This thesis treats type synthesis and kinematic analysis of translational parallel manipulators (TPMs) from a group theoretical point of view. / For the TPM type synthesis, based on displacement group theory (DGT), TPM leg motion is represented by a series of displacement subgroups. For symmetrical three-legged TPMs, three categories are classified and a total of 90 architectures are proposed. For asymmetrical three-legged TPMs, 13 cases of possible leg combinations are presented and some possible constructions of TPMs are shown. The advantages of DGT in mechanism synthesis are described. / For the kinematic analysis of symmetrical TPMs, geometric elements are employed to denote the workspace of different leg links. Solutions of the inverse and direct kinematics of a certain example are obtained. Moreover, the proposed approach is applied to a class of TPMs, and the corresponding geometric representations are listed. The applicability of the proposed approach is discussed as well. / This thesis provides a theoretical approach to design TPMs and analyze their kinematics for practical applications.

The goal of this thesis is to model the plastic sheet reheat and forming process for thermoforming, and to measure the exact value of heat capacity for a set of plastic materials. / A series of experiments with heat flux and temperature sensors are performed on thermoforming machines at the Industrial Materials Institute, National Research Council, for the purpose of determining the heat capacity for several materials. The ambient air velocity, air temperature and the sheet internal temperatures at different depths inside a plastic sheet were measured. Based on these on-line measurements, a methodology for evaluating heat transfer coefficients and heat capacities was developed. Moreover, two different approaches were used to tune the experimental results through sensitivity analysis. Additionally, in order to improve simulation precision and better control the heating process, an approach for measuring internal sheet temperature is discussed. (Abstract shortened by UMI.)

As humanity prepares to migrate to the frontiers of the Moon and other planets, the area of mining in space must go along for the purpose of exploration and in-situ resource utilization. In the present work the literature that has been developed over the years in the area of mining in space as applicable to Lunar and Martian environments is reviewed. Subsequently, the key mining technologies that are most suitable for Lunar and Martian environments are identified. From the literature review, it is concluded that an optimal combination of both mechanical methods and novel energy (lasers, microwaves, nuclear energy) methods for rock destruction drawing a trade off between the energy and mass would be the most ideal option for space applications. / One such technique of applying low power microwaves to the rocks to thermally weaken them without actually melting them before employing mechanical methods of rock destruction is investigated. Finite element simulations were carried out to simulate microwave heating of a calcareous rock to determine the temperature profiles and thermal stresses at different microwave heating times and powers. Preliminary experiments were carried out in order to determine the microwave susceptibility of terrestrial basalt (which has similar composition as Lunar and Martian rocks). Temperature and strength of the rock sample before and after microwaving was measured. / The results of the finite element simulation indicated that a calcareous rock with microwave responsive phase and a microwave non-responsive phase developed thermal stresses of large magnitudes exceeding the actual strength of the rock. The simulation methodology can be applied to other rock types as well, provided the thermal, electrical and structural properties of constituent mineral phases are available. / The preliminary experimental results showed that the basalt rock specimens used were quite susceptible to the low power microwaves. There was a decreasing trend in terms of the point load index of the rock samples as the microwaving exposure times were increased, with some rock samples showing visible cracks at higher microwaving times.

In order to elucidate the propagation mechanism of detonations, the failure mechanism of detonations propagating in tubes with porous walls is investigated experimentally. Two distinct failure mechanisms were identified depending on the type of detonating mixture. Experiments in mixtures characterized by piecewise laminar reaction zone structures with weak three-dimensional effects revealed that the attenuation and failure is caused by the global mass divergence to the porous, permeable walls. The limits observed in these mixtures agreed very well with the theoretical limiting conditions for the existence of curved ZND detonations subjected to lateral expansions. / Experiments were also conducted in a second class of mixtures, characterized by irregular cellular structures and turbulent reactions zones. When detonations in this class of mixtures are attenuated, transverse waves re-amplify from local instabilities in the reaction zone. This re-amplification permits the detonation wave to continue to propagate and overcome the effects of global mass divergence and transverse wave attenuation at the porous walls. Ultimately, when the rate of transverse wave re-generation is surpassed by the rate of transverse wave damping at the walls, the detonation can no longer be self-sustained and fails. The limits obtained in these irregular structure mixtures were found significantly wider than predicted by the ZND formulation, thus further confirming the important role of the three-dimensional turbulent structure in these detonations in providing a more efficient mechanism of gas ignition and propagation than detonations that exhibit a regular structure. / The implications of the present study are that the ZND model is not valid in describing the reaction zone structure, the ignition mechanism and thus the propagation mechanism in these turbulent detonations. In these mixtures, the ignition mechanism is a combination of the classical mechanism of adiabatic shock compression leading to ignition as well as the intense turbulent mechanism usually attributed to deflagrations only.

