<p>Exhaust gas recirculation (EGR) cooiing devices are used in EGR systems to significantly reduce NOx emissions from diesel engines. However, the thermal and hydraulic performances of these devices change over time during operation due to the deposition of soot from the diesel exhaust gas in these devices. The objective of this work was to investigate in detail the heat transfer and the soot deposition characteristics in diesel engine EGR cooling devices. Controlled soot deposition tests were performed on a series of generic single-tube and three-tube EGR cooling devices that were exposed to exhaust gas for ifferent periods of times in the diesel exhaust test facility to investigate the time-dependent heat transfer and pressure drop characteristics of these devices. A non-destructive technique was developed to characterize the three dimensional soot layers that occurred in the EGR cooling devices using the neutron radiography test facility of the McMaster Nuclear Reactor. The results showed that the deposited soot layer in the single-tube devices was much larger for the higher flow rate which corresponded to turbulent flow, particularly in the entrance region. The thickness of the deposited soot in the entrance region of the tube was much larger than the remainder of the tube, and appeared to restrict the flow area by up to 75% over a length of 4 tube diameters. The variation in the deposited soot thickness throughout the rest of the tube is similar in magnitude to the layer itself indicating that the soot layer in the tube was three- imensional. The results also showed that the profiles of the deposited soot had a wave-like structure in all cases. The thermal effectiveness of the devices decreased from approximately 82% to % over 5 hours for the laminar flow test case, and from 70% to 35% for the turbulent flow test. The pressure drop across these devices increased by up to 320% during the 5 hours of testing. Measurements were also performed using three-tube EGR cooling devices with 45° and 60° expansion angle inlet headers exposed to diesel exhaust for 3 hours. The soot deposition was more evenly distributed in the tube bundle with the 60° expansion angle inlet header suggesting that the flow may be more evenly distributed in this device. There was more soot deposited in the center tube than outer tubes in the three-tube bundle with the 45° expansion angle inlet header suggesting that the flow was not evenly distributed. The thermal effectiveness in both cooling devices decreased during the 3 hours of testing, but the change in the effectiveness was larger for the device with the 45° expansion angle inlet header. The pressure change in the exhaust gas across the cooling devices increased more rapidly for the device with the 45° expansion inlet header, reaching a nearly constant value of nearly 5 times the initiai value after 3 hours.</p> / Doctor of Philosophy (PhD)

<p>Coordinate measuring machines (CMM) are widely used in inspection and reverse engineering. There is increasing interest in using optical sensors, such as laser digitizers, to take advantage of their high data acquisition rates and the fact that they are a non-contact measurement suitable for soft materials or delicate parts. The issue to be addressed is the fact that the sensors and CMMs are made by different manufacturers, so there is no integration of the error compensation scheme. As well, unlike conventional touch trigger probes, correction of the laser digitizer pose must be explicitly incorporated. The purpose of this study was to develop a mathematical algorithm that would improve the accuracy of data collected with a laser digitizer mounted on a CMM. To this end, an algorithm was developed to compensate for the variations in digitizer pose. It compensates for both positional errors and angular errors, whether caused by CMM geometric component error or quasi-static thermal error. To implement and verify the error compensation, tests were made using a Hymarc laser digitizer mounted on a DEA Iota 1102 CMM. The thesis proposes using an optical ballbar, an adaptation of the traditional ballbar recommended by ASME Standard B89.4. The scan plane coordinates of the digitizer. the CMM axis scale positions and temperatures were recorded simultaneously in real time. The CMM error compensation algorithm was then used to post process the data to obtain improved global part coordinates. Accuracy of a laser equipped CMM is improved from 67 um to 17 um. The benefits of using optical sensors can be improved by incorporating this error compensation scheme into an integrated optical metrology system.</p> / Doctor of Philosophy (PhD)

