This doctoral thesis is concerned with fatigue life of welded structures. Several topics related to fatigue of welded structures are treated such as; weld defects and their influence on fatigue performance of welded structures, fatigue life prediction using LEFM (Linear Elastic Fracture Mechanics), fatigue testing, welding simulation, residual stress prediction and measurement and their influence on fatigue life. The work that is reported in this doctoral thesis is part results of the Nordic R&D project QFAB (Quality and Cost of Fabricated Advanced Welded Structures) and the Swedish R&D project LOST (Light Optimized Welded Structures). One of the main objectives is to compare different welding processes for the fatigue performance, weld quality and gain understanding of the weld defects, their appearance in different welding processes and their effect on fatigue life. Another main objective is to study welding residual stresses and their effect on fatigue. The design rules are in some cases conservative and especially on the weld root sides the knowledge about the residual stress field may improve the life prediction. The aim is to develop simplified procedures for analysis of residual stresses, their relaxation and influence on fatigue life. Fatigue testing of Hybrid Nd: YAG laser/MAG and MAG welded (tandem arc solid wire, flux cored wire, tandem flux cored wire) non-load carrying cruciform joints was carried out. Four batches were produced, tested and the results were compared. The local weld geometry of the cruciform welded joints was measured and analyzed. Residual stress measurement was carried out close to the toe region using X-ray diffraction. Weld defects, in most cases cold laps, in the cracked specimens were measured. Further fatigue testing, weld defect assessment and residual stress and local weld geometry measurements were carried out on joints welded with flux cored and metal cored arc wires. Two-and three dimensional LEFM crack growth analysis were carried out in order to predict the influence of weld defects, local weld geometry and residual stresses. Residual stresses in multi-pass welded tube-to-plates were studied for two different tubular joint configurations; a three-pass single-U weld groove for maximum weld penetration and a two-pass fillet (no groove) welded tube-to-plates for minimum weld penetration. Torsion fatigue tests were performed in order to study crack propagation from the weld root. Mode III propagation from the lower and upper weld toe on the same tubular joints was also studied. Some tubes were stress relieved (PWHT) and some were fatigue tested with internal static pressure. A three dimensional finite element welding simulation of the multi-pass welded tubular joint was carried out. The calculated temperatures in the transient thermal analysis were compared with measured temperatures. The FE predicted residual stresses in the as-welded conditions were verified with hole drilling strain gage measurements. The residual stresses were used as internal stresses in the finite element model for the torsion fatigue simulation in order to study the cycle by cycle relaxation of the residual stresses in constant amplitude torsion loading. A two dimensional finite element welding simulation procedure was developed in order to predict welding residual stress. The predicted residual stresses were used together with a developed 2D LEFM subroutine to predict the fatigue life, crack path and the effect of residual stresses on weld root defects. The developed simulation subroutines were validated with results found in the literature. Residual stresses measurement, two-and three dimensional welding simulations were carried out in fillet welded joints in order to study the three dimensional effects of the welding process, boundary conditions and modelling technique on the formation of residual stresses. / QC 20100706

The studied material is an A356 Al alloy, used to produce engine cylinder heads for the automotive industry by die casting process. The material displays a quite coarse dendritic microstructure in a eutectic matrix, with a mean grains size of 25 microns, intemetallic precipitates and porosities. The tensile properties are strongly affected by testing temperature, with a quite sensitive drop of the Young's modulus, the Yield stress as the temperature was raised. The isothermal fatigue life dropped of markedly (approximately 10 times) when the testing temperature is raised from 120 to 280 °C, under strain control. From the themomechanical in-phase cyclic tests, with temperature varying from (120 to 280 oC), it was possible to observe that life is quite similar to the isothermal fatigue test at 280 oC. In this case, the more sensitive damage caused the in-phase mechanical and thermal cycle take place at the highest temperature. Relaxation tests indicated two distinct behaviors, with the temperature of 240°C being a threshold. At lower temperatures, the material hardens cyclically whereas it softens cyclically at higher temperatures. From the fatigue crack growth results, it was observed that temperature and wave shape has a strong influence on the crack growth rate as well as on the stress intensity threshold. Considering sinusoidal wave shape (10 Hz), as the temperature increased the DKth decreased and the crack propagation rate increased. However, the rate as da/dN change with temperature is quite similar, as an indicative that the micromechanism of crack growth has not changed due to the high frequency used, and it was due only to loss of mechanical strength. An elastic-visco-plastic non-isothermal constitutive law was identified for the material. For the cast material studied in this work, the mechanical behavior parameters are statistically distributed. However, it was shown that the model was able to reproduce, with a reasonable approximation, the stress - strain relationship at different temperatures, for the isothermal and anisothermal cases.