Methods of structural optimization have been studied and developed over the last three decades. An important aspect of structural optimization pertains to the conditions under which the loads are applied. Problems in which the applied load is single and deterministic are well documented in the literature, whereas those in which the load is deterministic but varying, stochastic, or uncertain have been less studied. Although methods of structural optimization under single-loading conditions can be extended to multiple-loading cases, the computational burden imposed by these methods is often extremely high, to the point of becoming inapplicable. / In this thesis, a novel approach is introduced. We first propose the use of the concept of stiffness ellipsoid, whose boundary defines the field of displacements of a structure subject to a load with a constant magnitude, but applied in an arbitrary direction. Based on the study of the stiffness ellipsoid, an optimality criterion that determines the worst case of the structure subject to a load with a given magnitude, but applied in an arbitrary direction, is proposed. The underlying ideas are applied to the structural optimization of the roller-carrying disk of a novel class of cam-follower speed-reduction devices---Speed-o-Cam (SoC). The optimum structures obtained in this thesis are compared with the existing prototype, which was designed using established design criteria only. Results indicate an improvement of almost twice the stiffness with a mass reduction of 40% of the original prototype. / Also reported in the thesis is a novel algorithm of mesh generation, named the contour mesh generation algorithm (CMGA). The CMGA, based on the concept of penalty functions, is applied throughout the examples given in the thesis and shows a great advantage in solving structural optimization problems.

Cardiovascular diseases are the leading cause of death worldwide. In most cases, cardiovascular diseases are caused by an underlying pathology: atherosclerosis. Numerical models are very helpful to study the biomechanics of atherosclerosis, assess the risks and more likely sites of rupture, to design and study the effect of interventions. Yet, the reliability of the simulations is directly dependent on the fidelity of the material properties, geometry and boundary conditions to represent the reality. However, data available to precisely model the mechanical behaviour of the plaque and its constituents are relatively limited, particularly for the inclusions. / The main objective of this project is to propose a method to determine the non-linear material parameters for an inclusion and for the surrounding tissues of atherosclerotic medium arteries using numerical and experimental models. To evaluate the proposed method, a mock artery with a stenosis and a lipid inclusion was designed and fabricated with polyvinyl alcohol gel for each of the two wall layers, and animal fat (lard) for the inclusion. PVA gel has a non-linear elastic behaviour that can be described with a Money-Rivlin model. / The experimental part of the approach consists in extension-pressurization of the intact vessel. For different states of axial extension, the vessel is pressurized while intravascular ultrasound images are acquired at regular intervals. Then, the images are segmented to extract the contours of the different constituents. The contours from before the pressurization are use to build the geometry of finite element models. Simulations are run for different pressure values. The resulting deformed contours are compared to the contours extracted from the images for the same pressure values, to optimize the material parameters for the inclusion and for the wall layers. The comparison is defined in terms of circumferences and thicknesses. / The estimated material parameters for the gels were applied to predict the responses in uniaxial and biaxial tensions, to compare them to experimental data from tension testing of samples made of the same gels. The estimated parameter for the lard was compared to shear test measurements. The predictions were good (R2 ≥ 0.90) for deformations in the range and slightly over the range of those used for the optimisation process. / Les maladies cardiovasculaires constituent la première cause de mortalité dans le monde. Dans la majorité des cas, elles sont causées par une pathologie sous-jacente, l'athérosclérose. Les modélisations numériques sont d'une grande aide pour étudier la biomécanique de la plaque d'athérosclérose, évaluer les risques et sites de rupture les plus probables, concevoir et étudier l'effet d'interventions. Cependant, les données disponibles pour modéliser avec précision le comportement mécanique de la plaque et ses composants sont relativement limitées, en particulier pour les inclusions. / L'objectif principal de ce projet est de proposer une méthode pour déterminer les paramètres des modèles de matériaux non-linéaires pour l'inclusion et tissus adjacents d'artères athérosclérotiques, à l'aide de modèles expérimentaux et numériques. Afin d'évaluer la méthode proposée, une pseudo-artère imitant une artère coronaire sténosée avec une inclusion lipidique a été conçue et fabriquée avec du gel d'alcool de polyvinyle (PVA) pour chacune des deux couches de la paroi et du saindoux pour l'inclusion. Le gel de PVA a un comportement élastique non-linéaire pouvant être décrit à l'aide d'un modèle de Mooney-Rivlin. / La partie expérimentale de l'approche consiste à soumettre le vaisseau à des tests d'extension-pressurisation. Pour différents états d'extension axiale, le vaisseau est pressurisé tandis que des images par ultrasons intravasculaires sont enregistrées. Ces images sont ensuite segmentées pour extraire les contours des différents constituants. Les contours avant pressurisation sont utilisés pour construire la géométrie de modèles par éléments finis. Des simulations sont exécutées pour différentes valeurs de pression. Les géométries déformées obtenues sont comparées aux contours extraits des images à ces même valeurs de pression afin d'optimiser les paramètres des modèles de matériaux pour l'inclusion et les deux couches de la paroi. Cette comparaison est définie en termes de circonférences et d'épaisseurs. / Les paramètres estimés pour les gels sont appliqués pour prédire les réponses en tensions uniaxiale et biaxiale, et comparer aux données expérimentales de tests tension réalisés sur des échantillons faits des mêmes gels. Le paramètre estimé pour le lard est comparé à des mesures obtenues aves des tests de cisaillement. Les prédictions étaient bonnes (R2 ≥ 0.90) pour des déformations de l'ordre de grandeur et légèrement supérieures à celles utilisées dans le processus d'optimisation.

