To capitalize on the efficiency and simplicity of wheeled robots, as well as the adaptability and maneuverability of legged robots, many hybrid leg-wheel designs have been developed. To date, none of these platforms have possessed the ability to execute dynamic maneuvers and thus have major shortcomings in their speed, efficiency and obstacle negotiating capabilities. A hybrid leg-wheel quadruped capable of such dynamic behaviour is introduced. Using an accurate model of this platform, a variety of dynamic behaviours and examples of their utility were successfully simulated. Passive leg compliance and the placement of wheels at the foot of the legs proved invaluable in achieving such high energy maneuvers on this power autonomous platform. A full systems design of a testbed capable of executing the presented dynamic behaviour was completed. From both a mechanical and control stand-point, it is a simple and robust robot. This prototype will prove the feasibility of such behavioural feats for autonomously powered platforms, demonstrate their wide utility and pave the way for their realization on ruggedized platforms.

In many exciting modern video games, one of the player's main tasks is to move and control the "hero" or alter-ego on a curved path. This is also a key issue in more serious applications like an automobile driver training simulator. In order to determine where the vehicle is at any time, while the speed is controlled with the trainee's accelerator pedal, we need a computationally efficient way to establish this position by, say, defining an arc length parameterized curve to represent the "road" or other trajectory constraint. To compute the are length of a parametric curve or define the arc length parameterized curve using the analytical methods compromises many real-time simulations where designers compete to continuously add realism enhancing visual and audio effects. In this thesis, an efficient and novel way to parameterize a cubic curve with respect to arc length is adapted and implemented. Its use in real-time application is demonstrated and its performance is compared to that of another recently proposed method. The main advantage of our approach is the simplicity. That is, Bezier segments are subdivided as required and arc length parameterization is obtained offline for fast look-up, thus requiring less computation in real-time.

In the present study, the instability resulting from the interaction of a spherical flame and a rarefaction is investigated. The goal is to understand the physical mechanisms governing the evolution of the flame front subsequent to the interaction. It is shown that the distinct acceleration characteristics of the rarefaction give rise to both Rayleigh-Taylor and Richtmyer-Meshkov types of instabilities. Furthermore, following the interaction, the formation of a funnel of unburned gas into the combustion products is observed, similarly to the one observed by Markstein in flame/shock interaction experiments. It is found that the growth rate of the funnel is constant and that it increases with augmenting pressure gradient at the leading edge of the rarefaction. / It is also demonstrated that the interaction between a flame and a rarefaction can lead to a greater rate of increase of pressure inside the combustion vessel. This effect is only observed for a relatively high pressure gradient at the leading edge of the rarefaction. Furthermore, it is found that augmenting the flame diameter prior to the interaction results in a higher pressure increase due to the interaction.

This thesis develops a quantitative quality function deployment (QFD) method and integrates it with the methodologies of value engineering and risk analysis. Numerical parameters and scales allow measurement of product development and onward adjustment as the process unfolds. In QFD, customer requirements are usually defined in terms of design characteristics. In the method developed here, customer requirements are related to product functions. The use of value engineering then allows costs to be quantified. The analysis of design uncertainties allows the determination of risk factors. By introducing the new concepts of customer satisfaction value and an integrated evaluation index, an appropriate balance between customer and enterprise satisfaction can be attained in the final product. The new QFD method provides a quantitative approach based on cost and risk for evaluating how well customer requirements are taken into account in the development of a product.

When an aircraft flies through clouds under icy conditions, supercooled water droplets at temperatures below the freezing point may impact on its surfaces and result in ice accretion. The design of efficient devices to protect aircraft against in-flight icing continues to be a challenging task in the aerospace industry. Advanced numerical tools to simulate complex conjugate heat transfer phenomena associated with hot-air anti-icing are needed. In this work, a 3D conjugate heat transfer procedure based on a loose-coupling method has been developed to solve the following four domains: the external airflow, the water film, conduction in the solid, and the internal airflow. The domains are solved sequentially and iteratively, with an exchange of thermal conditions at common interfaces until equilibrium of the entire system is achieved. A verification test case shows the capability of the approach in simulating a variety of anti-icing and de-icing cases: fully evaporative, running wet, or iced. The approach is validated against a 2D dry air experimental test case, because of the dearth of appropriate open literature 3D test data.

