<p>Eight tube arrays were tested to determine the effect of tube pattern and pitch on excitation phenomena in tube arrays in crossflow. These consisted of two different pitch-to-diameter ratios for each of the four standard heat exchanger tube array configurations. Measurements of vortex shedding frequencies showed that laminar vortex shedding may occur for some tube arrays at low Reynolds numbers and is associated with very high Strouhal numbers. Some arrays showed multiple values of Strouhal numbers at moderate Reynolds numbers, and over a small range of pitch ratios. Quite good agreement is found with some of the published results. Also, an improved correlation of the existing Strouhal numbers is obtained. The effect of tube pattern and pitch on the critical flow velocities at low damping parameter was determined. The results were used with other available results to determine the variation of the reduced velocity with the damping parameter for the different arrays tested.</p> / Doctor of Philosophy (PhD)

<p>A flow visualization technique has been developed to investigate the flow developments in tube arrays over a wide range of Reynolds numbers. The techniques is non-obtrusive and permits observations deep inside a tube bundle where the flow is unaffected by free surface or boundary effects. The technique has been used to examine flow behaviour in a rigidly and flexibly mounted square in-line and rotated square arrays in a water cross-flow. The important case of a single flexible tube in an otherwise rigid bundle, which received considerable attention in the literature, has also been examined. Results pertinent to vortex shedding, turbulence and fluidelastic instability are given which include response curves and frequency spectra together with flow visualization films and photographs.</p> <p>The results indicate that discrete alternate and symmetric vortex shedding can occur within the confinement of a tube bundle. Increased turbulence, as more rows are traversed by the flow, has a detrimental effect on the discrete vortex structures. Vorticity shedding and turbulence coexist as separate phenomena however, vorticity shedding is identified as the mechanism responsible for the observed resonance peaks in the response curves. Vortex shedding and the associated flow induced response in the square array is different from that in the staggered rotated square array. Fluidelastic instability is associated with marked increase in the transverse response amplitude and significant flow redistribution although it appears random in a full flexible bundle. Finally, a single flexible tube in an otherwise rigid bundle will become unstable at essentially the same flow velocity as that when the surrounding tubes are free to move. The response is predominantly in a transverse direction with significant flow redistribution. The motion of the redistributed flow lags behind the tube motion.</p> / Doctor of Philosophy (PhD)

<p>Accurate prediction of dynamic machining forces is essential in order to estimate the product and process quality, tool life and stability of the machining process. To achieve such a goal, all the aspects affecting the machining process dynamics should be considered. An integrated dynamic model that takes into consideration the dynamic effects of different machining process elements on the chip load was developed. This model considers the effect of cutter geometry, cutter initial position errors, spindle tilt, workpiece geometry, machine tool dynamics, and workpiece/fixture system dynamics. The open literature has no information on the effect of fixture dynamics on the chip load and the machining process stability. Proper modeling of the workpiece/fixture contact requires the modeling of the friction conditions because friction forces can be utilized to reduce the number of fixture components, thereby exposing more of the workpiece features to machining operations. Also, it provides a damping mechanism to dissipate input energy from machining forces out of the workpiece/fixture system. The literature lacks research on the tribological aspects of the workpiece/fixture contact. Since many factors influence the coefficient of friction, an experimental investigation was carried out to study the tribological aspects of the workpiece/fixture contact. Based on the findings of this study a velocity limited friction model (VLFM) was incorporated in the finite element analysis of the workpiece/fixture contact. The model utilizes NURBS curves and surfaces for the geometric modeling of the tool cutting edges and the workpiece geometry. A grid based meshing scheme for the finite element analysis was developed based on the NURBS iso curves. An in-house finite element code was developed for the analysis of the workpiece/fixture dynamics. Simulated cutting forces showed good agreement with the experimental validation. Two case studies are presented to demonstrate the practical application of the developed methodology. The integration of all of the above in one model provides an off-line tool to simulate and optimize the machining parameters and the fixture configuration. This cuts production time and cost.</p> / Doctor of Philosophy (PhD)

