Most turbulence studies up to the present are based on statistical modeling, however,the spatio-temporal flow structure of the turbulence is still largely unexplored. Tur-bulence has been established to have a multi-scale instantaneous streamline structurewhich influences the energy spectrum and other properties such as dissipation andmixing. In an attempt to further understand the fundamental nature of turbulence and itsconsequences for efficient mixing, a new class of flows, so called ?turbulent-like?, is in-troduced and its spatio-temporal structure of the flows characterised. These flows aregenerated in the laboratory using a shallow layer of brine and controlled by multi-scaleelectromagnetic forces resulting from a combination of electric current and a magneticfield created by a fractal permanent magnet distribution. These flows are laminar, yetturbulent-like, in that they have multi-scale streamline topology in the shape of ?cat?seyes? within ?cat?s eyes? (or 8?s within 8?s) similar to the known schematic streamlinestructure of two-dimensional turbulence. Unsteadiness is introduced to the flows bymeans of time-dependent electrical current. Particle Tracking Velocimetry (PTV) measurements are performed. The techniquedeveloped provides highly resolved Eulerian velocity fields in space and time. Theanalysis focuses on the impact of the forcing frequency, mean intensity and amplitudeon various Eulerian and Lagrangian properties of the flows e.g. energy spectrum andfluid element dispersion statistics. Other statistics such as the integral length and timescales are also extracted to characterise the unsteady multi-scale flows. The research outcome provides the analysis of laboratory generated unsteady multi-scale flows which are a tool for the controlled study of complex flow properties relatedto turbulence and mixing with potential applications as efficient mixers as well as ingeophysical, environmental and industrial fields.

The research focuses on the development of mathematical models for describing transient flows within and between connected fluid-filled vessels. The fluid mechanics of connected vessels is of broad interest and numerous examples may be found in industry, the built environment and the laboratory. This work focuses primarily on the interaction between three connecting vessels and considers two main areas of application: (i) the so-called `double-tank' method, as used by experimental fluid dynamicists to stratify environments, and (ii) the passive transient ventilation of multi-storey buildings. An analytical description of the double-tank method, a classic example of liquid exchanges under controlled (constant) flow rates between horizontally connected vessels, was developed. Subsequently, a new technique was proposed, modelled and tested which enabled a broader range of density stratifications to be set up and without the use of pumps. This technique enabled liquids to drain freely under gravity from one vessel to another - the rates of liquid transfer no longer constant but functions of the instantaneous liquid depths. Modelling the fluid mechanics of multi-storey building ventilation added additional tiers of complexity as air and heat exchanged between rooms drive turbulent mixing and there is complex feedback between the individual rooms. Three vessels were again considered, two storeys connected to a common atrium, and the development of the buoyancy- driven flow following the activation of heat sources was investigated. A description of the transient response of the ventilation in an atrium building leading to a steady state, as typically achieved during the course of a day, was developed. Wind pressure variations and solar heat gains in the atrium were also incorporated. The effect of atria geometry on the ventilation of adjoining rooms was established and shown to be analogous to either an `assisting' or an `opposing' wind. When `opposing', the ventilation flow rate reduced. For a strongly `opposing' atrium, a reversal in the direction of flow through the storey occurred, revealing the possibility of multiple flow regimes during the transients - the dynamics of which were explored. Finally, the building ventilation model was generalised to n storeys (n > 2) connecting to a common atrium. Controversially, the implications of the predictions indicate that current atrium designs do not guarantee enhanced flow as is generally accepted.

