This thesis presents three reluctance machines with flux assistance. These machines provide alternative novel geometries that provide high efficiencies with a reduction in the ampere turns in the armature windings for torque production, lowering armature winding switching losses and reducing the power electronic rating. The Dual Stack Variable Reluctance Machine is a switched reluctance variant of the homopolar inductor alternator topology. The Single Stack Variable Reluctance Machine is a simplification of the Dual Stack machine. Both machines use a toroidal field winding to provide additional flux. The methods of connecting armature coils on each stator pole to utilise the armature flux and the choice of power electronic circuitry are important. Testing shows that such machines favour unipolar excitation with single coil per pole for the armature windings. Use of the field winding in series with the armature windings improves torque production. The Dual Stack Variable Reluctance Machine can have the mechanical angular displacement between the two stator stacks varied to provide an improved back emf waveshape for smoother torque production. The Single Stack Variable Reluctance Machine has parasitic and axial air gaps that pose interesting design issues (e.g. end thrust). Magnets can be placed in steel sections where flux is unidirectional. An ideal candidate for magnet insertion is the Flux Switching Motor. A Permanent Magnet Flux Switching Motor has been built that replaces the field windings with ferrite magnets. The Permanent Magnet Flux Switching Motor achieves efficiencies of over 80%. It adds no additional cost to the fan application as cost savings in lower temperature rated thermoplastics offsets the cost of magnets. A prototyping circuit incorporating a novel micro-processor program to alter the commutation timings as the machine operates has been designed to allow fast optimisation of each machine for minimum input power.

This thesis describes the development of grid generation and numerical methods for predicting the flow in variable geometry, positive displacement screw machines. It has been shown, from a review of available literature, that the two main approaches available to generate deforming grids for the CFD analysis of 3D transient flow in screw machines are algebraic and differential. Grids that maintain the cell count and connectivity, during solution, provide the highest accuracy and customised grid generation tools have the capability to accommodate large mesh deformations. For the analysis of screw rotors with a variable lead or varying profile, these techniques are suitable but are required to be developed further with new procedures that can define the three dimensional variation of geometry of the rotors onto the computational grid. An algebraic grid generation method was used for deforming grid generation of variable lead and varying profile rotors. Functions were developed for correlating a specified lead variation along the rotor axis with the grid spacing. These can be used to build a continuously variable lead with linear, quadratic or higher order functions. For variable profile rotors, a novel approach has been developed for three dimensional grid structuring. This can be used to specify a continuously variable rotor profile, a variable lead, and both internal and external rotor engagement, thus making it possible to generate rotor domains with conical and variable lead geometries. New grid distribution techniques were developed to distribute boundary points on the rotors from the fixed points on the rack and the casing. These can refine the grid in the region of interlobe leakage gaps between the rotors, produce a one to one connected interface between them and improve the cell quality. Inflation layers were applied and tested for mesh refinement near the rotor boundaries. Case studies have been presented to validate the proposed grid generation techniques and the results have been compared with experimental data. Simulated results agreed well with measured data and highlighted the conditions where deviations are highest. Results with variable geometry rotors showed that they achieve steeper internal pressure rise and a larger discharge port area could be used. With variable lead rotors the volumetric efficiency could be improved by reducing the sealing line length in the high pressure zone. Calculations with inflation layers showed that local velocities were better predicted but there was no substantial influence on the integral performance parameters.

Vibratory bowl feeders are widely used in automation processes for the storage, feeding and orientation of identical components for presentation to workstations or other mechanical handling devices. The investigation described here has been directed at modelling the dynamiC behaviour of vibratory bowl feeders, both to improve understanding of their behaviour, and to facilitate improvements in their design. The work undertaken has involved the following stages: i) A numerical model for the prediction of the eigenvalues and eigenvectors of the bowl feeder was developed, modelling the structure as a lumped parameter eight degree-of-freedom system; ii) The natural frequencies and mode shapes predicted by the model were compared with those obtained from experimental modal analysis. There was good agreement for the first three natural frequencies. Differences in the higher frequency modes indicated an overconstrained model which could be accounted for by the flexural vibration of the bowl; iii) A numerical model of the forced response of a bowl feeder when driven by a harmonic excitation was developed using a spreadsheet package, and verified experimentally; iv) The spreadsheet package was developed further, varying the geometric parameters of the bowl and springs over specified ranges. Changes in spring angles were investigated experimentally to verify the predicted values; v) A customised design tool was developed using the spreadsheet package to enable engineers to investigate the behaviour of different configuration feeders; vi) An investigation of the causes of dead-spots was undertaken. These were shown to be due to the asymmetrical arrangement of the springs and electromagnetic coil relative to each other; and vii) Solutions proposed to the problem of dead-spots were the use of four spring banks instead of three, and the specification of an annular shaped pole piece for the electromagnetic coil.