The temporal pattern of the anterior-posterior force acting at the human knee joint during walking is determined in the present study. A three-dimensional anatomically based dynamics model, which includes the bones and the muscles working across the knee joint, is developed and the redundant inverse dynamics problem is formulated. The muscle redundancy is resolved by reducing the number of unknowns based on certain logical assumptions. In addition, a parametric study is conducted to study the effect of antagonistic muscle activity on the joint force. The developed dynamics model is then exercised to calculate the anterior-posterior knee joint force for normal and ACL-deficient knees. The anterior-posterior knee joint force obtained from the normal knee analysis indicates that a posterior force is present for the majority of the gait cycle with a maximum force of 0.25 BW (body weight), whereas anterior force acts shortly after the beginning of the stance phase reaching a maximum of 0.4 BW. Among the muscles working across the knee joint, the quadriceps, gastrocnemius and the tensor fascia latae produce an anterior force at the knee joint whereas the hamstrings, sartorius and the graclilis generate a posterior force. For the ACL-deficient knee, the results reveal that it is possible for patients to adapt to the injury by changing their walking pattern to avoid or minimize quadriceps activation in the stance phase, thus reducing or entirely avoiding the anterior force at the knee joint. The study also reveals that it is necessary to strengthen the hamstring muscles for the rehabilitation of the ACL-deficient knee patients to reduce the anterior force at the knee joint.

The interaction of two scalars (both temperature) emitted from concentrated line sources is studied in fully-developed, high-aspect-ratio, turbulent channel flow. The thermal fields are measured using cold-wire thermometry in a flow with a Reynolds number (Uh/nu) of 10200. / The transverse, total RMS temperature profiles are a function of the separation distance between the line sources (d/h), their average wall-normal position (ys_av/h), and the downstream location (x/h) (relative to the line sources). The non-dimensional form of the scalar covariance, the correlation coefficient (rho), is a function of the same parameters and quantifies the mixing of the two scalars. / In all cases, the transverse profiles of the correlation coefficient flatten and tend towards 1 as x/h increases. The minimum value of the correlation coefficient for a given transverse profile (rhomin) drifts towards the wall for line source combinations that are not symmetric about the channel centreline. (For profiles that are symmetric about the centreline, the minimum correlation coefficient always occurs at the centreline.) The initial downstream evolution of rhomin (and rhoys_av/h - the value of the correlation coefficient measured at the average wall-normal source location) depend mostly on d/h. Therefore, to an extent, the measured evolutions are similar to those in homogeneous turbulence. However, the dependence on Ys_av/h is never entirely negligible. At a certain downstream position, the (inhomogeneous) high turbulence intensity (found near the wall) serves to accelerate the scalar mixing. / For line source combinations that are not influenced by the above-mentioned, near-wall, high turbulence intensities, spectral analysis indicates that the degree of mixing across scales is relatively constant for the range of parameters studied herein. In the case of near-wall mixing, the large scales were found to evolve more rapidly. This presumably derives from the large-scale nature of turbulence production, which is strong in the nearwall region.

