Les contraintes résiduelles, introduites lors de la fabrication des pièces mécaniques, influent sur leur durée de vie en fatigue et doivent donc être prises en compte. Ce travail vise à développer une méthode de dimensionnement rapide en fatigue oligocyclique permettant leur prise en compte, aussi bien sous chargement de traction que de compression. L’idée principale est de calculer analytiquement le cycle stabilisé uniquement au point critique de la structure à l’aide d’une méthode simplifiée en plasticité confinée. Cette dernière est basée sur une loi de localisation, qui étend les méthodes énergétiques de type Neuber au cas général multiaxial. La loi de localisation permet de relier le chargement appliqué à la structure à celui observé localement au point critique. Elle est calibrée à l’aide d’une simulation éléments finis sous chargement monotone de la structure. Pour cela, le comportement élasto-plastique du matériau est d’abord caractérisé à partir d’un essai de traction avec charges-décharges. A l’aide de ce seul essai, la loi de comportement est identifiée de manière séquentielle, en utilisant des écrouissages cinématique et isotrope à seuil. Ce seuil permet d’améliorer la prévision de la contrainte moyenne stabilisée, utilisée dans le critère de fatigue proposé. Ce dernier a été identifié sur des essais purement alterné (Rε=-1), puis validé pour des chargements de traction alternée (Rε>-1) et de compression alternée (Rε<-1). L’application de la démarche en plasticité confinée est réalisée sur des éprouvettes à double encoche présentant initialement différents états de contraintes résiduelles. L’évolution des contraintes locales sous chargement cyclique nominal de traction répétée (Rσ=0) ou de compression répétée (Rσ=-∞) est mesurée. Cette évolution est ensuite comparée aux prévisions numériques et analytiques obtenues à l’aide de la loi de comportement et de la méthode simplifiée. L’application du critère de fatigue aux données stabilisées prévues par les modèles a permis d’obtenir des durées de vie avec un conservatisme indépendant du rapport de charge et de l’état initial. Enfin, une application de la démarche complète sur des joints soudés en T est réalisée afin d’élargir le spectre d’application de la méthode. / Residual stresses are inherent to the manufacturing processes and can have a strong effect on the fatigue life of structures. Therefore, they shall be taken into account in fatigue design. In this PhD thesis, a fast design method is developed to take residual stresses into account in low cycle fatigue, either under tensile or compressive loadings. The main idea is to calculate analytically the stabilised stress-strain curve, at the critical point, by using a simplified method for confined plasticity. This method is based on a localisation law that extends energetic methods like Neuber to general multiaxial stress states. The localization law links the applied load to the local load at the critical point. It has to be identified on a finite element analysis of the structure under monotonic load. For this purpose, the elasto-plastic behaviour of the material is characterised from one single tensile test with loadings-unloadings. The behaviour law, identified sequentially, includes isotropic and kinematic hardenings with thresholds. This improves the stabilised mean stress prediction, which is used in the proposed fatigue criterion to represent the influence of the load ratio on the fatigue life. This criterion is identified on purely alternated tests (Rε=-1), and then validated under various tensile load ratios (Rε>-1) as well as compressive load ratios (Rε<-1). The methodology is validated on plate specimens with two semi-circular notches, which initially present different residual stresses states. The experimental evolution of local stresses during cyclic compressive (Rσ=-∞) and tensile (Rσ=0) repeated applied loads have been studied experimentally and compared to the numerical and analytical predictions. The same work has been achieved on the fatigue lives predictions. It shows that a slight conservatism, independent from the load ratio and the initial state, is obtained. Finally, an application of the method on T-Joints is realised to enlarge the scope of the method.

