<p>Accelerated life testing (ALT) is commonly used to obtain the reliability (lifetime) of a system in a short period by applying severe stress factors (i.e. temperature, humidity, chemical etc.,) to the system. Accelerated life testing was done on 2N3904 N-P-N bipolar junction transistor. After accelerated life testing, the acceleration factor is calculated through statistical analysis to predict the lifetime of a system at a certain operating condition. Multiple temperature and water immersed experiments were conducted on transistors at various temperatures. With the obtained results, analysis was done on the data to predict lifetime of transistors at different environmental conditions. Effects of the environmental conditions on the transistors were also discussed. Corrosion effects on transistors were studied. ALTA software is used to analyze the experimental data and to calculate the lifetime of the transistors. From the obtained failure data prediction of mean time to failure at various temperatures was done. </p>

Die bei Textilbetonbauteilen üblicherweise zum Einsatz kommenden Bewehrungen können sowohl aus Carbon oder AR-Glas bestehen. Aufgrund des Preises und der Verfügbarkeit hat AR-Glas gegenüber Carbon allerdings eine deutlich größere Verbreitung am Markt. Trotz der chemischen Modifikation zur Verbesserung der Alkalibeständigkeit von AR-Gläsern gegenüber E-Glas, kann bei AR-Gläsern ein nennenswerter Festigkeitsverlust infolge der Alkalität des Betons innerhalb der Lebensdauer eines Bauwerks festgestellt werden. Eine Möglichkeit den langfristigen Festigkeitsverlust zu reduzieren, ist der Aufbau einer Diffusionsbarriere um die einzelnen Rovings durch eine polymere Tränkung. Im Rahmen der vorliegenden Veröffentlichung werden mögliche Tränkungsmaterialien sowie die Auswirkungen dieser auf die Dauerhaftigkeit der textilen Bewehrungen anhand von Versuchen bewertet und ein Modell zur Prognose des langfristigen Festigkeitsverlustes polymermodifizierter Bewehrungen vorgestellt. / The reinforcement in TRC (Textile Reinforced Concrete) is usually made of alkali-resistant glass rovings due to their low price and availability on the market, but even the glass composition is changed compares to E-glass to increase its alkali resistance the durability in concrete is known to be unsatisfactory. One possibility to increase the durability is the application of a diffusion barrier around the reinforcement. The diffusion barrier can be made out of reactive polymeric coatings. Within the paper the results regarding the durability of AR-glass reinforcement achieved with a variety of reactive materials, mainly epoxy resins, will be presented as well as discussed. In addition to the results achieved in the laboratory, a model which allows long term predictions of the durability of polymer impregnated AR-glass reinforcements will be presented.

This research project was a step forward in developing an efficient data mining method for estimating the service life of metallic components in Queensland school buildings. The developed method links together the different data sources of service life information and builds the model for a real situation when the users have information on limited inputs only. A practical lifetime prediction system was developed for the industry partners of this project including Queensland Department of Public Works and Queensland Department of Main Roads. The system provides high accuracy in practice where not all inputs are available for querying to the system.

La prédiction des durées de vie des pièces de structures aéronautiques a toujours été une problématique à la fois complexe et capitale dans ce domaine de l’industrie. Néanmoins, la majorité des tests existant à l’heure actuelle cherche à évaluer la capacité des matériaux à résister aux sollicitations mécaniques notamment en fatigue. Les problématiques liées à l’endommagement causé par l’environnement comme la corrosion sont encore mal comprises. En effet, bien que certains tests permettent de détecter et de caractériser cet endommagement, aucun outil fiable de prédiction des vitesses de propagation des défauts de corrosion intergranulaire n’existe. Ainsi, actuellement, un défaut de corrosion détecté induit systématiquement un changement de la pièce. Les travaux de cette thèse s’inscrivent dans cette problématique ; ils ont pour but de comprendre les phénomènes de corrosion intergranulaire sur l’alliage d’aluminium 2024, le plus utilisé dans le secteur aéronautique, et d’étudier les cinétiques de propagation des défauts de corrosion. L’étude s’appuie sur une approche multi-échelle des processus de corrosion, des états microstructuraux et de l’influence de l’hydrogène. Ce projet s’inscrit également dans une dynamique de collaboration avec Airbus Group et l’Université de Bourgogne dans le cadre du projet ANR M-SCOT (Multi-Scale Corrosion Testing ANR-14-CE07-0027-01). / Nowadays, cracks kinetics is a key point in aircraft risk and reliability analysis. In particular, the propagation of corrosion defects is of special interest and could promote an early mechanical crack initiation. However, today most of the tests are calibrated to control mechanical damage and do not take into account the propagation of the corrosion defects. Indeed, when a corrosion defect is observed, the airplane part is automatically changed which leads to high manufacturing costs. The aim of this work is to understand the intergranular corrosion mechanisms and to study the propagation kinetics of the corrosion defects in an aeronautical reference alloy, i.e. the 2024-T351 aluminium alloy. A multiscale approach of the corrosion processes, the microstructural states as well as hydrogen influence was performed. This work is supported by ANR-14-CE07-0027-01 – M-SCOT: Multi Scale COrrosion Testing.

