A study of two-phase flow in annular channels with annular gaps of less than 1 mm is useful for the design and safety analysis of high power density systems such as accelerator targets and nuclear reactor cores. Though much work has been done on pressure drop in two-phase flow, designers rely mostly on empirical models and correlations; hence, it is valuable to study their applicability for different channel sizes, geometries, and gas qualities. The pressure drop along a concentric annular test section was measured for cases of either constant quality or variable quality along its length (such as in sub-cooled and flow boiling). A porous tube was used to inject gas along the inner surface of the annular channel, thereby simulating the case of flow boiling along the inner surface. The data were compared to predictions of various models and correlations. Additionally, the effect of wall vibrations on the pressure drop was examined. Experiments were conducted by imposing vibrations of known amplitudes and frequencies on the outer tube of the annulus. Wall vibrations were thought to be important for flow in microchannels where the vibration amplitudes may be significant compared to the channel hydraulic diameter. The results obtained in this investigation indicate that the pressure drop correlation given by Beattie and Whalley provides the best agreement with the data for both porous tube gas injection (i.e. variable quality) and constant quality two-phase flow within the narrow annulus. Furthermore, the results show that there is a minimal effect of vibrations on two-phase pressure drop over the range of frequencies and amplitudes studied.

This thesis is focused on two aspects related to wheeled mobility: 1) Evaluating the impact of a new tire design on powered mobility, and 2) Instrumentation that permits better monitoring and assessment of wheeled mobility in everyday use. The Tweel technology tires developed by Michelin USA are comprised of an outer polyurethane ring supported by polyurethane fins instead of metal spokes, which allow the tire to deflect under pressure. As a wheelchair tire they offer a potential breakthrough as they have deflection properties similar to a pneumatic tire while maintaining the low maintenance of a solid foam-core tire. A study was conducted to compare the Tweel technology tires to standard solid foam-core tires for vibration transmission, traction and overall life span. The Tweel technology tires failed produce any significant difference in vibration transmitted to the user compared to solid foam-core tires. Additionally, the Tweel technology tires showed significant signs of deterioration after a month long field trial, thus indicating a short life span. However, Tweel technology tires provided better traction on both dry and wet concrete. Overall, Tweel technology tires have to be re-engineered to provide better damping properties, leading to lower vibrational levels transmitted to the user. The second section this thesis addressed the need to develop a methodology of measuring mobility in everyday usage. This section is part of a greater ongoing research project at CATEA (Center for Assistive Technology and Environmental Access) aimed at understanding everyday wheelchair usage. Methodology was developed to measure bouts of mobility that characterize wheelchair usage; which includes the number of starts, stops, turns and distance traveled through the day. Three different technologies which included, Accelerometer unit on the rim of the drive wheel, Gyro-Accelerometer unit on the frame of the chair and Reed switches, were tested. Testing included various criteria for accuracy, durability and compatibility for measuring bouts of everyday wheeled mobility. Although a single technology could not be used to measure all aspects of mobility, the Accelerometer unit on the rim met the design criteria for measuring starts stops and distance, while the Gyro-Accelerometer unit met the requirements for measuring turns.

Misalignment in turbomachinery has been commonly known to produce twotimes- running-speed (2N) response. This project aimed to investigate the source of the 2N vibration response seen in misaligned vibrating machinery by simulating misalignment through a coupling. Three flexible disc-pack couplings (4-bolt, 6-bolt, and 8-bolt coupling) were modeled, and parallel and angular misalignments were simulated using a finite element program. The stiffness terms obtained from the coupling simulations had 1N, 2N, and 3N harmonic components. The 4-bolt coupling had large 1N reaction components under angular and parallel misalignment. The 6-bolt coupling model only had a 1N reaction component under angular misalignment, and both cases of parallel misalignment showed a strong 2N reaction component, larger than both the 1N and 3N components. The 8-bolt coupling model under angular misalignment produced large 1N reaction components. Under parallel misalignment, it produced 1N, 2N, and 3N components that were similar in magnitude. All the couplings behaved linearly in the range studied. A simple model predicted that the 2N frequency seen in the response is caused by the harmonic (1N) term in the stiffness. The amplitude of the 2N component in the response depends on the amplitude of the 1N term in the stiffness compared to the average value of the stiffness and the frequency ratio. The rotordynamic response of a parallel and angular misaligned system was completed in XLTRC2. When the frequency ratio was 0.5, the system response with the 4-bolt and 6-bolt coupling had a synchronous 1N component that was much larger than the 2N component. The response did not have a 2N component when the 8-bolt coupling was used but the response did have a 1.6N component that was considerably larger than the 1N component. When the frequency ratio was 2, the system response with the 4-bolt and 6-bolt coupling had a synchronous 1N component and a relatively small ½ frequency component. The response with the 8-bolt coupling had a 0.4N component that was larger than the 1N component. A 5-tilting pad journal bearing was also tested to better understand its behavior under misalignment because some experts attribute the 2N response to the nonlinear forces produced by bearings with high unit loads. The response of the 5-tilting pad bearing did not produce any 2N components while the bearing was subjected to unit loads of up to 34.5 bars.