O aumento da vida em fadiga em peças submetidas a esforços cíclicos devido à aplicação de processos de tratamento mecânico superficial já é bastante conhecido, tanto no meio industrial quanto no meio acadêmico. Para molas, o processo de shot-peening se torna etapa essencial no processo de fabricação, porém um estudo sistemático do efeito do shot-peening na vida em fadiga se faz necessário. O objetivo deste trabalho é aprofundar o conhecimento nos processos de jateamento superficial do tipo shot-peening para molas em lâminas de veículos, através da análise de tensões residuais por difração de raios-x e ensaios de fadiga em uma série de amostras que sofreram dez diferentes receitas de processos de jateamento. Dos dez diferentes processos, o de jateamento com granalha esférica fundida de aço com 1,0mm de diâmetro seguido de um segundo jateamento com granalha esférica fundida de aço de 0,3mm de diâmetro levou a uma maior vida em fadiga das amostras. A análise por difração de raios-x comprovou que a importância das tensões residuais de compressão se dá até a uma profundidade de 0,05mm, influenciando diretamente no modo de nucleação de trincas de fadiga. Acima desta profundidade, as tensões residuais induzidas por shot-peening não têm influência no modo de propagação destas trincas, e por consequência na vida em fadiga das peças. Correlações entre os resultados e a alta dureza e diminuída ductilidade do material utilizado são discutidas. / The improvement of fatigue life in parts subjected to cyclic stresses by application of mechanical surface treatment processes is already well known, both in the industry and in the academy. Dealing with automotive springs, the shot peening process becomes an essential step in manufacturing these parts. In the case of leaf springs, however, a systematic investigation of the effect of shot peening on fatigue life is still required. The aim of this work is to improve the knowledge of shot peening on leaf springs for vehicles, through the analysis of residual stresses by x-ray diffraction and fatigue tests on a series of samples that were subjected to ten different peening schedules. Among the investigated processes, the usage of 1.0 mm diameter cast steel shot followed by a second peening with 0.3 mm diameter cast steel shot leads to better performance, regarding fatigue life. X-ray diffraction analysis shows that this improved performance may be attributed to residual compressive stress maintained until a depth of 0.05 mm below the surface, which directly influences the fatigue crack nucleation. Residual stresses induced by shot-peening in larger depths, have no influence on sample fatigue life, showing that crack propagation is not affected by the induced residual stresses. Consequently, the durability of parts is improved by shot-peening exclusively due to this influence on crack nucleation at samples surface. Correlations with the increased hardness and decreased ductility of the employed material are discussed.

Atualmente, a exigência da indústria na construção de estruturas complexas é consequência da acirrada concorrência, principalmente, entre empresas dos segmentos rodoviários. Com o auxílio de ferramentas computacionais, baseadas no Método de Elementos Finitos, é possível dimensionar essas estruturas submetidas a solicitações aleatórias variando no tempo. Nesse caso, é comum a realização de simulações dinâmicas no domínio do tempo que se empregue um modelo de Elementos Finitos de placa ou viga e, assim, obter as tensões nos locais críticos de interesse. Porém, análises no domínio do tempo capturam tensões e deformações com boa precisão, mas costumam ser caras computacionalmente. Na maioria dos casos estudados, a vida em fadiga é fundamental para prever falhas estruturais de elevadas proporções, como o colapso da viga principal da estrutura chassi de um implemento rodoviário. Diante disso, são de interesse tecnológico e industrial as análises no domínio da frequência, a fim de diminuir o custo computacional de simulação dinâmica independente do modelo de Elementos Finitos. Com esse propósito, o presente trabalho propõe uma comparação entre metodologias capazes em estimar a vida em fadiga no domínio do tempo e da frequência, de um caso simples e um segundo caso, mais complexo, que é o da indústria. Os diferentes modelos de Elementos Finitos e métodos de cálculos da vida em fadiga no domínio da frequência são comparados e discutidos, usando-se como referência o método clássico de cálculo da vida em fadiga, domínio do tempo. A metodologia leva em conta os diferentes Elementos Finitos empregados, as simplificações dos modelos estruturais usados e potenciais restrições cinemáticas aplicadas ao modelo. A comparação é feita com a ajuda de ferramentas comerciais que executam as análises modais, harmônicas, transientes, espectrais e rotinas de cálculos. Assim, o analista pode selecionar malhas com quantidade viável de graus de liberdade tornando possível as simulações dinâmicas e, além disso, prever a vida em fadiga no domínio do tempo para modelos simples e complexos. A partir dos resultados, verificou-se que as análises de estimativa da vida em fadiga no domínio da frequência, para modelos complexos, ainda não apresentam resultados confiáveis e, assim, são sugeridas recomendações. / Currently, the industry demand for the construction