The development of improved numerical methods for computer simulation of high velocity impact dynamics is of importance in a variety of science and engineering fields. The growth of computing capabilities has created a demand for improved parallel algorithms for high velocity impact modeling. In addition, there are selected impact applications where experimentation is very costly, or even impossible (e.g. when certain bioimpact or space debris problems are of interest). This dissertation extends significantly the class of problems where particle-element based impact simulation techniques may be effectively applied in engineering design. This dissertation develops a hybrid particle-finite element method for a general hexahedral mesh. This work included the formulation of a numerical algorithm for the generation of an ellipsoidal particle set for an unstructured hex mesh, and a new interpolation kernel for the density. The discrete model is constructed using thermomechanical Lagrange equations. The formulation is validated via simulation of published impact experiments. / text

For several years, research work has been carried out at the University of Canterbury aimed at the development of a rapid prototyping and manufacturing process referred to as Robotic Foam Sculpting (RFS). This system uses a six-axis industrial robot and electrically-heated hot-wire and hot-blade tools to sculpt desired parts from blocks of polystyrene foam. The vision for this system is that it will be able to rapidly create large volume foam models at low cost, for a range of potential applications. Parts produced by the RFS system can potentially be used as investment casting patterns, cores for sculptures and architectural details, demonstration and testing models, wind tunnel test models, and many other potential applications. At the beginning of the work reported in this thesis, there was very little understanding of the nature of the surfaces produced by hot-tool cutting of foam, very little knowledge of the range of input cutting conditions that affected the surface quality, and almost no understanding of the relationships between the cutting strategy and the nature of the surfaces being produced. In addition, there was little evidence of published work on these subjects that was sufficiently robust to be applicable to the RFS system. This research was concerned with rectifying this gap in the existing knowledge. There were a number of different focal areas for this research. These included the surface texture of surfaces cut with hot tools, the effects of cutting strategy on the surface quality in single-pass cutting of foam, the effects of cutting strategy on the surface quality in multi-pass cutting, and the application of a current-control system to control the surface quality in real time during a cut. In each of the focal areas the goal was to develop a detailed understanding of the nature of the different aspects of surface quality, to map the factor interactions and dependencies that controlled these aspects of surface quality, to develop methods for predicting the expected surface quality based on cutting strategy (and vice versa) and to develop techniques for minimising the surface errors. The detailed investigation of the surface texture of surfaces produced with hot-tool cutting is presented in Chapter 4. This chapter explores the characteristic nature of foam surfaces, presents the development of a method of measuring the surface texture of foam, and investigates the usefulness of a range of standard texture parameters for assessing foam surface quality. It is concluded in this chapter that common texture parameters based on the relative heights of surface features are not capable of reliably discriminating between different foam surfaces, so a new texture parameter (the 10%-Height Contiguous Diameter) is developed and implemented. Using this parameter, it is possible to reliably predict the surface texture to be expected for a given set of cutting conditions. Investigations of the cutting strategy in single-pass cutting are presented in Chapter 5. This chapter identifies the two key aspects of surface quality in single-pass cutting, the kerfwidth and the surface barrelling. Experimental work is carried out to investigate the relationships between these errors and the cutting strategy, and the factors that influence each of them are identified. In addition, statistical models are developed for the kerf along the length of a cut so that the kerf can be predicted based on cutting conditions. This chapter also includes a study of the cutting force in single-pass cutting, and develops models that allow the prediction of the expected cutting force for a given cutting strategy. A detailed study of the cutting strategy for multi-pass cutting is presented in Chapter 6. This study identifies the most significant surface errors in multi-pass cutting and determines the causes of each of these errors and the factor interactions and dependencies that have to be considered when developing a multi-pass cutting strategy. Once again, statistical models that allow the prediction of these surface errors based on cutting strategy, or the evaluation of cutting strategy parameters to achieve a desired surface quality, are developed. The models for cutting force in single-pass cutting are applied to multi-pass cutting, and it is found that these models can accurately predict the force in multi-pass cutting as well. The characterisation of the acoustic output in hot-tool cutting forms the subject matter of Chapter 7. This study establishes that the magnitude of the acoustic output is proportional to the cutting force experienced during the cut, and is therefore potentially suitable for use as a trigger signal for feedback current control. This would allow an acoustic signal to be used instead of the current force signal, which has a number of drawbacks that will be discussed in Chapter 2, the Background Material chapter. The specific trigger signal identified as being of most use is the acoustic output in the 4 – 12 kHz band, where the presence of any non-zero acoustic output above background noise is a reliable and repeatable indicator of the presence of thermomechanical cutting. The work presented in this thesis provides a detailed, quantitative, evidence-based and reliable understanding of the nature of the cutting strategy in hot-tool cutting of foam. The key cutting strategy parameters and the important aspects of surface quality for different cutting types are identified, the relationships between all these parameters are mapped, and quantitative models are developed that allow the output metrics like the surface quality or the cutting force to be predicted with a high degree of accuracy based on the input cutting strategy conditions. Armed with this understanding, it is possible to determine the most suitable cutting strategy for sculpting a given part, and to assess whether a given part can be sculpted with the RFS system. As such, the research problem posed at the start of this thesis has been largely solved, and the stage is set for further research to optimise the cutting strategy for sculpting different parts and to correct the remaining drawbacks of the RFS system to complete the development of a commercially-useful manufacturing system.

