Les travaux présentés dans ce mémoire concernent le développement de méthodes de pronostic à base de modèles. Le pronostic à base de modèles a pour but d'estimer le temps qu'il reste avant qu'un système ne soit défaillant, à partir d'un modèle physique de la dégradation du système. Ce temps de vie restant est appelé durée de résiduelle (RUL) du système.Le pronostic à base de modèle est composé de deux étapes principales : (i) estimation de l'état actuel de la dégradation et (ii) prédiction de l'état futur de la dégradation. La première étape, qui est une étape de filtrage, est réalisée à partir du modèle et des mesures disponibles. La seconde étape consiste à faire de la propagation d'incertitudes. Le principal enjeu du pronostic concerne la prise en compte des différentes sources d'incertitude pour obtenir une mesure de l'incertitude associée à la RUL prédite. Les principales sources d'incertitude sont les incertitudes de modèle, les incertitudes de mesures et les incertitudes liées aux futures conditions d'opération du système. Afin de gérer ces incertitudes et les intégrer au pronostic, des méthodes probabilistes ainsi que des méthodes ensemblistes ont été développées dans cette thèse.Dans un premier temps, un filtre de Kalman étendu ainsi qu'un filtre particulaire sont appliqués au pronostic de propagation de fissure, en utilisant la loi de Paris et des données synthétiques. Puis, une méthode combinant un filtre particulaire et un algorithme de détection (algorithme des sommes cumulatives) a été développée puis appliquée au pronostic de propagation de fissure dans un matériau composite soumis à un chargement variable. Cette fois, en plus des incertitudes de modèle et de mesures, les incertitudes liées aux futures conditions d'opération du système ont aussi été considérées. De plus, des données réelles ont été utilisées. Ensuite, deux méthodes de pronostic sont développées dans un cadre ensembliste où les erreurs sont considérées comme étant bornées. Elles utilisent notamment des méthodes d'inversion ensembliste et un observateur par intervalles pour des systèmes linéaires à temps discret. Enfin, l'application d'une méthode issue du domaine de l'analyse de fiabilité des systèmes au pronostic à base de modèles est présentée. Il s'agit de la méthode Inverse First-Order Reliability Method (Inverse FORM).Pour chaque méthode développée, des métriques d'évaluation de performance sont calculées dans le but de comparer leur efficacité. Il s'agit de l'exactitude, la précision et l'opportunité. / In this manuscript, contributions to the development of methods for on-line model-based prognosis are presented. Model-based prognosis aims at predicting the time before the monitored system reaches a failure state, using a physics-based model of the degradation. This time before failure is called the remaining useful life (RUL) of the system.Model-based prognosis is divided in two main steps: (i) current degradation state estimation and (ii) future degradation state prediction to predict the RUL. The first step, which consists in estimating the current degradation state using the measurements, is performed with filtering techniques. The second step is realized with uncertainty propagation methods. The main challenge in prognosis is to take the different uncertainty sources into account in order to obtain a measure of the RUL uncertainty. There are mainly model uncertainty, measurement uncertainty and future uncertainty (loading, operating conditions, etc.). Thus, probabilistic and set-membership methods for model-based prognosis are investigated in this thesis to tackle these uncertainties.The ability of an extended Kalman filter and a particle filter to perform RUL prognosis in presence of model and measurement uncertainty is first studied using a nonlinear fatigue crack growth model based on the Paris' law and synthetic data. Then, the particle filter combined to a detection algorithm (cumulative sum algorithm) is applied to a more realistic case study, which is fatigue crack growth prognosis in composite materials under variable amplitude loading. This time, model uncertainty, measurement uncertainty and future loading uncertainty are taken into account, and real data are used. Then, two set-membership model-based prognosis methods based on constraint satisfaction and unknown input interval observer for linear discete-time systems are presented. Finally, an extension of a reliability analysis method to model-based prognosis, namely the inverse first-order reliability method (Inverse FORM), is presented.In each case study, performance evaluation metrics (accuracy, precision and timeliness) are calculated in order to make a comparison between the proposed methods.

