The automotive industry is increasingly using resin transfer moulding (RTM) to produce composite body panels at high volumes and low costs. However, the processing parameters affect the physical and mechanical properties of the final finished part, which may have negative impact on the desired mechanical properties and surface finish. Unsaturated polyester (UP) mixed with organo nanoclay provides improved processing parameters which is a major concern for surface finish in RTM parts. Organo clays are silicate layers treated suitable to make them organophilic in order to be compatible with resin systems. UP systems give rise to numerous possible approaches in synthesizing nanocomposites. Three different techniques to synthesize nanocomposite systems, sequential mixing, solvent mixing and simultaneous mixing were studied and a suitable synthesizing technique was selected and used to make the nanocomposite systems. Two mixing approaches, three roll mill (TRM) machine and ultrasonication, were introduced to mix a pre-promoted UP resin/Low Profile Additive (LPA)/St system with organoclay. The effect of using these two approaches on the dispersion of the clays within the resin mix was studied using X-ray diffraction. Based on the results obtained from the X-ray analysis and viscosity measurements, ultrasonication was chosen as the optimum mixing method (OMM) for this system. One of the factors that influenced these results was the uncontrolled evaporation of styrene during the application of TRM. 0-5 wt% of Cloisite 20A was then added to the resin and mixed using optimum mixing method (OMM). The effect of adding clays on the viscosity, cure kinetics, storage modulus and glass transition temperature (Tg) was investigated. It was seen that the addition of clays increased the storage modulus of the resin samples, increased the viscosity, reduced the glass transition temperature very slightly and reduced the cure activity of the resin-clay system. Shrinkage tests on these / Le procédé d'injection sur renfort (RTM) est de plus en plus utilisé en industrie automobile pour produire de grands volumes de carrosseries à faible coût. Cependant, les paramètres du procédé peuvent avoir un effet négatif sur les propriétés mécaniques et le fini de surface de la pièce produite. Une résine polyester insaturé (UP) mixée à des nano-charges organiques confère de meilleurs paramètres de production, ce qui est une préoccupation majeure pour le fini de surface des pièces RTM. Les nano-charges sont des particules faites de couches de silicate traitées pour assurer leur compatibilité avec différentes résines. Plusieurs techniques pour synthétiser les nanocomposites sont possibles avec les systèmes UP. Trois techniques de mixage ont été étudiées : mixage séquentiel, mixage avec solvant et mixage simultané. La méthode la plus efficace a ensuite été sélectionnée et utilisée pour fabriquer des échantillons. Deux types de mixage, un moulin à trois cylindres et l'ultrasonification, ont été introduits pour mixer un système UP/agent anti-retrait (LPA)/styrène avec nano-charges. L'effet de ces deux approches sur la dispersion des nano-charges a été observé par la diffraction à rayons X. Suite à ces résultats et des mesures de viscosité, l'ultrasonification a été choisie comme étant la méthode optimale de mixage (OMM) pour le système étudié. Un des principaux facteurs qui a influencé ce choix était l'évaporation non-contrôlée de styrène durant la méthode TRM. 0% à 5% de nano-charges a ensuite été ajouté à la résine en utilisant la méthode OMM. L'effet de l'ajout des charges sur la viscosité, la cinétique de réaction, le module de conservation et la température de transition vitreuse (Tg) a été investigué. Il a été observé que l'ajout des nano-charges augmente le module de conservation et la viscosité, mais diminue légèrement la température de transition vitreuse et l'activité$

This thesis presents the theory required to estimate the available test time in the expansion tube with a reactive gas assuming perfect conditions. The superiority of the expansion tube over the non-reflected and reflected shock tubes will be shown. Further, a method of increasing testing times of combustible flows in the expansion tube through use of a chemical inhibitor is proposed and validated theoretically. The compound selected to illustrate the concept is CF$ sb3$Br. / In order to carry out the theoretical analysis of the expansion tube with reactive gas in the presence of trace amounts of inhibitor, the following empirical relation relating the induction period, $ tau,$ to the concentration of the reactants and the local gas temperature is used: tau=0.235 lbrack CH sb4 rbrack sp{0.33} lbrack O sb2 rbrack sp{-1.03} lbrack Ar rbrack sp0 lbrack CF sb3Br rbrack sp{1.35}e{ sp{25.82 over Re T}} / It will be shown that at free stream Mach numbers of approximately 6 with velocities on the order of 3000 m/s in the test region, an uninhibited mixture of argon diluted methane-oxygen cannot sustain the elevated temperatures of the post normal primary shock region without experiencing premature combustion. With trace amounts of CF$ sb3$Br (1% or 2% of total mixture), however, sufficient lengthening of the induction delay period occurs so as to afford testing time in the test region of the expansion tube. / Comparisons of three different uninhibited combustible mixtures--CH$ sb4$-O$ sb2$-Ar, C$ rm sb3H sb8$-O$ sb2$-Ar and C$ rm sb2H sb4$-O$ sb2$-Ar will also be made at lower operating conditions to demonstrate the marching action of the explosion line on the t-x wave diagram for the operation of the expansion tube with reactive gas. (Abstract shortened by UMI.)

