The validity of the quasi-static assumption was tested for the case of the leeward cylinder of a pair of parallel cylinders staggered in cross-flow. Static force coefficients were used to predict dynamic force coefficients according to the quasi-static assumption. These predicted values were then compared against actual measured values. / Tests were done for non-dimensional cylinder streamwise spacing, L/d, between 2 and 5, and corresponding cross-stream spacing, T/d, between 0.17 and 1. The range of reduced velocities U/fd, which were studied, was between 15 and 250. / Three regions were identified: one where the quasi-static assumption was always true, one where the quasi-static assumption was never true, and one where the quasi-static assumption improved with increasing U/fd. The experimental data indicates that necessary conditions for the application of the quasi-static assumption are a high U/fd as well as low dC$ sb{ rm L}$/dy and dC$ sb{ rm D}$/dy.

A pair of equilateral tetrahedra is formed by drawing all the face diagonals of a cube. Holding one tetrahedron fixed, the motion of the other one is investigated, subject to the constraint that edges formed by the diagonals of the same cube face remain coplanar, i.e., intersect or coincide. / Based entirely on the solution of equations arising directly from this constraint, displacement of the movable tetrahedron is expressed as rigid body rotation with translation. This original result forms the foundation upon which kinematic and dynamic analyses are built: (1) Edge pair intersections are expressed in closed, parametric form so as to facilitate further investigation. (2) A closed-form solution for the two degree-of-freedom workspace boundary is obtained by evaluating conditions under which these parameters approach zero or one. This important extension of previous research is necessary to control the motion of any hardware implementation of a double tetrahedral mechanism. (3) Inverse kinematics, which concerns the motion of such a mechanism, is presented. Two kinds of singularity, basic position singularity and overlapping singularity, are discussed from the point of view of the Jacobian matrices. (4) Given a wrench applied to the movable frame, the system of reactive forces or torques needed at the joints, i.e., the edge intersections, is determined. (5) Using the Lagrangian formulation, the inverse dynamics of this mechanism are derived and simulation results are presented. (6) The possible application of the double tetrahedral mechanism as a two degree-of-freedom manipulator joint is discussed.

No description available.

This work investigated the factors leading to the generation of process-induced stresses and deformations in woven composites manufactured by Resin Transfer Moulding (RTM). Both intrinsic and extrinsic mechanisms were examined. First, a comprehensive methodology was applied to characterize the thermal, chemorheological and thermomechanical properties of the CYCOM 890RTM epoxy resin. The developed models were essential to establish a clear understanding of the resin behaviour during the processing conditions. Then, the thermomechanical properties of 5-Harness satin woven fabrics were investigated using a micromechanical approach. Different fibre volume fractions between 50% and 62%, corresponding to typical volume fractions used in the RTM process, were examined. These thermomechanical properties were compared to the properties of equivalent unidirectional crossply laminates in order to validate the use of crossply configuration to model the behaviour of woven fabric laminate. The developed resin models and the fibre properties were then implemented in a finite element software, ABAQUS/COMPRO in order to predict the evolution of the laminate properties during the RTM process. Fibre Bragg Grating sensors were then used to characterize experimentally the tool-part interaction occurring during the RTM process. The separation of the composite from the mould during the cool down period was measured by the optical sensors. The maximal shear stress allowed by the laminate before the debonding was estimated to be 140 kPa for a steel mould. The observed tool-part interaction was simulated using frictional contact conditions at the composite/mould interface in a finite element analysis. Finally, using the ABAQUS/COMPRO interface, the manufacturing of a composite structure by RTM was modelled, taking into account simultaneously the intrinsic and extrinsic mechanisms examined. This case study demonstrates the capacity of the process mod / Ce travail étudie les facteurs conduisant au développement de contraintes et déformations résiduelles induites par le procédé d'injection sur renforts dans le cas de composites à renforts tissés. Les mécanismes intrinsèques ainsi qu'extrinsèques ont été pris en considération. Tout d'abord, une méthodologie a été utilisée pour caractériser les propriétés thermo-chimiques et thermo-mécaniques de la résine époxy CYCOM 890RTM. Le développement de modèles mathématiques pour décrire les différentes propriétés est une étape essentielle pour comprendre clairement l'évolution du comportement de la résine lors du procédé de fabrication. Ensuite, les propriétés thermo-mécaniques d'un tissu satin de 5 ont été examinées en utilisant une approche de micromécanique sur un motif élémentaire. Différentes fractions volumiques de fibre entre 50% et 62%, correspondant aux fractions volumiques de fibre typiquement utilisées pour le procédé d'injection sur renfort, ont été étudiées. Les propriétés ont été comparées aux propriétés obtenues pour des laminés unidirectionnels croisés équivalents. En général, les propriétés des deux configurations sont comparables. Ceci justifie l'utilisation des propriétés de fibre unidirectionnelle dans une configuration croisée pour simuler le comportement d'un laminé à renfort tissé satin de 5. Les modèles des propriétés de la résine et des fibres ont été ensuite implémentés dans un logiciel d'éléments finis, ABAQUS/COMPRO, afin de prédire l'évolution des propriétés du laminé pendant le procédé d'injection sur renforts. Des fibres optiques à réseaux de Bragg ont été également utilisées pour caractériser expérimentalement les interactions entre le composite et le moule générées pendant le procédé. Les fibres optiques saisirent la séparation entre le composite et le moule pendant le refroidissement. La contrainte