The mechanical behavior of a high pressure die cast AM60B Mg alloy is studied. Constant load amplitude fatigue tests were conducted at room, elevated and cold temperatures, with a stress ratio of R=0.1, and frequency of 30 Hz. The objective was to identify the possible effects of temperature on fatigue life cycle. In addition, fatigue crack propagation tests were conducted to ascertain the fatigue response of the alloy and determine its fatigue crack growth rate as a function of the applied stress ratio, experimentally, analytically and computationally, using Walker’s model. The results demonstrated that temperature had a significant influence on the fatigue life, and that the life increased at cold temperature but decreased at elevated temperature as compared to that evaluated at room temperature. In this study, the limit for applicability of LEFM was established for AM60B magnesium alloy. In addition, fatigue crack propagation test results were used to evaluate the coefficients of the Paris model.

The main objective of the project was to establish a fracture-mechanics energy-based approach for the design of structural adhesive joints under cyclic loading. This required understanding how an adhesive system behaved near its fatigue threshold, and how the key factors affected this behavior in a fresh undegraded joint. The investigated factors were mode ratio (phase angle), substrate material, surface treatment and surface roughness (both affecting the bond strength), bondline thickness and load ratio. It was first required to understand how the adhesive system behaved under quasi-static loading by examining a fracture mechanics-based design approach for adhesive systems with different substrate materials and geometries. Experiments were initially performed to characterize the strength of aluminum and steel adhesive systems based on the fracture envelope, critical strain energy release rate as a function of the mode ratio. Ultimate failure loads of aluminum and steel adhesive joints, having different overlap end conditions and different geometries were then experimentally measured. These values were compared with the failure loads extracted from the fracture envelope. Considering the toughening behavior of the adhesive in the fracture mechanics analyses, a very good agreement (average of 6%) was achieved between the predictions and experiments for all types of overlap end conditions and geometries. Different fatigue threshold testing approaches, which are commonly used in the literature or suggested by the ASTM standard, were evaluated for the cracked and intact fillet joints. Based on the experimental and analytical studies, the most appropriate technique for fatigue testing and characterization of adhesive systems was suggested. Comparing the mixed-mode near-threshold behavior of different adhesive systems with the fracture behavior and fatigue mode-I and mixed-mode high crack growth rates showed the high sensitivity of the mixed-mode near-threshold fatigue to the subtle changes in the interfacial bond strength. In order to make a baseline for the design of adhesive joints under cyclic loading, similar to the previous fracture tests and following the energy-based approach, fatigue behavior was characterized as a function of the loading mode ratio for aluminum and steel adhesive joints. The effect of substrate material, surface treatment, bondline thickness, surface roughness and fatigue testing load ratio on the near-threshold fatigue behavior of adhesives joints was evaluated experimentally. The experimental observations were then explained using finite element modeling. To generalize the conclusions, the majority of experiments and studies covered a broad range of crack growth rates, as low as fatigue threshold and as high as 10-2 mm/cycle. Having understood the significant testing and design parameters, an adhesive system can be designed based on a safe cyclic load that produces an insignificant (for automotive industry) or reasonably low but known crack growth rate (for aerospace industry).