of complex structures is a consequence of the fierce competition, mainly between companies from semi-trailers segments. With the support of computational tools, based on the Finite Element Method, it is possible to size these structures submitted to random requests varying in time. In this case, it is common to perform dynamic simulations in the time domain to use a Finite Element model of plate or beam and thus obtain the stresses at the critical sites of interest. However, time domain analyses capture tensions and deformations with good accuracy, but they are often computationally expensive. In most of the studied cases, fatigue life is fundamental to predict structural failures of high proportions, such as main beam collapse of the chassis structure of a semi-trailer. Therefore, the semi-trailer industry segments is interested in the frequency domain, in order to reduce the computational cost of dynamic simulation no matter of the Finite Element model. With this purpose, the present work proposes a comparison between methodologies which are able to estimate the fatigue life in time domain and frequency, of a simple and second case, more complex, that of the industry. The different Finite Element models and methods of calculating fatigue life in frequency domain are compared and discussed using the classic time domain fatigue life calculation method as reference. The methodology takes into account the different Finite Elements are used, the simplifications of the structural models used and potential kinematic constraints applied to the model. The comparison is made with not only the following analyses: modal, harmonic, transient, spectral, but also computational routines. Thus, the analyst can select meshes with a viable amount of degrees of freedom making dynamic simulations possible, in addition, to its predict the fatigue life in time domain for simple and complex models. Based on the results, it was verified that the fatigue life estimates in frequency domain, for complex models, do not present reliable results yet. Therefore, recommendations are suggested.

O uso de ligas de alumínio pela indústria aeronáutica é sabidamente consagrado. Neste tipo de uso os carregamentos e seu histórico acabam por ter uma importância ímpar na determinação da vida em fadiga das peças estruturais. Este trabalho busca estudar e apresentar a modelagem em uma série de pacotes fechados de análise de fadiga para carregamentos de amplitude constante e variável, em ligas de alumínio usadas na indústria aeronáutica. Neste contexto, modelam-se carregamentos simples ou complexos espectros de voos comerciais e militares. Outro aspecto importante da pesquisa consiste em aliar as técnicas da formulação de elementos finitos, no estudo da vida em fadiga de corpos de prova e demais propriedades do material, com os modelos bem consagrados para problemas de propagação de trinca, comumente presentes na indústria em questão. Obtiveram-se boas respostas, no tocante à representação dos fenômenos estruturais e de material, sobretudo aqueles relacionados à propagação. Ademais, o Método dos Elementos Finitos Estendidos ajuda significativamente na melhoria da resposta numérica, contudo sem aumentar de forma restritiva o custo computacional das análises. Os principais resultados ao final do trabalho foram uma contribuição ao estudo deste material relacionado à modelagem via elementos finitos conseguindo-se representar bem campos de tensões e propagações de trinca e, em especial, a representação obtida do ensaio de KIC. Por outro lado, com o uso de pacotes fechados e abertos no estudo da vida em fadiga, conseguiu-se simular de forma satisfatória problemas de fadiga, em situações envolvendo carregamentos de amplitude constante e variável, tais como picos e espectros de voos civis e militares. / The use of aluminum alloys by the aeronautical industry is well known. In this application, load time-histories resulting important for fatigue life determination of aircraft structural components. Present work aims to study and present the modeling in some closed packages of fatigue analysis for both constant and variable amplitude loads in aluminum alloys used for aeronautical purposes. In this context, simple or complex loading spectra are modeled, for commercial and military flights. Another important issue of the research is to combine techniques of finite element formulation in order to predict the fatigue life of test specimens and properties of the material, using well-established models for crack propagation commonly used in the industry of concern. Were obtained good results, regarding the representation of stress states, especially those related to crack propagation. Moreover, the Extended Finite Element Method (X-FEM) significantly helps in improving the numerical response, however saving the computational cost of the analyzes. The main result of present work is the increase of the knowledge about modeling aeronautical alloys via finite elements, where were possible to represent accurately tension fields, crack propagations and especially the KIC test. On the other hand, with the use of closed and open packages for fatigue life prediction, it was possible to simulate satisfactorily problems involving constant and variable amplitude loading, such as peaks and spectra of civil and military flights.