An investigation into turbulent reacting flows in an opposed jet geometry and a sudden expansion duct has been performed. For the opposed jet geometry, measurements of the velocity and reaction progress variable were obtained in lean premixed flames. Both velocity and scalar measurements were taken using PIV (Particle Image Velocimetry). Three gaseous fuels (methane, propane and ethylene) and three liquid fuels (JP-10, cyclopentane and cyclopentene) were considered for a range of equivalence ratios. The broad range of fuels enabled an investigation of the effect of different fuel reactivities on the velocity field and flame location and also allowed the effect of the Lewis number on flame extinction to be investigated. Preliminary work included isothermal measurements of the flow between and inside the nozzles. The use of fractal grids inside the nozzle increased turbulence intensities at the nozzle exit by 100% and turbulent Reynolds numbers between 50 - 220 were achieved. Velocity and normal stress components were measured with attention focused on the inlet boundary, along the burner centreline and the stagnation plane. A circular duct, incorporating a sudden expansion step, was also used to investigate the effect of swirl on pressure oscillations within the duct, the lean flammability limits and the NOx emissions. Measurements were performed for stratified flow conditions using methane as a fuel. The results show that excessive swirl leads to an increase in local strain in the vicinity of the expansion step and makes the flame more prone to local extinction. Moderate swirl was found to lower the amplitudes of the pressure oscillations close to global extinction and also to decrease the lean extinction limit of the stratified flow conditions. However, it did not decrease the overall equivalence ratio of flows with a richer core and a leaner annulus. Flows with only air in the core flow led to an overall equivalence ratio as lean as 0.3 for methane compared with 0.6 for the uniform flow. Stratification with a fuel rich core flow and a leaner annular flow led to an increase in NOx emissions due to locally increased temperatures. The addition of moderate swirl enhanced mixing of the annular and the core flows, which resulted in a more uniform fuel distribution close to the step and a reduction in NOx-levels up to 50%.

Both experiments and numerical simulations have provided evidence that an initially fully developed two-dimensional boundary layer, subjected to a sudden spanwise forcing, exhibits a decrease of turbulent quantities such as the Reynolds shear stress, turbulent kinetic energy and turbulent friction drag. In past experiments and investigations, such forcing has traditionally been in the form of spanwise wall oscillations, spanwise travelling Lorentz forcing, superimposed spanwise pressure gradients and spanwise travelling waves of an inplane flexible wall. The aim of this work is to take the idea a step further and develop an active surface which locally executes the motions described above making such a system more readily deployable. Two surfaces were developed: both executing in-plane local oscillations with amplitude close to or larger than the mean streak spacing in a turbulent flow, but based on two different technologies, electroactive polymers in the dielectric form of actuation and electromagnetic motor forcing. The effect of these two surfaces was confined to wall-normal heights on the order of the linear sublayer of the turbulent boundary layer, and frequency and wavelength similar to those reported in literature. Extensive hot-wire measurements, some PIV measurements and direct measurement of friction drag using a bespoke drag balance are presented for the systematic variation of the relevant parameters for turbulent friction drag reduction. Electroactive polymers (EAP) are able to undergo relatively large deflections at high frequencies. Developments in the field of EAP such as static and dynamic characterisation of the EAP membranes in use in this work, development of robust electrodes and their characterisation, in-house manufacturing of thin silicone membranes and post-processing of pre-built silicone membranes are presented. Numerical studies of the optimum pre-strain values and of the optimum electrode to passive portions width ratios are presented. Actuator development techniques including EAP membrane pre-stretch in a bespoke jig, EAP membrane pre-conditioning to go past the Mullins' effect, electrode preparation procedure and deposition, and frame preparation are presented. Actuator characterisation results including analysis of multi-flash photographs and laser profilometer scans for in-plane and out-plane deflections at different frequencies are also presented.

