No description available.

The main goal for this work was to produce microbubbles for a wide range of applications with sizes ranging between 10 to 300 μm in a capillary embedded T-junction device. Initially the bubble formation process was characterized and the factors that affected the bubble size; in particular the parameters that reduce it were determined. In this work, a polydimethylsiloxane (PDMS) block (100 x 100 x 10 mm3) was used, in which the T-shaped junction was created by embedded capillaries of fixed outer diameter. The effect of the inner diameter was investigated by varying all the inlet and outlet capillaries’ inner diameter at different stages. In addition, the effect of changes in the continuous phase viscosity and flow rate (Ql) as well as the gas pressure (Pg) on the resulting bubble size was studied. Aqueous glycerol solutions were chosen for the liquid phase, as they are widely used in experimental studies of flow phenomena and provide a simple method of varying properties through dilution. In addition, the viscosity could be varied without significantly changing the surface tension and density of the solutions. The experimental data were then compared with empirical data derived from scaling models proposed in literature, which is widely used and accepted as a basis of comparison among investigators. While the role of liquid viscosity was investigated by these authors, it was not directly incorporated in the scaling models proposed and therefore the effect of viscosity was also studied experimentally. It was found that bubble formation was influenced by both the ratio of liquid to gas flow rate and the capillary number. Furthermore, the effect of various surfactant types and concentrations on the bubble formation and stability were investigated. Preliminary studies with the current T-junction set-up indicated that producing microbubbles with size ranging from 50-300 μm was achievable. Subsequently, the study progressed to optimise the junction to produce smaller bubbles (~ 20 μm) by directly introducing an electric field to the T-junction set-up and assisting the bubble breakup with the combination of microfluidic and electrohydrodynamic focusing techniques. Finally, in this thesis, a novel method that combines microfluidics with electrohydrodynamic (EHD) processing to produce porous BSA scaffolds from microbubble templates with functional particles and/or fibres incorporated into the scaffolds’ structure is presented.

With an ever increasing need for the reduction of carbon and nitrogen based emissions, gas turbine technology has evolved over the years. Lean burning (i.e. high air to fuel ratio) has proven to be an effective method in reducing the nitrous oxide emissions. However, these flames are susceptible to combustion oscillations that could lead to excessive heat transfer, oscillatory thermal and mechanical stresses and flame blow-off or flashback. The addition of hydrogen fuel to turbulent flames has been studied in the past and been reported to have reduced combustion oscillations and nitrous oxide formation, however, there are studies in literature that show combustion oscillations triggered by the addition of hydrogen. This project aimed to further investigate the effects of local hydrogen addition on the dynamics of turbulent lean premixed flames in a model gas turbine combustor. The flames were subjected to controlled reactant mass flow oscillations (created by acoustic speakers) that led to flame surface roll-up, thus causing heat release perturbations (i.e. combustion oscillations). Under these conditions hydrogen fuel was locally added into the flow to study the effects on the response of the flame. A secondary objective was to develop a laser technique for imaging the laser induced fluorescence of atomic hydrogen in turbulent flames. While earlier studies had only focused on laminar flames, this project concentrated on imaging large sections of turbulent premixed flames which were subjected to acoustic forcing. The results showed that the local addition of hydrogen reduced the heat release oscillations of the flame under certain operating conditions. The change in the response of the flame was due to the alteration of the flame surface area. While there were benefits of reducing heat release oscillations, the addition of hydrogen to the flame had a negative impact on the nitrous oxide exhaust emissions.

Computing the price and risk of financial derivatives is a necessary activity for many financial market participants and is often undertaken by large and costly computing farms. This thesis seeks to explore the use of parallel computing, with particular focus on graphics processing units (GPUs), to improve the speed per cost ratio of such computation. This thesis addresses three distinct layers of high performance parallel financial derivatives computation: the first layer is related to the formulation of parallel algorithms that are generally used in the context of derivatives. The second layer is related to the optimum computation of pricing models, which consist of a series of computational steps or algorithms, where such pricing models are used to calculate the price and risk of individual derivatives. The third and final layer is related to deploying several pricing models within large scale infrastructures with particular focus on optimal scheduling approaches. Several contributions are made within this thesis: (i) with regard to the formulation of parallel algorithms, we introduce novel approaches for evaluating the normal cumulative distribution function (CDF), calculating option implied volatility, calibrating SABR (stochastic-αβρ) volatility models and generating CDF lookup tables. (ii) With regard to pricing models, we explore the computation of two dominant fixed income pricing models, namely non-callable bullet options and callable bond options. (iii) With regard to the computation of many such pricing models within large scale infrastructures, we devise and verify novel scheduling approaches that are able to optimally allocate tasks between a heterogeneous mix of CPU and GPU processors.