In this thesis, we consider the problem of estimating the regression function in location-scale regression models. This model assumes that the random vector (X,Y) satisfies Y = m(X) + s(X)e, where m(.) is an unknown location function (e.g. conditional mean, median, truncated mean,...), s(.) is an unknown scale function, and e is independent of X. The response Y is subject to random right censoring, and the covariate X is completely observed. In the first part of the thesis, we assume that m(x) = E(Y|X=x) follows a polynomial model. A new estimation procedure for the unknown regression parameters is proposed, which extends the classical least squares procedure to censored data. The proposed method is inspired by the method of Buckley and James (1979), but is, unlike the latter method, a non-iterative procedure due to nonparametric preliminary estimation. The asymptotic normality of the estimators is established. Simulations are carried out for both methods and they show that the proposed estimators have usually smaller variance and smaller mean squared error than the Buckley-James estimators. For the second part, suppose that m(.)=E(Y|.) belongs to some parametric class of regression functions. A new estimation procedure for the true, unknown vector of parameters is proposed, that extends the classical least squares procedure for nonlinear regression to the case where the response is subject to censoring. The proposed technique uses new `synthetic' data points that are constructed by using a nonparametric relation between Y and X. The consistency and asymptotic normality of the proposed estimator are established, and the estimator is compared via simulations with an estimator proposed by Stute in 1999. In the third part, we study the nonparametric estimation of the regression function m(.). It is well known that the completely nonparametric estimator of the conditional distribution F(.|x) of Y given X=x suffers from inconsistency problems in the right tail (Beran, 1981), and hence the location function m(x) cannot be estimated consistently in a completely nonparametric way, whenever m(x) involves the right tail of F(.|x) (like e.g. for the conditional mean). We propose two alternative estimators of m(x), that do not share the above inconsistency problems. The idea is to make use of the assumed location-scale model, in order to improve the estimation of F(.|x), especially in the right tail. We obtain the asymptotic properties of the two proposed estimators of m(x). Simulations show that the proposed estimators outperform the completely nonparametric estimator in many cases.

Current board technologies are inherently performance-limited (FR-4) or cost-prohibitive (Al2O3/AlN). Next-generation high-density packaging applications would necessitate a new base substrate material to achieve ultra-fine pitch solder-joint reliability and multiple layers of fine-line wiring at low cost. The NEMI 2000 roadmap defines the need for 4-8 layers of 5-10 m wiring for future system boards. The 2003 ITRS roadmap calls for organic substrates with less than 100-m area-array pitch in the package or board by year 2010. Solder-joint reliability at such fine-pitch is a matter of concern for the industry. Use of underfills reduces thermal stresses but increases cost and, in addition, their dispensing becomes increasingly more complicated with the shorter gaps required for future interconnects. Therefore, there is a pronounced need to evaluate board materials with CTE close to that of Si for reliable flip-chip on board without underfill. Recently, a novel manufacturing process (using polymeric precursor) has been demonstrated to yield boards that have the advantages of organic boards in terms of large-area processability and machinability at potentially low-cost while retaining the high stiffness (~250 GPa) and Si-matched CTE (~2.5 ppm/㩠of ceramics. This work reports the evaluation of novel SiC-based ceramic composite board material for ultra-fine pitch solder-joint reliability without underfill and multilayer support. FE models were generated to model the behavior of flip-chips assembled without underfill and subjected to accelerated thermal cycling. These models were used to calculate solder-joint strains which have a strong direct influence on fatigue life of the solder. Multilayer structures were also simulated for thermal shock testing so as to assess via strains for microvia reliability. Via-pad misregistration was derived from the models and compared for different boards. Experiments were done to assemble flip-chips on boards without underfill followed by thermal shock testing so as to get the number of cycles to failure. To assess microvia reliability, 2 layer structures containing vias of different diameters were fabricated and subjected to thermal cycling. Via-pad misalignment was also studied experimentally. Modeling and experimental results were corroborated so as to evaluate thermomechanical suitability of C-SiC for high-density packaging requirements.