I present a new finite element (FE) simulation method to simulate pore-scale flow. Within the pore-space, I solve a simplified form of the incompressible Navier-Stoke’s equation, yielding the velocity field in a two-step solution approach. First, Poisson’s equation is solved with homogeneous boundary conditions, and then the pore pressure is computed and the velocity field obtained for no slip conditions at the grain boundaries. From the computed velocity field I estimate the effective permeability of porous media samples characterized by thin section micrographs, micro-CT scans and synthetically generated grain packings. This two-step process is much simpler than solving the full Navier Stokes equation and therefore provides the opportunity to study pore geometries with hundreds of thousands of pores in a computationally more cost effective manner than solving the full Navier-Stoke’s equation. My numerical model is verified with an analytical solution and validated on samples whose permeabilities and porosities had been measured in laboratory experiments (Akanji and Matthai, 2010). Comparisons were also made with Stokes solver, published experimental, approximate and exact permeability data. Starting with a numerically constructed synthetic grain packings, I also investigated the extent to which the details of pore micro-structure affect the hydraulic permeability (Garcia et al., 2009). I then estimate the hydraulic anisotropy of unconsolidated granular packings. With the future aim to simulate multiphase flow within the pore-space, I also compute the radii and derive capillary pressure from the Young-Laplace equation (Akanji and Matthai,2010)

An experimental study concerning the rise height and morphology of a negatively buoyant fluid projectile, produced when a finite volume of saline solution is forced vertically upwards into a quiescent and uniform fresh water environment, is presented. Unlike the much studied continuous injection of high Reynolds number negatively buoyant fluid giving rise to a turbulent fountain, the behaviour of a finite volume negatively buoyant release remains, until now, unstudied. The work presented herein is based on the development of an experimental set up and methodology specifically tailored to the study of the negatively buoyant projectile. We commence by identifying the two source parameters governing the behaviour of the negatively buoyant projectile; namely the source Froude number FrD, expressing the ratio of source momentum to source buoyancy, and the aspect ratio of release L/D. relating the length L of the column of saline solution dispensed to the nozzle diameter D. In doing so, we note the link to turbulent fountains (continuous negatively buoyant releases) whose behaviour is governed solely by FrD and to vortex rings (finite volume neutrally buoyant releases), whose behaviour is governed solely by L/D. Based on its differing rise height behaviour and morphology of release (for varying FrD and L/D), we classified the negatively buoyant projectile into one of three regimes: the weak-fountain regime, where the rise height behaviour of the negatively buoyant projectile adhered to very weak fountain predictions; the vorticity-development regime, where the development of an internal vortical structure within the head of the negatively buoyant projectile inhibited its vertical propagation; and the forced-release regime, where the rise height behaviour of the negatively buoyant release adhered, under certain source conditions, to forced fountain predictions. This adherence of the rise height behaviour of negatively buoyant projectiles to fountain rise height predictions led to a study of the source conditions (in terms of FrD and L/D), separating finite volume behaviour (negatively buoyant projectile) from continuous behaviour (fountain), at least in terms of initial fountain rise heights. This study led to a classification of the FrD and L/D vales marking the transition from finite volume behaviour to continuous behaviour for negatively buoyant releases, linking our work on the negatively buoyant projectile to existing fountain literature. Finally, we studied the time dependent volume of the head of the negatively buoyant projectile as it propagated, and identified two stages: a growth stage and a decay stage. We established that the volume of the head of the negatively buoyant projectile is subject to an absolute limit. This finding compares favourably with the absolute limit on the volume of fluid contained within a neutrally buoyant vortex ring (the formation number), as detailed in vortex ring literature. On comparing our findings on the negatively buoyant projectile to those on vortex rings, we developed a new method for estimating the formation number of negatively buoyant projectiles (possibly also for vortex rings), determined the dependence of the formation number on FrD, and linked our work on the negatively buoyant projectile to the existing literature on vortex rings.

