L’instrumentation de plus en plus poussée des systèmes mécaniques (aéronautique, automobile,…) impose une utilisation croissante des connecteurs électriques. Cependant, leur environnement de fonctionnement (sollicitations chimiques et vibratoires) peut entrainer une dégradation plus ou moins sévère des contacts électriques limitant ainsi le passage du courant. Pour limiter cette dégradation et assurer la stabilité des connexions, des revêtements d’or sont couramment appliqués au niveau des contacts. Cependant, la conjecture économique et le coût très élevé de l’or nécessite de trouver une alternative moins chère. Parmi les métaux conducteurs, l’argent est aujourd’hui le meilleur candidat. L’objectif de cette thèse est donc d’étudier la réponse électrique et l’endommagement de dépôts argent soumis à des sollicitations de fretting. Pour cela, ces travaux de recherche ont été abordés selon trois axes. Le premier axe a permis une étude complète d’un contact homogène argent/argent afin d’identifier les mécanismes de dégradation responsables de la rupture électrique aussi bien en fretting qu’en glissement alterné. Il a aussi été possible, par une approche énergétique, de mettre en place un modèle prédictif permettant d’extrapoler les durées de vie du contact selon différents paramètres de chargement. Une étude complémentaire a également montré l’impact d’une atmosphère corrosive à base de soufre sur les contacts électriques en argent. Le second axe a permis, quant à lui, d’étudier le comportement tribologique et électrique de nouveaux matériaux à base d’argent développés dans le but de remplacer les dépôts dorés. L’analyse de ces contacts homogènes a permis de mettre en évidence les mécanismes de dégradation et les comportements mécaniques des contacts soumis à des environnements humides. Dans le dernier axe, une étude a été menée sur ces mêmes matériaux à base d’argent mais en configuration hétérogène contre un dépôt d’or afin d’identifier le comportement tribologique et électrique de ces contacts quand ils sont composés par des matériaux avec des propriétés similaires ou opposées. / Advanced instrumentation in mechanical systems (aeronautical, automobile etc…) goes hand in hand with an ever increased use of electrical connectors. However, the unfavorable operating environment (chemical attack and vibrational loads) causes more or less severe degradation of electrical contacts, which in turn perturbs their electrical conductivity. Gold plating is usually applied in electrical contacts in order to limit damage and to ensure connector stability. However, economic constraints and the high cost of gold require cheaper alternatives. Amongst conductive metals, silver is the best candidate. Hence, the purpose of this PhD project is to investigate the electrical response and the degradation of silver coatings when subjected to fretting loadings. The study is divided into three main research axes. The first axis consists in realizing a complete study of a homogeneous silver/silver contact in order to identify the degradation mechanisms that are responsible for the electrical failure, both in fretting loadings and reciprocating sliding. It was possible to formalize a predictive model, using an energy density approach, allowing to extrapolate the lifetime of the contact as a function of various loading parameters. A complementary study also showed the impact of a corrosive sulfur atmosphere on these electric contacts. As part of the second research axis, an investigation of the tribological and electrical behavior of novel silver-based materials, solely synthesized as a gold replacement, was performed. The analysis of these homogeneous contacts allowed to explain the degradation mechanism and the mechanical behavior of these contacts when subjected to a wet environment. In the last research axis a study was led on the same silver-based materials but in a heterogenous configuration against a gold coating in order to identify the tribological and electrical behavior of these contacts when composed by materials with similar or opposite properties.