Le laminage est un procédé de mise en forme à chaud permettant d’obtenir des barres de différents diamètres en partant de blooms issus de la coulée continue. Dans les bruts de coulée se répartissent des inclusions qui sont de natures, de formes et de tailles différentes. Le laminage va permettre de réduire le diamètre de la barre et d’agir sur la microstructure du matériau, notamment, en fragmentant et dispersant les inclusions. L’objectif des travaux de thèse est de contribuer à la compréhension des effets du chargement thermomécanique sur l’évolution des défauts de coulée en laminage. Vu la dimension des installations, les investigations expérimentales directes sur les moyens industriels ne sont pas envisageables. L’idée développée dans les travaux de thèse concerne la mise en place et la validation d’un essai de caractérisation à échelle réduite reproduisant le chemin thermomécanique subi par la matière au cours du laminage. Pour ce faire la ligne de laminage industrielle est modélisé afin d’obtenir le chargement thermomécanique de la barre au cours du laminage. Seules les sollicitations ayant un rôle majeur sur le comportement des défauts sont reproduites. Un essai dit de « forgeage libre » a ainsi été conçu, réalisé, mis en place sur les moyens de mise en forme de la plateforme VULCAIN de l’ENSAM. L’essai a été utilisé avec des défauts artificiels présentant des similarités comme la malléabilité avec les défauts réels. Une campagne expérimentale paramétrique a été menée sur les différents types de défauts. Les échantillons forgés ont été analysés par des méthodes non destructives comme les ultrasons, la radiographie et la tomographie X. Ces analyses ont permis de suivre le comportement du défaut et d’observer certains phénomènes mis en évidence dans la littérature comme l’apparition de cavité à l’interface défaut-matrice. / Rolling is a hot forming process dedicated to manufacture bars with different diameters. The initial product is a bloom from the continuous casting. The bloom contains inclusions which have different forms, sizes and distributions. The rolling reduces the diameter of the bar and acts on the material microstructure by fragmenting and dispersing the inclusions. The aim of the thesis work is to understand the behavior of the inclusions with the rolling thermo mechanical loading effects. The direct investigations on the rolling mill are not possible because of the size of the installations. To address this problem, the solution is to develop and validate a small scale characterization test reproducing the thermo mechanical loading of the rolled bar. To achieve this, the rolling mill is modeled. Only the solicitations which have a major role on the behavior of the defects are reproduced. An open-die forging test is designed, manufactured and implemented on the VULCAIN installation of the ENSAM. The artificial defects which have a similar malleability to the real defects are integrated into the sample. A parametric experimental campaign has been conducted on the different defects. The forged samples have been analyzed with non-destructive methods such as ultrasound, radiography and tomography. These analyses allowed to follow the behavior of the defects and to observe certain phenomena illustrated in the literature such as the emergence of a cavity on the defect-matrix interface.