The effect of retrogression and reaging heat treatment on corrosion fatigue crack growth behavior on AA7050 T73651 aluminum alloy is investigated. CT (Compact Tension) specimens are prepared in LS direction for fatigue crack growth tests . Samples are solution heat treated at 477 &deg / C and aged at 120 &deg / C for 24 h (T6 condition). After that, samples are retrogressed at 200 &deg / C for times of 1, 5, 30, 55 and 80 minutes in a circulating oil bath. Then, samples are re-aged at 120 &deg / C for 24 h (T6 condition). Hardness measurements are taken at different retrogression times and at the end of the heat treatment. Fatigue crack growth tests are performed at as received condition and at different retrogression times with sinusoidal loading of R=0.1 and f=1 in both laboratory air and corrosive environment of 3.5% NaCl solution. The highest fatigue crack growth resistance is observed for 30 min. and 5 min. retrogression for laboratory air and corrosive environment respectively. It is concluded that RRA can successfully be used to improve fatigue performance of this alloy.

Les essais de fatigue peuvent être effectués dans une gamme de fréquences similaire, mais les fréquences les plus élevés sont généralement préférables afin de réduire le temps de test et d'obtenir des résultats dans une période raisonnable. La question reste à savoir si un effet de fréquence existe. Cette étude traite de l'effet de la fréquence de chargement sur la résistance à l'amorçage et à la propagation des fissures sur 3 matériaux (M800HY, CP1000 et DP1180) utilisés dans le domaine automobile. Deux rapports de charge (R=-1 et 0.1) et cinq fréquences de chargement (25 Hz, 30 Hz, 150 Hz et 20 kHz) ont été étudiés. Tout d'abord, nous avons réalisé des calculs par élément finis en dynamique pour dimensionner les éprouvettes et calculer l'étendue du facteur d'intensité de contrainte durant la propagation de la fissure de fatigue. Ensuite, nous avons réalisé des essais d'amorçage et de propagation. Indépendamment de la fréquence de chargement,l'amorçage de la fissure de fatigue est toujours surfacique à cause des concentrations de contraintes dues au processus de laminage ou de l'électro-galvanisation. L'observation des faciès de rupture sur éprouvettes testées à 30 Hz et 20 kHz montre que le mode de propagation est transgranulaire avec quelques localisations de rupture intergranulaire ou par clivage sur la nuance DP1180. Les stries de propagations sont présentes sur les faciès d'éprouvettes M800HY et absentes sur les faciès des deux autres nuances. La nuance DP1180 présente une meilleure résistance à l'amorçage et à la propagation de la fissure de fatigue. Concernant l'effet de la fréquence de chargement, globalement l'amorçage de la fissure de fatigue est dépendante de la fréquence de chargement, par contre, la propagation est indépendante de la fréquence. / Fatigue tests can be performed in a similar frequency range, but the higherfrequencies are generally preferable to reduce the test time and obtain results within areasonable time. The question remains whether a frequency effect exists. This study dealswith the effect of the loading frequency on fatigue strength and crack growth of 3 steels(M800HY, CP1000 and DP1180) used in automotive industry. Two load ratios (R = -1 and0.1) and five load frequencies (25 Hz, 30 Hz, 150 Hz and 20 kHz) were studied. First, weperformed dynamics finite element method to calculate the dimensions of the specimens(to vibrate with a resonance at 20 kHz), and, the stress intensity range as a function of the length of crack. Then, we have done some test on fatigue crack initiation and propagation. Whatever the loading frequency, the initiation of the fatigue crack is always on the surfacebecause of the stress concentrations due to the rolling process or the electro-galvanization.The observation with SEM of fracture surface of the specimens tested at 30 Hz and 20 kHzshows that the mode of propagation is transgranular with some locations of intergranularfailure or by cleavage on DP1180 grade. The fatigue striation on fracture surface arepresent on the M800HY and absent on the other two materials. The grade DP1180 has abetter resistance to the initiation and propagation of the crack. Regarding the effect ofloading frequency, overall the initiation of the fatigue crack is dependent on the loadingfrequency, however, the propagation is independent of the loading frequency.