No description available.

Metal loss due to corrosion is a serious threat to the integrity of pressurised oil and gas transmission pipes. Pipe metal loss defects are found in either single form or in groups (clusters). One of the critical situations arises when two or more defects are spaced close enough to act as a single lengthier defect, causing major impact on the pressure containing capacity of a pipe and leading to rupture rather than localised leak at the centre of defects. There have been many studies conducted to determine the distance needed for defects to interact leading to a failure pressure lower than that when the defects are treated as single and not interacting. Despite such efforts, there is no universally agreed defect interaction rule and pipe operators around the world have various rules to pick and choose from. In this work, the effects of defect shapes and orientations on closely spaced defects are tested experimentally and further analysed using finite element analysis. Burst pressures of commonly used ductile steel pipes in the oil and gas industries, namely X52 and X60, are measured under internal pressure loading. The pipes were machined with circular and curved boxed defects at different orientations to simulate actual metal loss defects. The burst pressure results were compared with those obtained using existing analytical methods. Comparison of the results showed conservatism in the existing analytical methods which may potentially lead to unnecessary plant shutdowns and pipe repairs. A failure criterion for both single and interacting defects was proposed and validated numerically using the experimental data obtained in this research work. The numerical results when using the proposed failure criterion showed that defect shapes and orientations have a great influence on the failure pressure of pipes containing interacting defects. A simplified mathematical model based on the parametric results and relevant to the cases studied is proposed with the objective of reducing the known conservatism in the existing pipe standards when it comes to the assessment of defect interaction.

The piston-cylinder liner interface comprises more than half of the total engine rubbing friction. With current demand being for internal combustion engines with better fuel economy, lower exhaust emissions and higher performance, reducing this form of friction is the subject of much study. The research reported in this thesis is concerned with the development of an existing floating liner rig to measure the friction in this region. The performance features of the modified setup have also been assessed. Parametric studies have been undertaken with the modified setup to identify the potential means of friction reduction. Modifications undertaken in the sealing method and driveline assembly has facilitated friction measurement at higher gas loading of up to 80 barg. The modified sealing assembly with a sealing ring overcame the problem of arbitrary force interruption through irregular liner and seal contacts. Addition of five times higher inertia flywheel aided the motor to support the rig with adequate torque during high gas loading operations. Calibrations have been performed at each different build of piston-liner combination to reduce build to build variability in measurement. Experimental studies have been undertaken to assess the friction characteristics for different factors such as operating pressure, temperature and speed, lubricant oil formulation, piston-liner clearance, piston material etc. Tests have been undertaken at a range of operating conditions; peak pressure of 0 to 80 barg, speed of 1000 to 2000 rpm and average mid-liner temperature of 40 to 90 ˚C. Peak cylinder pressure has been observed to be dominating the friction followed by temperature and speed. Friction spikes were observed near the top dead centre for pressurised operations; where normal load on the rings are highest in a cycle. Higher speed generally results in a higher total frictional loss. However at higher temperature and peak pressure, contrasting effect of speed on total friction has been reported. The study further identified that piston motion play important roles in determining mixed/boundary friction along with the local gas pressure, velocity and oil film temperature. Friction reductions have been obtained by using a lower viscosity oil and higher piston-liner clearance. Maximum friction reduction of 18% has been reported in this study by using SAE 0W-30 oil in place of SAE 5W-30. Diametric clearance of 80 μm obtained a maximum reduction of 12% compared to a lower clearance of 20 μm. The use of steel piston has shown potential in reducing friction over aluminium piston but the design and weight of piston played a dominant role in the frictional loss.

No description available.

No description available.

Digitized by Kansas Correctional Industries

Call number: LD2668 .T4 1950 S84 / Master of Science