Les composites 3D C/C sont utilisés dans l’industrie aérospatiale ou nucléaire pour leurstrès bonnes propriétés mécaniques à haute température. Afin d’assurer une intégrité optimaledes structures, la connaissance du comportement mécanique du composite, et plus particulièrementde ses mécanismes de rupture, est essentielle.Dans ce but, ce travail présente une modélisation de l’endommagement d’un composite3D C/C à température ambiante. Pour cela, une approche de modélisation à l’échelle mésoscopiquea été adoptée. A cette échelle, le composite 3D C/C présente deux types deconstituants : les baguettes de fibres de carbone et la matrice carbonée. Le comportement dechacun de ces méso-constituants est modélisé par une loi de comportement élastique endommageable(isotrope pour la matrice, orthotrope pour les baguettes) nécessitant un nombrede paramètres restreint. L’identification de ces paramètres repose sur des données expérimentalestirées de travaux antérieurs ainsi que celles issues d’une campagne expérimentale,menée durant la thèse, visant à compléter la connaissance du comportement mécanique ducomposite 3D C/C à l’échelle mésoscopique. Par ailleurs, des essais macroscopiques ont étéréalisés afin de valider le modèle développé. Les réponses expérimentales d’essais de flexion4 points et de flexion 3 points sont notamment bien reproduites par le méso-modèle. / 3D C/C composites are commonly employed in aerospace industry due to their outstandingmechanical properties at high temperatures. In order to ensure the integrity of structures,knowledge of the composite mechanical behaviour and fracture mechanisms is crucial.For this purpose, damage modeling of a 3D C/C composite, at room temperature, isproposed in which a meso-scale approach is considered. At this description scale, 3D C/Ccomposites are made of two materials : carbon fibers yarns and carbon matrix. Each materialbehavior is modeled by an elastic damage law (isotropic for matrix, orthotropic for yarns)with a limited number of parameters.The parameters identification process is based on experimentaldata obtained from previous work and from an experimental campaign carried outthrough this thesis work. This campaign aimed to a greater understanding of the materialmechanical behavior at mesoscopic scale. Furthermore, experimental tests were carried outto validate the composite modeling. It is shown that experimental reponses obtained fromfour-point and three-point bending tests are particularly well described from the proposedmesoscopic model.

AN ABSTRACT OF THE DISSERTATION OF TITLE: (OPTIMIZATION PROTECTION OF CARBOB-CARBON COMPOSITES DISC-BRAKES MATERIAL BY COATING WITH ANTI-OXIDANTS) Developing glass enhancer mixture solutions (Ki’s), which promote the formation of a stable glass layer, homogenous clear liquid solution, and low viscosity liquid form, are easy to apply, and penetrating. They are compatible with ceramic liquid glass based anti-oxidants for treating surfaces of carbon/carbon composites material, and significantly increase the rate of protection against oxidation. Ki’s’ are comprised of mixing chemical compositions at standard temperature and pressure conditions from group one and two such as Na, K, Ca, Mg, etc. of 5 to 25 wt. %, deionized water from 95 to 75 % by weight, and adding up to 1 % by weight of surfactants such as DF-16, DF-20, and CF-10 with specific proportions, and followed by thorough stirring to produce a homogeneous blend of mixture solution. The glass enhancers, which are aqueous mixture solutions, are applied to the surfaces of carbon/carbon (C/C) composites by dipping, brushing, spraying, or other painting application techniques, followed by annealing, or a heat-treating range of 80 to 110 ℃ for a minimum of 8 hours, and allowing cooling time of the coated C/C composites of a minimum of 12 hours to room temperature. Preferential compatibility of the glass enhancer mixture solutions (Ki's) is with liquid glass former's, anti-oxidants comprised mostly of borate and phosphate glasses. The glass enhancer solution mixtures (Ki’s) are supplemental additions to ceramics’ liquid anti-oxidants coatings used for carbon-carbon composites protection against oxidation, and it will increase the rate of protection against oxidation for low, and moderate temperature’s range from 400 to 900 ℃. The glass enhancer Ki’s mixture solutions should be used with liquid glass former's’ anti-oxidants, such as SiO₂, GeO₂, B₂O₃, and P₂O₅. A series of glass enhancer’s Ki’s, heat treatment cycle (char-cycle) ranged between 700 to 900 ℃, and application methods, were developed and tested experimentally. Two arbitrary isothermal temperatures of 650 ℃, and 871 ℃ were selected for thermal oxidation testing, and a temperature of 650 ℃ was selected, and tested against catalytic thermal oxidation. Additions of glass enhancer Ki’s improved protection of C/C composites disc-brakes against oxidation by double, and triple amount of time in hours versus the use of anti-oxidant coatings alone.