As recentes descobertas de petróleo e gás na camada do Pré-sal representam um enorme potencial exploratório no Brasil, entretanto, os desafios tecnológicos para a exploração desses recursos minerais são imensos e, consequentemente, têm motivado o desenvolvimento de estudos voltados a métodos e materiais eficientes para suas produções. Os tubos condutores de petróleo e gás são denominados de elevadores catenários ou do inglês \"risers\", e são elementos que necessariamente são soldados e possuem fundamental importância nessa cadeia produtiva, pois transportam petróleo e gás natural do fundo do mar à plataforma, estando sujeitos a carregamentos dinâmicos (fadiga) durante sua operação. Adicionalmente, um dos problemas centrais à produção de óleo e gás das reservas do Pré-Sal está diretamente associado a meios altamente corrosivos, tais como H2S e CO2. Uma forma mais barata de proteção dos tubos é a aplicação de uma camada de um material metálico resistente à corrosão na parte interna desses tubos (clad). Assim, a união entre esses tubos para formação dos \"risers\" deve ser realizada pelo emprego de soldas circunferenciais de ligas igualmente resistentes à corrosão. Nesse contexto, como os elementos soldados são considerados possuir defeitos do tipo trinca, para a garantia de sua integridade estrutural quando submetidos a carregamentos cíclicos, é necessário o conhecimento das taxas de propagação de trinca por fadiga da solda circunferencial. Assim, neste trabalho, foram realizados ensaios de propagação de trinca por fadiga na região da solda circunferencial de Inconel® 625 realizada em tubo de aço API 5L X65 cladeado, utilizando corpos de prova do tipo SEN(B) (Single Edge Notch Bending) com relações entre espessura e largura (B/W) iguais a 0,5, 1 e 2. O propósito central deste trabalho foi de obter a curva da taxa de propagação de trinca por fadiga (da/dN) versus a variação do fator de intensidade de tensão (ΔK) para o metal de solda por meio de ensaios normatizados, utilizando diferentes técnicas de acompanhamento e medição da trinca. A monitoração de crescimento da trinca foi feita por três técnicas: variação da flexibilidade elástica (VFE), queda de potencial elétrico (QPE) e análise de imagem (Ai). Os resultados mostraram que as diferentes relações B/W utilizadas no estudo não alteraram significantemente as taxas de propagação de trinca por fadiga, respeitado que a propagação aconteceu em condições de escoamento em pequena escala na frente da trinca. Os resultados de propagação de trinca por fadiga permitiram a obtenção das regiões I e II da curva da/dN versus ΔK para o metal de solda. O valor de ΔKlim obtido para o mesmo foi em torno de 11,8 MPa.m1/2 e os valores encontrados das constantes experimentais C e m da equação de Paris-Erdogan foram respectivamente iguais a 1,55 x10-10 [(mm/ciclo)/(MPa.m1/2)m] e 4,15. A propagação de trinca no metal de solda deu-se por deformação plástica, com a formação de estrias de fadiga. / Recent oil and gas discoveries in the Pre-Salt layer represent a huge exploration potential in Brazil, however, the technological challenges for the exploitation of these mineral resources are immense and therefore have motivated the development of studies looking for efficient methods and materials for their productions. The oil and gas pipellines, called risers, are elements that are necessarily welded and have fundamental importance in the production chain, since they transport oil and natural gas from the sea bed to the platforms and are subject to dynamic loads (fatigue) during operation. Additionally, one of the central problems in the production of oil and gas in the Pre-Salt reserves is directly associated with a highly corrosive media, such as H2S and CO2. A cheaper way to protect the pipelines from these medias is applying a protective layer of a corrosion resistant metal on the inner diameter of these pipes, creating a cladded pipe. Thus, a joining process of these pipes to form the risers must be carried out by the use of girth welds with a corrosion resistance material similar to the clad metal. As the welded structures are seen as potential location of \"crack like\" defects, to ensure the structural