Tests are performed on a flexure-pivot-pad tilting-pad bearing with a series integral squeeze film damper in load-between-pads configuration, with both active and locked damper. The damper effects are negated when locked, resulting in a flexure-pivot-pad bearing only. Experimental tests provide static performance data and dynamic stiffnesses from which rotordynamic coefficients are extracted. The following two excitation schemes are implemented: (1) multi-frequency, single direction excitation and (2) single-frequency, rotating load excitation (or "circular excitation"). The XLTRC2 Rotordynamics Software Suite provides stiffness and damping coefficient, eccentricity, and power loss predictions for the locked damper bearing. Test conditions include the rotor-speed range of 4000-12000 rpm and the unit-load range of 0-862 kPa (0-125 psi). Dynamic tests utilizing the multi-frequency excitation for the locked and active damper bearing configurations both show that the real portion of the dynamic stiffness is well modeled by a quadratic curve fit, and the imaginary portion representing the damping is a linear function of excitation frequency. This means that frequency independent coefficients can be obtained when an added mass term is included. While stiffness coefficients are lower for the active damper bearing, damping coefficients remain almost constant between the locked and active damper configurations. A simulation shows that, although the damping coefficients do not change significantly, the reduced stiffness provided by the damper results in greater effective damping. Static performance tests for the locked and active damper bearing indicate low cross-coupling, as shown by the eccentricity and low attitude angle measurements. Pad metal temperature measurements show a smaller temperature differential along the pad arcs for the active damper bearing, than observed for the locked damper case. Frictional power loss is estimated based on lubricant temperature rise and does not differ significantly for the two bearing configurations.

Self-excited chatter vibration in machining is one of the most important limitations on utilizing the increasing productivity of modern machine tools. In order to predict stable depth of cuts at high cutting speeds, the stability lobe diagram for a spindle-tool holder-tool combination must be developed. The frequency response function (FRF) of the system must be known for analytical prediction of the stability lobe diagrams. When the flexibility of the workpiece is important, the workpiece itself should be included in the system model by considering the variation of its dynamics at different stages of the machining process. In this thesis, an exact structural modification method is used to find the frequency response functions of the workpiece to be machined at every stage of the machining process. In order to obtain the system matrices and the modal parameters of the original structure, a commercial finite element program MSC. Marc&copy / is used. The frequency response functions of workpiece are calculated by using the computer program developed in this thesis, and are compared with the ones found by MSC. Marc&copy / . The stability lobe diagram of the system is obtained by combining the FRFs of the tool with those of the workpiece. The effects of the dynamic of the workpiece on the stability lobe diagrams are studied extensively by using the results of case studies presented in this thesis. In order to increase productivity, minimum chatter-free machining times are also calculated for different cases. For this purpose the effects of the different radial depth of cuts and different cutting strategies on the stability and the machining time are examined with various case studies.

The objective of this study is to develop and implement practical and accurate methods for prediction of the workpiece dynamics during a complete machining cycle of the workpiece, so that FRFs of the workpiece can be used in chatter stability analysis. For this purpose, a structural modification method is used since it is an efficient tool for updating FRFs due to structural modifications. The removed mass is considered as a structural modification to the finished workpiece in order to determine the FRFs at different stages of the process. The method is implemented in a computer code and demonstrated on representative parts such as turbine blades. The predictions are compared and verified with the data obtained using FEA. The FRFs are used in chatter stability analyses, and the effect of part dynamics on stability is studied.

