This thesis describes an experimental and theoretical study of natural convection heat transfer from a downwardfacing constant-heat-flux inclined flat plate in water. The theoretical investigation, which comprised an analysis of the integrated boundary layer equations for steady two-dimensional laminar and turbulent flows, produced analytical solutions for a heated plate inclined at any angle of inclination. Experiments, using two rigs, investigated heat transfer in the laminar, transition and turbulent regimes. The major outcome is a correlation for the local Nusselt Number over a range of modified Rayleigh Number from 10 6 to 10 15. In addition, boundary layer temperature and velocity profiles were measured using traversing probes and the flow outside the boundary layer was investigated using flow visualization techniques.

The aim of the project is to investigate the detail of cavitation loading and erosion process using a submerged jet cavitation technique. Large size cavitating jet apparatus in the University of Nottingham was used with an long orifice nozzle and experiments were carried out using tap water as a test liquid with upstream pressure ranging from 8D-120bar. Distribution of the mean pressure, cavity clouds and cavitation damage on a specimen have been obtained and their mutual relation was discussed. Effects of pressures and stand off distances on the characteristics of the erosion produced by the cavitating jet were studied and the results were compared with previous investigations. These include not only the weight loss but also the size of the damage and the jet length both related with the optimum stand off distance. Indentations on soft aluminium produced by the cavitating jet were investigated. Their size distributions were obtained for various pressures and stand off distances. Variations of the total number and the average size of indentations with stand off distance were also presented. The cavitation loading pulses were successfully measured by a novel piezoelectric pressure transducer using PYDF polymer and the pulse height measurement system, both of which were developed in the present project. During the process to determine the size of the sensitive area of the transducer, its effect on the pulse height was found. Then, the loading pressure was estimated from the pulse height and the indentation size distribution. The value estimated is around 2GPa and compared with results of the other investigators using similar method with different, vibratory and water tunnel, cavitation facilities. All of them show the similar magnitude. Good correlations of the indentation counting and the pulse height analysis with erosion results were obtained in terms of the intensity of cavitation loading. Simple calibration apparatus for the pressure transducer which utilises a pencil lead break.

The separation of gas-liquid flows is an integral part of many industrial processes. Traditionally, such separations are performed in large vessels under the effect of gravity. However, such vessels can contain inherently large inventories of potentially flammable and/or toxic material. The main objective of this thesis is to combine the knowledge of partial phase separation at T-junctions with control strategies to enhance the development of continuous compact partial phase separators. Such applications would form an integral part of more intensive phase separation systems that allow for smaller downstream separator vessels. This would be especially beneficial to the petroleum industry where safety, space, weight and cost are all issues related to off-shore oil platforms. For such applications a simple definition for a partial phase separator would be one that produced two streams, one rich in gas and the other rich in liquid, each containing less than 10% v/v of the unwanted phase. A series of optimisation experiments produced the final T-junction configuration. This comprised of two horizontal T-junctions placed in series, the first with a vertically upwards side-arm, the second with a vertically downwards one. The addition of control valves on the exit streams of the T-junctions extended previous fundamental studies, incorporating the concept of control and flexibility. An automatic liquid level control on the down leg provided a physical barrier against gas entrainment by maintaining a constant liquid presence within that pipe. A further control valve beyond the second junction then optimised the liquid hold-up above this down leg. Experiments showed that the run valve setting was only dependent on the approaching flow regime and independent of the inlet phase flowrates. A simple active control strategy was developed based around these control valves such that for stratified flows the run arm control valve was set at 20% open, while for slug flows the valve was required to be 55% open. Under this control scheme it was possible to obtain a liquid only stream and a gas-rich stream which always satisfied the simple separation criterion of less than 10% v/v liquid-in-gas. Within industrial situations it is rare to operate under steady-state flow conditions continuously and there will be at least one time dependent variable. Examples of general transient situations involve plant shutdown and start-up, changes in flowrates in response to planned operating conditions and emergency situations. Even more relevant to the petroleum industry however, is bringing an additional well on line. Within the petroleum industry the problem of multiphase transient flows has lead to the development of many commercially available prediction packages but none that handle branched pipe networks. A series of experiments were performed to compare the outlet phase mass flowrate responses for a straight pipe and the T-junction separator. The results indicate that in general the T-junction responses are analogous to those observed in a pipe. However, the existence of pipe branches adds another level of complexity as the flow splits exhibit a very non-linear nature.