integrity of such component when subjected to repetitive loading conditions, it is necessary to know the fatigue crack growth rates for the girth weld. Therefore, in this work it was carried out fatigue crack propagation tests in the weld region of an API 5L X65 cladded pipe with Inconel® 625, girth welded using Inconel® 625 electrodes. From the welded region, Single Edge Notch Bending specimens, SEN(B), were removed with different thickness and width ratios (B/W= 0.5, 1, and 2). From the fatigue tests, the crack propagation rates (da/dN) as function of the variation of the stress intensity factor (ΔK), were determined for the weld metal, using different crack size measurement techniques: the elastic compliance (EC), electric potential drop (EPD) and image analysis (IA). The results showed that the different B/W ratios used in study did not modified significantly the fatigue crack growth rates, considering that crack propagation took place under small scale yielding conditions. The results of fatigue crack growth tests allowed to obtain the regions I and II of da/dN x ΔK curves for the weld metal. The ΔKth value obtained for the weld metal was around 11,8 MPa.m1/2 and the found values of the experimental constants C and m of Paris-Erdogan\'s equation were respectively equal to 1,55 x10-10 [(mm/cycle)/( MPa.m1/2)m] and 4.15. The micromechanism of fatigue crack growth took place by plastic deformation, with the formation of fatigue striations.

Aços microligados pertencem à classe dos aços ARBL contendo baixa ou média quantidade de carbono e pequena adição de elementos de liga tais como Mn, Nb, Mo, V e Ti. A variedade microestrutural desses aços pode ser obtida dependendo da temperatura de conformação, taxa de resfriamento e composição química. Os tratamentos intercríticos e isotérmicos produzem microestruturas multifásicas com diferentes quantidades de ferrita, martensita, bainita e austenita retida. A presença de diferentes fases nestes materiais, com morfologias distintas, pode afetar de modo significativo seu comportamento mecânico, afetando, por exemplo, o fechamento da trinca e resultando em mudanças na taxa de crescimento da mesma. O objetivo deste trabalho é avaliar as propriedades de tração e a resistência ao crescimento da trinca por fadiga de um aço microligado RD 480 com 0.08%C-1, 5%Mn (p), correlacionando-as com suas características microestruturais. Esse aço, desenvolvido recentemente pela CSN (Companhia Siderúrgica Nacional), é considerado promissor como alternativa para substituir o aço de baixo carbono utilizado em componentes de rodas na indústria automotiva. Distintas condições microestruturais foram obtidas por meio de tratamentos térmicos seguidos de resfriamento em água. As condições de tratamento intercrítico e têmpera simples foram escolhidas para se avaliar a resistência à propagação da trinca por fadiga. Os resultados dos ensaios foram sintetizados em termos da taxa de crescimento da trinca (da/dN) versus a variação do Fator Intensidade de Tensão (_K) no ciclo de carregamento. Para descrever o comportamento das trincas foram utilizados dois modelos: a equação convencional de Paris e um novo modelo exponencial que mostra o comportamento não linear das curvas de fadiga. Os resultados mostraram que uma microestrutura combinando ferrita de aspecto acicular e fases duras (martensita/bainita) resultou em menores taxas de crescimento da trinca. No entanto, a melhor combinação entre as propriedades de tração (limite de escoamento, resistência e ductilidade) e fadiga foi obtida com uma microestrura bifásica contendo martensita dispersa em uma matriz ferrítica. Observou-se uma transição nas curvas de crescimento da trinca para todas as condições tratadas termicamente e, por conta disto, as curvas das condições microestruturais bifásicas e multifásicas foram melhores modeladas quando divididas em duas regiões. As superfícies de fratura dessas amostras, bem como o caminho percorrido pela trinca, foram analisados via MEV e MO. / Microalloyed steels are a class of HSLA steels with low or medium carbon content and small additions of alloy elements such as Mn, Nb, Mo, V and Ti. A variety of microstructures in microalloyed