<p>The determination of principal finite strains from measurements made on a pair of deformed line elements is discussed. The deformation process is assumed to be either homogeneous or pure homogeneous in nature. Emphasis is placed on the pure homogenous mode since it leads to a simpler finite strain tensor and this technique is applied to determine the strain over the surface of an industrial stamping. The present work draws the distinction between homogeneous and pure homogeneous deformation. In the latter mode, an orthogonal triad can be identified (the principal axes), which remains orogonal throughout the deformation. An appropriate strain measure in such processes is that of logarithmic strain. Furthermore, its material derivative equals the rate of deformation tensor. Such a simple expression does not hold when the deformation gradient tensor is nonsymmetric, as in homogeneous processes.</p> <p>The material derivative of the tensor logarithm is no longer simply related to the rate of deformation tensor, and this is exemplified herein. The resulting expression involves the spin of the triad of the Eulerian and Lagrangian ellipsoids.</p> <p>Stress components vary as a result of material rotation and constitutive equations whereby rotational effect of material has been accounted for must be formulated. In finite deformation various "rotation" tensors can be defined. Consequently a wide choice of objective stress rates is available for adoption in constitutive equations, and a number of objective stress rates are examined herein. The utility of the resulting expressions is demonstrated for the case of a hypoelastic material undergoing finite deformation in simple (rectilinear) shear.</p> <p>Another aspect of this work has been an attempt to establish an approximate computer-aided technique for blank development, referred to as Geometric Modelling, and the investigation of possible strain distributions in forming sheet metal components. The technique is based on the initial assumption (this can be refined at a later stage) that a sheet metal component is transformed from a fiat sheet into a non-developable surface without change in thickness. Although few practical forming process occur in this way, many traditional die design procedures are based on similar notions; either there is no change in surface area or that a line length on the undeformed blank is unchanged during forming. Simple plasticity theory also suggests that the membrane stresses in a sheet would be minimized (in the absence of a normal stress) if no change in thickness occurred, therefore given the opportunity, the deformation is likely to take place in this ideal manner. The method of geometric modelling simulates the traditional manual calculations performed by experienced tool designers. The technique does not aim to replace the skill and experience of the designers but rather to enhance them.</p> <p>The present work describes the formulation of the fundamental theories of the method which comprise of the element-by-element mapping and remapping procedures and techniques of surface adjustment. The basic geometric assumptions employed in the development are also described. Two particular automotive stampings have been considered; one is the corner section of a car seat panel and the other is an inner deck-lid of a mid-size vehicle.</p> <p>A computer-aided design package for tool/die designers has been developed and the detailed analytical procedure is implemented in Fortran code. The analysis has been performed without access to advanced computer graphics. However, it is carried out in a way that future modelling using interactive computer graphics may well be attainable.</p> / Doctor of Philosophy (PhD)

<p>This thesis presents the results of pulsing flow experiments on solid-liquid suspensions. Specifically, the hydraulic energy requirement in pulsing flows for a given slurry flowrate is compared with the steady state energy consumption for the same flowrate. It is also shown theoretically that, for the same time-averaged flowrate as in the steady state case, it is feasible to reduce the time-averaged pressure gradient and hence power requirement by pulsing.</p> <p>The following five different types of flow systems were the subject of study: water, bentonite clay water, fine sand water, fine sand bentonite clay water and coarse sand water.</p> <p>For the above five types of flow systems the following pertinent flow variables were studied: pulsation frequency, pulsation amplitude, average slurry velocity, solids volume concentration, mean particle size and the presence of a small amount of ultra fine particles in coarse solid liquid mixtures.</p> <p>The experimental results of pulsing flow of water alone indicated that the pulsing to steady state power ratio (Jᴘ/Js) was always greater than unity indicating that for a Newtonian fluid the hydraulic energy in pulsing flow is greater than the steady flow, the increase of energy being used to maintain pulsation.</p> <p>Results of pulsing flow experiments with bentonite slurry have indicated that the power ratio (Jᴘ/Js) was observed to be less than unity for certain combinations of frequency, amplitude and slurry velocity. The minimum value of (Jᴘ/Js) found was about 0.75 at a frequency of about 0.80 Hz, amplitude of 51 mm and velocity of 1.9 m/s.</p> <p>In the ranges investigated, results with sand-water slurries have indicated that the power ratio (Jᴘ/Js) was a function of the variables listed above. The minimum value of the power ratio (Jᴘ/Js) for sand water slurries was found to be about 0.50 at the following values of the variables: pulsation frequency 0.45 Hz, amplitude 32 mm, slurry average velocity 1.8 m/s, solids volumetric concentration 12.5% and particle size 0.47 mm.</p> <p>The reduction in the power ratio in the case of pulsing flow of sand water slurries is attributed to the formation of a particle free layer near the pipe wall while the bulk of the particles flow as a plug in the middle of the annulus of the lubricating particle free layer.</p> <p>In the case of bentonite clay water slurry the reduction in the power ratio is attributed to the formation of a low viscosity layer near the pipe wall. The bulk of the slurry flows as a plug in the middle of the annulus of the low viscosity layer.</p> <p>Results of this thesis indicate that power consumption for transport of slurries using pulsing flow is considerably less than power required to transport the same amount of slurry by a steady flow method.</p> / Doctor of Philosophy (PhD)