This thesis examines the propagation of small relative errors in poles in discrete-time domains $z, varepsilon={{z-1} over {T}}, z sp prime=z-1, w sp prime={2 over {T}}{{z-1} over{z+1}}$ and z = ${{z-1} over{Tz}},$ where T is the sampling period, to the continuous-time domain. By prescribing pole locations in the discrete-time domains or usable sampling periods in a continuous-time context, sensitivity specifications in time constant of a real pole, natural frequency, damping ratio and uncertain relative region of a complex pole can be achieved. It is shown in this thesis that the alternative discrete-time operators, $ varepsilon, z sp prime, w sp prime$ and z, provide superior performance in the propagation of errors coming from coefficient quantization of first order control laws than the z operator at fast sampling rates, and possess sensitivity properties converging to those of an equivalent continuous-time system as the sampling interval approaches zero. A two-stage least squares identification process of a high order plant is studied with emphasis placed on sensitivity effects as well as on the effect of the accuracy of the digital-to-analog and analog-to-digital converters. The Euler form identification process is shown to yield the most accurate continuous-time pole estimates among the operator forms examined in this work and the accuracy of the converters is shown to bring an upper limit on the sampling rate, at which the data are captured for identification, so that relatively accurate pole estimates are obtained.

An analytical and experimental investigation was conducted to study the design and fabrication of carbon fiber track bicycle frames. A finite element software was used for the geometry development, laminate configuration, and for predicting failure using the maximum stress criteria. A load case and boundary conditions simulating actual riding conditions were developed. The stresses in each of the composite layers were found to be lower than the allowable stresses because of a properly designed geometry and laminate. Two composite frames were fabricated using the hand lay-up technique, using unidirectional and woven carbon fiber pre-preg material over an internal foam core. Using static testing techniques and comparisons with traditional tubular frames, the carbon fiber prototypes were shown to be better in all rigidity aspects. Combining the experimental and theoretical results, a good understanding of the critical problems related to composite monocoque bicycle frame design was obtained.

Flash X-ray and high speed photography were used to experimentally investigate the fragmentation process of single tin drops (0.5 g) in water during a steam explosion. Water velocities ranging from 9 to 28 m/s and a melt temperature of 800$ sp circ$C were used in order to study the effect of the water velocity on the drop fragmentation. Tests were also conducted in which the drop break-up was due entirely to hydrodynamic effects to evaluate the influence of the thermal energy on the fragmentation mechanism. / Results showed that there was a transition from thermal to hydrodynamic fragmentation through an intermediate stage in which the drops initially underwent hydrodynamic fragmentation followed by the formation of a vapour bubble. For low velocities (9 m/s) this bubble collapsed, fragmenting the remainder of the drop while at great velocities (15 m/s) the drop broke up within the bubble before it condensed. At 22 and 28 m/s there was no vapour formation and the drop fragmented due to hydrodynamic effects. The total non-dimensional break-up time was 2.0-2.5 for all cases.

In this thesis, we study the problem of large wrench application using robotic systems with limited force or torque actuators. It is shown that such systems may be able to apply a wrench in some configurations only; therefore their useful Force Workspace is limited, and may be smaller than their reachable workspace. / To improve the force capabilities of a system, base mobility or redundancy can be employed. A planning algorithm is proposed which results in proper base positioning relative to large-force quasi-static tasks. Similarly, the Force Workspace can be used to position such tasks relative to a robotic system. An efficient numerical algorithm is proposed to generate the Force Workspace, based on the 2$ sp{n}$-tree decomposition of the Cartesian space. Its efficiency stems from the variable resolution nature of the Cartesian space representation, and also from the use of four test levels that restrict the search to valid regions of the Cartesian space only. Examples of Force Workspaces are given for redundant and non-redundant planar manipulators, and spatial manipulators. / Next, the case of tasks requiring the application of a wrench along a given path is considered. The Task Workspace, the set of Cartesian space locations that are valid starting positions for such tasks, is shown to be a subset of the Force Workspace. / To plan redundant manipulator postures during large force-tasks, a new method based on a mini-max optimization scheme is developed. (Abstract shortened by UMI.)

Non-linear vibration of a spinning tether is studied in this thesis. The tether is thought to be a part of a spinning tethered satellite system in the station-keeping phase so that the tether has a constant nominal length and is forced to spin at a constant rate about its nominal axis. Using the extended Hamilton's principle the governing equations of motion are derived retaining non-linear terms up to the third order that originate from geometric non-linearity. They are discretized by the assumed-modes method, truncated to one-mode equations, and transformed to the phase-space form. Then the method of averaging is applied. / When the tether has high nominal tension, averaging with two variables results in a closed form solution, which shows dependence of the frequency contents on the initial amplitude parameters of the system. In the case of very low nominal tension, averaging with a single variable is useful to obtain the steady state and the limit steady state solutions, both of which result in a circular whirling motion like a skip-rope. Without damping, a general transverse mode appears to be quasi-periodic but it can be periodic under certain initial conditions. Numerical investigations reveal that the material damping through the longitudinal mode derives the steady state to the limit steady state. Also, several interesting shapes are observed in phase plots.