This study develops a reliable intelligent non-destructive evaluation (NDE) expert system for carbon-carbon (C/C) composites based on thermography, ultrasonic, computed tomography and post processing by means of fuzzy expert system technique. Data features and NDE expert knowledge are seamlessly combined in the intelligent system to provide the best possible diagnosis of the potential defects and problems. As a result, this research help ensure C/C composites' integrity and reliability. Four types of orthotropic aerospace composite material groups, which include 2-D pitched based commercial aircraft disc brakes and asmolds, 3-D PAN based C/C composites, and carbon fiber reinforced plastic (CFRP) panels, were tested. Based on the performance testing results of thermography, air-coupled ultrasonic, and x-ray computed tomography, the testing data pattern corresponding to feature and quantification of defects were found. This NDE knowledge databases were transformed to fuzzy logic expert system models. The models succeefully classified and indicated the defect's size and distribution and the intelligent systems perform NDE better than human operators. These fuzzy expert systems not only eliminate human errors in defect detection but also function as NDE experts. In addition, fuzzy expert systems improve the defect detection by incorporating fuzzy expert rules to remove noises and to measure defect size more accurately. In the future, the expert system model could be continuously updated and modified to quantify the size and distribution of defects. The systems developed here can be adapted and applied to build an intelligent NDE expert system for better quality control as well as automatic defect and porosity detection in C/C composite production process.

Le présent travail est réalisé dans le cadre de l’étude d’un procédé industriel de densification de matériaux composites carbone/carbone (C/C) destinés aux freins d’avion. Une préforme poreuse de fibres de carbone baigne dans un précurseur liquide et elle est chauffée par induction électromagnétique radio-fréquence. Le précurseur porté à ébullition dans l’espace poral crée un dépôt de carbone dans les zones les plus chaudes ; ce dépôt constitue la matrice du composite. On propose une modélisation physico-chimique de ce procédé afin d’en assurer le contrôle et l’optimisation. Le travail a consisté à développer un solveur couplant l’induction électromagnétique avec les transferts de masse, de chaleur, de mouvement et d’espèces chimiques, en incluant l’ébullition et le dépôt chimique. Le modèle inclut le circuit électrique complet permettant d’effectuer le chauffage : il permet donc de suivre en temps réel et de façon non destructive l’avancement de la densification par l’évolution des grandeurs électriques. Une formulation originale adaptée à la représentation simultanée du liquide, du gaz et de la zone en ébullition a été développée et implémentée avec succès dans un logiciel commercial d’éléments finis. Les résultats de la simulation sont comparés avec les données obtenues sur le moyen expérimental, avec un bon accord. Enfin, la simulation est utilisée pour proposer des pistes d’amélioration du procédé, en altérant la géométrie du dispositif de chauffage par induction et en modifiant la stratégie de pilotage en puissance. / This work has been carried out in the frame of the study of an industrial process for the manufacturing of carbon/carbon (C/C) composite aircraft brake discs. A porous preform made of carbon fibres is immersed in a liquid precursor and is heated by Radio-Frequency electromagnetic induction. The boiling precursor enters the porous preform and yields a carbon deposit in the hottest zones; this deposit will be the carbon matrix of the composite. A physico-chemical process model is proposed in the aim of ensuring its control and optimisation. The work consisted in developing a numerical solver coupling electromagnetic induction heating with heat, mass and species balances accounting for boiling, diffusion and chemical deposition reactions. The model includes the complete electrical circuit of the heating device: it therefore allows real-time, non-destructive monitoring of the infiltration progress through the evolution of the electrical properties. An original formulation has been designed to simultaneously describe the liquid, the gas and the boiling zone; it has been implemented in a commercial Finite Element software package and validated physically with respect to experimental data, with a good agreement. Finally, the simulation software has been used to propose directions for process improvements, through alterations of the inductive heating device geometry or of the heating power supply program.