<p>A simple theoretical model has been developed from first princples for the cross-flow induced fluid-elastic instability in heat exchanger tube bundles. A series of experiments were conducted to verify the basic assumption that only a single tube need be modelled in a flow channel which preserves the basic geometry of the array. The mechanism of dynamic instability is found to be one of flow redistribution due to transverse tube motion and a phase lag resulting from fluid inertia. Static instabilities in both streamwise and transverse directions are also predicted, but at higher reduced flow velocities than the transverse dynamic instability. Quite good agreement is found with available experimental data, particularly for parallel triangular and square arrays, without the need for empirical fluid force coefficients. The model includes the effects of tube array pattern and pitch.</p> / Doctor of Philosophy (PhD)

<p>A Flexible Manufacturing System is an integration of machine tools, material handling devices and computers into a system designed to improve productivity in the mid-volume production.</p> <p>Although a number of such systems have been designed, relatively little is know about the managing and optimal operating conditions of such systems. Even less is known about how future systems should be designed for optimal efficiency. The main objective here is to develop a user-oriented general purpose simulator which can be used as an experimental tool to gain insight into the problems posed by such complex systems.</p> <p>The developed FMSSIM simulator uses the discrete event simulation approach, incorporating the philosophy of the GASP simulation package in event monitoring. It is capable of simulating different configurations and topology including bidirectional tracks. It can check blockage of route due to interference of carts. It can also simulate random failures and repairs of the various components in the system. It provides the user with a wide range of priority rules to select from. It also allows the user to define his own priority rules.</p> <p>The main features of this simulator are: (1) Ease of use: FMSSIM is a user-oriented package where simulation is transparent to the user. Only a set of input data needs to be supplied. The program will perform the simulation and produce a report on various vital system performance statistics.</p> <p>(2) Graphical Output: The program produces a graphical output on a visual display terminal showing the movement of parts through the system.</p> <p>(3) Data Input: The program has a systematically organised data structure where information can be easily modified and retrieved.</p> <p>A series of experiments were conducted using this simulator to test its versatility and study the behaviour of a hypothetical Flexible Manufacturing System when various design parameters are changed. The response of the system to changes in transporter speed, in- and out-shuttles capacity and number of pallets was studied. Two different types of material handling systems viz. the cart and the conveyor were compared. The effects of introducing an additional machine and changing the system topology were also examined. The general trends of results obtained from the case study are in agreement with the findings reported in other published literature.</p> <p>The ability to simulate a wide range of systems with varied design parameters has made the FMSSIM simulator a very useful design aid which enables the system designer to analyse various parameters and arrive at a more efficient system.</p> / Master of Engineering (ME)

<p>The first part of this thesis contains a theoretical formulation and solution for unsteady flows (pulsatile and start-up flows) of non-Newtonian time-independent fluids through rigid pipes. The approach was based on the use of the equation of motion for axisymmetric unsteady flow of fluids in cylindrical coordinates. In the case of pulsating now, the unsteady behaviour of the pressure gradient was considered to be described by a periodic function of time, of sinusoidal form, added to a stationary pressure gradient, while for start-up flow the fluid was assumed to start its transient motion from rest due to instantaneous and sudden imposition of a stationary pressure gradient. The constitutive equation of generalized Bingham fluids was used since it represents the majority of time-independent fluids.</p> <p>A grid was imposed on the flow field in order to obtain a system of equations in finite difference form. The use of finite difference techniques provided detailed information about the time deformation of the pulsating and start-up velocity profiles as well as valuable information about energy consumption and flowrates under different pulsating flow conditions. The results are presented in the most general form so that they are widely applicable to any case where the assumptions and the boundary conditions are all satisfied. The main conclusion which can be drawn from the theoretical results is that the hydraulic power required to transport a fluid in pulsating flow is never less than that required for the same flowrate under steady flow conditions for all fluids except those which exhibit yield stress, i.e., Bingham fluids.</p> <p>On the experimental side, an investigation of pulsating flow of solid-liquid mixtures is presented. Solid-liquid mixtures are divided into two types, homogeneous suspensions and pseudohomogeneous/heterogeneous slurries. The terms homogeneous and pseudohomogeneous are used when solids concentration gradient along pipe vertical axis is constant for homogeneous flows and almost constant for pseudohomogeneous flows; while the term heterogeneous represents the cases where appreciable solids gradient along pipe vertical axis exists.</p> <p>The experiments were carried out for two types of solid-liquid mixtures. The first was a bentonite clay-water suspension with weight concentration ranges from 2.97% to 11.2%, while the second has coal/water slurry with weight concentrations between 5.34%-53.7%.</p> <p>The main aim of the experimental rig was to create a sinusoidal pressure gradient. The experimental set-up allowed different ranges of different flow parameters to be adopted, these were:</p> <p>1. Pulsing frequency ranging between 0-1.25 Hz.</p> <p>2. Pulsing amplitude (axial deformation of rubber bellows) of 34.6, 52.1 and 76.2 mm.</p> <p>3. Average flow velocity at 1.63, 2.18, and 2.63 m/sec.</p> <p>The effect of different combinations of these parameters on the ratio of the hydraulic power in pulsating flow to that in steady flow for the same throughput was studied.</p> / Doctor of Philosophy (PhD)