Areas of stress concentration around welded structures are likely to lead to fatigue cracks and corrosion pitting during the life time of technical machinery. Performing periodical non-destructive testing of the critical area is crucial for the maintenance of structural integrity and the prevention of unforeseen shutdowns of the system. Low frequency guided ultrasonic waves can propagate along thin structures and allow for the efficient testing of large components. Structural damage can be localized using a distributed array of guided ultrasonic wave sensors. Guided waves might be employed to overcome the accessibility problem for stiffened plate structures where access to some parts of the inspected structure is not possible. The transmission and reflection of the A0 Lamb wave mode for a variation of the stiffener geometry and excitation frequency was investigated numerically and verified experimentally. The dispersive behaviour of the guided waves has been studied to ascertain a frequency thickness product that provides limited pulse distortion. The limitations of the plate geometry as well as the excitation and monitoring locations were discussed. The radial spreading of the incident, transmitted and reflected waves from a stiffener has been investigated. The efficient quantification of the transmitted and reflected waves from the stiffener for a wide range of angles has been obtained from a single Finite Element model containing two parallel lines of nodes in front of and past the stiffener. The research outcomes have shown the dependency of the scattered wave on the incident angle and stiffener dimensions. Reasonably good A0 wave mode transmission was obtained from the oblique wave propagation (up to an angle of 45o) across realistic stiffener geometries. The choice of an optimum excitation frequency, which can ensure maximum transmission across the stiffener for specific plate geometry, was recommended. The ability for defect detection in inaccessible areas has been investigated numerically and validated experimentally. The possibility of detecting and characterizing the reflection of a guided wave pulse (A0 mode) from a through-thickness notch located behind the stiffener has been discussed. Two different approaches, based on the access to the sides of the stiffener on the plate, were employed. The limitations of the detectable defect size and location behind the stiffener have been investigated. The energy of the transmitted wave across the stiffener was adequate to detect simulated damage behind the stiffener. The evaluation has shown that defect detection in inaccessible areas behind stiffeners is achievable if the signal-to-noise ratio is high enough. In experimental measurements the noise level was of similar magnitude to the observed reflections at the defect. Thus, there is necessity to enhance the signal-to-noise ratio in experimental measurements.

The recent advances in medical imaging and in computer technologies have improved the prediction capabilities of biomechanical models. In order to replicate physiological, pathological or surgically corrected portions of the cardiovascular system, several engineering methodologies and their combinations can be adopted. Specifically, in this thesis, 3D reconstructions of patient-specific implanted devices and cardiovascular anatomies have been realised using both volumetric and biplanar visualisation methods, such as CT, MR, 4D-MR Flow and fluoroscopy. Finite Element techniques have been used to computationally deploy cardiovascular endoprosthesis, such as stents and percutaneous pulmonary valve devices, under patient-specific boundary conditions. To analyse pressure and velocity fields occurring in patient-specific vessel anatomies under patient-specific conditions, Lumped Parameter Networks and Computational Fluid Dynamics simulations have been employed. The above mentioned engineering tools have been here applied to address three clinical topics: 1 - Percutaneous pulmonary valve implantation (PPVI) Nowadays, more than 5,000 patients with pulmonary valve dysfunctions have been treated successfully with a percutaneous device, consisting in a bovine jugular venous valve sewn inside a balloon expandable stent. However, 25% of the treated patients experienced stent fracture. Using a novel methodological patient-specific approach that combines 3D reconstructions of the implanted stent from patients’ biplane fluoroscopy images and FE analyses, I carried out a risk stratification for stent fracture prediction. 2 - Transposition of the Great Arteries (TGA) Patients born with the congenital heart defect TGA need a surgical correction, which however, is associated with long term complications: the enlargement of the aortic root, and the development of a unilateral pulmonary stenosis. These may originate a complex hemodynamics that I tried to investigate by using patient-specific LPN and CFD models. 3 - Aortic Coarctation (CoA) Finally, combinations of FE and CFD-LPN models have been used to plan treatment in a patient with CoA and aberrant right subclavian.

This thesis analyses the chemistry and physics behind acidic jets and plumes. The research was motivated from the discussions between industry and regulatory bodies concerning the dispersion of highly acidic discharges from exhaust gas scrubbers on ships into seawater. The industrial problem is simplified in an analytical model for acidic jets and plumes, which is then validated through an experimental study. The analytical model allows for the construction of an optimisation tool that considers the acidity of the discharge, the alkalinity of local seawater and the required scrubber flow rate to propose optimal discharge pipe configurations. This tool can be used for designing discharge pipe configurations in compliance with regulation MEPC 59/24/Add.1 Annex 9. The analytical model was then extended to also take into account the effects of ambient flow and buoyancy on the discharge trajectory. Existing regulatory compliance tests for scrubber discharges assume that no deflection occurs, however, the experimental study shows that an offset of one jet radius leads to an overestimation of pH recovery by one unit. Simplified expressions are developed to improve the accuracy of regulatory compliance tests by taking into account the effects of buoyancy and ambient flow. A general purpose computational fluid dynamics code was written to study the dispersion of contaminants in the wake of a ship. The study suggests that rapid dilution occurs in the near field as a result of the turbulence generated by propulsion, and further dilution occurs more slowly through the widening of the ship wake. Different velocity profiles are generated when the ship is either decelerating, accelerating or moving at a steady pace, but the widening of the wake is relatively insensitive to these factors in the near field.