Chemical kinetics is the science of modeling the steps involved in a chemical reaction at the molecular level. This investigation is concerned with the chemical reactions that occur during combustion. There is ample room for improvement in most presently accepted chemical kinetic models and many reactions are still not modeled. The key to creating and improving chemical kinetic models of combustion reactions is gathering sufficient experimental data and solidifying the foundations upon which more complicated models may be built. / The experimental apparatus that is used to gather data in this study is a stagnation flame burner. Particle image velocimetry, a laser-based flow visualization technique, is used to measure velocity in both the axial and radial directions. The results are analyzed first to confirm the validity of the one-dimensional, axisymmetric model commonly used to describe impinging-jet flow. Experimental results are found to have good agreement with theoretical approximations and numerical models. Tests are then conducted to reproduce previously published data for methane-air flames at lean, stoichiometric and rich conditions in order to prove the reliability of the experimental apparatus. There is agreement between published data and the new experimental results. / Experiments with oxygen enrichment are then conducted at equivalence ratios not normally within the flammability limits of methane-air mixtures, from the very lean, phi = 0.55, to the very rich, phi = 1.45. Results show marked disagreement between model and experiment at equivalence ratios far from stoichiometric. These experimental data will allow the chemical kinetic model for premixed methane combustion to be improved -- by correcting divergence in the model at very lean and very rich conditions, the model should be improved appreciably over the narrower range of equivalence ratios typically seen in methane-air combustion. / La cinétique chimique est la science de la modélisation des étapes d'une réaction chimique au niveau moléculaire. Ce mémoire s'interesse aux réactions chimique qui se produisent durant la combustion. Les modèles cinétique présentement acceptés sont en besoin d'amélioration et d'autres enplus, il existe plusieurs réactions qui n'ont aucun modèle. Pour mieux développer les modèles et pour en créer des nouveaux, il fault rassembler assez de données expérimentales pour produire une base solide sur laquelle des modèles de plus en plus compliqués peuvent etre construits. / Pour générer des données, cette étude utilise un appareil expérimental avec une géométrie de point d'arrêt, dans laquelle une flamme aplatie peut exister. Les vitesses dans la direction axiale et radiale sont mesurées par la vélocimétire image-particule, une technique à base de laser. Les résultats sont d'abord analisés pour vérifier que le modèle unidimensionnel est valide. Une bonne concordance est observée entre la théorie et les résultats et entre les modèles numériques et les résultats. Subséquemment, des expériences sont effectuées pour reproduire des données déjà publiées pour le méthane et l'air aux rapports d'équivalence oxydants, stoechiométriques et réducteurs, afin de prouver la fiabilité de l'appareil expérimental. Il ya une bonne concordance entre les données publiées et les résultats expérimentaux. / Des expériences dans lequelles l'air est enrichi avec de l'oxygène sont effectuées à des rapports d'équivalence de phi = 0.55 juste qu'à phi = 1.45; ces rapports d'équivalence ne sont normalement pas dans les limites d'inflammabilité des mélanges de méthane-air. Ici, les résultats montrent un désaccord marqué entre le modèle et les expériences pour les rapports d'équivalence très loin de phi = 1.0. Les données permettent le modèle cinétique chimique pour la combustion du méthane prémélangée d'être amélioré. En corrigeant les divergences entre le modèle et les resultats experimentales pour rapports d'équivalence très oxydants et très réducteurs, le modèle sera amélioré pour rapports d'équivalences généralement vus dans la combusiton de méthane dans l'air.

