The thesis presents three new computer correction methods for measuring immittances and reflection coefficients using an automatic network analyser. The first correction method called the "Three Short Circuit Correction Method" was invented to overcome the difficulty of measuring components embedded in microstrip or in confined environments where sliding matched terminations required by normal correction methods cannot be used. With this new method, error correction is carried out based on calibration measurements produced using three unevenly spaced short circuits. This method was subsequently published in the I.E.E. journal "Electronic Letters" of 22nd March, 1973. The second correction method using four evenly spaced short circuits was introduced to overcome the difficulty of accurately specifying electrical lengths in microstrip or in a transmission line with non-hmnogenous medium. The third correction method was evolved to overcome the practical difficulties of constructing good short circuits within microstrip lines. With this method, direct knowledge of the terminations of the calibration standards is unecessary. Rigorous theoretical derivations for the last two correction methods have been published in the I.E.R.E. journal "Radio and Electronic Engineer" issue of May, 1978. Detailed measurement results and comparisons of the different types of correction measurements have been published in the "Microwave Journal" issue of June, 1978. The last two correction methods also provide a means of measuring the attenuation and phase changes of transmission lines, and the effective dielectric constant of a nonhomogerous line may be calculated if its physical length is known. Comparative measurements carried out using the new correction systems as opposed to the older established correction syst.ems are included in this thesis. They agree most favourably. Results obtained using a.ny of the new correction systems have shown that measurement accuracies of + 0.5% with a confidence level of 99% are attainable. These results have been confinned by statistical data obtained from repeated measurements (100 in some cases) of an accurately defined test standard.

No description available.

This thesis addresses the problem of autonomous object segmentation. To do so the proposed segementation method uses some prior information, namely that the image to be segmented will have a low depth of field and that the object of interest will be more in focus than the background. To differentiate the object from the background scene, a multiscale wavelet based assessment is proposed. The focus assessment is used to generate a focus intensity map, and a sparse fields level set implementation of active contours is used to segment the object of interest. The initial contour is generated using a grid based technique. The method is extended to segment low depth of field video sequences with each successive initialisation for the active contours generated from the binary dilation of the previous frame's segmentation. Experimental results show good segmentations can be achieved with a variety of different images, video sequences, and objects, with no user interaction or input. The method is applied to two different areas. In the first the segmentations are used to automatically generate trimaps for use with matting algorithms. In the second, the method is used as part of a shape from silhouettes 3D object reconstruction system, replacing the need for a constrained background when generating silhouettes. In addition, not using a thresholding to perform the silhouette segmentation allows for objects with dark components or areas to be segmented accurately. Some examples of 3D models generated using silhouettes are shown.

This thesis investigates effervescent atomisation, a liquid fuel atomisation technique with wide industrial applications, and one which offers several important advantages over conventional atomiser types. An “inside-out” type atomiser rated at 2MW equivalent power (based on mass flow rate) was designed and tested using a state-of-the-art 2-D Phase Doppler Anemometry (PDA) system which allowed for simultaneous real-time droplet size and velocity data to be obtained. High quality data was achieved, with data rates up to 10 kHz and validation rates over 90% in 2-D PDA coincident mode in the high density sprays. Droplet diameters up to 600 μm could be measured. The parameters investigated included operating parameters (air-to-liquid by mass ratio, pressure drop across the nozzle), geometric parameters (exit orifice diameter, nozzle length-to-diameter ratio, mixing chamber diameter, mixing length and air injection geometry) and fluid viscosity. The parameter ranges investigated included 1.83-11.11% air-to-liquid by mass ratio, 4.64-7.05 barG pressure drop across the nozzle, 2-2.8mm exit orifice diameter, 60-136 mm mixing length, 20-30 mm mixing chamber diameter, 0.5-2 nozzle length-to-diameter ratio and 1-18 x10-6 m/s2 kinematic viscosity. In addition 3 air injector geometries were studied which allowed the influence of air injector hole radial symmetry and aerating hole diameter to be determined. Water and air were used as the operating fluid and assist-medium, respectively, for the operating parameter and geometric parameter tests. However, the use of water-glycerol mixtures in the fluid viscosity tests allowed the viscosity of the operating fluid to be controlled. Altering the fluid viscosity allowed the production of a range of simulated fuels (that will encompass Bio-Fuels). The effervescent atomiser designed was compared to an industrial type Y-Jet atomiser frequently used in steam-assisted boiler combustion applications. It was found that the Y-Jet atomiser performed slightly better than an effervescent atomiser without any optimisation, but that improvements in effervescent atomiser performance were possible once atomiser geometry had been fully optimised. Comparisons were also made with the droplet SMD, coefficient of discharge and spray angle predicted by correlations from the literature (obtained using earlier versions of the hardware or alternative sampling techniques). These were found to provide poor agreement with the present experimental data. Finally, global spray SMD correlations were developed; these were shown to agree well with the present experimental data.