Clinical investigations of the effects of dehydration on vocal fold tissue rheology have suggested that vocal fold hydration level has a significant influence on vocal fold vibration, voice quality and vocal health in general. Most previous investigations of the influence of hydration on mechanical properties have been qualitative. Few have used a quantative approach to analyze and estimate water transport from the human larynx. A numerical study of airflow and water transport was thus conducted. The numerical methods were verified by performing experiments and numerical simulations of water transport in a rectangular channel. Two-dimensional and three-dimensional computer models of the human larynx were created to simulate and analyze airflow through the larynx during inbreathing, and estimate water transport due to convection. The results were found to be consistent with available data in the literature, and experimental data. Parameters such as the inlet pressure, the glottal angle and the orifice width were varied and their influence on water transport was quantified. The results may be useful in assessments of the role of water convection on vocal fold dehydration. / Des études récentes des effets de la deshydratation sur les propriétés mécaniques des cordes vocals indiquent une forte influence sur les vibrations des tissus, l'hygiène vocale, et la santé en général. La plupart des etudes publiées sur ce sujet sont qualitatives. Le but de ce travail de recherche fut d'évaluer de façon quantitative la mas se d'eau transportée par convection dans l'écoulement d'air passant par la glotte durant la production vocale. Des methods de simulation numérique ont été utilisées, validées par des mesures expérimentales pour un cas simple d'écoulement en conduite rectangulaire. Les modèles numériques bi- et tri-dimensionels furent alors appliqués à des cas d'orifices avec des géometries similaires à la glotte. Les résultats semblent corroborer les données éxistantes publiées à ce jour. L'influence de parameters tels que la pression pulmonaire, la géométrie de l'orifice, et sa largeur fut étudiée. Ces resultats permettent une meilleure evaluation de la quantitée d'eau perdue par les cordes vocales pendant le discourse du à la convection massique.

This thesis elaborates three aspects in the field of flow-induced vibrations associated with annular geometries. / A method to find the unsteady fluid forces on a cylinder oscillating in annular turbulent flow is developed by considering the superposition of the turbulent fluctuating quantities on potential flow. The theory is compared with experiments. / Then, the unsteady fluid forces acting on the vibrating cylinder walls of non-uniform annular configurations are computed by a method which performs the accurate time integration of the Navier-Stokes equations. It is the extension for unsteady flows of the method of artificial compressibility used for steady flows. A time-discretization of the momentum equation using a three-point-backward implicit scheme is introduced, and the addition of pseudo-time derivative terms to the semi-discretized equations, including artificial compressibility in the continuity equation, allows to use time-marching solution techniques thereafter. / Finally, the integration method used for the Navier-Stokes equations is combined with the equation governing the dynamical behavior of a structure in order to perform the fluid-structure stability analysis of this system in the time domain.