The experimental investigation described in this thesis is primarily aimed at acquiring digital particle-tracking velocimetry (DPTV) measurements of a free-surface granular flow down the upper inclined surface of a wedge-shaped static pile of the same material, contained in a narrow channel between two parallel vertical glass plates. The glass plates are 610 mm long and 350 mm high. Three different values of the normal separation distance between them were investigated: 25.4 mm, 38.1 mm, and 50.8 mm. For each of these separation distances, the following three values (nominal) of the mass flow rate per unit width were considered: 0.81 kg/s-m, 1.85 kg/s-m, and 3.33 kg/s-m. The granular material used in this work consists of slightly polydisperse, almost spherical, ceramic (zirconium silicate) beads: mean effective diameter of 1.59 mm and mass density of 4071 kg/m3. For these ranges of dimensional parameters, statistically-steady fully developed flows were established in a reliable and repeatable manner. A high-speed camera system was used to acquire digital images of the granular flows of interest in the fully developed region. Image processing and a commercial PTV software package (DiaTrackPro 2.3) were used to obtain the particle trajectories. Special routines were written in Matlab to obtain the corresponding instantaneous and ensemble-averaged velocity distributions. These results were then used to compute the corresponding distributions of granular temperature and a related dimensionless parameter that is commonly referred to as the Savage-Jeffrey parameter. The aforementioned PTV and data processing procedures, the results in dimensional and dimensionless forms, and the applicability of some recently proposed scaling laws are discussed in this thesis.

The thesis centres on the dynamic behaviour of woven composite materials. The increase in the use of these materials in the aerospace industry demands a reliable constitutive damage model to predict their response under high strain rate loading. The current available models do not include rate effects, are too complicated to use, are unable to provide a clear procedure for characterizing the model parameters, or are unable to accurately predict response when compared to experimental tests. / The present work focuses on developing a rate-dependant continuum damage mechanics model which considers all in-plane damages, i.e., damage due to matrix cracks and fibre failure in normal and shear directions. A physical treatment of growth of damage based on the extensive experimental results is combined with the framework of continuum damage mechanics models to form the foundation of the model for materials whose response is governed by elastic deformation coupled with damage. / The developed model simulates the non-linear and rate dependant behaviour of woven composite materials due to damage evolution. The model is implemented into a commercial finite element code (ABAQUS) as a two-dimensional user material subroutine. / Detailed study of the behaviour of these materials under high velocity loading and several experiments were required to establish material model parameters. Uniaxial tension as well as bias extension shear tests were carried out at both static and dynamic rates on three different woven composite materials. / A tensile version of the Hopkinson bar setup, in conjunction with the designed specimen and specimen fixture, are used for testing laminated woven composite materials at high rates of strain and a peak strain rate of up to 560/sec. Comparison of Hopkinson bar results with results of tests performed at a quasi-static rate using a servo-hydraulic testing machine clearly show that strain rate has an effect on both the stress and the strain to failure for these carbon woven composite materials. A higher stress is observed for dynamic tests, whereas the strains at the maximum stress are higher for the static experiments. Also, an increase in the initial undamaged elastic modulus is observed for both tensile and shear response with increasing strain rate. / Finally, the simulation results presented in this work demonstrate that the model can well predict the dynamic and the static response obtained from the experimental measurements. Simulation results for higher than tested strain rates are also presented to illustrate the response at these high rates. It is shown that the tensile strength increases with the increase in loading rate, and also a strain softening phenomenon is observed after the maximum tensile strength. The loading-unloading, as well as compression response of the materials are also simulated to ensure the capability of the model to capture these responses. / Cette thèse se concentre sur le comportement dynamique des matériaux composites tissés. L'augmentation de l'utilisation de ces matériaux dans l'industrie aéronautique requiert le développement de modèles constitutifs fiables pour prédire leurs réponses sous des conditions de chargement à haute vitesse de déformation. Actuellement, les modèles disponibles n'incluent pas les effets de vitesse de déformation, sont compliqués à utiliser ou ne sont pas capables de prédire avec précision le comportement expérimental de ces matériaux. / Ce travail focalise sur le développement d'un modèle de mécanique continue de la rupture tenant en compte la dépendance avec le taux de déformation. Ce modèle considère tous les dommages dans le plan : dommages provoqués par la fissuration de la matrice ainsi que ceux dus à la rupture de la fibre dans les directions normale et tangentielle. Un traitement physique de la propagation du dommage, basé sur un vaste ensemble de résultats expérimentaux, est combiné avec les modèles de mécanique continu de la rupture afin de développer un modèle pour ces matériaux tissés, dont la réponse comportementale est gouvernée par un couplage ente la déformation élastique et la rupture. / Le modèle développé dans cette étude simule le comportement non-linéaire et dépendant du taux de chargement des matériaux composites tissés provoqué par l'évolution des dommages. Ce modèle est implémenté dans un logiciel commercial d'élément fini (ABAQUS) en tant qu'un sous-programme définissant les propriétés bidimensionnel des matériaux. / Une étude détaillée du comportement de ces matériaux, soumis à des chargements à hautes vitesses, et plusieurs expériences ont été requises pour établir les paramètres du modèle. Des tests en tension uni-axiale et en cisaillement à bi-extension ont été réalisés à la fois pour des cas statiques et dynamiques sur trois matériaux composites tissés différents. / Une version en tension du dispositif à barres de Hopkinson, utilisant les spécimens et fixations conçus, a été utilisée pour tester des laminés de matériaux composites tissés dans des conditions de haut taux de déformation. Un taux maximum de déformation a été obtenu à 560/sec. La comparaison des résultats obtenus avec les barres de Hokinson, avec ceux obtenus avec des tests quasi-statique utilisant une machine de test servo-hydraulique, démontre clairement l'effet du taux de déformation sur les contraintes et déformations à la rupture pour ces matériaux composite à fibres de carbone tissés. Des contraintes plus importantes sont observées pour les tests dynamiques, tandis que les déformations pour les contraintes maximales sont plus grandes pour les tests statiques. De plus, une augmentation du module élastique initial non endommagé est observée avec l'augmentation du taux de déformation, à la fois pour la réponse en tension et en cisaillement. / Finalement, les résultats des simulations présentés dans ce travail démontrent que le modèle peut prédire correctement la réponse statique et dynamique obtenue par les mesures expérimentales. Les résultats des simulations, pour des taux de déformations plus grands que ceux testés expérimentalement, sont aussi présentés pour illustrer la réponse du matériau à ces haut taux de déformation. Cette étude montre que la résistance en tension augmente avec l'augmentation du taux de chargement. De plus, un phénomène d'adoucissement des déformations est aussi observé après la résistance maximum en tension. Le chargement-déchargement, ainsi que la réponse en compression de ces matériaux sont aussi simulés pour assurer la capacité du modèle à capturer ce type de réponses.