<p>An adaptive forging schedule procedure that targets forging simple 3-D external profiles on a flexible forging cell is presented. The cell is composed of a numerically controlled press, a robot for workpiece manipulation, a shape feedback system and a main supervisory computer. The scheduling procedure utilizes the shape error, computed from the required and the initial workpiece shapes, to plan and execute a set of forging steps. The feedback system is then used to capture the deformed shape of the workpiece and update the shape error to adaptively plan the next set of forging steps. This adaptive control loop iterates until the workpiece shape is forged within a preset tolerance. The reduction in height and the bite values used in the schedule are limited by a set of technological bounds. Several of those bounds were identified and modelled as constraints on the adaptive scheduling procedure. Illustrative case studies showed that the adaptive scheduling procedure successfully forged external profiles (given an arbitrarily set tolerance limit) when a-priori approaches failed. The studies also showed that the new scheduling procedure could forge 2-D and 3-D profiles as easily (in terms of programming effort) as forging a uniform billet. Three new spread prediction models are developed, of which the newly developed Incrementally Updated Upper Bound model proved to provide the most accurate spread estimate compared to other developed and traditional spread prediction models. This model was then chosen to be integrated with the adaptive scheduling procedure to form an intelligent forging schedule procedure. The integration provided predictive capabilities to the adaptive scheduling module and enabled the generation of better quality forging schedules. In a typical example the number of forging steps in the forging schedule were reduced from six (using the adaptive scheduling procedure) to two (using the intelligent scheduling one).</p> / Doctor of Philosophy (PhD)

<p>The cam-follower mechanism is a most important machine element used to control the complicated motion of machine parts, and the accuracy of the motions and subsequent machine performance depends on the cam-follower output. Imperfections in the system affect this output, and an investigation of all possible effects is carried out here through the analytical and simulation means developed.</p> <p>The imperfections considered here include all factors that cause a design to deviate from the theoretical model. These factors have generally been neglected in the past because of their complexity. They are classified in the following three categories:</p> <p>1. Geometrical Imperfections</p> <p>a) Machining or manufacturing errors: Tolerances</p> <p>b) Backlashes or clearances</p> <p>2. Kinematic Imperfections</p> <p>a) Non-constant angular velocity of cam</p> <p>b) Impossibility of a perfectly manufactured cam profile</p> <p>c) Asymmetry between rise and return period</p> <p>3. Dynamic Imperfections</p> <p>a) Inertial mass</p> <p>b) Flexibility of system elements</p> <p>c) Energy dissipation</p> <p>The performance of the system is indicated by its responses, such as time responses (displacements, velocities and accelerations) and inertial responses (contact forces or stresses, torques and inertia forces).</p> <p>To investigate the effects of the imperfections, the mechanism of a translating roller follower-cam system driven by a motor through a rigid coupling is simulated on a digital computer dynamically and stochastically. The dynamic simulation, which will produce the effects of the kinematic and dynamic imperfections, is accomplished on the basis of an eleven degree of freedom model which was carefully devised to include all possible properties or factors concerning the real system behaviour, such as flexibility of cam shaft, non-constant angular velocity of the cam due to torsional vibration, jump phenomenon caused by inertia force, preloading of the retaining spring, cantilever effects of follower due to pressure angle, nonlinear damping, and the exact contact condition between cam and roller follower. In the model, an analytical method to calculate the spring constant of the interface between cam and follower is introduced by using Hertzian deflection characteristics. To analyse the effects of geometric imperfections which involve random characteristics, the ground and finished cam profile is modelled stochastically by generating normally distributed random numbers and applying a cubic polynomial spline function to obtain the cam profile.</p> <p>Thus, the compounded effects of tolerances and flexibility in the system can be investigated.</p> <p>The motion equations derived from the dynamic model and stochastic model consist of simultaneous nonlinear differential equations in which the factors having a random nature are implicitly included. To solve the motion equations the refined Runge-Kutta algorithm is utilized so that the computing accuracy can be controlled.</p> <p>The PDP-11/34 minicomputer and its graphic peripheral devices are exploited by the overlay technique. Intermediate results are transferred to subsidiary memory while renewing the previously executed memory, so as to diminish the processing cost of the dynamic and stochastic simulation program as well as to compensate for the insufficient main memory.</p> <p>Finally, the results of the simulation are analyzed and compared with the work of other researchers, in which the effect of an imperfect profile has usually been neglected.</p> / Doctor of Philosophy (PhD)

<p>Discontinuous yielding in the form of Lüder's bands in frequently observed in tensile testing to mild steel sheets. In these materials, instability or kinking may or may not be observed in bending operations and various theories have been proposed to relate the bending and tensile phenomena.</p> <p>In this work, an extensive experimental investigation was performed on five samples of low carbon sheet steels. The main objective was to develop a new analytical model, based on the uniaxial tensile stress-strain curve that will predict the behaviour of these materials in pure bending. Two models appeared possible and the experiments performed suggested that one of these is an appropriate model for steels of this kind.</p> / Master of Engineering (ME)