Environmental fluctuations in operating fuel cells impose significant biaxial stresses in the constrained proton exchange membranes (PEM). The PEM's ability to withstand cyclic environment-induced stresses plays an important role in membrane integrity and consequently, fuel cell durability. In this thesis, pressure loaded blister tests are used to study the mechanical durability of Gore-Select® series 57 over a range of times, temperatures, and loading histories. Ramped pressure tests are used with a linear viscoelastic analog to Hencky's classical solution for a pressurized circular membrane to estimate biaxial burst strength values. Biaxial strength master curves are constructed using traditional time-temperature superposition principle techniques and the associated temperature shift factors show good agreement when compared with shifts obtained from other modes of testing on the material. Investigating a more rigorous blister stress analysis becomes nontrivial due to the substantial deflections and thinning of the membrane. To further improve the analysis, the digital image correlation (DIC) technique is used to measure full-field displacements under ramped and constant pressure loading. The measured displacements are then used to validate the constitutive model and methods of the finite element analysis (FEA). With confidence in the FEA, stress histories of constant pressure tests are used to develop linear damage accumulation and residual strength based lifetime prediction models. Robust models, validated by successfully predicting fatigue failures, suggest the ability to predict failures under any given stress history whether mechanically or environmentally induced - a critical step in the effort to predict fuel cell failures caused by membrane mechanical failure. / Master of Science

Adhesive bonding is one of the most attractive joining techniques for any structural application, including high profile examples in the aerospace, automotive, marine construction and electrical industries. Advantages of adhesive bonding include; superior fatigue performance, better stress distribution and higher stiffness than conventional joining techniques. When the design of bonded joints is considered, fatigue is of critical importance in most structural applications. There are two main issues that are of importance; a) in-service damage characterisation during fatigue loading and, b) lifetime prediction under both constant and variable amplitude fatigue loading. If fatigue damage characterisation is considered, there has been some work to characterise damage in-situ using the backface strain (BFS) measurement technique, however, there has been little investigation of the effects of different types of fatigue behaviour under different types of geometry and loading. Regarding fatigue lifetime prediction of bonded joints, most of the work in the literature is concentrated with constant amplitude fatigue, rather than variable amplitude fatigue. Fatigue design of a bonded structure based on constant amplitude fatigue, when the actual loading on the structure is of the variable amplitude fatigue, can result in erroneous lifetime prediction. This is because of load interaction effects caused by changes in load ratio, mean load etc., which can decrease the fatigue life considerably. Therefore, the project aims to a) provide a comprehensive study of the use of BFS measurements to characterise fatigue damage, b) develop novel techniques for predicting lifetime under constant amplitude fatigue and c) provide an insight into various types of load interaction effects. In this project, single lap joints (SLJ) and compound double cantilever beam geometries were used. Compound double cantilever beams were used mainly to determine the critical strain energy release rate and to obtain the relationship between strain energy release rate and fatigue crack growth rate. The fatigue life of SLJs was found to be dominated by crack initiation at lower fatigue loads. At higher fatigue loads, fatigue life was found to consist of three phases; initiation, stable crack propagation and fast crack growth. Using these results, a novel damage progression model was developed, which can be used to predict the remaining life of a bonded structure. A non-linear strength wearout model (NLSWM) was also proposed, based on strength wearout experiments, where a normalised strength wearout curve was found to be independent of the fatigue load applied. In this model, an empirical parameter determined from a small number of experiments, can be used to determine the residual strength and remaining life of a bonded structure. A fracture mechanics approach based on the Paris law was also used to predict the fatigue lifetime under constant amplitude fatigue. This latter method was found to under-predict the fatigue life, especially at lower fatigue loads, which was attributed to the absence of a crack initiation phase in the fracture mechanics based approach. A damage mechanics based approach, in which a damage evolution law was proposed based on plastic strain, was found to predict the fatigue life well at both lower and higher fatigue loads. This model was able to predict both initiation and propagation phases. Based on the same model, a unified fatigue methodology (UFM) was proposed, which can be used to not only predict the fatigue lifetime, but also various other fatigue parameters such as BFS, strength wearout and stiffness wearout. The final part of the project investigated variable amplitude fatigue. In this case, fatigue lifetime was found to decrease, owing to damage and crack growth acceleration in various types of variable amplitude fatigue loading spectra. A number of different strength wearout approaches were proposed to predict fatigue lifetime under variable amplitude fatigue loading. The NLSWM, where no interaction effects were considered was found to over-predict the fatigue life, especially at lower fatigue loads. However, approaches such as the modified cycle mix and normalised cycle mix approaches were found to predict the fatigue life well at all loads and for all types of variable amplitude fatigue spectra. Progressive damage models were also applied to predict fatigue lifetime under variable amplitude fatigue loading. In this case a fracture mechanics based approach was found to under-predict the fatigue life for all types of spectra at lower loads, which was established to the absence of a crack initiation phase in this method. Whereas, a damage mechanics based approach was found to over-predict the fatigue lifetime for all the types of variable amplitude fatigue spectra, however the over- prediction remained mostly within the scatter of the experimental fatigue life data. It was concluded that, the damage mechanics based approach has potential for further modification and should be tested on different types of geometry and spectra.