Cardiac arrests in the infant population result in undesirable rates of both mortality and morbidity. Cardiopulmonary resuscitation (CPR) is a potentially life-saving emergency procedure, performed during cardiac arrest, which aims to promote blood flow to vital organs through the provision of external chest compressions. To improve cardiac arrest outcomes, current international resuscitation guidelines emphasise the provision of high quality chest compressions during infant CPR. The principle goal of this research was to monitor, assess and engineer an improvement in the quality of chest compressions performed during simulated infant CPR. This was investigated in three experimental stages that evaluated: (i) the current quality of chest compressions performed on a commercially available manikin, (ii) the effects of a more ‘physiological’ infant CPR training manikin design on chest compression quality and thoracic over-compression and (iii) the effects of a real-time performance feedback program, developed to aid resuscitators in achieving high quality chest compressions. Performance was benchmarked against infant specific evidence based quality targets, with highly trained resuscitators achieving these targets in <9% of chest compressions during unassisted simulated infant CPR. The potential for thoracic over-compression in clinical practice was also highlighted when investigating the effects of the more ‘physiological’ infant CPR manikin design. The provision of real-time performance feedback considerably improved chest compression quality, with resuscitators achieving quality targets in >70% of all chest compressions. This research indicates that unassisted chest compressions rarely comply with evidence based quality targets during simulated infant CPR. Real-time performance feedback, in combination with a more ‘physiological’ infant CPR manikin, can considerably improve the quality of chest compressions performed by resuscitators. Importantly, these findings provide an excellent foundation for translation into clinical practice, particularly for assisting resuscitators in providing high quality chest compressions during infant CPR, which may contribute to improving the future outcomes of infant cardiac arrest.

This thesis presents two new approaches in optical interferometry: phase difference determination by fringe pattern matching and a spatial phase-shifting interferometry (spatial PSI) algorithm. These two approaches are both theoretically described and experimentally illustrated in this thesis. The method of phase difference determination by fringe pattern matching is capable of detecting the phase difference between two interferograms with subpixel resolution. In this method, the phase curves are obtained from mean-square difference calculations of any two fringe patterns shifted pixel by pixel, and the phase difference between the interferograms can be achieved by linear interpolation or polynomial curve fitting from the phase curves. The signal to noise ratio is significantly improved due to the region-based matching and its effect of averaging noise. The equations derived from the statistical analysis of matching process clearly explain the reason that the larger image patches have a better accuracy in the measurement of phase difference. The three applications of fringe pattern matching, measurement of electrostatic force displacement, displacement measurement based on Youngs experiment, and phase-shifting interferometry with arbitrary phase steps, are also investigated in this thesis. Computer simulation and experimental results have proved that fringe pattern matching is a powerful technique for measuring some basic parameters in optical interferometry such as phase difference, fringe spacing and displacement. In the algorithm of spatial PSI, one fringe pattern is captured by a CCD camera, and the other two shifted fringe patterns with the phase steps of 90oC and 180oc are generated by computer, according to the features of the light intensity distributions and the method of interpolation. The phase is then calculated by a standard three-step algorithm of phase-shifting interferometry. Experimental results have shown that it is a useful approach to spatial PSI.

An initial review of the various theories of frost heave indicated that Miller's theory of secondary heave was the most convincing. The crucial area in this is the representation of the behaviour in the partially frozen region, known as the frozen fringe, which exists below the lowest ice lens. However, the computational difficulties of the associated mathematical model were likely to limit its application. A simpler quasi-static approach for a semi-infinite region had therefore been initiated, for a restricted range of conditions, by Holden. The work described in this thesis traces the development of the quasi-static approach and its application to the unidirectional freezing of a finite soil column. The resulting generalised model successfully predicts the freezing behaviour under a wide range of conditions. In particular, it is applicable to all overburden pressures, including zero. At low overburdens the frozen fringe disappears, but the final phase is nevertheless modelled to its ultimate equilibrium state. The predictions of the model agree with published experimental data from a number of investigators, and thus support the validity of Miller's theory. Parametric studies with the model have highlighted the importance of the hydraulic conductivity and the relationship between suction, temperature and ice content in the frozen fringe. Simulations are relatively insensitive to variations in thermal conductivity. The model has proved to be robust and stable and should form a sound basis for further studies. However, its full application will depend on the development of experimental techniques to determine the hydraulic conductivity in the frozen fringe.