This doctoral thesis deals with the determination of the influence of plasma electrolytic oxidation (PEO) on fatigue behaviour of extruded AZ61 magnesium alloy in air and in the 3.5% NaCl solution. The coatings were formed in the silicate-phosphate electrolyte under pulsed current conditions at a frequency of 50 Hz. The influence of current density on coating formation was examined under current densities of 70, 130 and 200 mA cm-2 for different durations up to a maximum of 1800 s. 8 g dm-3 of KF were added to the electrolyte to study the influence of fluoride ions in plasma electrolytic oxidation. It is shown that fluoride ions inhibit localised oxidation in the initial stage of the process, associated with the secondary particles based on Al–Mn. The presence of fluoride also modified the sparking characteristics, decreased the rate of coating growth and changed the morphologies of the coatings. The influence of fluoride on the coating hardness, and the corrosion resistance of the alloy during exposure to salt spray, was negligible. Based on previous optimised PEO conditions, coatings formed under a current density of 130 mA cm-2 for 300 s in the electrolyte containing KF were chosen for fatigue testing. The high-cycle fatigue tests were carried out on cylindrical samples under a force controlled sinusoidal tension-tension cycle with asymmetry parameter R=0. The experimental data were fitted with Kohout & Věchet function. The fatigue limit of uncoated alloy in air for 107 cycles was determined at 145.4 MPa and the combination of PEO coating with chloride ions caused a reduction of ~55 %. Attention was paid to the fatigue crack initiation in different conditions of cyclic loading. The fracture surfaces underwent detailed fractography analysis including secondary crack observation on the gauge length. The contribution of Al–Mn particles were confirmed on the uncoated alloy in air and the presence of chloride ions were observed as another influential contributor to local corrosion attack. The cyclic loading caused spalling of the outer layer, and the multiple initiation was observed on PEO coated alloy, caused by cracks in the coating and stress transferring to the alloy.

No description available.

Finite element analyses are frequently used to model growing fatigue cracks and the associated plasticity-induced crack closure. Two-dimensional, elastic-perfectly plastic finite element analyses of middle-crack tension (M(T)), bend (SEB), and compact tension (C(T)) geometries were conducted to study fatigue crack closure and to calculate the crack opening values under plane-strain and plane-stress conditions. The loading was selected to give the same maximum stress intensity factor in both geometries, and thus similar initial forward plastic zone sizes. Mesh refinement studies were performed on all geometries with various element types. For the C(T) geometry, negligible crack opening loads under plane-strain conditions were observed. In contrast, for the M(T) specimen, the plane-strain crack opening stresses were found to be significantly larger. This difference was shown to be a consequence of in-plane constraint. Under plane-stress conditions, it was found that the in-plane constraint has negligible effect, such that the opening values are approximately the same for the C(T), SEB, and M(T) specimens. Next, the crack opening values of the C(T), SEB and M(T) specimens were compared under various stress levels and load ratios. The effect of a highly refined mesh on crack opening values was noted and significantly lower crack opening values than those reported in literature were found. A new methodology is presented to calculate crack opening values in planar geometries using the crack surface nodal force distribution under minimum loading as determined from finite element analyses. The calculated crack opening values are compared with values obtained using finite element analysis and more conventional crack opening assessment methodologies. It is shown that the new method is independent of loading increment, integration method (normal and reduced integration), and crack opening assessment location. The compared opening values were in good agreement with strip-yield models.

No description available.

No description available.

Thermomechanical fatigue is a fatigue failure caused by combined thermal and mechanical loading cycles where both stresses and temperature can vary with time. A cohesive zone model simulates material failure within the finite element method by describing material behaviour and damage evolution through a traction-separation relation. In this work, a viscoplastic cohesive zone model is developed in order to describe the fatigue and fracture behaviour of high performance superalloys undergoing thermomechanical fatigue loading. A micromechanically motivated cohesive potential, incorporating fatigue and creep damage variables, is proposed and its thermodynamic consistency is established. The characteristic behaviour of the cohesive law during cyclic loading is described in terms of both traction and energy. The developed cohesive zone model is applied to simulate the lifetime behaviour in a smooth specimen. The effect of loading parameters such as temperature, dwell period, and strain rate on isothermal, in-phase and out-of-phase thermomechanical fatigue is successfully demonstrated. Using crack growth simulations in a boundary layer model, the relation between the local behaviour of the cohesive zone and fracture parameters like stress intensity factor, J-integral and crack tip opening displacement is established. The three stages of fatigue crack growth rate curves that are experimentally observed were obtained through simulations. Crack growth in a corner crack specimen is simulated under cyclic stress-controlled loading to obtain the experimentally observed crack profiles. Material parameters for the Nickel based superalloy MAR-M247 are identified using experimental data from the literature.