Modern battle tanks are equipped with gun stabilization systems using gyro and encoder data to stabilize the gun barrel, although these systems are very sensitive and reliable, these systems assume the gun barrel as a rigid beam, and do not use information from

Les processus de mise en forme des matériaux par enlèvement de matière s'accompagnent de nombreux phénomènes vibro-acoustiques à la fois source de nuisances mais aussi révélant des informations utiles sur la coupe et sur la qualité du processus en cours. Dans le cas du déroulage, les opérateurs sont capables d'ajuster "à l'oreille" certains réglages de la machine en cours d'usinage. Une enquête auprès de professionnels a permis de cibler des modalités expérimentales, susceptibles de créer les situations caractéristiques pour les opérateurs. Des essais de déroulage ont été conduits sur la microdérouleuse du LABOMAP à partir de deux essences homogènes (hêtre et peuplier) et d'un matériau de référence (PTFE). Cette première étude a permis, dans des conditions contrôlées, l'identification de la signature de l'ouverture des fissures de déroulage dans les domaines temporel et fréquentiel. Les influences des paramètres de coupe sur le mécanisme de fissuration ont pu être clarifiées ou confirmées. La caractérisation de la qualité du bois à l'état vert demeure un problème, tant du point de vue technique que normatif. Le principe d'un outil simple de mesure de la position effective des fissures sur le placage a été proposé à partir de données vibratoires. Le comportement vibratoire et acoustique de la microdérouleuse en fonction des paramètres de coupe (épaisseur du placage, angle de dépouille, vitesse de coupe, taux de compression de la barre, essence, usure de l'outil) a été analysé dans les domaines temporels et fréquentiels. La prochaine étape passe par la validation de résultats obtenus à l'échelle d'un billon sur une dérouleuse de type industriel.

Le travail de cette thèse s'intéresse à l'identification des efforts (effort tranchant ou moment fléchissant)<br />présents aux limites des structures (plaques ou poutres) à partir de déplacements mesurés. Les expressions analytiques de ces<br />efforts contiennent des termes proportionnels aux dérivées spatiales du déplacement. L'approximation de ces dérivées aux<br />limites d'une structure est problématique car ces dernières sont très sensibles aux erreurs de mesure et les méthodes classiques<br />de calcul (différences finies, méthodes modales, transformée de Fourier spatiales, ...) et de régularisation (filtrage en nombre<br />d'onde, troncature modale) pour les obtenir ne sont pas adaptées pour une identification aux limites d'un domaine spatial.<br />Dans ce travail, une approche mathématique est proposée. Il s'agit de calculer les efforts aux limites, sans<br />utiliser directement leurs expressions analytiques qui contiennent les dérivées spatiales. La méthode est basée sur un calcul<br />intégral de l'équation de mouvement de la structure considérée, multipliée par une fonction test, qui vérifie des conditions aux<br />limites particulières. Ces conditions permettent d'extraire, lors du développement du calcul, les efforts recherchés à la limite du<br />domaine (efforts tranchant ou moment fléchissant).<br />La technique est développée dans les cas mono dimensionnel (poutre) et bidimensionnel (plaque). Des<br />simulations numériques illustrent la méthode, testent la robustesse de la méthode, les effets de différent bruits de mesure sur<br />l'identification, et établissent les limites spatiales et fréquentielles de l'approche. Des expérimentations ont été menées,<br />montrant la faisabilité de la méthode d'identification en utilisant des mesures réelles.

Cette thèse s'appuie sur une méthode multi-échelle dédiée à la modélisation et au calcul des vibrations moyennes fréquences : la "Théorie Variationnelle des Rayons Complexes", qui suppose initialement que la structure étudiée soit un assemblage de sous structures homogènes. Elle a déjà été validée dans le cas des assemblages de plaques et de coques tridimensionnels. Cependant, la plupart des structures industrielles présentent des inhomogénéités structurales comme des trous de passage de câbles, qui modifient sensiblement la réponse vibratoire en moyennes fréquences. L'objet de la thèse est de permettre leur prise en compte. La démarche proposée consiste enrichir l'espace des fonctions de base qu'exploite la méthode par des fonctions de correction adaptées à l'inhomogénéité.