Discontinuous carbon fibre composites are becoming increasingly popular in the automotive and aerospace sectors, as an alternative to textile-based fibre reinforced composites for both semi-structural and structural components. Materials are highly heterogeneous, with the random architecture leading to uncertainties when modelling and predicting mechanical performance. The microscopic characteristics are known to dominate the strength of the composite, which need to be correctly represented to improve mechanical property predictions at the macroscale. This thesis presents a multi-scale modelling approach that captures the effects of microstructural (filament level) parameters at the macro scale (component level) to predict the mechanical properties of discontinuous composite materials. In the present work, a continuum damage approach has been used to initiate and monitor failure in the models at all scales, via a user defined material (UMAT), allowing strength predictions to be made for the discontinuous material within the ABAQUS solver. Experimental testing of the material constituents (fibre bundle and matrix materials) has been performed to provide input data for the finite element analyses. Micromechanical models have been developed to calculate the properties of fibre bundles, which are used directly at the meso and macroscale. Debonding criterion has also been established and validated which has been used to demonstrate that a small interface, with a thickness of only 1% of the fibre radius, can strongly influence the stress transfer between fibre and matrix materials. Interactions between multiple fibre bundles have been considered at the mesoscale, at a range of bundle orientations and separation distances. As the separation distance between the fibre bundles decreased there was an increase in stiffness 0 f the unit cell (~1.9%) across the bundle orientations considered, however, this also coincided with greater stress concentrations (up to 9.6%) being found in the bundles aligned to the direction of loading. These stress concentrations have been used to produce a comprehensive stiffness reduction scheme at the macro scale to account for the 3D nature of the bundle interactions. A 2D macro scale model is presented for generating discontinuous random fibre architectures consisting of high filament count bundles, with interfacial debonding permitted between the bundle and matrix materials. The fibre bundles are deposited randomly in a 2D plane to provide a representative material. The model has shown that the interface between the bundle and matrix material is critical at short fibre bundle lengths (~5mm) when determining the mechanical properties of the material, with reductions in strength of up to 40% observed at low interfacial shear strengths. The results from the macro scale analysis for discontinuous materials provide predictions within ~10% for tensile stiffness and ~18% for tensile strength when compared with experimental validations.

The thesis is submitted on the basis of papers published between April, 1974 and July, 1976 (chapters 2-8). Minor editing of the papers has been carried out for the sake of continuity of the thesis as a whole. Since each of the papers was written to be largely self-contained, however, a certain amount of repetition, particular to introductory remarks, is unavoidable. For this reason the "contents" lists only those sections which introduce new material. Chapter appendices have been placed immediately following the papers to which they refer. In the majority of cases these were not included with the original texts, but have been added here to expand on points mentioned in the papers or dealt with during presentation. The report supporting the submission is given in two sections: an introduction reviewing the behaviour and analytical requirements of tension structures which relates subsequent chapters to other publications, and a conclusion correlating the papers into an overall context of interactive design and analysis. For the sake of completeness as a thesis, main appendices A-D have been included at the end of the thesis to give comprehensive reviews of published work relating respectively to static analysis, form: -finding, dynamic analy1sis, and the development and mathematical basis of dynamic relaxation. References in these appendices and the introduction (chapter 1) are contained in the main bibliography, appendix E.

The interface between the piston, piston-rings and the cylinder liner represents one of the most hostile environments the crankcase lubricant experiences in reciprocating engines. The role of piston rings is also becoming more complex in line with the increasing requirements for lower oil consumption and friction. A test rig has been developed to simulate under idealised conditions the lubrication action between the piston-ring and the cylinder-liner in reciprocating engines. This approach attempts to separate the tribological conditions from some of the fundamental and unsteady conditions occurring in production engine piston assemblies such as lubricant starvation, ring and piston dynamics, thermal and elastic deformations and blow by, thus allowing the lubricant film characteristics to be examined in isolation. The current work proposes a methodology in assessing the rheological behaviour of various base oils and their additive chemistry with a view to establishing the likely field performance in formulated lubricating oil. The experiment comprised of testing different types of engine oils where the lubricant film thickness, oil film pressure and friction were simultaneously measured throughout the stroke as a function of speed, load and temperature. Furthermore, laser induced fluorescence (LIF) was used to identify the onset of cavitation occurring in the diverging part of the lubricant film between the piston ring specimen and the liner wall. Recognition that lubricant films under certain condition may cavitate opens up a new area for research, focusing on possible links between cavitation, oil consumption, friction and wear. Oil flow visualisation in the test rig using a high speed video camera allowed observation of cavitation in the diverging part of the ring through an optical liner. The high recording framing rates of the camera have permitted the initiation, development and disintegration of the cavities throughout the stroke to be revealed. An attempt was made to extend the oil film visualisation to a motored diesel engine, with parts of the liner cut and fitted with quartz windows, to confirm the presence of cavitation in the piston-rings interface. Moreover the identification of sub-atmospheric pressure in the rings provided additional support to the visualisation of cavitation found in the engine, albeit of much inferior resolution to that of the idealised piston-ring assembly.