Le but principal de cette thèse est de développer un modèle numérique pour modéliser les phénomènesde bifurcation et du branchement des fissures. Pour réaliser cet objectif, il était indispensablede posséder un modèle permettant un couplage fort entre le modèle de Plasticité cristalline etcelui de l'Endommagement régularisé. Dans un premier temps, quelques outils de post-traitement ont été développés pour analyser les systèmes de glissement actifs. Ces outils ont été utilisés surdes simulations d'éprouvettes réelles, et comparés à des résultats expérimentaux. Par ces comparaisons, l'application du modèle de Plasticité cristalline aux superalliages monocristallins a été validée. Ce modèle a ensuite été couplé avec le modèle d'endommagement régularisé. Le couplage a été réalisé dans les deux sens, c'est-à-dire que l'évolution de la plasticité a une influence sur l'endommagement et vice-versa. Le nouveau modèle peut être implémenté simplement, avec la méthode traditionnelle des Éléments Finis. Des expériences étudiant la propagation de fissure sous des chargements de types différents ont été simulées à l'aide de ce nouveau modèle :éprouvettes CT,fissuration en Mode II et rupture en fluage. Une méthode pour l'identification des paramètres matériaux a également été proposée. / The main goal of this dissertation was to develop a model to simulate the processes of crack bifurcation and crack branching in anisotropic materials. To achieve this goal, a thorough coupling of crystal plasticity and regularised damage models was deemed necessary. Firstly, post-processing tools were developed to better analyse the results obtained from standard Crystal Plasticity simulations. These were then compared with experiments, thereby validating the use of Crystal Plasticity models for Nickel base single crystal superalloys. The validated Crystalplasticity model was then coupled with a regularised microdamage model such that the evolution of plasticity influenced damage and vice versa. The newly developed model allows for the simulation of cracks using the standard Finite Element Approach. Experiments studying crack propagation under different types of loads were simulated using the newly developed model, including CT, shear andcreep specimens. A methodology was also proposed for the identification of the newly introduced material parameters.

L'évaporation d'une goutte reposant sur un support plat semble être un système relativement simple à étudier et a fait l'objet d'études scientifiques depuis plus d'un siècle. Cependant, l'étude de l'évaporation de gouttes sessiles est toujours d'actualité aujourd'hui avec l'essor de nouvelles techniques de visualisation ou de l'apparition de nouveaux types de fluides complexes.Cette étude expérimentale sera focalisée sur deux aspects distincts :- L'étude sur l'évaporation de fluides purs permettra d'étudier la dynamique d'évaporation et les ondes hydrothermales qui apparaissent dans les gouttes de fluides volatils lors du changement de phase. L'influence du type de fluide (différents alcools et alcanes) et du niveau de gravité (terrestre, lunaire et martienne) seront étudiés. De plus, l'utilisation de différents niveaux de gravité permet de développer un modèle empirique afin de prendre en compte dans le modèle quasi-stationnaire limité par diffusion de la vapeur la convection naturelle qui augmente fortement le débit d'évaporation.- Si les fluides complexes présentent une dynamique de séchage similaire à celle des fluides purs, d'autres mécanismes entrent en jeux, comme la gélification, l'organisation des particules et l'apparition de craquelures (voir Figure 2). Le mouillage et les différents groupes fonctionnels graphés sur les particules seront étudiés en regard du motif final de craquelures. / Sessile droplets are widely found in day to day life: it might be a coffee spilt, rain onto a waterproof raincoat or again, water falling onto a cooking plate. However, despite the vast number of studies devoted to droplets for almost half a century, the fundamental phenomenon of the evaporation of sessile droplets is still a field that attracts a high level of interest due to its wide applicability and the development of new visualisation techniques or new types of complex fluids. This experimental study is focused two distinct aspects:- The evaporation of pure fluids has allow to study hydrothermal waves that appear in the droplets of volatile fluids during phase change. The influence of the type of fluid ---different alcohols and alkanes--- and the gravity levels ---Terrestrial, Lunar and Martian--- is investigated to have a better understanding of the flow motion inside droplet. Moreover, the use of different gravity levels allows to experimentally evidence the contribution of the atmospheric convective transport to sessile droplet evaporation. This investigation has allowed to develop an empirical model to take account of natural convection which greatly increases the evaporation rate in the quasi-steady diffusion-controlled evaporation model.- If complex fluids exhibit an evaporation dynamic similar to pure fluid, other mechanisms come into play, such as gelation, particles organisation and cracks formation. Wetting and different functional groups on the particles graphs will be studied in relation to the final pattern of cracks.