M. Tech. (Mechanical Engineering, Faculty of Engineering and Technology): Vaal University of Technology / Presently there exist a lot of controversies about the mechanical properties of nanomaterials. Several convincing reasons and justifications have been put forward for the controversies. Some of the reasons are varying processing routes, varying ways of defining equations, varying grain sizes, varying internal constituent structures, varying techniques of imposing strain on the specimen etc. It is therefore necessary for scientists, engineers and technologists to come up with a clearer way of defining and dealing with nanomaterials’ mechanical properties. The parameters of the internal constituent structures of nanomaterials are random in nature with random spatial patterns. So they can best be studied using random processes, specifically as stochastic processes. In this dissertation the tools of stochastic processes have been used as they offer a better approach to understand and analyse random processes. This research adopts the approach of ascertaining the correct mathematical models to be used for experimentation and modelling. After a thorough literature survey it was observed that size and temperature are two important parameters that must be considered in selecting the relevant mathematical definitions for nanomaterials’ mechanical properties. Temperature has a vital role to play during grain refinement since all severe plastic deformation involves thermomechanical processes. The second task performed in this research is to develop the mathematical formulations based on the experimental observation of 2-D grains and 3-D grains deformed by Accumulative Roll-Bonding and Equal Channel Angular Pressing. The experimental observations revealed that grains deformed by Accumulative Roll-Bonding and Equal Channel Angular Pressing are elongated when observed from the rolling direction, and transverse direction, and equiaxed when observed from the normal direction. In this dissertation, the different experimental observations for the grain size variants during grain refinement were established for 2-D and 3-D grains. This led to the development of a stochastic model of grain-elongation for 2-D and 3-D grains. The third task was experimentations and validation of proposed models. Accumulative Roll-Bonding, Equal Channel Angular Pressing and mechanical testing (tensile test) experiments were performed. The effect of size on elongation and material properties were studied to validate the developed models since size has a major effect on material’s properties. The fourth task was obtaining results and discussion of theoretical developed models and experimental results. The following facts were experimentally observed and also revealed by the models. Different approaches of measuring grain size reveal different strains that cannot be directly obtained from plots of the corresponding grain sizes. Grain elongation evolved as small values for larger grains, but became larger for smaller grains. Material properties increased with elongation reaching a maximum and started decreasing as is evident in the Hall-Petch to the Reverse Hall-Petch Relationship. This was alluded to the fact that extreme plastic straining led to distorted structures where grain boundaries and curvatures were in “non-equilibrium” states. Overall, this dissertation contributed new knowledge to the body of knowledge of nanomaterials’ mechanical properties in a number of ways. The major contributions to the body of knowledge by his study can be summarized as follows: (1) The study has contributed in developing a model of elongation for 2-D grain and 3-D grains. It has been generally reported by researchers that materials deformed by Accumulative Roll-Bonding and Equal Channel Angular Pressing are generally elongated but none of these researchers have developed a model of elongation. Elongation revealed more information about “size” during grain refinement. (2) The Transmission Electron Microscopy revealed the grain shape in three directions. The rolling direction or sliding direction, the normal direction and the transverse direction. Most developed models ignored the different approaches of measuring nanomaterials’ mechanical properties. Most existing models dealt only with the equivalent radius measurement during grain refinement. In this dissertation, the different approaches of measuring nanomaterials’ mechanical properties have been considered in the developed models. From this dissertation an accurate correlation can be made from microscopy results and theoretical results. (3) This research has shown that most of the published results on nanomaterials’ mechanical properties may be correct although controversies exist when comparing the different results. This research has also shown that researchers might have considered different approaches to measure nanomaterials’ mechanical properties. The reason for different results is due to different approaches of measuring nanomaterials’ mechanical properties as revealed in this research. Since different approaches of measuring nanomaterials’ mechanical properties led to different obtained results, this justify that most published results of nanomaterials’ mechanical properties may be correct. This dissertation revealed more properties of nanomaterials that are ignored by the models that considered only the equivalent length. (4) This research has contributed to the understanding of nanomaterials controversies when comparing results from different researchers.