Im Rahmen des Forschungsprojekts ”TMF-Rissverlaufsberechnung für ATL-Heißteile“ des FVV e.V. wurden Untersuchungen zur Entwicklung eines Rissfortschrittsgesetzes unter thermomechanischer Wechselbeanspruchung (thermo-mechanical fatigue, TMF) für das austenitische Gusseisen Ni-Resist D-5S durchgeführt. Ziel der experimentellen Arbeiten war es, mit einseitig gekerbten Proben (SENT) eine Datenbasis für Risswachstum unter TMF-Belastung zu schaffen. Das Werkstoffverhalten des betrachteten austenitischen Gusseisens Ni-Resist D-5S wurde mithilfe eines validierten viskoplastischen, temperaturabhängigen Materialmodells modelliert, das zur Berücksichtigung großer Verzerrungen und Rotationen am Riss auf große Deformationen erweitert wurde. Zur Beurteilung des Rissfortschritts unter TMF wurde die zyklische Rissspitzenöffnungsverschiebung (ΔCTOD) als geeigneter Beanspruchungsparameter verwendet. Für die Simulation der Rissausbreitung wurde ein automatischer FEM-Algorithmus mit inkrementeller, adaptiver Neuvernetzung entwickelt. Dabei wurden Verformungen und inelastische Zustandsvariablen auf das neue Netz übertragen. Der Einfluss verschiedener Parameter innerhalb der Simulationsstruktur wurde analysiert. Ein verbesserter Mapping-Algorithmus zur Übertragung der Zustandsvariablen wurde entwickelt. Mithilfe begleitender 2D FEM-Simulationen wurden die Rissfortschrittskurven des Werkstoffs basierend auf den experimentellen Daten ermittelt und unter Anwendung des ΔCTOD Konzepts in parametrisierter Form dargestellt. Zusätzlich wurden Rissfortschrittsgesetze durch den Einsatz von Machine-Learning-Konzepten bestimmt. Dies ermöglicht erstmals eine quantitative Vorhersage der Rissentwicklung unter Beanspruchungsbedingungen mit Großbereichsfließen unter Berücksichtigung von TMF. / As part of the research project ”TMF crack propagation calculation for ATL hot parts“ of the FVV e.V., an investigation was carried out to develop a crack propagation law under thermomechanical fatigue (TMF) for the austenitic cast iron Ni-Resist D-5S. The aim of the experimental work was to create a database for crack growth under TMF loading with single edge notch tension specimens (SENT). The material behavior of the austenitic cast iron under consideration, Ni-Resist D-5S, was modeled using a validated viscoplastic temperature dependent material model, which was extended to large deformations to account for large distortions and rotations in the crack. The cyclic crack opening displacement (ΔCTOD) was used as a suitable loading parameter to assess crack propagation under TMF. An automatic FEM algorithm with incremental adaptive remeshing was developed for the simulation of crack propagation. In the process, the deformations and inelastic state variables were transferred to the new mesh. The influence of different parameters within the simulation structure was analyzed. An ideal mapping algorithm for the transfer of state variables was developed. With the help of accompanying 2D FEM simulations, the crack propagation curves of the material were determined on the basis of the experimental data and presented in parameterized form using the concept ΔCTOD. Furthermore, crack propagation laws were determined using machine learning concepts. This allows for the first time a quantitative prediction of crack development under loading conditions with large scale yielding taking TMF into account.

Les outillages de mise en forme à chaud sont soumis à des sollicitations thermomécaniques transitoires. Ils sont donc confrontés à la fissuration par fatigue. Le but de ce travail est d'étudier le comportement et la durée de vie en fatigue des aciers d'outillage, notamment leurs résistances à la propagation de fissure. Dans un premier temps, les chargements thermiques imposés aux outillages de fonderie sous pression de magnésium et d'emboutissage à chaud ont été évalués. Ensuite, le comportement et la durée de vie de l'acier à haute conductivité thermique HDC1 ont été étudiés et comparés à l'acier AISI H11 (acier de référence) en fonction de la température. L'acier HDC1 présente un adoucissement cyclique stable à 20°C et 300°C. Par contre, l'intensité d'adoucissement est plus importante à hautes températures. La durée de vie a été étudiée en utilisant les lois de Manson-Coffin et de Basquin. A hautes températures, l'oxydation devient un mécanisme d'endommagement primordial pour l'acier HDC1 et provoque des durées de vie plus courtes que celles observées sur l'acier AISI H11. La résistance à la propagation de fissure de fatigue a été déterminée dans des aciers à la température ambiante par le biais de critères : l'amplitude de facteur d'intensité de contrainte (∆K) et l'amplitude d'ouverture en pointe de fissure (∆CTOD). La méthode de corrélation d'images a permis de mesurer (∆COD) et d'évaluer (∆CTOD). L'ensemble de ces résultats a permis de mettre en évidence l'effet de fermeture de fissure et le comportement plastique en pointe de fissure. ∆CTOD présente un bon critère pour rationaliser la propagation de fissure dans les aciers étudiés.Enfin, la simulation numérique de la propagation de fissure en fatigue a été menée dans l'acier AISI H11 à 600°C par la méthode de relâchement des nœuds en éléments finis. L'effet des modèles de comportement monotone (élastoplastique) et cyclique (élastoviscoplastique) a été étudié sur le calcul de l'ouverture de fissure et la plasticité en pointe de fissure. L'influence du modèle de comportement est faible sur le calcul de l'amplitude d'ouverture de fissure ∆COD, du fait d'une plasticité confinée en pointe de fissure. Alors que, le modèle de comportement cyclique est mieux adapté pour décrire la plasticité en pointe de fissure. / Hot metal forming tools are subjected to cyclic thermomechanical loading and damage by complex fatigue/wear/oxidation interactions. Thermal solicitations were measured on high pressure die casting and hot stamping processes. Based on thermal measurements, the isothermal fatigue behaviour and lifetime of a new high conductivity steel HDC1 were investigated at different temperatures and strain amplitude then compared to AISI H11 steel. As AISI H11, continue cyclic softening was observed in HDC1 at all temperatures. The Manson-Coffin and Basquin laws were used for life prediction models under different temperatures. It was observed that the fatigue/oxidation interaction was a principal damage mechanism of the HDC1 steel at high temperature. Fatigue crack propagation in steels was investigated at room temperature in SENT specimens. A digital image correlation technique was used to evaluate crack opening (∆COD) and crack tip opening displacement (∆CTOD) ranges. Crack growth rate were investigated using ∆K (Paris law) and ∆CTOD criteria. It was observed that the cyclic crack tip plasticity control the crack propagation resistance. Crack closure could be evaluated by ∆CTOD.Finite element method by debond technique was used to model the crack propagation of AISI H11 at 600°C using both monotonic elasto-plastic (EP) and cyclic elasto-viscoplastic (EVP) constitutive laws materials. The comparison of ∆COD calculated and measured had shown that monotonic EP and cyclic EVP had no significant effect on the ∆COD, mainly due to the small-scale yielding conditions. It is however observed that the cyclic constitutive law was the best suitable model for the crack tip plasticity effect.