Abstract Recently, the expansion in power system capacities leads to the development of large-scale steam turbine generator units. As a result, a fault on the power system may induce large fault current and give rise to serious torque vibrations on turbine shafts and blades, which ought to be improved in order for the reliable operation of a turbine-generator system. In the thesis, countermeasures are proposed from electrical viewing-point and from mechanical viewing-point respectively. Based on electrical viewing-point, the apparatus in the generator stator side and in the rotor side respectively is applied to suppress the induced disturbing source. The high temperature superconductive fault current limiter bank introduces a large normal-state resistance to restrict the dc component of stator fault current. The choke coil acts as a low pass filter to restrict the system-frequency component of field fault current. Both of them lead to the reduction in electromagnetic torque of system-frequency and effectively improve the vibrating behavior of blades. Based on mechanical viewing-point, it is found from the electromechanical analysis that the Generator/LP-Turbine shaft stiffness and the Generator rotor inertia constant determine the responses of all turbine blades. Once the stiffness on this shaft section is reduced by replacing the rigid shaft coupling with a flexible one or the inertia constant is augmented by a system-frequency mechanical filter, the blades become intrinsically less responsive to electrical disturbances. As a result, the blades will bear less stress impact and can be designed with smaller safety factor. On the other hand, LP-turbine long blades operated in corrosive environment and underwent the statistical stress impact due to randomly distributed negative sequence current is studied also. In such situation, the blades may be subjected to corrosion fatigue and the long term effects of power system unbalance may become the cause of fatigue damage on blades though the negative sequence current is still within the limitation of generator thermo-rating. As a result, turbine blades are possibly unprotected by traditional system unbalance protection scheme. Therefore, it will depend on the operating environments and the blade materials whether such long-term stress can be neglected or not. If there is the potential of blade damage, one has to reconsider the I2 protection settings and rearrange the load distribution to limit the system unbalance.

This research aims to understand the reliability of Sn-Ag-Cu solder interconnects used in plastic ball grid array (PBGA) packages using microstructure evolution, laser moiré interferometry and finite-element modeling. A particle coarsening based microstructure evolution of the solder joint material during thermal excursions was studied for extended periods of time lasting for several months. The microstructure evolution and particle coarsening was quantified, and acceleration factors were determined between benign field-use conditions and accelerated thermal cycling (ATC) conditions for PBGA packages with different form factors and for two different lead-free solder alloys. A new technique using laser moiré interferometry was developed to assess the deformation behavior of Sn-Ag-Cu based solder joints during thermal excursions. This technique can used to estimate the fatigue life of solder joints quickly in a matter of few days instead of months and can be extended to cover a wide range of temperature regimes. Finite-element analysis (FEA) in conjunction with experimental data from the ATC for different lead-free PBGA packages was used to develop a fatigue life model that can be used to predict solder joint fatigue life for any PBGA package. The proposed model will be able to predict the mean number of cycles required for crack initiation and crack growth rate in a solder joint.