The flow of immiscible liquids in pipelines has been studied in this work in order to formulatea one-dimensional model for the computer analysis of two-phase liquid-liquid flow in horizontalpipes. The model simplifies the number of flow patterns commonly encountered in liquid-liquidflow to stratified flow, fully dispersed flow and partial dispersion with the formation of one ortwo different emulsions. The model is based on the solution of continuity equations for dispersedand continuous phase; correlations available in the literature are used for the calculation of themaximum and mean dispersed phase drop diameter, the emulsion viscosity, the phase inversionpoint, the liquid-wall friction factors, liquid-liquid friction factors at interface and the slipvelocity between the phases. In absence of validated models for entrainment and depositionin liquid-liquid flow, two entrainment rate correlations and two deposition models originallydeveloped for gas-liquid flow have been adapted to liquid-liquid flow. The model was appliedto the flow of oil and water; the predicted flow regimes have been presented as a functionof the input water fraction and mixture velocity and compared with experimental results,showing an overall good agreement between calculation and experiments. Calculated valuesof oil-in-water and water-in-oil dispersed fractions were compared against experimental datafor different oil and water superficial velocities, input water fractions and mixture velocities. Pressure losses calculated in the full developed flow region of the pipe, a crucial quantity inindustrial applications, are reasonably close to measured values. Discrepancies and possibleimprovements of the model are also discussed. The model for two-phase flow was extended to three-phase liquid-liquid-gas flow withinthe framework of the two-fluid model. The two liquid phases were treated as a unique liquidphase with properly averaged properties. The model for three-phase flow thus developed wasimplemented in an existing research code for the simulation of three-phase slug flow with theformation of emulsions in the liquid phase and phase inversion phenomena. Comparisons withexperimental data are presented.

Thin-film flows are central to a number of industrial, biomedical and daily-life applications, which include coating flow technology, enhanced oil recovery, microfluidics, and surfactant replacement therapy. Though these systems have received a lot of attention in a variety of settings, the understanding of the dominant physics governing the flows is not completely thorough; this is especially true in cases where the free surface of the film or, in two-layer flows, the fluid-fluid interface is susceptible to instabilities leading to the break-up of the film and the formation of fingering patterns. The elucidation of the underlying mechanisms behind the onset of these instabilities is of utmost importance to several industrial processes. The work in this thesis focusses on modelling the dynamics of thin-film flows in the presence of complexities; the latter arise from the presence of surface-active chemicals and spatial confinement. The lubrication approximation, which is valid in the limit of small film aspect ratios, is used to simplify the governing equations; this facilitates the derivation of an evolution equation for the interfacial position. This methodology is employed extensively in the present thesis to examine co- and counter-current two-layer flows in a closed, rectangular channel and the dynamics of a thin film laden with surfactant, driven to climb up an inclined substrate. In the two-fluid case, the dynamics of the flow are described by a single, two-dimensional, fourth-order nonlinear partial differential equation. Analysis of the one-dimensional flow demonstrate the existence of travelling-wave solutions which take the form of Lax shocks, undercompressive shocks, and rarefaction waves. In unstably-stratified cases, a Rayleigh-Taylor mechanism spawns the formation of large-amplitude capillary waves. A wide range of parameters is studied, which include the density and viscosity ratios of the two fluids, the flow configuration (whether co- or counter-current), the heights of the films at the channel ends and the channel inclination. The stability of these structures to perturbations in the spanwise direction, is also examined through a linear stability analysis and transient, two-dimensional numerical simulations. These analyses demonstrate successfully that some of the structures observed in the one-dimensional flow are unstable to fingering phenomena. In the case of the climbing film, two configurations are examined, namely, constant flux and constant volume whereby the evolution equation for the interface is coupled to convective-diffusive equations for the concentration of surfactant, present in the form of monomers and micelles. The former are allowed to exist at the gas-liquid and liquid-solid interfaces, and in the bulk; the latter can only be present in the bulk. For the constant flux case, the flow is simulated by a continuously-fed uncontaminated fluid and surfactant at the flow origin allowed to spread on a solid substrate which has been prewetted by a thin, surfactant-free precursor layer. The constant volume configuration simulates the deposition of a finite drop, laden with surfactant, spreading on a thin, uncontaminated film. In the absence of spanwise disturbances, the one-dimensional solutions demonstrate how the climbing rate and the structural deformation of the film are influenced by gravity, and physico-chemical parameters such as surfactant concentration (whether above or below the critical micelle concentration), and rates of adsorption of monomers at the two interfaces. The stability of the flow is examined through linear theory and transient solutions of the full, nonlinear, two-dimensional system of equations revealing the growth of spanwise perturbations into full-length fingers. A brief introduction to the experimental design of an apparatus, aimed at validating channel flow results, is also described. The objective of the experiment was to investigate the physical features associated with the counter-current, pressure-driven flow of a gas-liquid system. Preliminary experimental results revealed that upon perturbing the flow, an initially uniform liquid film becomes unstable, resulting in the formation of fingers which elongated downstream as time progressed. Finally, we conclude with recommendations for future work, representing natural extensions to the theoretical work described in the present thesis.