Les polyamides sont des thermoplastiques techniques qui entrent dans la conception de pièces destinées à remplacer certains composants métalliques des moteurs automobiles. En dépit de propriétés mécaniques initiales satisfaisantes, leur tenue à long terme est limitée par leur sensibilité à l'oxygène conduisant à des réactions de thermo-oxydation. Ces réactions ont été largement étudiées dans le cas des polyoléfines mais peu dans le cas des polyamides, rendant nécessaire l'élaboration d'un modèle cinétique susceptible de prédire la fragilisation donc la durée de vie des polyamides. Cette thèse est une contribution à la compréhension du processus d'oxydation dans le cas des polyamides aliphatiques et à la construction d'un modèle cinétique. La démarche cinétique réside tout d'abord dans la caractérisation physico-chimique multi-échelle de films de PA11 oxydés dans différentes conditions de températures (90 à 165 °C sous air) et sous différentes pressions partielles d'oxygène. Un modèle cinétique couplant oxydation et post-polycondensation est proposé ici : il permet de simuler les données expérimentales (hydroperoxydes, carbonyles et masse molaires) quelles que soient les conditions d'exposition. Parallèlement, un critère intrinsèque gouvernant la fragilisation du PA11 est identifié afin de prédire cette dernière à partir du modèle cinétique. Enfin, l'influence de l'ajout d'antioxydants phénoliques et des sels de cuivre sur la cinétique d'oxydation est caractérisée. Un premier modèle cinétique prenant en compte la stabilisation du PA11 décrit les tendances spécifiques de la stabilisation du PA11 comme l'apparition de la pseudo-période d'induction contribuant à une augmentation significative de la durée de vie du PA11. / Some metal components of automotive engine are bound to be replaced by polyamide parts. However, despite their thermal resistance polyamides are sensitive to oxygen leading to thermal oxidation chain reactions responsible for their long-term properties. While durability is critical for polyamide users, only a few studies deal with the elaboration of a kinetic model capable of predicting polyamide lifetime (time to embrittlement) in contrary to polyolefins (especially polyethylene). This PhD thesis is a contribution to the understanding of aliphatic polyamide thermal degradation by considering chemical and physical aspects of oxidation process in order to build a kinetic model. Our approach is based on a multi-scale physicochemical characterization of oxidized PA11 film samples under air between 90 and 165 °C but also under oxygen pressure. The proposed kinetic model coupling oxidation and solid state polymerization is able to simulate the whole experimental data (hydroperoxides, carbonyls and molar mass changes). In a same time, an intrinsic criterion for embrittlement is assessed to predict lifetime whatever the exposure conditions. Finally, the influence of phenols and copper salts on the oxidation kinetic is investigated. A first kinetic model including the phenol stabilizing effect is capable of simulating the main observed trends for stabilized PA11 such as the appearance of the pseudo induction period which contributes to the significant improvement of PA11 durability.