The aim of this Thesis is to obtain time-accurate solutions of the Navier-Stokes equations for laminar incompressible unsteady flows generated by oscillating boundaries in an annular region made from two concentric cylinders. For this, a time-dependent coordinate transformation is first used to obtain a fixed computational domain. The resulting governing equations in the fixed domain are discretized in real time based on a three-time-level implicit scheme. A pseudo-time integration with artificial compressibility is then used to reduce the resulting coupled discretized equations in delta form to set a decoupled scalar tridiagonal systems. / The method of solution has been applied to various 3-D unsteady flows in annular geometries, as well as to 2-D annular flows. The numerical results obtained are compared with those based on a mean position analysis, without transformation, for small-amplitude oscillation. This comparison shows that the time-dependent coordinate transformation is necessary to obtain accurate solutions for larger-amplitude oscillations. / The mean-position approach has also been applied to the analysis of axially variable annular configurations. The results obtained show more pressure recovery after a diffuser section with 6$ sp circ$ half-angle than in the case of 20$ sp circ$ half-angle. / A comprehensive experimental study was conducted to validate the theoretical results in the range of laminar flow. The results obtained were in good agreement with the numerical results, specially with those obtained by the time-dependent coordinate transformation. Experiments were also conducted for turbulent flow. / Based on the theoretical models developed, a computational method has been used to study fluid-structure interaction phenomena. It was applied to several cylindrical annular configurations in which one side of the annulus, the outer cylinder, is assumed to be flexibly supported, and thus to be susceptible to flow-induced instabilities. The structural and N-S equations were solved simultaneously by employing the numerical method developed for the unsteady flow and a fourth-order Runge-Kutta scheme for the structural motion. The numerical results thus obtained have predicted the stability of the structure for different annular geometries. The structure having a uniform annular geometry was shown to be more damped, while the annular geometry with a backward facing step is less damped. The study of the structure for a uniform annular geometry in the case of the rocking motion of the outer cylinder predicts an instability in the form of flutter of the outer cylinder.

This thesis is concerned with an experimental and modeling study of a novel drum dryer for black liquor utilizing multiple impinging jets of superheated steam. Drying of black liquor to solids contents in excess of 92% is a pre-requisite for some of the new black liquor recovery cycles being examined. / An experimental drum dryer was designed and built to evaluate the performance characteristics and effects of various operating parameters thereon. Appropriate ranges of parameters such as steam jet temperature and velocity were examined experimentally to quantify the optimal operating conditions for the formation of black liquor film on the drum surface as well as the drying kinetics. / To permit scale up of the laboratory scale data to industrial size a two-region mathematical model was developed to simulate the impinging jet flow and the drying of the black liquor film. Both polar and three dimensional cylindrical coordinate systems were employed for solving the governing transport equations. A modified low Reynolds number version of the $k sim varepsilon$ turbulence model was selected after a careful evaluation of the predictions of various $k sim varepsilon$ models for impingement flow and heat transfer. The computed heat transfer distribution on the film surface was applied as the boundary condition for the quasi unsteady one-dimensional diffusion model for the liquor film. Agreement between experimental data and results of the model was found to be satisfactory.

Inverse kinematics (IK) of serial manipulators is related to reflection-free Euclidean spaces where distances and orientations remain invariant. These direct isometries can be derived from simpler forms known as quaternions and dual quaternions. Combining them with Hermann Grassmann's Extension Principle, the IK of serial manipulators is formulated entirely in projective geometry free of any metric. The only rules governing this geometry is the preservation of ratios and incidence. Actually, the holonomicity of these IK problems can be described using incidence relations alone. / The algebraic constraints, derived from incident relations, define a manifestly holonomic system as opposed to general holonomic systems that need only satisfy Frobenius' Theorem for integrability using Pfaffian forms. The solution to these algebraic equations will require an introduction to algebraic geometry and commutative algebra. Definitions of basic geometric and algebraic objects along with a study of their respective properties are included. Pertinent theorems are proved and illustrated with simple examples to establish a dictionary between algebra and geometry. By this means, kinematic analysis is conveniently subjected to the theories of algebraic geometry and commutative algebra. / More precisely, the inverse kinematics (IK) of a 6R serial manipulator (6Rsm) is formulated in the homogeneous projective space of dual quaternions (DQ). This leads to a robust algorithm because the 8-space of DQ together with the Grassmannian extension ensures that only valid solutions, which satisfy all of the constraint equations, are admitted. Numerical examples, based on two real 6Rsm architectures, are presented to illustrate the efficacy of the new algorithm. Its fiber (solutions) and critical loci (singularities) are described.