Cette thèse, qui s’intègre dans le cadre du projet européen EMCOMIT, a pour objectif de contribuer à l’étude des circuits imprimés haute densité d’intégration comportant un nombre de couches important et des composants enterrés. La qualification de cette technologie est effectuée en conduisant des simulations et des mesures électriques sur des véhicules de tests spécifiques. L’analyse des résultats électriques permet d’évaluer l’aptitude de ces matériaux à répondre aux exigences des applications de télécommunication et de technologie de l’information rapide. La fiabilité d’un assemblage de BGA de grande taille sur un circuit imprimé a été évaluée. Des simulations thermomécaniques ont été effectuées afin de calculer les contraintes résiduelles accumulées pendant le procédé d’assemblage puis l’énergie dépensée dans les parties critiques des joints au cours d’un cycle thermique. Simultanément, des BGA reportés sur des circuits imprimés ont été placés dans une chambre climatique et ont subi des variations de températures. / This thesis, which is part of the European EMCOMIT project, aims at contributing to the study of high density printed circuit board including a great number of internal layers and embedded components. The qualification of this technology is done by the way of simulations and electrical measurements on specific test vehicles. The electrical results allow estimating the performance of materials for telecommunication applications and speed data transfer. The reliability of the assembly of the large BGA on a printed circuit board has been evaluated. Thermomechanical simulations have been done in order to compute residual stresses stored during the assembly process and the deformation energy density in the solder joints during one thermal cycle. Simultaneously BGA soldered on printed circuits have been positioned in climatic chamber and have been subjected to temperature variations.

This thesis deals with processing of nanocellulose originating from pulps, with focus on mechanical pulp fibres and fines fractions. The nanocellulose materials produced within this research project were tested for different purposes ranging from strength additives in paper and paperboard products, via composite films to foam materials. TAPPI (Technical Association of Pulp &amp; Paper Industry) has recently suggested a standard terminology and nomenclature for nanocellulose materials (see paper I). In spite of that we have decided to use the terms nano-ligno-cellulose (NLC), microfibrillated cellulose (MFC), nanofibrillated cellulose (NFC) and nanocellulose (NC) in this thesis . It is well-known that mainly chemical pulps are used as starting material in nanocellulose production. However, chemical pulps as bleached sulphite and bleached kraft are quite expensive. One more cost-effective alternative can be to use fibres or fines fractions from thermo-mechanical pulp (TMP) and chemi-thermomechanical pulp (CTMP).   In paper II-IV, fractionation has been used to obtain fines fractions that can easily be mechanically treated using homogenisation. The idea with this study was to investigate the possibility to use fractions of low quality materials from fines fractions for the production of nanocellulose. The integration of a nanocellulose unit process in a high-yield pulping production line has a potential to become a future way to improve the quality level of traditional products such as paper and paperboard grades.   Paper III describes how to utilise the crill measurement technique as a tool for qualitative estimation of the amount of micro- and nano-material produced in a certain process. The crill values of TMP- and CTMP-based nanocelluloses were measured as a function of the homogenisation time. Results showed that the crill values of both TMP-NLC and CTMP-NLC correlated with the homogenisation time. In Paper V pretreating methods, hydrogen peroxide and TEMPO are evaluated. Crill measurement showed that hydrogen peroxide pretreatment (1% and 4%) and mechanical treatment time did not improve fibrillation efficiency as much as expected. However, for TEMPO-oxidised nanocelluloses, the crill value significantly increased with both the TEMPO chemical treatment and mechanical treatment time. In paper V-VII TEMPO-mediated oxidation systems (TEMPO/NaBr/NaClO) are applied to these fibres (CTMP and Sulphite pulp) in order to swell them so that it becomes easy to disrupt the fibres into nanofibres with mechanical treatment.   