This thesis details the development and evaluation of a procedure to predict the fuel consumption and pollutant emissions of spark ignition engines during cold-started drive cycles. Such predictions are of use in the early development and optimisation of an engine and vehicle combination with regard to legislated limits on vehicle performance over defined drive cycles. Although levels of pollutant emissions are the main focus of legislation, reducing fuel consumption is also of interest and drive cycle fuel consumption figures provide a useful benchmark of vehicle performance appraisals. The procedure makes use of a combination of engine friction models and experimentally defined correction functions to enable the application of fully-warm engine test bed data to cold-start conditions. This accounts for the effects of engine temperature on friction levels, mixture preparation and start-up transient behaviour. Experimental data to support the models and assumptions used are presented and discussed. Although not an essential part of the procedure, neural networks have been used to characterise the fully-warm engine mapping data. These are shown to provide an effective way of interpolating between engine mapping points. To facilitate the prediction of tail-pipe emissions, a simple catalyst efficiency model has been included and the complete procedure incorporated into a single software package enabling second-by-second fuel and emissions flow rates to be predicted for a given engine and vehicle combination over a defined drive cycle. This package is called CECSP or the Cold Emissions Cycle Simulation Program. The program has been designed to run on PC machines. The procedure has been validated by application to a typical 1.8 litre medium sized vehicle driven over the ECE+EUDC drive cycle and the predictions found to be within the target accuracy of +/-5% for fuel consumption and +/-10% for engine-out emissions. Envisaged applications of the procedure to rank the sources of increased fuel consumption and emissions due to cold-starting and engine and vehicle details are outlined.

Creep failure of welds in high-temperature power plant steam piping systems is known to be a potential cause of plant failure. Creep behaviour of plain pipes with circumferential welds and cross-weld specimens have received fairly extensive attention. However, research into the creep behaviour of welded thick-walled branched steam pipes has received less attention. Consequently, this thesis addresses improving the understanding of the creep behaviour for this type of geometry. Numerical and analytical methods are used to assess the creep behaviour of typical power plant branched pipe geometries. The effects of various geometric and material parameters on the creep stress and creep life behaviour of the connections are studied. In particular, the effect of the differing creep properties associated with the various material regions of the weld are investigated. The importance of incorporation of weld properties in creep life assessments is thus assessed. Finite element steady-state and continuum damage mechanics creep analyses have been used to identify the relative creep strength of typical connections compared to plain pipes. The work identifies typical creep rupture locations within branched pipe welds and the associated damage accumulation at and around these positions. Various creep life assessment methods/procedures are used in practise: these are mainly the British Standard codes, British Energy's R5 procedure, steady state creep approaches and continuum damage mechanics approaches. The relative accuracy and conservatism of these distinct approaches are addressed for the application to typical branched pipes. The general formulation of steady-state creep stress is applied to the parametric study of weld materials in a typical multi-material welded branched pipe. An approximate interpolation technique for power-law creep is implemented to reduce the number of analyses needed to span a wide range of material parameters. The method is used to estimate the creep stresses and lives at several critical regions within the various material zones of the weld. The advantages of the technique are related to the small number of analyses required and the simple and compact way of presenting the results for weld design and life assessment purposes.

This thesis describes the application of novel swirl inducing pipe to various pipe configurations, when pumping a range of fluids and fluid / particle mixtures. An extensive experimental programme, incorporating particle image velocimetry and photography, was implemented using a pipe flow loop designed specifically for the purpose. Experimental data was obtained on the effect of a 4-lobe near-optimal swirl pipe on coal-water, sand-water and magnetite-water slurries of various particle size. Results indicated that swirl induction produced greater benefit for denser slurries and higher concentrations, and that swirl induced into slurries containing larger and denser particles decayed more rapidly. At low velocity, experimental data highlighted a reduction in the total pressure drop experienced across a 3.0m horizontal pipe section, a downward sloping section and vertical pipe bends, when the swirl-inducing pipe was present. PIV was used to measure the axial and tangential velocity of swirling flows downstream of a near-optimal swirl-inducing pipe. It was confirmed that a significant tangential velocity was generated when pumping water in the turbulent regime, however, when the fluid viscosity was increased, leading to laminar flow, no significant tangential velocity was detected.