Violent wave attacking offshore and coastal structures is a complex phenomenon frequently involving air entrapment. A study on fluid/structure impact with air cavity effect is carried out in the framework of velocity potential theory. The purpose is twofold. One is to develop methodologies to tackle the technical difficulties involved. The other is to achieve a better insight into the impact dynamics and the subsequent structure/water/air interaction process, as well as the associated air cavity effect and its acting mechanism. The study starts with axisymmetric problems. Impact by a liquid column on a rigid plate is studied analytically and numerically. The initial singularity at the body-free surface intersection is analysed in detail. The feature of the resulting long thin jet is revealed: providing field solution over larger wetted area without influencing the main impact dynamics. This is favourable in the study of some problems (e.g. steady state solution or local impact over a tiny region), and thus a decoupled shallow water approximation scheme is developed for the computation with long jet. Impact with air cavity of various parameters is studied systematically. Wave impact with air entrapment in practical engineering situations is then focused. A domain decomposition method together with a dual-system technique is developed to provide fully nonlinear simulation on the early impact stage by a plunging wave crest, tackling the large variation in scales involved. Local pressure peak is found to be generated by the sharp turn of the wave surface along the wall. The trapped cavity, governed by an adiabatic law, is found to cause oscillating loading on the wall. The local free jet drawn from the upper cavity surface in each re-contraction stage reveals its distortion and fragmentation mechanism. The initial dimensionless potential energy of the air cavity is found to largely influence its maximum pressure, and the scaling law revealed could be applied to the prediction of impact pressure in practical situations from a laboratory experiment.

Failure in composite structures due to low-velocity impact damage raises a significant maintenance concern because it can lead to a barely visible and difficult-to-detect damage. Depending on the severity of the impact, fiber and matrix breakage or delaminations can occur, reducing the load carrying capacity of the structure. Efficient structural health monitoring (SHM) of composite structures can be achieved by using low- frequency guided ultrasonic waves as they have advantages of propagating over large structure and being sensitive to defects located at any thickness position. This work focuses on the use of first antisymmetric guided wave mode (A0) for health monitoring in laminated composite plates. The first part of this work is to investigate the propagation of A0 mode in undamaged composite plates experimentally and compare the results to Finite Element simulations and semi-analytical analysis. This study is essential in order to improve understanding of the guided waves behavior in composite plates and would benefit the interpretation of received signals particularly for defect characterization. To gain a good understanding of the A0 mode interaction with defects in composites, a full three- dimensional (3D) Finite Element (FE) analysis is used. A systematic study of the influence of defect geometry and range of situations on guided wave scattering is demonstrated. Combined delamination with material degradation to simulate mixed- modes defect is shown. Two dimensional FE simulations used for analysis of large delamination are also presented. The final part of this thesis presents the scattering of guided waves at the impact damage using a non-contact laser interferometer. In this study, the results were quantified and compared to baseline measurements on undamaged composite panels. Significant scattering activities were observed, allowing for the detection of impact damage in composite plates. The impact damage was further characterized using standard ultrasonic C-scans. Good agreement between experiments and predictions was found.

This thesis describes the development of a multi-scale and multi-physics model of the left ventricle in silico. The model presented here, provides a computational model that allows further insight into the events and mechanism describing the ventricular contraction and relaxation (excitation contraction coupling process) at low computational costs. The formalisms and methods describing the electro-mechanical coupling (from intracellular, to electrical, to flow, to physiology mechanisms) across the scales are presented in a novel manner, coherently combining the various processes of all scales from a biological and mathematical point of view. The multiple scales involved in the model encompass the protein, cellular and organ level. In order to achieve this, at each scale the typical characteristics and mechanisms, such as the Action Potential, cross-bridge kinetics, pressure-volume relationship are simulated, extrapolating the behaviour of a ventricular cardiomyocyte to the whole ventricle. The coupling between the three scales presented is achieved via two links, the intracellular calcium concentration and the cross-bridge kinetics. The generation of force calculated at the organ level gives further insight on the cardiovascular haemodynamics, such as changes in pressure, flow and volume. In collaboration with TU/e, Netherlands, the presented multi-scale and multi-physics model of the left ventricle (developed in MATLAB) is expanded to a whole heart model in SIMULINK, enabling to investigate the behaviour of a healthy heart. Following, a case study of idiopathic dilated cardiomyopathy is conducted. While mainly the effects of idiopathic dilated cardiomyopathy (IDC) are presented in literature, it lacks of quantitative data to describe these effects. To simulate the effects of IDC, as shown in corresponding literature, key parameters across all scales were chosen and modified in the multi-scale model of the heart. A second collaboration with ANSYS UK demonstrates the feasibility of the ventricular multi-scale and multi-physics model as a boundary condition, being coupled to a 3D Model in ANSYS. The interaction and exchange of ventricular pressure and mitral flow between MATLAB and ANSYS, respectively, drives the local haemodynamics of the mitral valve in a CFX model. The 0D-3D coupling sets a foundation and coupling technique that can be further expanded to other models and conduct case studies on pathologies of the heart.