The demand for paperboard and other packaging materials are steadily increasing. Paper strength properties are crucial when the paperboard is to withstand high load. A solution that are investigated in papers IV and VI, is to use MFC as an alternative paper strength additive in papermaking. However, if one wish to target extremely higher strength improvement results, particularly for packaging paperboards, then it would be fair to use MFC or cationic starch (CS). In paper VI CS or TEMPO-based MFC was used to improve the strength properties of CTMP-based paperboard products. Results here indicate significant strength improvement with the use of different levels of CS (i.e., 20 and 10 kg t–1) and 5% MFC. The strengthening impact of 5% MFC was approximately equal to that of 10 kg t–1 of CS.   In paper VII, NFC and nanographite (NG) was used when producing composite films with enhanced sheet-resistance and mechanical properties. The films produced being quite stable, flexible, and bendable. Realising this concept of NFC-NG composite film would create new possibilities for technological advancement in the area of high-yield pulp technology.  Finally, in paper VIII, a new processing method for nanocellulose is introduced  where an organic acid (i.e., formic acid) is used. This eco-friendly approach has shown to be successful, a nanocellulose with a uniform size distribution has been produced. / <p>Vid tidpunkten för disputationen var följande delarbeten opublicerade: delarbete 5 och 7 inskickade, delarbete 6 och 8 manuskript.</p><p>At the time of the doctoral defence the following papers were unpublished: paper 5 and 7 submitted, paper 6 and 8 manuscripts.</p>

La relaxation des contraintes résiduelles de trempe au cours du revenu dépend fortement des évolutions des microstructures et du comportement thermomécanique. Ainsi, la prévision de l’évolution des contraintes résiduelles nécessite le couplage entre les calculs des champs de température, des transformations de phases, des contraintes et des déformations au cours du traitement. Dans ce travail, nous avons développé un modèle métallurgique pour décrire les cinétiques de précipitation au cour du revenu : la germination, la croissance et la coalescence (ou dissolution) du carbure epsilon et de la cémentite. Ce modèle permet de calculer les évolutions de la composition chimique de la matrice, du type, des distributions de taille des précipités lors du revenu de la martensite d’aciers faiblement alliés. Dans une deuxième étape, nous avons développé un modèle thermomécanique permettant de prédire l’évolution de la contrainte d’écoulement de la martensite revenue, basé sur les évolutions des paramètres microstructuraux obtenus par le modèle métallurgique. Le modèle thermomécanique prend en compte les différents mécanismes de durcissement (durcissement par solution solide, durcissement par les précipités et durcissement par dislocations), et l’effet de la température de déformation. Une validation expérimentale de ces modèles a été menée dans le cas de l’acier 80MnCr5 en s’appuyant sur des résultas obtenus par microscopie électronique à transmission et dilatométrie thermomécanique. Les modèles métallurgique et thermomécanique ont finalement été introduits dans le logiciel de calcul par éléments finis ZeBuLoN pour simuler l’évolution des contraintes internes lors de la trempe et du revenu. Une première analyse de la relaxation des contraintes au cours du revenu a été effectuée. / The relaxation of the residual stress during the tempering process is strongly related to the evolutions of the microstructure and thermomechanical properties. Thus, the simulation of the evolution of residual stresses requests the coupling of the temperature field, the phase transformations and the stress and strain evolutions all along the process. In this work, firstly, a metallurgical model has been developed to describe the tempering kinetics: nucleation, growth (or dissolution) and coarsening of the epsilon carbide and the cementite. This model allows to predict the evolutions of the matrix composition, the type, size and distribution of the precipitates during the tempering of martensite for low and middle alloyed steels. Secondly, based on the microstructure parameters calculated by the metallurgical model, a thermomechanical model has been also developed to predict the evolution of the flow stress of tempered martensite. The thermomechanical model not only takes into account the different hardening mechanisms (solution hardening, precipitate hardening and the dislocation hardening), but also the effect of deformation temperature. An experimental validation of the two models has been performed for a 80MnCr5 steel using the results obtained by electron transmission microscopy and thermomechanical dilatometry. Finally, the metallurgical model and the thermomechanical model have been implemented into the finite element calculation software ZeBuLoN to simulate the evolution of internal stress during the quenching and tempering process. A first analysis of the stress relaxation during tempering has been performed.