Main extrinsic and intrinsic modifiers factors of crack growth rate in AA7050-T7451 were assessed in order to provide tools for aeronautical structures designers. These tools cover most necessary information to project aircraft\'s structures using the studied alloy, under damage tolerance philosophy. The experimental methodology consisted of use CT specimens, on TL and LT rolling direction to test its behavior under different conditions of stress ratio, force waveform, and the environment. The stress ratio values were 0.1 and 0.5, the force waveform used were sine and trapezoidal or Dwell under normal air laboratory conditions and salt fog 3.5%NaCl weight in order to simulate the marine environment. In Dwell tests, results were checked with the electrical potential drop technique (DCPD) in addition to the crack opening displacement (COD) method. Using the Walker coefficients, calculated on the present research, could be projected accurately the crack propagation behavior on Paris region and do fatigue life predictions using da/dN and S-N diagrams for different stress ratio values. The corrosion environment increases both crack growth rate and ΔKth due to oxides formation on the crack path that generates a crack closure effect. Dwell carrying makes decrease the crack growth rate by decreasing the slope of the Paris line on log (da/dN) versus log (ΔK) curve, instead of shifting down the line as occurs on titanium alloys. Rolling direction change from LT to TL increase the FCG rate in both threshold and Paris region, where the rate change use to be small. / Os principais fatores modificadores extrínsecos e intrínsecos da taxa de propagação de trincas na liga AA7050-T7451 foram avaliados para fornecer subsídios para projetistas de estruturas aeronáuticas, com base na filosofía de tolerância ao dano. A metodologia experimental consistiu em ensaiar corpos de prova do tipo compact tension (CT) da liga nas direções de laminação TL e LT, para verificar seu comportamento sob diferentes razões de tensões, forma de onda e condição ambiente. Os valores de razão de tensão estudados foram 0,1 e 0,5, as formas de onda foram senoidal e trapezoidal ou de Dwell, em condições normais de laboratório, ao ar, e névoa salina 3,5% NaCl, em massa, para simular um ambiente marinho. No caso dos ensaios Dwell, os resultados foram conferidos pelo método de queda de potencial eléctrico (QPE), além do método de flexibilidade elástica. Usando os coeficientes de Walker calculados a partir dos resultados obtidos, pôde-se projetar com precisão o comportamento da propagação de trinca na região de Paris e prever a vida em fadiga usando os diagramas da/dN e S-N para diferentes valores da razão de tensões. O ambiente corrosivo aumenta tanto a taxa de propagação de trinca, quanto o valor de ΔKth por causa da formação de óxidos na trajetória da trinca, que geram um efeito de fechamento sobre a mesma. Quanto à forma de onda, verificou-se que o carregamento Dwell diminui a taxa de propagação de trinca, diminuindo a inclinação das curvas log (da/dN) versus log (ΔK) na região de Paris, ao invés de deslocá-la paralelamente como ocorre com ligas de titânio. A mudança da direção de laminação de LT para TL aumenta a taxa de propagação de trinca por fadiga (PTF) tanto na região de threshold, quanto na região de Paris, onde a mudança de taxa é pequena.