This thesis provides unique insights into the application of a geometrically induced swirl by a three-lobed helix pipe on a lean phase of particulate suspension in air along a horizontal pipe section. A series of experimental and computational studies were applied to three flow conditions employing high speed photography, Particle Image Velocimetry (PIV) and Laser Doppler Anemometry (LDA), as well as Computational Fluid Dynamics (CFD) techniques. The CFD simulation predictions were validated both qualitatively and quantitatively against the experimental data and were then used to obtain further insights into the characteristics of the flow behaviour. The LDA measurements of u, v and w velocities were shown to be in good agreement with the predicted CFD velocity components. Additional pressure loss caused by the swirl pipe was found to be proportional to the Reynolds number of the flow and increased further with an addition of particles to the swirling flow. It was concluded that the swirl pipe imparts a wall jet type swirl to both an air-only flow and a lean pneumatic flow with velocity and momentum shifts from axial to tangential closer to the wall. The cusps and ridges of the twisted three lobe surfaces were shown to create a primary flow parallel to the flow axis, and secondary flows of a circulatory motion perpendicular to the primary flow. As a result, the trajectories followed by particles were observed to be affected by their size. The generated turbulence was shown to impart higher core axial velocity for both air and particles. The swirl was found to decay proportionally with the distance downstream of the swirl pipe and inversely to the flow's Reynolds number. The major conclusions drawn from the study were that the swirl pipe locally increases the conveying velocity and produced an improved particle distribution across the downstream section of the pipe.

In two-phase gas/liquid flow, the phenomenon of maldistribution of the phases occurring downstream of a splitting T-junction has been the topic of investigation of several authors. The negative consequences of this maldistribution on the operation of downstream unit have often led to the conclusion that T-junctions in two-phase pipelines are to be avoided. However, the large degree of segregation of the phases obtained at the outlets of a T-junction for certain flow rates and geometries, has encouraged Industry and researchers to exploit this simple device as a partial phases separator. In this work, experiments and interpretations are carried out in two experimental rigs, one with a horizontal main pipe (0.127 ID) and the other with a vertical main pipe (0-076). These consist of measurement of the split characteristic and, in the case of horizontal annular flow, of film thickness. Comparison with predictive models is carried out for the horizontal geometry. For the vertical main pipe experiments, interpretation and semi-empirical correlations are proposed to fit a large database including the present data and previous findings.

The thesis investigates ways to use internal pressure in a favourable way. It focuses on the structural effects of internal pressurisation of mechanical components. The buckling phenomenon of shell structures is analysed in depth and the conducted work confirms the long known beneficial influence of the internal pressure on buckling and suggests how to exploit this to the utmost extent. Changes in failure modes, stiffness and dynamic response due to pressurisation are also considered. Given the nature of the problem, Finite Element Analysis (FEA) is an essential part of the PhD project. The state-of-the-art FEA techniques are described and employed. Geometric imperfections are introduced in the FE models and, to this regard, a novel FEA technique ensuring high-accuracy simulations is developed. Parametric studies on thin-walled structures are carried out. The studies concern both straight and curved cylindrical shells, as well as more complex geometries. These were subjected to different combinations of loads including bending loads, compressive loads and internal pressure. Among the main findings, it is found that internal pressurisation can change the failure mechanism of the structure and, for a given geometry and material, an optimal value of internal pressure exist. This value allows to fully exploit the material capabilities and to maximise the mechanical performance of the structure. Moreover, it is found that internal pressurisation, as well as pipe curvature, modifies the stiffness and this is significant for structures wherein deflections must be kept to a minimum. Exploitation of internal pressurisation is especially attractive in applications wherein weight minimisation is a key objective. Therefore the content of the thesis is particularly relevant to the aerospace sector. A possible application consisting in the use of pressurised members within aircraft wings is here proposed. With regard to the above application, a prototype of a UAV wing with an internal pressurised structure was built. The structure is made of composite material for performance maximisation and its manufacturing process and related considerations are described. Experimental tests were performed with the aim of measuring the effects of internal pressurisation in the component stiff-ness and natural vibrational frequencies. Experimental results were compared to numerical results. Results confirms the potential of internal pressurisation to enhance mechanical properties.

Impact moling is an effective method of pile driving and percussive drilling to bore underground tunnel for various civil applications such as pipe, cable and ducts installation. An effective electro-vibroimpact system has been built on the basis of interactions between two sources of electromagnetic force. A vertical downward progression of mechanism into hard or brittle material required an increased magnitude of impact force within a compact geometry. Horizontal progression into clay is tested by combining periodic impact and static forces that produces an effective progression rate. As a consequence of this experimental work, a prototype electro-vibroimpact system is tested. Electrical circuitry consists of a timer and batteries which is a compact arrangement, functioning as waveform generator, and power supply. A cylindrical hollow aluminium tube houses the main components such as electromagnetic solenoids and oscillating bar within. This protects the main components from clay while progressing into soil and also reduces soil resistance with a minimal surface area. A mathematical model has also been numerically solved for both single and two degreeof-freedom system. Correlation has been achieved to a certain extent, and it is possible either deploy or further optimise this system.