Avec l’apparition de moteur électrique dans le secteur automobile, les réducteurs mécaniques fonctionnent à des vitesses de rotation de plus en plus élevées. Pour ce type de réducteur, les pertes de puissance au niveau des paliers à roulements peuvent être prédominantes pour des vitesses de rotation élevées. Ces pertes sont fortement dépendantes du comportement thermique du palier à roulement. D’où l’intérêt de développer des outils numériques permettant d’estimer les puissances générées au sein d’un palier à roulement. Cette estimation permettra en phase de pré-étude de développer un dispositif de refroidissement adapté afin d’éviter tout risque de dégradation du palier à roulement. Cette thèse présente une nouvelle approche permettant d’étudier le comportement thermomécanique de palier à roulement. Cette approche intermédiaire requiert un minimum de paramètres d’entrée (géométrie externe du palier à roulement et condition de fonctionnement). Elle utilise la méthode dite « des réseaux thermiques » pour obtenir une interconnexion entre les pertes de puissance et le comportement thermique du palier à roulement. Enfin cette approche permet de calculer la puissance dissipée au sein d’un roulement et les températures des bagues et du lubrifiant principalement pour une application hautes vitesses. Cette approche est développée pour des roulements à billes à gorge profonde et contact oblique. Des valeurs expérimentales sont nécessaires pour valider cette nouvelle approche. C’est pourquoi, un nouveau banc d’essai modulaire dédié à l’étude du comportement thermomécanique de palier à roulement a été développé. Sur ce nouveau moyen d’essai, une grande variété de paliers à roulements peut être testée pour différentes conditions de fonctionnement. Pour les premiers essais, un roulement à billes à gorge profonde lubrifié par injection est testé pour un fonctionnement hautes vitesses (produit (n×D_m) supérieur au million). / Rolling Element Bearing (REB) is an essential component in mechanical transmission to reduce friction between rotating parts. Now, with the development of electrical motor in mechanical industry, REBs may work at very high rotation speed. It leads to an increase of REB power losses and temperatures. Theses power losses are strongly coupled with the REB thermal behaviour. The oil temperature has a significant impact on the kinematic viscosity which in turn affects the REB power losses. Based on thermal network approach, an intermediate model is developed in this study. This new model allows obtaining lumped information (temperature of rings) with a minimum of input data (REB external geometry and operating condition only) and by using global power loss models. This intermediate model is developed for angular contact ball bearing and deep groove ball bearing under oil jet lubrication for high speed application. Experimental data are required to validate this new approach. That why, a modular test rig is designed to obtain information on the REB thermomechanical behaviour. The new test rig developed in this study is dedicated to a wide range of REB dimensions and for different operating conditions. For the first test, a deep groove ball bearing under oil jet lubrication is studied for high speed application ((N.dm) product is higher than one million.)

Les propriétés thermomécaniques étonnantes des alliages à mémoire de forme (AMF) sont mises à profit dans de nombreux domaines. Ce matériau est capable de mettre en mouvement une structure suite à un changement de température. Or les façades de bâtiments contemporains, pour s’adapter à des conditions climatiques variables, doivent réguler le passage de la lumière et de l’énergie thermique, par exemple au moyen de systèmes motorisés. Le potentiel de fils AMF pour l’actionnement de protections solaires en façade est exploré ici. La modélisation du couplage mécanique induit par l’introduction de tels matériaux dans une structure a été peu étudiée : l’AMF agit sur la structure qui en retour modifie le comportement de l’AMF. La première étape de ce travail a consisté en une contribution à la modélisation du comportement thermomécanique de ce matériau reposant sur le choix d’une énergie libre, d’un potentiel de dissipation et de plusieurs variables internes. Deux modèles unidimensionnels ont été proposés : un premier modèle monocristallin reproduit de façon simplifiée le comportement du matériau, et un second modèle polycristallin propose une description plus fidèle. En parallèle un dispositif d’essai original à température contrôlée a été développé, il a permis une caractérisation fiable de fils Nickel-Titane et l’identification des paramètres des modèles. Dans un second temps ces modèles ont permis de résoudre des cas de couplage élémentaires (fil AMF + ressort, lame élastique + fil AMF noyé) pour des chargements thermomécaniques simples, et des solutions analytiques ont été établies. Les modèles ont été implémentés numériquement via un script matériau utilisateur (UMAT) pour le logiciel éléments finis ABAQUS et au moyen d’un algorithme d’optimisation sous contraintes. Ceci permet de simuler la réponse couplée de systèmes structuraux a priori quelconques intégrant des AMF, connectés à ou noyés dans, une structure. Dans un troisième temps, divers actionneurs ont été conçus, réalisés et testés dans le cadre de l’occultation solaire des façades. Le