[Truncated abstract. Please see pdf format for complete text.] Over the years much work has been done on studying sequence effects under multilevel loading. Yet, the underlying fatigue mechanisms responsible for such interactions are not fully understood. The study of fatigue under complex loading histories begins by investigating strain interaction effects arising from simple 2-step loading sequences. Fatigue for all investigations were conducted under uniaxial push-pull mode in strain-control. Fatigue is traditionally classified as either low or high cycle fatigue (LCF and HCF respectively). The boundary for LCF and HCF is not well-defined even though the fatigue life of LCF is typically dominated by crack “initiation”, while for HCF, fatigue life is usually dominated by stable crack growth. The terms LCF and HCF, apart from referring to the low and high number of fatigue cycles required for failure, also bear little physical meaning in terms of describing the state of fatigue imposed. As a result, conventional definitions of the two distinct regimes of fatigue are challenged and a new method of classifying the boundary between the two regimes of fatigue is proposed. New definitions are proposed and the terms plastically dominant fatigue (PDF) and elastically dominant fatigue (EDF) are introduced as suitable replacements for LCF and HCF respectively. PDF refers to the condition of a material undergoing significant reverse plasticity during cyclic loading, while for EDF, minimal reverse plasticity is experienced. Systematic testing of three materials, 316 L stainless steel, 6061-T6 aluminium alloy and 4340 high strength steel, was performed to fully investigate the cycle ratio trends and “damage” accumulation behaviour which resulted from a variety of loading conditions. Results from this study were carried over to investigate more complex multilevel loading sequences and possible mechanisms for interaction effects observed both under 2-step and multi-step sequences were proposed. Results showed that atypical cycle ratio trends could result from loading sequences which involve combinations of strain amplitudes from different fatigue regimes (i.e. PDF or EDF). Mean strain effects on fatigue life were also studied. The objective of this study was to identify regimes of fatigue which are significantly influenced by mean strains. Results indicated that mean strains affected EDF but not PDF. 2-step tests, similar to those performed in earlier studies were conducted to investigate the effects of mean strain on variable amplitude loading. Again, atypical cycle ratio trends were observed for loading sequences involving combinations of PDF and EDF. It is understood that fatigue crack growth interaction behaviour and mean stress effects are two dominant mechanisms which can be used to explain cycle ratio trends observed. The significance and importance of proper PDF/EDF definition and specification are also stressed. The study of fracture mechanics is an important component of any fatigue research. Fatigue crack growth in 4140 high strength steel CT specimens, under conditions of plane stress and plane strain were studied. In this investigation, the effects of R and overload ratios were also studied for both plane stress and plane strain conditions. Results indicate that differences in the point of crack “initiation” under both plane stress and plane strain conditions decrease with increasing load range, while the extent of crack retardation as a result of overloading, is greater under plane stress than plane strain conditions. The extent of crack growth retardation increases with decreasing R ratios and increasing overload ratios. The final phase of this project involves the proposal of two practical models used to predict cumulative “damage” and fatigue crack propagation in metals. The cumulative “damage” model proposed takes the form of a power law and the exponent which governs “damage” accumulation can easily be calculated by knowing the failure life, Nf, for a given strain or load level. Predictions for the “damage” model performed better when compared to other popular cumulative “damage” models. The second model proposed predicts fatigue crack growth behaviour from known monotonic and smooth specimen fatigue data. There are several benefits of having a model that can predict fatigue crack growth from monotonic and smooth specimen fatigue data: a) traditionally, engineers had to rely on expensive and time-consuming crack propagation tests to evaluate and select materials for maximum fatigue resistance, and b) monotonic and smooth specimen fatigue data are readily available. The crack propagation model is proposed to alleviate the material selection process by providing engineers a means to rapidly eliminate and narrow down selections for possible material candidates.

Simplifications regarding calculations of fatigue life due to vibrational loading is based on weakly stationarity assumptions which is a time independent method. The hypothesis was based on the uncertainty of these assumptions. The aim of this study was to examine whether the analysed data fulfilled the assumptions of weakly stationarity. It was determined that the assumption was not valid for most signals and a comparison of time dependent methods should be performed to evaluate the difference compared with the time independent method. Two time dependent methods were constructed and implemented on the signals based on the results of performed stationarity tests. The result determined that a decrease in fatigue life of an investigated weld might occur for the two time dependent methods compared with the time independent method. The method which was considered to produce the most accurate results was also the most constrained as to the amount of data that fulfilled its requirements. A conclusion was drawn that signals containing more data was necessary to achieve conclusive results of the fatigue life. The hypothesis was proven to be mostly true since most of the analysed signals were found to be piecewise weakly stationary.