<p>Modern trends in gas turbine design have been geared to producing a compact high work output engine as a power unit. To achieve this, the turbine rotor must be comprised of high turning angle blades. Secondary flow has always accounted for a major portion of the losses within a blade row and high turning angle blades tend to reinforce these losses. A review of the current literature on secondary flows shows that there is a void of data for two-dimensional high turning angle turbine blade cascades.</p> <p>The object of this study was to construct and experiment upon a large scale two-dimensional high turning angle turbine blade cascade. Each of the four blades within the cascade had a chord and span of 214 mm and 188.5 mm respectively. The overall test rig was an open circuit wind tunnel which was operated at a Reynolds number of 3.68 X 10⁵ (based on the blade chord) which corresponded to a mean blade inlet velocity VI of 26.5 m/s (87 ft/s). The rig was designed with the facility to adjust the inlet flow in the pitchwise direction to ensure periodicity. Suction chambers were installed on the inlet channel endwalls to control the boundary layers and hence observe their effect on the measured secondary losses. The flow parameters were measured and recorded at the inlet bellmouth throat, at the inlet and the exit from the cascade and at a distance of 1.5C in the streamwise direction from the cascade exit.</p> <p>A blade passage surface static pressure distribution is presented and the exit flow parameters are given in the form of contour of total pressure loss coefficients and static pressure coefficients. The spanwise distribution of pitchwise averaged flow parameters is given and the measured losses are reported and compared with existing correlations. The measured losses were found to be significantly lower than those predicted by these correlations.</p> / Master of Engineering (ME)

<p>This research develops a new methodology for the control of two link flexible arms. The methodology is referred to here as "Nonlinear Integrated Tabular" NIT Control. Starting with a new trajectory, the inverse kinematic problem is solved and adequate algorithms are presented to evaluate the commanding voltages to guarantee that the wrist point will track the desired trajectory. Continuous and discontinuous paths are tried with models which were developed to account for the flexibility in the joints and the flexibility in the links. Lagrange's equation and finite element formulation were used to construct realistic models. Computer simulations were conducted to assess the control qualities for tracking using NIT. The accuracy, smoothness as well as the levels of the command signals during tracking were quantified to serve as criteria to evaluate the quality of tracking.</p> <p>A number of classical controls were developed and applied to the elbow arm to compare NIT with the current available approaches. A proportional plus derivative plus gravity, a feedback linearization as well as a VEE-MOD-VEE controllers were tried. It was shown that NIT compares well with the most successful conventional methods. In addition, it was found to offer simple, smooth, robust, computationally economic controls. NIT compares well with the majority of the conventional methods.</p> <p>An experimental setup was constructed with the objectives of verifying the analytical predictions. It was built in a modular form to allow continuity of research in this area at McMaster's laboratories. Great care and long hours were spent in the design, fabrication, assembly, debugging of the hardware for the mechanical parts. The setup is referred to here as FLEXROD. The up-to-date technologies available today were implemented in the design of FLEXROD which include Harmonic Drives for speed reduction and transputers for the controller. Enormous efforts went into streamlining the software and in developing the interfaces between the sensory system and the controller. Experiments were conducted to validate the transient dynamics of the elbow arm when the motors are brought to a sudden stop by applying the brakes simultaneously. Reasonable agreement was found between the analysis and the experiment. The thesis is concluded with recommendations for future research in this area to implement and develop NIT using FLEXROD.</p> / Doctor of Philosophy (PhD)