principe est d’utiliser un cycle de température permettant à l’AMF de déformer la structure, puis à l’énergie élastique de déformation de la structure d’assurer le retour à la forme originale. Le comportement réel de ces actionneurs a été comparé aux calculs analytiques et éléments finis. Des tests cycliques ont également été réalisés / The surprising thermomechanical properties of shape memory alloys (SMA) are harnessed in many engineering fields. This material is able to set a structure in motion upon a temperature change. Today, contemporary building facades must adapt to variable climate conditions as well as to evolving building use and occupancy. In particular, they must regulate light and thermal energy passing through the facade, with motorized systems, for example. We explore the potential of SMA wires for putting in motion solar shading devices in facades. The modelling of the mechanical coupling induced by the introduction of such materials in a structure has received little attention as of now. The SMA acts on the structure which in return modifies the SMA behavior. The first step of this work is a contribution to modelling the thermomechanical behavior of this material through the choice of a free energy, a dissipation potential and internal variables. We propose two one-dimensional models: a first monocrystalline model reproduces the material behavior in a simplified way, and a second polycrystalline model offers a more accurate description of it. An original temperature-controlled testing apparatus was developed in parallel. This led to a reliable characterization of Nickel-Titanium wires and the identification of the model parameters. In a second stage, these models allowed to solve elementary coupling cases (SMA wire + Spring, Elastic plate + Embedded SMA wire) for simple thermomechanical loadings and we established analytical solutions. The models were then numerically implemented via a user-material script (UMAT) for the finite elements software ABAQUS, by using a constrained optimization algorithm. This enables the simulation of the coupled response of, in principle, any structural system including SMA wires, connected or embedded in the structure. Finally, we designed, fabricated and tested different actuators in the context of sunlight control in facades. The working principle lies in using a temperature cycle which allows the SMA to deform the structure, and then allows the elastic strain energy in the structure to ensure the return to the original shape. The real behavior of these actuators have been compared to analytical and finite element calculations. We also performed cyclic tests

L’objectif de ce travail a été de modéliser les écoulements de polymères dans les co-malaxeurs de type BUSS. Le co-malaxeur est une extrudeuse particulière utilisée comme une boite noire depuis des dizaines d’années par des industriels. Sur le fourreau sont fixés des doigts de malaxage et la vis principale dispose d’interruptions de filets et oscille pour permettre le passage des doigts de malaxage dans les différents chenaux. Cet outil est utilisé pour gélifier du PVC ou disperser des charges dans des matrices polymères. Ce travail de thèse répond donc à un besoin industriel puisque, actuellement, mettre au point un produit nécessite de nombreuses heures d’essais coûteux.Un modèle thermomécanique instationnaire en deux dimensions fondé sur les approximations d’Hele-Shaw a été développé. La discrétisation du domaine du co-malaxeur a été faite par éléments finis. La résolution numérique s’effectue par éléments finis et éléments finis stabilisés par SUPG. La modélisation nous permet d’obtenir en tout point du domaine de calcul la pression, les vecteurs débits, les vitesses de cisaillement, la viscosité et la température.Les résultats issus des simulations nous permettent d’obtenir des ordres de grandeur de pressions et températures et des indices de mélange pour différentes conditions procédés. Une comparaison entre les résultats issus de la modélisation et des essais expérimentaux montre que, dans les zones remplies, le modèle donne des résultats satisfaisants. / The aim of this study was to model molten polymers flow in BUSS type co-kneaders. The BUSS co-kneader is a particular single-screw extruder. It is composed of a rotating screw like standard single screw extruders, but with interrupted flights and mixing pins fixed to the barrel. The screw has also an axial reciprocal movement. It has been used for decades in industry for its mixing capacities, specifically for PVC gelification or polymer compounding with fibres, additives or carbon black. This work is therefore answering to nowadays industrial needs, as developing new products is expensive and time consuming.A 2D time-dependent thermomechanical model based on Hele Shaw approximations was developed and the co-kneader domain was discretized by finite elements. The numerical problem was solved by finite elements and SUPG stabilized finite elements. This model provides, at every point of the calculation domain, the pressure, throughput vectors, shear rates, viscosity as well as temperature.Simulation results provide pressure and temperature orders of magnitude, as well as information on polymer mixing depending on process parameters. A comparison between the model and experimental trials shows a satisfactory agreement in the filled zones.