<p>This dissertation details a fundamental study of bubble growth dynamics for both spherical vapour bubbles in a uniform temperature field and hemispherical bubbles on a heated plane surface in a non-uniform temperature field in microgravity. The governing equations were solved numerically in order that accurate predictions of bubble growth could be generated for a wide range of system conditions for two fluids, water and R113. One dimensional spherically symmetric bubble growth in an initially uniformly superheated liquid was investigated by first developing grid and time step independent solutions on a computational grid with uniform spacing between adjacent nodes. These solutions were utilized as benchmark solutions for subsequent bubble growth models. A clustered grid arrangement was then implemented to reduce the computational time with insignificant loss of accuracy. The numerical predictions of the uniform and clustered grid arrangements predict the available analytic theories and experimental data with sufficient accuracy. The mechanisms which govern the growth of a spherical vapour bubble in an unbounded liquid were exposed by investigating the complex interaction between the heat transfer and the fluid flow surrounding a bubble as it grows from inception, through the various growth stages, to diffusion controlled growth. The influence of system pressure and liquid superheat were also investigated. Utilizing similar physical arguments, the model developed for spherically symmetric growth was extended to simulate hemispherical growth at a heated plane surface in microgravity. The theory was able to accommodate both spatial and temporal variations in the temperature and velocity fields in the liquid surrounding the bubble as it grows. Utilizing the present theory, the complicated thermal and hydrodynamic interactions between the vapour, liquid and solid have been manifested for a single isolated bubble growing on a heated plane surface from inception.</p> / Doctor of Philosophy (PhD)

<p>One important field of "Structural Dynamics" is "Modal Analysis Testing". This field, concerns the modes of vibration which constitute the link between experimental and analytical methods, like Finite Elements.</p> <p>The present work deals with the use of the experimental data obtained from Modal Analysis Testing in building a mathematical model of the structure, identifying its parameters and predicting the effects of possible design changes on its dynamics. It also deals with the use of these experimental data in predicting the behaviour of the structure according to a certain criterion. In this work, in particular, machine tool structures are considered.</p> <p>A systematic formulation of identifying the structural parameters from the Modal Analysis Testing is presented. It depends basically on the accessibility of the relevant coordinates for measurement. If all these coordinates are accessible then the formulation based on the equation of motion of the system and on the orthogonality relationships leads to linear equations. On the other hand if some of the relevant coordinates are missing, like coordinates on the bearings inside a headstock, then nonlinear optimization is used to minimize the errors between experimental and estimated modal parameters. These identification formulations are applied here to theoretical structures as well as actual machine tools.</p> <p>As a special exercise a procedure is suggested to be used in predicting the dynamics of a lathe with different workpieces using the modal data measured on a single workpiece. Such a procedure can help estimate the limit of stability against machining chatter beforehand and consequently could be implemented in the postprocessors of Numerically Controlled Turning Centers.</p> <p>The thesis deals also with further development of the theory of machining chatter. For the first time digital simulation in the time domain of the cutting process including chatter is carried out using mathematical models of machine tools established through Modal Analysis Testing. Cutting tests carried out in this work have shown that the digital simulation approach to machining chatter represents the reality very closely. Thus it could be used in formulating acceptance test procedures of machine tools as well as in designing the cutters to achieve higher metal removal rates.</p> / Doctor of Philosophy (PhD)