Une modélisation numérique des grands barrages en enrochement à masque amont (CFRD) a été entreprise pour mieux comprendre les pathologies observées sur ces ouvrages récemment construits, principalement des fissures horizontales et verticales sur le masque amont en béton en phase de construction et de mise en eau. Le but de cette modélisation est ensuite de confirmer, préciser ou compléter les solutions proposées par les experts pour éviter ces désordres.D‟abord, une loi de comportement développée au sein d‟EDF-CIH appelée L&K-Enroch a été présentée. Il s'agit d'un modèle élastoplastique qui prend en compte les déformations irréversibles de l‟enrochement sous sollicitation déviatorique et isotrope dans les grands CFRD. Cette thèse propose aussi une méthode de calage des paramètres de ce modèle.Deux modélisations des barrages sont également présentées dans cette thèse afin de vérifier la fiabilité de la loi de comportement L&K-Enroch et de mieux comprendre les pathologies sur le masque amont en béton du CFRD. La modélisation en déformation plane (2D) est d‟abord entreprise et a comme support le barrage d‟Aguamilpa (180,5 m), au Mexique. Une modélisation plus poussée en 3D prenant en compte de l'effet de la vallée est ensuite entreprise. Elle a comme support le barrage de Mohale (145 m) au Lesotho.L'effet d‟échelle de l'enrochement est également étudié en comparant les résultats de simulation pour les matériaux de différentes tailles. L'impact de l‟effet d‟échelle sur le comportement du barrage est aussi analysé.Les simulations présentées en 2D et 3D sont une contribution à l'analyse du comportement mécanique des grands CFRD. Les résultats de la modélisation 2D montrent généralement une bonne correspondance avec les mesures in-situ. La modélisation 3D explique, d'une manière plus convaincante, l‟apparition des fissures sur le masque amont en béton des CFRD. Certaines mesures constructives sont finalement proposées afin de limiter ou d'éviter ces désordres dans le masque. / A numerical modeling of high concrete face rockfill dam (CFRD) has been considered in order to better understand the pathology of cracks observed in the recently constructed high CFRD, especially the horizontal and vertical cracks detected on the concrete face slab during the construction and the impoundment. The purpose of this modeling is then to confirm, clarify or complete the solutions proposed by the experts to avoid these disorders.This PhD thesis presents a constitutive model developed by EDF-CIH known as L&K-Enroch. It is an elastoplastic model that takes into account the irreversible deformations of rockfill under deviatoric and isotropic loading in high CFRD. A first attempt of proposing a parameter calibration approach has been made as well.Two dam modellings are also presented in this thesis in order to verify the reliability of the constitutive model and to better understand the pathology of cracks observed on the concrete face slab of CFRD. A 2D modeling is compared to the behaviour of the 180.5 m Aguamilpa Dam, Mexico; a 3D modeling taking into account the valley effect is compared to the behaviour of the 145 m Mohale Dam, Lesotho.This thesis also studies the scale effect of rockfill through a comparison of the simulation results of materials of different sizes. The impact of the scale effect on the dam behavior is analyzed as well.The presented 2D and 3D simulations are a contribution to the analysis of the mechanical behaviour of high CFRD. The results of the 2D modeling show a relatively good agreement with the monitored data. The 3D modeling explains, in a rather convincing way, the pathology of cracks on the slab face of CFRD. Some measures are proposed to limit or avoid these different kinds of cracks.