<p>The U-bend region of nuclear steam generators tube bundles have suffered from two-phase cross flow induced vibrations. Tubes in this region have experienced high amplitude vibrations leading to catastrophic failures. Turbulent buffeting and fluidelastic instability have been identified as the main causes. Previous investigations have focused on flow regime and two-phase flow damping ratio. However, tube bundles in steam generators have vapour generated on the surface of the tubes, which might affect the flow regime, void fraction distribution, turbulence levels and tube-flow interaction, all of which have the potential to change the tube vibration response. A cantilevered tube bundle made of electric cartridge heaters was built and tested in a Freon-11 flow loop. Tubes were arranged in a parallel triangular configuration with pitch to diameter ratio of 1.48. The bundle was exposed to two-phase cross flows consisting of different combinations of void from two sources, void generated upstream of the bundle and void generated at the surface of the tubes. Tube tip vibration response was measured optically and void fraction was measured by a gamma densitometry technique. It was found that the ratio of tube vibration amplitude in the transverse direction was reduced by a factor of 8 for void fraction generated at the surface of the tubes only, when compared to the response observed under the upstream only void generation case. The main explanation for this effect is a reduction in the correlation length of the turbulent buffeting forcing function. Theoretical calculations of the tube vibration response due to turbulent buffeting under the same experimental conditions predicted a similar reduction in tube amplitude. The void fraction for the fluidelastic instability threshold in the presence of tube bundle void fraction generation was higher than that for the upstream void fraction generation case. The first explanation of this difference is the level of turbulent buffeting forces to which the tube bundle is exposed. Increased values of turbulence will lower the void fraction for instability. The second explanation is related to the flow regime. In this study, it was clear that flow regime for bundle void generation was at all times bubbly and homogeneous, while the upstream void fraction generation cases showed a clear tendency to churn flow. A change in flow regime from bubbly to churn flow will produce the same effect as an increase in turbulence buffeting levels, and hence it seems difficult with the present knowledge to distinguish between the two causes. In as much as turbulence levels are related to flow regime, it is essential to have a clear knowledge of the flow regime in steam generators in order to predict the fluidelastic instability threshold of the tubes.</p> / Doctor of Philosophy (PhD)

<p>A unique rotary machine, which can be modified to function as an air motor, a hydraulic pump or a rotary combustion engine, has been developed. This thesis is concerned with both theoretical modelling of the engine aspects of the machine and the testing of a model and prototype. The machine is a simple rotational machine with an internal parallel axis arrangement. It is unique in the sense that there are two rotors turning in the same direction on their own stationary centres of gravity and working with the fluid around a full 360° degree path.</p> <p>Using CAD/CAM techniques a model of the air motor and a prototype hydraulic pump were built and tested. Both of them have demonstrated very good performance features.</p> <p>The engine described in this work is a true rotary engine with a unique semi-internal/external combustion system. It has a large power/weight ratio with a unique eight power strokes per revolution feature. Its multi fuel capability is a special contribution to the development of internal combustion engines where and when it is necessary to use alternative fuels.</p> <p>A three zone turbulence eddy entrance thermodynamic model has been developed to predict the performance of the engine and optimize some of the structural and operational parameters. The results of the theoretical analysis indicate that this engine has excellent features which are improvements over those of the reciprocating engines as well as the Wankel rotary engine.</p> / Doctor of Philosophy (PhD)

<p>Fundamental problems in two different flow regimes; subcooled water-steam bubbly condensing flow and subcooled flow boiling, in vertical conduits under low pressure and mass flux conditions, were investigated.</p> <p>For subcooled water-steam condensing bubbly flow, experiments were carried out to obtain a data base for the axial distribution of area-averaged void fraction, interfacial area concentration, interfacial condensation heat transfer and bubble relative velocity. The interfacial area concentration was obtained by measuring the distributions of bubble volume and surface area, using high speed photography and digital image processing techniques, as well as the area-averaged void fraction, using a single beam gamma densitometer, at various axial locations. The interfacial heat transfer coefficient was obtained by monitoring the rate of change of the volume of individual bubbles as a function of local conditions. The data was used to develop new correlations for interfacial area concentration and interfacial condensation heat transfer coefficients. The applicability of the proposed correlations, as closure equations in existing two-fluid model based computer codes, was checked by including them in a two-fluid model to predict the axial void fraction profiles. The prediction of the current two-fluid model was compared with the measured axial void fraction profiles as well as available data from literature with good agreement.</p> <p>For subcooled flow boiling, experimental data on the axial profile of void fraction, wall superheat and liquid subcooling along the test section were generated for various levels of mass flux, heat flux and inlet subcooling. The high speed photographic results confirmed the fact that the bubble parallel, or normal, detachment from the heating surface is not the reason of the NVG phenomenon. The digital image processing technique was used to measure the bubble size distributions as function of the local conditions. A correlation for the mean bubble diameter as a function of the mass flux, heat flux and local subcooling was obtained. Physical mechanisms causing the NVG phenomenon were investigated using the high speed photographic results. A net vapour generation model was proposed. The proposed model was based on the balance between the vapour generation and condensation rates at this point. A two-fluid model for the axial void fraction profile in subcooled boiling flow was introduced. A heat flux division model was proposed. The proposed two-fluid model was reasonably capable of predicting the axial void fraction profiles in subcooled flow boiling, without the need for prior identification of the location of the NVG point.</p> / Doctor of Philosophy (PhD)