In this thesis, the viscous fingering instability of radial immiscible displacement is analysed numerically using novel mesh-reduction and interface tracking techniques. Using a reduced Hele-Shaw model for the depth averaged lateral flow, viscous fingering instabilities are explored in flow regimes typical of subsurface carbon sequestration involving supercritical CO2 - brine displacements, i.e. with high capillary numbers, low mobility ratios and inhomogeneous permeability/temperature fields. A high accuracy boundary element method (BEM) is implemented for the solution of homogeneous, finite mobility ratio immiscible displacements. Through efficient, explicit tracking of the sharp fluid-fluid interface, classical fingering processes such as spreading, shielding and splitting are analysed in the late stages of finger growth at low mobility ratios and high capillary numbers. Under these conditions, large differences are found compared with previous high or infinite mobility ratio models and critical events such as plume break-off and coalescence are analysed in much greater detail than has previously been attempted. For the solution of inhomogeneous mobility problems, a novel meshless radial basis function-finite collocation method is developed that utilises a dynamic quadtree dataset and local enforcement of interface matching conditions. When coupled with the BEM, the numerical scheme allows the analysis of variable permeability effects and the transition in (de)stabilising mechanisms that occurs when the capillary number is increased with a fixed, spatially varying permeability. Finally, thermo-viscous fingering is explored in the context of immiscible flows, with a detailed mechanistic study presented to explain, for the first time, the immiscible thermo-viscous fingering process.

A droplet to film interaction modelling Computational Fluid Dynamics (CFD) technique is presented in this work. The eventual target application is an aeroengine bearing chamber where oil is used to lubricate and cool the bearings and the bearing chamber walls. Inside the chamber, the oil is found as jets/filaments, film and as droplets in the highly rotational environment. Of particular interest in this work is the formation of the continuous film from the droplets. Spray-film is another relevant application with droplets forming film as it cools the wall. In this work, the liquid and gas continua are modelled using an enhanced Volume of Fluid (VoF) technique. The droplets in the core-air are tracked using a Lagrangian framework that treats them as discrete particles and are coupled to the Eulerian VoF film upon impact using source terms. In finite volume CFD techniques, a prohibitively large number of computational cells would be required to describe, in details, the droplet-film impact phenomenon. The proposal here is that finer mesh, sufficient to capture the film physics, is used only close to walls or where film is expected to form. Simple droplet train to complex spray-film setups are used to verify and validate for mass, momentum and energy transfer. The technique was also applied to experimental rigs representative of aeroengine bearing chambers; and as with every CFD problems, the choice of boundary conditions determines the final output. A parametric study of the bearing chamber flows shows that film thickness increases with flow rate. The film thickness increases with a reducing shaft speed for same flow rate. The heat transfer coefficient results show that higher flow rates provide better heat transfer at higher shaft speeds.

This project aims to provide an insight into the behaviour of cemented sand under high pressures, and to further the understanding of the role of particle crushing. The discrete element method is used to investigate the micro mechanics of sand and cemented sand in high-pressure triaxial tests and one-dimensional normal compression. Using the software PFC3D, a new triaxial model has been developed, which features an effective flexible membrane that allows free deformation of the specimen and the natural failure mode to develop. The model is capable of exerting and sustaining high confining pressures. Cementation has been modelled using inter-particle bonds, and a full investigation of the bond properties is presented, highlighting their influence on the macroscopic behaviour (e.g. failure mode and volumetric response). A simple particle breakage mechanism is used to model the one-dimensional normal compression of sand. By considering the stresses induced in a particle due to multiple contacts, and allowing particles to fracture without the use of agglomerates, this work aims to explain the mechanics of normal compression. The influence of the mechanics of fracture on the slope of the normal compression line is investigated, and the normal compression is linked to the evolution of a fractal particle size distribution. A new equation for the one-dimensional normal compression line is proposed, which includes the size-effect on average particle strength, and demonstrates agreement with experimental results. It is shown that this new equation holds for a wide range of simulations. The time dependence of particle strength is incorporated in to this model to simulate one-dimensional creep tests, leading to a new creep law. The normal compression of cemented sand is investigated, and the results show that bonding reduces particle crushing, and that it is both the magnitude and distribution of bond strengths that influence the compression curve of the structured material. Simulations are also presented that show that it is possible to capture the effects of particle crushing in high-pressure triaxial tests on both sand and cemented sand. Particle crushing is shown to be essential for capturing realistic volumetric behaviour, and the intrusive capabilities of the discrete element method are used to gain insight into the effects that cementation has on the degree of crushing.