In oil industrial applications, the modem practice in the drilling of oil wells, deviated drilling, results in inclinations from the vertical to the horizontal being present in such wells. Their design requires an accurate knowledge of the pressure drop/flowrates/physical-properties relationships. The measurement of wet-gas streams can be improved significantly by the use of a Venturi flow meter with an "overreading" correction. Moreover, at high gas mass fractions, knowledge of the liquid distribution about the well tubing cross section is required to inform policy on the use of inhibitors to protect the tubing from corrosion. Therefore, the aims of this study are to address aspects of two-phase gas/liquid flow in the pipe upstream the Venturi and in the Venturi. The main thrust of the project is to examine the effects of stratification that occur in annular flow when the pipe inclination is from horizontal to much higher inclinations. The study of annular flow includes the prediction of the three principal dependent variables – film flowrate, film thickness and pressure drop - as a function of position along the channel. All experiments were carried out with air and water in an inclinable rig. It consisted of a 5 m long stainless steel pipe of 38 mm internal diameter. The pipe could be positioned at any angle between horizontal and vertical in intervals of 5 degrees. The Venturi, located downstream the pipe, had a 19 mm i.d. throat and angles of convergent and diffuser respectively of 32° and 4°. Measurements on liquid film flowrate and liquid film thickness were carried on with two conductance probe techniques and sintered porous wall units. Measurements on pressure drop were conducted with the use of two differential pressure cells. In all experiments described, gas and liquid flowrates and pipe orientation were varied. Another aim of the study was to develop computer modelling for the prediction of air-water pressure gradient and liquid film thickness along the Venturi. The models of Azzopardi et al. (1991) regarding pressure drop and the models of Fukano and Ousaka (1988) for film thickness circumferential variations have been analysed and modified according to the characteristics of the system.

This thesis is primarily concerned with the design and optimisation of transition ducts for lobed swirl-inducing pipes. Single-phase swirl-inducing pipe flows were modelled and optimised using Computational Fluid Dynamics (CFD). Optimised pipes were manufactured using rapid prototyping and an experimental investigation examines their effect on settling slurries of different densities. The CFD model was successfully validated by experimental measurements of pressure loss and tangential velocity. An optimum transition geometry was determined for use as an entry and an exit duct with optimised swirl inducing pipe. Transition pipes either before or after the swirl inducing pipe reduced entry and exit pressure losses by providing a gradual transition from circular to lobed cross-section. They also increased induced swirl and reduced swirl decay. CFD simulations with carboxymethyl cellulose (CMC) instead of water as the flow medium indicated that as the viscosity increased, a smaller pitch, thereby a tighter twist, is required in the swirl-inducing pipe to achieve effective swirl induction. Settling slurry experiments showed that swirl induction resulted in better particle distribution and prevented solids dragging along the bottom of the pipe. This suggests reduction in localised erosion and provides an opportunity to operate at lower flow velocities without blockage. Lower velocities mean lower energy costs and further erosion reduction. When transitions were incorporated pressure losses across the swirl inducing pipe were reduced and the length of particle suspension increased. It was proven, by CFD and experimentation, that entry and exit transition should be an integral part of the swirl inducing pipe. This results in an efficient swirl induction which reduces energy costs from high pressure losses that otherwise occur due to sudden changes in flow geometry.

The current work begins by considering a range of common high power ultrasonic components in order to establish a standardised approach to tool design for optimum performance. The vibration behaviour of tuned components resonating longitudinally at ultrasonic frequencies around 35 kHz is modelled via finite element analysis and measured by experimental model analysis. Significant improvements in experimental validation of the models are achieved by the use of a 3D LDV, which allows modal analysis from both in-plane and out-of-plane measurement, which is critical in proposing alternative designs. The vibration characteristics of complex multiple-component systems used in ultrasonic cutting of food products are also investigated. Commonly, the design approach for ultrasonic systems neglects to account for the mutual effects of physically-coupled components in the system vibration. The design of systems also neglects the nonlinear dynamic effects which are inherent in high power systems due to the nonlinearities of piezoelectric transducers. The first issue is tackled by considering the vibration behaviour of the whole system and the influence of individual components and, particularly, offers design improvements via modification of block horns and cutting blade components, which are modelled and validated. The issue of nonlinearity is addresses by identifying the mechanisms of energy leakage into audible frequencies and characterising the common multimodal responses. For this study, design modifications focused on reducing the number of system modes occurring at frequencies below the tuned system frequency. As a consequence of these approaches, insights for the design of multiple-component systems in general are provided.

This thesis reports on the design of ultrasonic bone cutting blades and the effect of various cutting parameters such as applied load, blade tip vibration velocity and frequency on cutting speed and temperature, two performance indicators used by orthopaedic clinicians. A range of high gain blades was developed to investigate the correlation between the frequency response predicted by finite element analysis (FEA) and the frequency response measured using an experimental model analysis (EMA) technique. It has been found that FEA frequency predictions are within 1.5% of measured frequencies. FEA has also been used to develop two novel ultrasonic cutting models which allow the effect of blade progression on cutting performance to be investigated. The models have been used to predict the relationship between applied load and cutting speed in single layer and multi-layer materials, and have shown that cutting speed decreases as cortical layer thickness increases, a trend also found from cutting experiments. Ongoing developments to predict temperature from both cutting models have produced a preliminary result which locates regions of maximum cutting temperature. The result influenced the design of blades with modified tip geometries that have been used to reduce cutting temperature. Ultrasonic cutting experiments were performed on bovine bone, two bone substitute materials and various grades of wood. Deep incisions were made for a range of applied loads and cutting speeds to investigate the effect of various cutting parameters on cutting temperature. Ultrasonic cutting has been successfully applied to perform deep incisions in bone whilst maintaining substrate temperature to within critical levels. Two innovative modelling techniques have been used to simulate ultrasonic cutting and demonstrate their potential for revolutionising blade design, and surgical trials.