AQUA is an underwater hexapod robot which uses its paddles to propel itself and to control its orientation. The use of oscillating paddles for propulsion and control represented a novel and challenging problem, which motivated the need for a simulation of the motion of the robot based on its paddle oscillations. The most difficult aspect of this simulation was the characterization of the forces generated by the paddles oscillating in the water. In this work, a model predicting the forces produced by an oscillating rigid paddle was developed and validated experimentally. Also, the forces produced by a flexible fin were determined experimentally and were compared to those generated by the rigid paddle. Rigid paddle and flexible fin experiments were performed on an experimental setup, which was designed and built to measure the forces and torques produced by a paddle oscillating in a water tank. Finally, a simulation of the AQUA robot was developed, based on the validated rigid paddle model.

The precursor shock wave appearing in high-explosive channel detonation can be prevented at high initial fill pressures as has been previously demonstrated by experiments. However, the theoretical and numerical results from previous works have only achieved qualitative agreement. It was suggested to introduce real-gas equations into the analytical model and numerical simulation to account for the high density, pressure, and temperature nature of the problem. In this thesis, two parametric studies are performed on the ideal gas analytical model developed in previous works, and the parameters being examined are the specific heat ratio (γ) and the total pressure loss of air and detonation products. The results show that (1) the real-gas effects are important in low air-to-explosive area ratio, and (2) the analytical model has to be modified to include the shock reflections for better predictions. Following the suggestions from previous works, real-gas equations are introduced into the analytical model. The selected equations are the Jones-Wilkins-Lee equation of state (JWL EOS) for detonation products and the polynomials with NASA Glenn coefficients for specific heats in air. An algorithm is developed to determine the critical area ratio at a given initial pressure, and results are similar to the ones from previous works. However, numerical simulations performed by an Euler solver incorporating the JWL EOS for the detonation products show improvement in results and confirmed the real-gas effects are important for the high-explosive detonation studies in small area ratio channels. / L'onde de choc des précurseurs apparaissant à la détonation d'explosifs de grande puissance peut être empêché par la haute pression de remplissage initial et a été démontrée par des expériences précédement. Toutefois, les résultats théoriques et numériques des travaux précédents ont seulement atteint un accord qualitatif. Il a été suggéré d'introduire des équations réelles du gaz dans le modèle analytique et le simulation numérique pour confronter le problème. Dans cette thèse, deux études paramétriques sont effectuées sur le modèle analytique de gaz parfait développé dans les travauxs précédentes, dont les paramètres examinés sont le taux de chaleur spécifique (γ) et la pression de stagnation de l'air et les produits de détonation. Les résultats montrent que (1) les effets réels du gaz sont importantes où le rapport de surface air/explosif est plus petit, et (2) le modèle analytique doit être modifiée pour tenir compte des réflexions de choc afin d'obtenir de meilleures prévisions. Suite aux suggestions des travaux précédents, les équations réelles du gaz sont introduites dans le modèle analytique. Les équations retenues sont l'équation d'état de Jones-Wilkins-Lee (EOS JWL) pour les produits de détonation, et les polynômes avec les coefficients de NASA Glenn pour l'air. Un algorithme est dévéloppé pour déterminer le rapport de surface critique à une pression initiale donnée, et les résultats sont similaires à celles du modèle des gaz parfaits. Toutefois, les simulations numériques effectuées par un solveur Euler intégrant les EOS JWL pour les produits de détonation démontrent une amélioration dans les résultats et confirment que les effets réels du gaz sont importantes dans les canaux avec un petit rapport de surface.