Les systèmes d’imagerie médicale, principalement les systèmes à rayons X, sont devenus incontournables dans le diagnostic et le traitement des maladies complexes. Le générateur à rayons X fait partie des sous-systèmes critiques d’un système à rayons X. La technologie des générateurs à rayons X se complexifie et les contraintes vues par les composants augmentent. L’évaluation de la fiabilité du générateur à rayons X est par conséquent nécessaire afin d’optimiser la durée de vie de ce dernier. Dans ces travaux de thèse, une méthodologie d’évaluation de la fiabilité d’un générateur à rayons X est proposée. La méthodologie repose sur l’évaluation de la fiabilité allant du composant au système. Des essais de vieillissement sont d’abord réalisés au niveau des composants critiques du générateur afin d’identifier les mécanismes de défaillance et de construire les courbes de durée de vie permettant d’effectuer une prévision de fiabilité. Les paramètres du recueil de fiabilité FIDES ont aussi été utilisés pour construire les courbes de durée de vie des composants critiques. Une méthode de prévision de la fiabilité basée sur l’hypothèse du dommage cumulé avec la règle de Miner est proposée pour évaluer la durée de vie des composants critiques sous contraintes thermomécaniques. Cette méthode utilise les règles de comptage rainflow pour obtenir une distribution des différences de température vues par les composants critiques. Une association de fiabilité permet enfin d’estimer la durée de vie de chaque sous système du générateur à rayons X à travers ses composants critiques. / Medical imaging systems, mainly X-rays imaging systems, have become essential in the diagnosis and treatment of complex diseases. X-rays generator is one of the critical subsystems of a medical system. Its technology became more complex and constraints seen by the components increase. An assessment of X-rays generator reliability is therefore necessary to optimize its lifetime. In this thesis, a reliability assessment method of an X-rays generator is proposed. The methodology is based on the assessment of the reliability from component to system. Aging tests are first performed for X-rays generator critical components in order to identify failure mechanisms and build lifetime curves for performing reliability prediction. FIDES guide parameters were also used to construct critical components lifetime curves. A reliability prediction method based on the assumption of cumulative damage with Miner's rule is proposed to evaluate critical components lifetime under thermomechanical stresses. This method uses rainflow counting rules for the temperature cycles distribution of critical components. A reliability block diagram is finally used to estimate the lifetime of each X-ray generator subsystem through its critical components.

Developing numerical tools for polymer lifetime prediction constitutes a promising opportunity for shortening the duration of material certification procedures in the automotive industry without decreasing their reliability. This PhD thesis aims at modeling the photothermal oxidation of isotactic polypropylene (iPP), which is responsible for the alteration of both its mechanical and aspect properties. The adopted approach consists in coupling the kinetics of photo- and thermo-oxidation reactions with physical phenomena, such as oxygen transport and UV-light attenuation in the material thickness, in order to describe all the physico-chemical changes. Upper-scale properties, from which will be defined the end-of-life criteria, will be calculated afterwards by applying the suitable structure-property relationships. The main challenge was to extend the pre-existing kinetic model of thermal ageing to photothermal ageing by taking into account initiation reactions of photolysis. Heavy campaigns of ageing and characterization tests made on a reference iPP, as well as an exhaustive capitalization of literature data of other iPPs, have allowed elaborating a kinetic model of photothermal oxidation and to generalize it to the whole iPP family in large domains of oxygen partial pressure (from 0.2 to 50 bars), temperature (from 40 to 230°C) and UV-light exposure (variable intensities and light sources) describing both natural and accelerated ageing conditions. The experimental validation of the model has allowed substantiating the kinetic approach and showing its limitations, as well as highlighting some numerical issues. The model has been designed in order to be an upgradable numerical tool which will allow, at term, optimizing the representativeness of the ageing testing devices and the performance of commercial iPP formulations. All these theoretical and numerical developments are prone to be applied to the photothermal degradation of other types of polymer substrates, but also in other application fields of the macromolecular photochemistry such as UV-photopolymerization.Keywords: Polypropylene, photothermal oxidation, oxygen diffusion control, screen effect, kinetic modeling, lifetime prediction.

In this thesis it has been studied how operational and workshop data can be used to improve the handling of field quality (FQ) issues for electronic units. This was done by analysing how failure rates can be predicted, how failure mechanisms can be detected and how data-based lifetime models could be developed. The work has been done on an electronic control unit (ECU) that has been subject to a field quality (FQ) issue, determining thermomechanical stress on the solder joints of the BGAs (Ball Grid Array) on the PCBAs (Printed circuit board assembly) to be the main cause of failure. The project is divided into two parts. Part one, "PCBA" where a laboratory study on the effects of thermomechanical cycling on solder joints for different electrical components of the PCBAs are investigated. The second part, "ECU" is the main part of the project investigating data-driven solutions using operational and workshop history data. The results from part one show that the Weibull distribution commonly used to predict lifetimes of electrical components, work well to describe the laboratory results but also that non parametric methods such as kernel distribution can give good results. In part two when Weibull together with Gamma and Normal distributions were tested on the real ECU (electronic control unit) data, it is shown that none of them describe the data well. However, when random forest is used to develop data-based models most of the ECU lifetimes of a separate test dataset can be correctly predicted within a half a year margin. Further using random survival forest it was possible to produce a model with just 0.06 in (OOB) prediction error. This shows that machine learning methods could potentially be used in the purpose of condition based maintenance for ECUs.