Die bei Textilbetonbauteilen üblicherweise zum Einsatz kommenden Bewehrungen können sowohl aus Carbon oder AR-Glas bestehen. Aufgrund des Preises und der Verfügbarkeit hat AR-Glas gegenüber Carbon allerdings eine deutlich größere Verbreitung am Markt. Trotz der chemischen Modifikation zur Verbesserung der Alkalibeständigkeit von AR-Gläsern gegenüber E-Glas, kann bei AR-Gläsern ein nennenswerter Festigkeitsverlust infolge der Alkalität des Betons innerhalb der Lebensdauer eines Bauwerks festgestellt werden. Eine Möglichkeit den langfristigen Festigkeitsverlust zu reduzieren, ist der Aufbau einer Diffusionsbarriere um die einzelnen Rovings durch eine polymere Tränkung. Im Rahmen der vorliegenden Veröffentlichung werden mögliche Tränkungsmaterialien sowie die Auswirkungen dieser auf die Dauerhaftigkeit der textilen Bewehrungen anhand von Versuchen bewertet und ein Modell zur Prognose des langfristigen Festigkeitsverlustes polymermodifizierter Bewehrungen vorgestellt. / The reinforcement in TRC (Textile Reinforced Concrete) is usually made of alkali-resistant glass rovings due to their low price and availability on the market, but even the glass composition is changed compares to E-glass to increase its alkali resistance the durability in concrete is known to be unsatisfactory. One possibility to increase the durability is the application of a diffusion barrier around the reinforcement. The diffusion barrier can be made out of reactive polymeric coatings. Within the paper the results regarding the durability of AR-glass reinforcement achieved with a variety of reactive materials, mainly epoxy resins, will be presented as well as discussed. In addition to the results achieved in the laboratory, a model which allows long term predictions of the durability of polymer impregnated AR-glass reinforcements will be presented.

Le « Lithium-Ion Capacitor » (LiC) est un supercondensateur hybride dont les caractéristiques peuvent être placées entre un condensateur à double couche électrique (supercondensateur) et une batterie lithium-ion. Il possède des densités d’énergie et de puissance intermédiaires grâce à sa composition hybride à base d'une électrode positive en charbon actif identique à celle d’un supercondensateur et d'une électrode négative en carbone pré-lithié identique à celle d’une batterie lithium-ion. L'objectif de cette thèse est d'étudier le vieillissement des LiC industrialisés aussi bien dans le cadre d’un vieillissement en stockage (calendaire) qu’en utilisation (cyclage). Un de leur spécificité principale concerne l’évolution particulière de leur capacité en fonction de la tension à leurs bornes (C(V)). Le premier type de vieillissement qu’est le vieillissement calendaire permet de représenter le comportement des LiC lorsqu’ils sont stocker avant utilisation ou lorsqu’ils sont en veille. La dégradation de leurs paramètres liée au vieillissement, est alors essentiellement influencée par leur tension et la température. Des essais de vieillissement à trois tensions caractéristiques et deux températures différentes sont étudiés. L’évolution des impédances des cellules a été suivie tout au long du vieillissement afin d’identifier un modèle électrique de suivi du vieillissement dont les paramètres sont liés aux phénomènes électrochimiques. Par ces essais, la meilleure tension de stockage des LiC, permettant la prolongation de leur durée de vie a été mise en évidence. Par ailleurs, des mécanismes de vieillissement différents d’une tension caractéristique à l’autre sont révélés et soulignent la spécificité de fonctionnement des LiC. Ces résultats ont été confirmés par des analyses post-mortem. Le second type de vieillissement étudié est le vieillissement par cyclage qui prend en compte l'impact du courant sur la durée de vie des LiC. Le choix des profils de courant de cyclage a été effectué en considérant le principe de fonctionnement électrochimique des LiC. Les évolutions des impédances et des courbes C(V) des cellules sont comparées et interprétées. Les mécanismes de vieillissement prenant naissance lors du cyclage continu sont abordés. Ils dépendent de la fenêtre de potentiel sur laquelle les LiC fonctionnent pendant leur utilisation. La fenêtre de tension optimale qui assure une longue durée de vie des LiC est aussi mise en évidence / Lithium-Ion Capacitors (LiCs) are the new emerging technology of hybrid supercapacitors that combines the advantages of conventional supercapacitors and lithium-ion batteries. They provide intermediate energy and power densities due to their hybrid composition based on a positive electrode made of activated carbon similar to that of supercapacitors and a negative electrode made of pre-lithiated carbon similar to that of lithium-ion batteries. The aim of this thesis is to study the aging of commercial LiCs using two accelerated aging procedures: calendar aging and cycle aging. One of their main particularities concerns the nonlinear capacitance evolution with respect to their voltage (C(V) curve). The first accelerated aging test is related to the calendar life of LiCs that represents their behavior independently of their usage. The degradation of their parameters due to aging is mainly affected by the voltage and the temperature only. These tests were applied to several cells at three different voltage values and two temperatures. The evolution of their impedances were followed during the whole aging period in order to identify an electrical model that can accurately describe the progress of aging and that possesses electrochemically meaningful parameters. The best voltage value that ensures the extension of the lifetime of LiCs was identified using the results of these tests. In addition, aging mechanisms that extremely depend on the applied voltage value were identified. They highlight the particularity of the functioning of LiCs. These results were confirmed using post-mortem analyses. The second accelerated aging test is the cycle aging that assesses the impact of the current on the life cycle of LiCs. The choice of current profiles was based on the electrochemical operating principle of LiCs. The evolution of the impedances and the C(V) curves of LiCs were compared and analyzed. Aging mechanisms produced during cycle aging were also evaluated. They depend on the voltage range in which the LiC operates. The optimal voltage window that guarantees a long lifetime of LiCs was highlighted

