing residual stresses of compression on a certain depth while preserving the initial state of the treated part. The numerical simulation becomes necessary to determine the best physical phenomena involved. Thus, the semi-analytical method offers a lot of advantages, in particular the simplicity of the models and the computation times saving. This method was never extended to the dynamic problems. In this thesis the semi-analytical method was extended to the dynamic problems and the model implemented is applied for the simulation of the Laser process of shock.

Dans l’industrie nucléaire, différents composants des réacteurs à eau pressurisée sont soudés par le procédé TIG (Tungsten Inert Gas). Pour obtenir une résistance à la corrosion et une tenue en service suffisante, ces pièces sont réalisées en alliage base nickel NY690. La compréhension des phénomènes physiques impliqués durant le soudage est nécessaire afin de connaître l’évolution de la microstructure et leur influence sur le comportement mécanique. Pour améliorer la qualité des produits soudés, il est crucial de déterminer les contraintes résiduelles après soudage ce qui implique de connaître le comportement mécanique de l’alliage NY690. En effet, le soudage de ces pièces massives dont l’épaisseur est comprise entre 50 et 250 mm nécessite un grand nombre de passes et donc grand nombre de cycles thermo-mécaniques. Pour reproduire ces chargements complexes se produisant dans la zone affectée thermiquement, des essais mécaniques ont été réalisés en utilisant la machine Gleeble 3500 et une machine de traction/torsion MTS-809. Des essais de traction et des essais cycliques ont été réalisés pour différents états microstructuraux à température ambiante et à 750°C. Plus précisément, l’influence de la taille de grain et de la présence des carbures de chrome sur le comportement mécanique a été étudiée dans le but de déterminer les paramètres à prendre en compte pour modéliser le comportement mécanique. Le comportement mécanique en chargement monotone a ensuite été modélisé en fonction de la température et de la vitesse de déformation en utilisant le formalisme de Kocks-Mecking-Estrin. Ce modèle ensuite été implémenté sous Sysweld pour simuler des expériences de soudage. / Mechanical behavior of nickel alloy 690 (NY690) is characterized from 25°C to 1100°C and for a strain rate ranging from 10-4 to 5×10-3s-1. The effects of chromium carbides and grain size (50-150 µm) on the tensile properties of NY690, were studied at 25°C and 750°C. Chromium carbides have negligible influence on the yield stress and on the strain hardening whereas the grain size slightly decreases the yield stress and the hardening rate at room temperature. The grain size has little influence on the strain-hardening but increases the steady-state stress. The dislocation density is the major microstructural parameter governing the mechanical behavior of the alloy for the studied experimental conditions. The Kocks-Mecking-Estrin formalism is adapted to a wide range of temperature and strain rate to predict the mechanical behavior.