The thesis presents a general one-dimensional mathematical model to simulate two-phase, gas-liquid, annular flow in horizontal as well as vertical pipes, and to mechanistically predict the transition from stratified to annular flow in horizontal pipes. The method is based on the transient one-dimensional two-fluid model whereby the two phases are considered as (i) liquid layer and (ii) a mixture of the gas and liquid droplets in which the droplet concentration in the mixture is considered as a flow variable. The model entails the introduction of a scalar transport equation for the conservation of mass of liquid droplets accounting for liquid transfer to and from the film liquid layer. The interface curvature is modelled by a double circle geometric configuration incorporating a new empirical relation for the specification of wetted angle. The droplet exchange rate between the liquid film and gas core is modelled by employing droplet entrainment and deposition rates derived from modifications of models existing in the literature. Using the new model the droplet entrained fraction (E), which is defined as the ratio of the droplet mass flow rate to the total liquid mass flow rate, is computed and validated against different experimental data for both horizontal and vertical pipes. The predictions show good agreement with most of the measurements, being within 30% of the data. This is a significant development since, unlike all other exist- ing models, both horizontal and vertical annular flows can be predicted well with the same model. Moreover, the transition point from the stratified to the annular regimes in horizontal flow can also be predicted and the transition points compare very well with the usual regime boundaries found in existing flow regime maps.

This thesis describes the application of multidimensional fluorescence imaging to microfluidic systems. The work focuses on time- and polarisation-resolved fluorescence microscopy to extract information from microchannel environments. The methods are applied to polymerase chain reaction (PCR) and a DNA repair enzyme, uracil DNA glycosylase (UDG). The fluorescence lifetimes Rhodamine B are calibrated over a thermal gradient using time correlated single photon counting. The dye is then introduced in solution into a novel microfluidic PCR device. Fluorescence lifetime imaging microscopy (FLIM) is then performed, and using the calibration curve, the temperature distributions are accurately determined. The device is subsequently optimised for efficient DNA amplification. A line-scanning FLIM microscope is used to characterise a rapid microfluidic mixer via a fluorescence quenching experiment. Fluorescein and sodium iodide are mixed in a continuous flow format and imaged in 3-D. The spatial distributions of the fluorescence lifetimes are converted to the concentrations of sodium iodide to quantify mixing. Computational fluid dynamic (CFD) simulations are validated by comparison to the quantitative concentrations obtained experimentally. The binding reaction between UDG and a hexachlorofluorescein (HEX) labelled DNA strand is characterised spectrally. As well as an increase in fluorescence polarisation anisotropy, a 700 ps increase in the fluorescence lifetime is measured. Confocal microscopy shows the same spectral properties when the reaction is performed in both simple and rapid microfluidic mixers. In the latter experiment, a concentration series allows the determination of kinetics, which agree with conventional stopped-flow data. A two-colour two-photon (2c2p) FLIM microscope is developed and applied to the UDG-DNA system. An oligonucleotide containing 2-aminopurine, a reporter of DNA base flipping, and HEX is mixed with UDG in a microfluidic Y-mixer. The 2c2p excitation allows FLIM of both fluorophores and hence detection of binding and base flipping. Comparison to CFD with known kinetic rate constants confirms the experimental observations.