Non-synchronous vibration is a common phenomenon in rotor-bearing systems which causes cabin comfort issues in mid size business jets. In rotating systems, the most significant source of non-synchronous vibration is the nonlinear contact in supporting bearings which is affected by internal radial clearance, contact materials, roller numbers and mass imbalance, etc. / In this study, a Jeffcott rotor-bearing system is modeled using the finite element method. / For solutions of the system's governing equations, both the time-marching technique and harmonic balance method are used. Moreover, a modal reduction technique is applied to save the computations. For bearing contacts, three treatment approaches are adopted: the Hertz theory, the Lagrange multipliers technique and the forcing functions. Finally, the non-synchronous vibration is investigated with respect to the bearing's parameters, and the internal radial clearance is proven to play the most significant role in controlling the non-synchronous vibration. / Les vibrations non-synchrones représentent un problème connu des machines tournantes et sont à l'origine de problèmes acoustiques dans les avions de taille moyenne. Dans les roulements de support, ces vibrations asynchrones sont initiées par les contacts non-linéaires qui dépendent du jeu de fonctionnement, du type de matériaux en contact, du nombre de rouleaux, etc. / Ce travail est consacré à la modélisation d'un système de Jeffcott et à son traitement numérique par la méthode des éléments finis. Les équations du mouvement sont résolues par intégration temporelle directe et par la méthode de l'équilibrage harmonique. Qui plus est, des gains informatiques sont réalisés grâce à la mise en oeuvre d'une méthode de synthèse modale. Le traitement des conditions de contact séffectue selon trois approches: la théorie de Hertz, la technique des multiplicateurs de Lagrange et par chargement équivalent. Les vibrations non-synchrones sont ensuite analysées en fonction des paramètres des roulements et il est mis en évidence que le jeu de fonctionnement au sein de ces roulements est le paramètre principal caractérisant ce type de vibrations.