Ce travail est consacré à l’étude de la fatigue des roulements hybrides (Si3N4 / M50 Duplex) soumis à des sollicitations extrêmes. Dans un premier temps, nous avons étudié le comportement expérimental du M50 trempé et du M50 Duplex sollicités en fatigue de roulement en présence d’indentation artificielle, dans le cas d’un contact hybride. Nous avons alors pu mettre en évidence le rôle prépondérant du glissement et nous avons pu caractériser et localiser l’endommagement relatif à ces conditions extrêmes. Aussi deux études portant sur la fissuration des éléments M50 Duplex, fissures radiales en bord d’indent, et sur la fissuration des éléments céramique Si3N4, sous forme de C-Cracks, ont été réalisées et ont permis d’évaluer la résistance de ces matériaux en sollicitation aggravée. Dans une seconde étape, nous avons mis au point une méthode d’évaluation du gradient de plasticité par mesure de la micro-dureté qui nous a permis de caractériser le M50 Duplex dans la zone de diffusion de la nitruration. Le processus d’indentation a alors été simulé par le logiciel éléments finis ABAQUS. Ces simulations ont permis d’analyser les mécanismes responsables de la fissuration des indents sur M50 Duplex. Enfin une étude numérique tridimensionnelle du roulement sur indent a été réalisée via le code semi-analytique ISAAC, développé au laboratoire LAMCOS de l’INSA de Lyon. Cette étude a permis d’évaluer le cycle de chargement au voisinage d’indents sollicités en fatigue de roulement. L’application du critère de Dang Van à ce cycle de sollicitation a alors permis de donner des indications quant à la prédiction de l’endommagement au voisinage d’indents. / This study focuses on the hybrid bearings fatigue (Si3N4 / Duplex M50) which are submitted to extreme stresses. Firstly, we studied the rolling contact fatigue of artificially dented rollers made of through-hardened M50 and Duplex M50, in the case of hybrid contact. We have been consequently able to highlight the prominent role of the sliding. We also characterized and located the damage due to the extreme conditions. What is more, two studies have been led: one about the cracking damage of Duplex M50 rollers, especially radial cracks on dent edges, the other about the cracking damage of silicon nitride rollers, especially C-shaped cracks. These studies allowed to assess the material strength in conditions of severized stresses. Secondly, we developed an assessment method of the plastic gradient through micro-hardness measurings. This process allowed us to characterize the Duplex M50 in the diffusion layer formed during nitriding. The denting process was then simulated using the ABAQUS finite element software. The simulations made it possible to analyze the mechanisms responsible for the cracking of dents made on Duplex M50. Eventually, a tridimensional numerical study focusing on dent over-rolling has been led using the ISAAC semi-analytical code, computed at the LAMCOS laboratory of INSA-Lyon (Lyon, FRANCE). This study made it possible to evaluate the loading path in the vicinity of dents submitted to RCF. The use of Dang Van criterion then allowed to give indications about damage prediction in the vicinity of dents.

Structural concrete is the most widely used material in the construction of bridges, pave-ments, runways, dams and other infrastructures which are subjected to uctuating loads during its service period. Concrete contains internal aws in the form of micro-cracks as an inherent property. When subjected to fatigue loading, distributed micro-cracks are formed at the sites of pre-existing aws, which subsequently, localize to form a major crack and propagates. The crack growth curve of a structural component when subjected to fatigue loading depicts a sigmoidal pattern. This curve is divided into three distinct regions namely sub-threshold crack propagation (short crack), stable crack propagation (long crack) and unstable crack propagation depending on the crack propagation rate. Most of the fatigue life is spent in the sub-critical stage (small crack) before the for-mation of long cracks. Hence, from the view of estimating the fatigue life, the crack initiation and early crack propagation (short crack stage) phase are the most important and correct concepts need to be developed. Hence, in this work, the behavior of propa-gation and non-propagationof short cracks in concrete when subjected to fatigue loading is addressed. Small non-propagating cracks are usually found at notch roots when the nominal stress range is below certain limits that depend on the notch sensitivity. Analysis is performed on geometrically similar three-point bend beams of three di erent sizes and subjected to fatigue loading in order to determine the important factors that a ect the notch sensitivity and to determine the minimum stress range required for the initiation and propagation of short cracks. A criterion for crack initiation and propagation is proposed based on linear elastic fracture mechanics. Using this criterion, the maximum length of non-propagating crack that can be formed from fatigue loading alone and the minimum stress range required to propagate a crack without arrest are computed. It is observed that the notch sensitivity increases with increase in beam size, decrease in notch-tip aspect ratio and increase in the fatigue limit of the material. Since the probability of formation of a non-propagating crack at a notch tip decreases with increase in notch sensitivity, and since it is desirable not to have a non-propagating crack in experimental investigations, it is essential to design a specimen with higher notch sensitivity. A crack spends a considerable amount of time in the short crack regime. The short cracks are found to propagate at higher rates than the long cracks at the same nominal stress intensity factor which is known as the short crack anomaly. It is important to consider this anomaly in the prediction of the residual life of damaged concrete structures. Hence, in the present work, an analytical model is developed using the principles of dimensional analysis and self-similarity in order to estimate the rate of short crack growth in concrete. The important parameters such as load range, threshold value of stress intensity factor range, modulus of elasticity, tensile strength, fracture energy, stress ratio, crack size and the maximum aggregate size are considered in the development of the short crack growth model. The model is calibrated and validated using the experimental results that are available in the literature. A probabilistic analysis is carried out to determine the sensitivity of each of the di erent parameters that has been considered on the crack growth rate using the coe cient of variation method. It is found that the crack length is the most sensitive parameter to short crack growth rate followed by the load range. A term called `characteristic fatigue life of short crack' is de ned as the number of fatigue cycles that can be applied such that not more than ve percent of the short cracks is expected to proceed to the long crack regime. Furthermore, the fatigue life of a crack spent in the short crack regime is determined through a reliability based study using the Monte Carlo technique. It is found that the smaller sized specimens have larger fatigue life in the short crack regime than the larger specimens.