Computational and experimental investigations of fluid flow and heat transfer aspects of loop heat pipes (LHPs) are presented in this thesis. The overall goal is to formulate and develop cost-effective mathematical models and numerical solution methods for computer simulations of LHPs. This work has five distinct parts. / First, a basic network thermofluid model of conventional LHPs operating under steady-state conditions is proposed, implemented, tested, and applied. It illustrates the main steps in the formulation of such models. The capabilities of this basic model are assessed by applying it to an LHP for which experimental results are available in the literature. The results show that this model is capable of at least qualitatively accurate predictions and could serve as a tool in preliminary designs of LHPs. / Second, experimental apparatus and procedures are designed and implemented for measurements of the following properties of sintered powder-metal porous plates that are used as wicks in LHPs: Porosity; maximum effective pore size; effective permeability; and effective thermal conductivity when saturated with a liquid (distilled water in this work). The aforementioned experimental apparatus and procedures are applied to two sintered powder-metal porous plates, one made of nickel 200 and the other of stainless steel 316, and the results are presented and discussed. / Third, an LHP with a flat (plate-type) evaporator and a fixed active mass of the working fluid (distilled water) is designed and constructed, and an experimental investigation of its steady-state operation is conducted. Full details of this LHP and the experimental setup and procedures are presented. The experimental results augment the available repertoire of experimental data on LHPs. They are used to test the predictions of the proposed network thermofluid model. / Fourth, an enhanced version of the aforementioned basic network thermofluid model is proposed and adapted for computer simulations of the above-mentioned LHP operating under steady-state conditions. / Fifth, results of experimental and computational investigations of the abovementioned LHP, operating with a stainless steel 316 wick and distilled water as the working fluid, are presented, compared, and discussed. The proposed network thermofluid model provides predictions that are within +/-10% of the experimental results.