This thesis presents experimental and theoretical results on the nonlinear dynamics and stability characteristics of thin-walled clamped-clamped circular cylindrical shells in contact with flowing fluid. Additionally, the large-amplitude vibrations of clamped shells subjected to an external radial harmonic force excitation are also investigated theoretically. / The experiments were conducted with three experimental set-ups: one for experiments with elastomer shells in annular air-flow, the second for elastomer shells with internal air-flow, and the last one for aluminum or plastic cylindrical shells with internal water-flow. In all cases the interaction between the shell and the fully developed flow gives rise to instabilities in the form of static or dynamic divergence at sufficiently high flow velocities. The experimental results show a softening type nonlinear behaviour, with a large hysteresis in the velocity for the onset and cessation of divergence. Only static solutions were observed. / In the theoretical study, the standard form of the Donnell's nonlinear shallowshell equations is used. The solution expansion includes both driven and companion modes, thus allowing for a traveling wave in the circumferential direction; it also includes axisymmetric modes to capture the nonlinear inward shell contraction and the correct type (softening) nonlinear behaviour observed in the experiments. The clamped-clamped beam eigenfunctions are used to describe the axial variations of the shell deformation, automatically satisfying the boundary conditions and the circumferential continuity condition exactly. The fluid is assumed to be incompressible and inviscid, and the fluid-structure interaction is described by linear potential flow theory. The theoretical results indicate loss of stability by divergence, and a strong nonlinear subcritical response. In the case of external excitation the shell exhibited a softening type of nonlinearity. / The results from the theoretical model are compared with experimental data, and in all cases good qualitative and quantitative agreement is observed. In addition, a parametric analysis investigates the effect of the length/radius and thickness/radius ratios and transmural pressure on the nonlinear response of the shell with internal flow.

The conditions that are required for DDT are studied in the present thesis by focusing on the final phase of the onset of detonation. A series of experiments were carried out where high speed turbulent deflagrations were generated downstream of a perforated plate by reflecting a CJ detonation upstream from the plate. Four different detonation tubes were used to investigate scaling effects. Four different fuels (acetylene, hydrogen, propane, and ethylene) were used with either pure oxygen or air as an oxidizer to investigate the effect of mixture sensitivity on the onset of detonation. In some cases, fuel-oxygen mixtures were diluted with either nitrogen or argon. Perforated plates with various hole diameters and hole spacings were used in order to control the turbulence parameters in the reacting downstream flow. A theoretical study was also carried out to determine the various flow fields as well as the ignition processes via turbulent mixing. / It is found that a high speed turbulent deflagration traveling at about the CJ deflagration speed of the mixture (around 800 to 1200 m/s, depending on the given mixture) is required for the onset of detonation in a smooth tube. This critical ve locity is found to be rather insensitive to boundary conditions, which supports the conclusion that critical deflagrations are driven by the expansion of the combustion products (similar to a CJ detonation) and are dependent on the energetics of the mixture rather than on transport rates. The ignition mechanism in a critical deflagration differs from that of a CJ detonation and is effected by the turbulent mixing of combustion products with the unburned mixture. Therefore, intense turbulence is required in the reaction zone. Since the initial plate generated turbulence decays, the sustained propagation of a critical deflagration wave in a smooth tube relies on "flame generated turbulence" where the high temperature sensitivity of the mixture can mantain the turbulence in the reaction zone. It is also found that the critical deflagration is comprised of a leading shock wave followed closely by an extended turbulent reaction zone rather than the conventional shock wave followed by a distinct thin flame. / Once a critical deflagration wave is established, the onset of detonation requires the generation of transverse waves that couple to the reactions and are sustained by the chemical energy release as in a cellular detonation. The onset of detonation can be effected by the amplification (via the Rayleigh criterion) of repeatedly reflected transverse waves resonating with the reaction zone in a confined tube. It can also occur via the formation of a local explosion centre where an overdriven detonation bubble and the accompanying retonation wave is formed in the turbulent reaction zone.