Sprintlyftet är ett vanligt lyft som används på byggarbetsplatser idag för att montera betongbalkar och det används dessutom när betongbalkarna hanteras i fabriken där de tillverkats. Allt fler prefabricerade byggelement tillverkas och det leder till att lyftens kapacitet behöver säkerställas. I den här studien provades kapaciteten experimentiellt i ett specifikt lyft, sprintlyftet. Det är ett lyft som Abetong använder till sina betongbalkar. Det som under studien kontrollerades var om betongbalkarna idag klarar mer än vad de gjorde vid motsvarande provning för 42 år sedan. En jämförelse mellan provningar som gjordes av Abetong år 1975 och den här studiens provning visar att sprintlyften som används idag klarar av mer last än vad provningarna från 1975 gjorde. Förklaringen till det är främst att kvalitén av betongen har utvecklats under åren. / The sprint lift is a common lift used on construction sites today to mount concrete beams and it is also used when the concrete beams are handled in the factory where they were manufactured. More and more prefabricated building elements are manufactured and this means that the lifts capacity needs to be ensured. In this study, capacity was experimentally tested in a specific lift, the sprint lift. It is a lift that Abetong uses for its concrete beams. What was checked during the study was if the concrete beams today manage more than they did in the corresponding test 42 years ago. A comparison of tests made by Abetong in 1975 and this study's test show that the sprint lift used today is capable of more load than what the 1975 tests did. The explanation for this is primarily that the quality of the concrete has developed over the years.

Concrete substructures are often subjected to environmental deterioration, such as sulfate and acid attack, which leads to severe damage and causes structure degradation or even failure. In order to improve the durability of concrete, the High Performance Concrete (HPC) has become widely used by partially replacing cement with pozzolanic materials. However, HPC degradation mechanisms in sulfate and acidic environments are not completely understood. It is therefore important to evaluate the performance of the HPC in such conditions and predict concrete service life by establishing degradation models. This study began with a review of available environmental data in the State of Florida. A total of seven bridges have been inspected. Concrete cores were taken from these bridge piles and were subjected for microstructural analysis using Scanning Electron Microscope (SEM). Ettringite is found to be the products of sulfate attack in sulfate and acidic condition. In order to quantitatively analyze concrete deterioration level, an image processing program is designed using Matlab to obtain quantitative data. Crack percentage (Acrack/Asurface) is used to evaluate concrete deterioration. Thereafter, correlation analysis was performed to find the correlation between five related variables and concrete deterioration. Environmental sulfate concentration and bridge age were found to be positively correlated, while environmental pH level was found to be negatively correlated. Besides environmental conditions, concrete property factor was also included in the equation. It was derived from laboratory testing data. Experimental tests were carried out implementing accelerated expansion test under controlled environment. Specimens of eight different mix designs were prepared. The effect of pozzolanic replacement rate was taken into consideration in the empirical equation. And the empirical equation was validated with existing bridges. Results show that the proposed equations compared well with field test results with a maximum deviation of ± 20%. Two examples showing how to use the proposed equations are provided to guide the practical implementation. In conclusion, the proposed approach of relating microcracks to deterioration is a better method than existing diffusion and sorption models since sulfate attack cause cracking in concrete. Imaging technique provided in this study can also be used to quantitatively analyze concrete samples.