Detonation is a branch of combustion that is initiated by an exothermic chemical reaction and it results in a supersonic shock wave called the “Detonation Wave”. Generally detonation occurs in homogeneous reacting gaseous, liquid or solid media. However sometimes, detonation is also observed in highly non-uniform medium which contains detonable sources (reacting gaseous, liquid or solid media) in a discrete pattern. This project focuses on the study of effect of discrete energy sources on the detonation wave velocity with the help of numerical modeling and simulations. The energy release as a result of detonation is continuous for a homogeneous gaseous mixture and generally results in a Chapman Jouguet (CJ) Detonation Velocity. However in the presence of discrete energy sources, the detonation velocity comes out to be higher than the CJ Detonation Velocity as observed theoretical by Higgins A. (Proc. 20th ICDERS, Montreal, 2005). This project aims to find numerical solutions for highly discrete detonation systems to verify the already existing theoretical results for discrete detonation systems. For the calculation of detonation wave velocities for medium with discrete sources, numerical analysis was carried out on an in-house one dimension Euler code. Modifications in the code were made to use it for continuous detonation purpose initially to verify the exact CJ detonation velocity through the code and then at a later stage, other modifications were made to the code to use it for discrete detonation phenomenon. Using the in-house numerical code for solving supersonic flow problems, the continuous detonation system was modeled and the results obtained were within 0.05% difference with respect to the exact CJ Detonation Velocity. After wards, discrete detonation system was modeled and numerical experiments were performed. It was observed that the detonation wave velocity increases with the increase in discreteness of energy source which is consistent with the / La détonation est une branche de la combustion qui est initiée par une réaction chimique exothermique duquel en résulte une onde de choc supersonique appelé «Detonation Wave". Généralement la détonation se produit dans une réaction homogène gazeuse, liquide ou solide. Mais parfois, la détonation est également observée en un milieu hautement non-uniforme qui contient des sources detonable (réaction gazeux, liquide ou solide ) dans un motif discret. Ce projet se concentre sur l'étude des effets des sources d'énergie discrets sur la vitesse de l'onde de détonation à l'aide de la modélisation numérique et des simulations. La libération d'énergie à la suite de la détonation est continue pour un mélange homogène de gaz et aboutit généralement à une vélocité de détonation appelé Chapman Jouguet (CJ). Toutefois, en présence de sources d'énergie distinctes, la vitesse de détonation sort d'être supérieure à la vitesse de détonation CJ observée théorique par A. Higgins (Proc. 20th ICDERS, Montréal, 2005). Ce projet vise à trouver des solutions numériques pour les systèmes de détonation très discrète afin de vérifier l'existance des résultats théoriques pour les systèmes de détonation discrets. Pour le calcul de vitesses des ondes de détonation pour les moyennes avec des sources discrètes, l'analyse numérique a été réalisée sur une maison d'une dimension de code Euler. Ces modifications du code ont été apportées à des fins de détonation continue . D'abord pour vérifier la vitesse de détonation CJ exacte dans lc code, puis à un stade ultérieur, d'autres modifications ont été apportées au code utiliser pour le phénomène de détonation discrets. Afin de pouvoir utiliser la maison en code numérique pour résoudre les problèmes d'écoulement supersoniques, le système continue de détonation a été modélisé et les résultats obtenus ont été de 0,05% de différence par rapport à la vitesse exacte d

The Automated Fiber Placement (AFP) process has good potential for manufacturing large composite structures and is used for large scale production. However, the AFP process reveals uncertainties associated with the induced defects. This investigation examines the ultimate strength variation induced by a single gap, overlap, half gap/overlap and twisted tow defects. Tests are carried out on three lamina properties (fiber tension, fiber compression and in-plane shear), two laminate properties (open hole tension and open hole compression) and each was compared with a baseline configuration without defects. It is observed that the effects are negligible at the lamina level (less than 5% reduction) compared to the laminate level (up to 13%). In addition, the fiber waviness induced by the defects has more impact on the compression properties due to induced micro-buckling. A global prediction method is finally presented to obtain a model that considers the effects of the defects under investigation. / Le procédé de placement de fibre automatisé a démontré un intérêt grandissant pour la fabrication de grandes structures en matériaux composites. Il fait néanmoins l'objet d'incertitudes dues aux nombreux défauts de fabrication. Cette étude se penche sur la variation de la contrainte ultime lorsque des défauts de type écartement, chevauchement, demi-écartement/chevauchement et torsadé sont présents. Des tests ont été éxécutés pour trois propriétés de pli (tension, compression et cisaillement) et deux propriétés de laminé (tension et compression avec un trou non contraint). Les propriétés sont ensuite comparées avec une configuration sans défaut. Les tests démontrent que les effets sont négligeables au niveau du pli (moins de 5%) comparé au niveau du laminé (jusqu'à 13%). De plus, l'ondulation des fibres induite par les défauts est plus néfaste en compression, puisqu'elle initie leur micro-flambage. Pour conclure, un modèle de prédiction global tenant compte de l'effet des défauts est proposé.