Diese Arbeit untersucht die Möglichkeiten, ortsaufgelöste (2D-) Lumineszenz als eine nicht-invasive Methode zur Charakterisierung des Ethylen-Vinyl-Acetat (EVA) Einkapselungsmaterials in Photovoltaik (PV) Modulen einzusetzen. Im Einzelnen wird gezeigt, dass aus nichtinvasiven Lumineszenzmessungen Aussagen erhalten werden können zu: (i) Auswirkungen der Alterungsbedingungen, Klima und Material, (ii) Nachweis bestimmter Fehlerarten, (iii) Überwachung des Produktionsprozess, (iv) Verbesserung von Lebensdauervorhersagen durch miniaturisierte Multi-Faktor-Alterung und (v) Aussagen über die Materialalterung. (i) Temperatur, UV-Strahlung und Sauerstoff werden als Hauptfaktoren identifiziert, die die Lumineszenz beeinflussen. Beschleunigte Alterung unter UV und Freibewitterung resultieren in ähnlichen Lumineszenzmustern und -Spektren, die sich von denen nach exklusiver thermischer Alterung unterscheiden. (ii) Insbesondere können Risse in Zellen nachgewiesen werden, darüber hinaus gibt es Anhaltspunkte, dass auch Korrosion und Delamination durch Lumineszenz nachweisbar sein könnten. (iii) Lumineszenz ist eine vielversprechende Methode, zerstörungsfrei den Vernetzungsgrad und die Homogenität zu bewerten. (iv) Für wirtschaftliche und effiziente Alterungstests wurde eine Multi-Faktor-Alterungs-Kammer entwickelt und getestet, die die Prüfung der gleichzeitigen Wirkung mehrerer Alterungsfaktoren ermöglicht. (v) Erstmals wird eine Korrelation der EVA-Lumineszenz zu anderen, insbesondere viskoelastischen Materialeigenschaften gezeigt. Es wird geschlossen, dass diffusionsbegrenzte Oxidation (DLO) im EVA zwischen Zelle und Glas eine inhomogene Materialdegradation verursacht. Perspektivisch werden Studien des komplexen Einflusses von Temperatur, UV-und Sauerstoff auf Materialalterung durch 2D-Lumineszenz in Kombination mit Multi-Faktor-Alterung aufgezeigt, ebenso wie die Möglichkeit einer Anwendung zur Bestimmung von apparenten Reaktionsraten aus Lumineszenzintensitätsprofilen. / This work examines the prospects of employing spatially resolved (2D-) luminescence as a non-invasive method for the characterization of the ethylene-vinyl-acetate (EVA) encapsulation material in photovoltaic (PV) modules. It is shown that information can be obtained related to the topics: (i) impact of aging conditions, climate and material, (ii) detection of specific types of failure modes, (iii) use to monitor the production process, (iv) improving service life prediction by miniaturized multi-factor aging and (v) statements about the material aging. (i) Temperature, UV radiation and oxygen are identified as main factors, which influence the luminescence. Accelerated aging under UV and outdoor weathering result in similar luminescence patterns and spectra, different from those after exclusive thermal aging. (ii) Specific failure modes, in particular cracks of the cells, can be detected. In addition, indications exist that corrosion and delamination might be detectable by luminescence. (iii) Luminescence is a promising method that allows non-destructive examination of the degree of crosslinking and the homogeneity in the PV module. (iv) For more economic and efficient aging tests a multi-factor-aging chamber was designed and tested which allows testing the simultaneous action of several aging factors. (v) For the first time, a correlation of EVA-luminescence to others, in particular viscoelastic, material properties, is demonstrated. From the inhomogeneous degradation patterns, it is concluded that diffusion limited oxidation (DLO) in the EVA between glass cell and causes an inhomogeneous material degradation. Interesting prospects emerge, e.g. studies of the complex influence of temperature, UV and oxygen on material aging by 2D-Luminescence using multi-factor-aging. In addition, the evaluation of intensity profiles, for example, regarding apparent reaction rates, is one of many promising applications.