Mechanical fatigue is a failure phenomenon that occurs due to repeated application of mechanical loads. Very High Cycle Fatigue (VHCF) is considered as the domain of fatigue life greater than 10 million load cycles. Increasing numbers of structural components have service life in the VHCF regime, for instance in automotive and high speed train transportation, gas turbine disks, and components of paper production machinery. Safe and reliable operation of these components depends on the knowledge of their VHCF properties. In this thesis both experimental tools and theoretical modelling were utilized to develop better understanding of the VHCF phenomena. In the experimental part, ultrasonic fatigue testing at 20 kHz of cold rolled and hot rolled stainless steel grades was conducted and fatigue strengths in the VHCF regime were obtained. The mechanisms for fatigue crack initiation and short crack growth were investigated using electron microscopes. For the cold rolled stainless steels crack initiation and early growth occurred through the formation of the Fine Granular Area (FGA) observed on the fracture surface and in TEM observations of cross-sections. The crack growth in the FGA seems to control more than 90% of the total fatigue life. For the hot rolled duplex stainless steels fatigue crack initiation occurred due to accumulation of plastic fatigue damage at the external surface, and early crack growth proceeded through a crystallographic growth mechanism. Theoretical modelling of complex cracks involving kinks and branches in an elastic half-plane under static loading was carried out by using the Distributed Dislocation Dipole Technique (DDDT). The technique was implemented for 2D crack problems. Both fully open and partially closed crack cases were analyzed. The main aim of the development of the DDDT was to compute the stress intensity factors. Accuracy of 2% in the computations was attainable compared to the solutions obtained by the Finite Element Method. / Very High Cycle Fatigue (VHCF) is considered as the domain of fatigue life greater than 10 million load cycles. Structural components that have service life in the VHCF regime include wheels and axles of high speed trains, gas turbine disks, and components of paper production machinery. Safe and reliable design, and the longevity, of these components depends on the knowledge of their VHCF properties. The overall aim of the experimental portion of this thesis was to gain in-depth knowledge of the VHCF properties of stainless steels. Fatigue test data in the VHCF regime was generated for different stainless steel grades using ultrasonic fatigue testing. The mechanisms for fatigue crack initiation and short crack growth were investigated using electron microscopes. Theoretical modelling of complex crack geometries involving kinks and branches was carried out by using the Distributed Dislocation Dipole Technique (DDDT). The main aim of this development was to compute the stress intensity factors and to analyse the stress state around the cracks. The results showed that accuracy of 2% was attainable compared to the solutions obtained by Finite Element Method (FEM). / <p>Artikel 4 publicerad i avhandlingen som manuskript</p>

Gypsum board assemblies are now widely used in buildings, as fire resistant walls or ceilings, to provide passive fire protection. The fire resistance of such systems is fundamentally due to the desirable thermal properties of gypsum. Yet there is wide variability in reported values of thermal properties of gypsum at high temperatures and a lack of understanding of its integrity in fire. To evaluate the fire protection performance of gypsum board assemblies, it is essential to quantify its thermal properties and obtain information on its mechanical properties at high temperatures. Gypsum boards shrink and crack at high temperatures, and this leads to collapse of parts of the gypsum boards in fire. Fall-off of gypsum in fire affects the fire resistance of the assembly considerably, and cannot be overlooked when evaluating the fire resistance of gypsum board assemblies. The current research proposes a model to define the temperature-dependent thermal properties of gypsum boards at high temperatures. Thermal conductivity of gypsum is considered as the most influential parameter in conduction of heat through gypsum, and a hybrid numerical-experimental method is presented for extracting thermal conductivity of various gypsum board products at elevated temperatures. This method incorporates a validated one-dimensional Finite Difference heat conduction program and high temperature test results on small samples of gypsum boards. Moreover, high temperature mechanical tests have been performed on different gypsum board products; thermal shrinkage, strength and stress-strain relationships of gypsum products at elevated temperatures are extracted for use in numerical mechanical analysis. To simulate the structural performance of gypsum boards in fire, a two-dimensional Finite Element model has been developed in ABAQUS. This model successfully predicts the complete opening of a through-thickness crack in gypsum, and is validated against medium-scale fire tests designed and conducted as part of this research. Gypsum fall-off in fire is a complex phenomenon; however, it is believed that delaying the formation of through-thickness cracking will delay falling off of gypsum in fire, and hence improve the fire resistance of gypsum board assemblies. Finally, a study has been performed on the effects of various detailing parameters in gypsum board wall assemblies, and recommendations are offered for improving the fire resistance of such systems.

Cracks in wood are considered to be one of the major problems for products, which have been made from this material. Crack detection and its propagation methods should be revised and improved with application of modern techniques. Nowadays new measuring techniques like digital camera image processing, and 3-D laser scanning are available. This work describes computer aided in-field methods for registration cracks in wood, its propagation, and tracing dimensional stability of glued laminated beams. The benefits of different methods for supervision of wooden element as well as its limitations are discussed. / <p>Validerat; 20120608 (anonymous)</p>