Laser welding, unlike conventional arc or gas welding, can be effectively utilised to produce high quality, clean and tough welds in high carbon steels. Results of welding high carbon steel are presented. The weld characteristics related to the fast cooling rate were critically evaluated and methods to reduce the rate of cooling were developed. The grain size produced in the fusion and narrow heat affected zones significantly affected the mechanical properties of the welded joint. Three lasers were used: Nd:YAG, CO2 and a high power laser diode (HPDL). The investigations were carried out using a pulsed, 400 W, Nd:YAG laser, a CW, 1.2 kW, CO2 laser and a CW, 1.4 kW high power diode laser. For the Nd:YAG laser, the dual beam delivery system was achieved with a step index fibre to produce in-line process heat-treatment during welding. The spatial and temporal temperature distribution was controlled in the weld region to generate the desired mechanical properties, without losing the benefits of this low distortion joining process. For the CO2 laser system, a dual beam system was successfully designed, fabricated and the performance of the multiple beam system was evaluated. The welding quality was characterised by quantifying the effect of different laser parameters and welding geometry, including flat, angular, clamped and unclamped. The welding performance of the Nd-YAG laser was dependent on the welding speed, pulse width and pulse repetition frequency (PRF). The effect of varying the laser parameters was quantified by measuring the hardness profiles, tensile strength, weld width, weld penetration and the rare of weld volume formation. Furthermore, microscopic examination was conducted at the welded joint. The quality of the welds was improved by increasing the pulse width and pulse repetition frequency (PRF), achieving a deeper penetration, wider weld width and greater weld volume formation rate and a tougher weld. At a slower welding speed, and for the higher pulse width and PRF, the hardness profiles were greatly reduced due to the greater spatial overlap of laser beam on the workpiece.

The nickel-base superalloys Waspaloy, INCO 901 and INCO , 718 have been machined with four different types of ceramic cutting tools. The cutting inserts were pure alumina (CC620), mixed alumina (CC650), composite ceramic (WG-300) and finally silicon nitride (Kyon 2000). Tests in the form of turning were carried out, in dry and wet conditions, in order to study the behaviour of the above mentioned cutting tools when machining superalloys. The cutting speeds employed for these tests were 90, 150, 215 and 300 m/min with feed rates of 0.125 and 0.18 mm/rev together with depths of cut of 1 and 2.5 mm. Machining in the presence of different atmospheres was also performed. In the majority of cases depth of cut notching (DOCN) proved to be the domineering factor controlling tool lives under different cutting conditions. However, flank wear, nose notch and surface roughness were the other tool failure modes. Attrition wear was predominant throughout the tests which was influenced by the cutting temperatures. The high temperatures also caused diffusion wear mechanisms to take effect.

Both a novel shearing interferometer and the first demonstration of particle image velocimetry (PIV) to the stator-rotor gap of a spinning turbine cascade are presented. Each of these techniques are suitable for measuring gas turbine representative flows. The simple interferometric technique has been demonstrated on a compressor representative flow in a 2-D wind tunnel. The interferometer has obvious limitations, as it requires a clear line of sight for the integration of refractive index along an optical path. Despite this, it is a credible alternative to schlieren or shadowgraph in that it provides both qualitative visualisation and a quantitative measurement of refractive index and the variables to which it is dependent without the vibration isolation requirements of beam splitting interferometry. The 2-D PIV measurements have been made in the stator-rotor gap of the MTI high-pressure turbine stage within DERA's Isentropic Light Piston Facility (lLPF). The measurements were made at full engine representative conditions adjacent to a rotor spinning at 8200 rpm. This is a particularly challenging application due to the complex geometry and random and periodic effects generated as the stator wake interacts with the adjacent spinning rotor. The application is further complicated due to the transient nature of the facility. The measurements represent a 2- D, instantaneous, quantitative description of the unsteady flow field and reveal evidence of shocks and wakes. The estimated accuracy after scaling, timing, particle centroid and particle lag errors have been considered is ± 5%. Non-smoothed, non-time averaged measurements are qualitatively compared with a numerical prediction generated using a 2-D unsteady flow solver (prediction supplied by DERA). A very close agreement has been achieved. A novel approach to characterising the third component of velocity from the diffraction rings of a defocusing particle viewed through a single camera has been explored. This 3-D PIV technique has been demonstrated on a nozzle flow but issues concerning the aberrations of the curved test section window of the turbine cascade could not be resolved in time for testing on the facility. Suggestions have been made towards solving this problem. Recommendations are also made towards the eventual goal of revealing a temporally and spatially resolved 3-D velocity distribution of the stator wake impinging on the passing rotor.