steels can be obtained depending on the deformation temperature, cooling rate and chemical composition. Heat treatments and isothermal transformation on these materials, with various temperatures and holding times, produce multiphase microstructures with different amounts of ferrite, martensite, bainite and retained austenite. These different phases, with distinct morphologies, are determinant of the mechanical behavior of the steel and can, for instance, affect crack closure or promote crack shielding, thus resulting in changes on its propagation rate under cyclic loading. The aim of this study is to evaluate the tensile properties and resistance to fatigue crack growth in a microalloyed steel RD 480 with 0.08%C-1, 5% Mn (wt), correlating with their microstructural characteristics. This steel, recently developed by CSN (Companhia Siderurgica Nacional), is being considered as a promising alternative to replace low carbon steel in wheel components for the automotive industry. Distinct microstructural conditions were obtained by means of heat treatments followed by water quench. The intercritical treatment and quenching conditions were chosen to evaluate the strength to crack propagation. The crack propagation test results were summarized in terms of FCG rate (da/dN) versus stress intensity factor range (?K) curves. In order to describe the FCG behavior, two models were tested: the conventional Paris equation and a new exponential equation developed for materials showing non-linear FCG behavior. The results showed that a microstructure combining aspect acicular ferrite and hard phases (martensite / bainite) resulted in lower rates of crack growth. However, the best combination between the tensile properties (yield stress, tensile strength and ductility) and fatigue was obtained with a dual phase steel microstructure containing martensite dispersed in a ferrite matrix. It was observed a transition in the crack growth curves for all heat treated conditions, so the curves of the dual and multiphase microstructural conditions were better modeled by dividing them in two regions. The fracture planes of the fatigued specimens, as well as the crack path, were examined using a scanning electron microscope (SEM) and optical micrography (OM).

Une méthode améliorant le calcul des facteurs d’intensité de contrainte en mode I par hybridation de deux fonctions de poids est présentée et appliquée aux cas de fissures elliptiques sous différents chargements. L'hybridation consiste à utiliser l'une ou l'autre des deux fonctions dans la zone de la fissure où la fonction est la plus efficace. La délimitation des zones est faite après optimisation des paramètres géométriques de la fissure. Afin d’étendre l’utilisation de cette approche à la modélisation des fissures semi-elliptiques, son couplage avec la PWFM (Point Weight Function Method) pour tenir compte de l’effet de la surface libre a été réalisé. L’utilisation des lois de propagation de fissure (de Paris et de Sih) a permis l’extension de l’application de l’approche d’hybridation aux problèmes de fatigue. La qualité des résultats trouvés pour les fissures elliptiques et semi-elliptiques est bonne aussi bien pour les chargements quasi-statiques que pour ceux de fatigue. / A method improving the evaluation of the stress intensity factor by hybridization of two weight functions is presented and applied for embedded elliptical cracks under various loadings. The hybridization consists in using one or another function in the zone of the crack where it is the most efficient. The delimitation of the zones is achieved after optimizing the geometrical parameters of the crack. In order to extend the use of this approach to the modeling of semi-elliptical surface cracks, its coupling with the PWFM (Point Weight Function Method) to take account of the free edge effect were carried out. The use of the fatigue crack growth laws (Paris law and Sih law) allowed the extension of the application of the hybrid approach to the fatigue problems. The quality of the results found for the elliptical and the semi-elliptical cracks is good for the static loads as well as for the fatigue ones.

No description available.