Two-phase flows occur regularly in nature and industrial processes and their understanding is of significant interest in engineering research and development. Various numerical methods to predict two-phase phase flows have been developed as a result of extensive research efforts in past decades, however, most methods are limited to Cartesian meshes. A fully-coupled implicit numerical framework for two-phase flows on unstructured meshes is presented, solving the momentum equations and a specifically constructed continuity constraint in a single equation system. The continuity constraint, derived using a momentum interpolation method, satisfies continuity, provides a strong pressure-velocity coupling and ensures a discrete balance between pressure gradient and body forces. The numerical framework is not limited to specific density ratios or a particular interface topology and includes several novelties. A further step towards a more accurate prediction of two-phase flows on unstructured meshes is taken by proposing a new method to evaluate the interface curvature. The curvature estimates obtained with this new method are shown to be as good as or better than methods reported in literature, which are mostly limited to Cartesian meshes, and the accuracy on structured and unstructured meshes is shown to be comparable. Furthermore, lasting contributions are made towards the understanding of convolution methods for two-phase flow modelling and the underlying mechanisms of parasitic currents are studied in detailed. The mesh resolution is of particular importance for two-phase flows due to the inherent first-order accuracy of the interface position using interface capturing methods. A mesh adaption algorithm for tetrahedral meshes with application to two-phase flows and its implementation are presented. The algorithm is applied to study mesh resolution requirements at interfaces and force-balancing for surface-tension-dominated two-phase flows on adaptive meshes.

The present Thesis investigates the quasi-static and cyclic-fatigue properties of nano-modified epoxy polymers. A diglycidyl ether of bisphenol-A was used as the base epoxy and polymer-based materials, block copolymers and core-shell rubber particles were used as the toughening agents. The morphologies formed by the tougheners were determined and the associated fracture and fatigue properties, including the toughening mechanisms, were investigated. The modified epoxy polymers with some of the best properties are described below. The unmodified epoxy polymer had a fracture energy, GIc, of 494 J/m2 and this was increased to 2380 J/m2, when 10 wt.% of a polyurethane toughener was employed. This modified epoxy polymer possessed a second-phase morphology, which was a nano-sized 'worm-like' structure. This nano-phase debonded from the epoxy, which enabled plastic void growth to occur in the epoxy polymer. This mechanism also led to an increase in the cyclic-fatigue threshold fracture energy, Gth, from 153 J/m2 of the unmodified epoxy to 444 J/m2. Nano-sized core-shell rubber particles (either 100 nm or 200 nm nominal diameters) were employed and they formed a well dispersed rubbery particulate phase. A maximum fracture energy of 2540 J/m2 was observed and the particles were found to undergo internal cavitation that induced plastic void growth. However, only the relatively large particles were able to cavitate under the cyclic-fatigue loading. Hence only these type of particles led to a meaningful increase in the fatigue threshold, Gth. 'Hybrid' epoxy polymers contained a polymer based toughener and core-shell rubber particles. Several of these hybrid formulations gave very significant improvements in both values of GIc and Gth. For example, one based upon a polyurethane and styrenebutadiene-core/functionalised polymethylmethacrylate-shell particles gave a value of 2570 J/m2 and 438 J/m2 for GIc and Gth, respectively.

Worldwide, BP operates many thousand kilometres of pipelines carrying valuable yet toxic and corrosive fluids. The structural integrity of these pipelines is crucial, as any failure may result in environmental damage, economic losses and injuries to personnel. Convention- ally, pipeline integrity is assessed on a time basis. This inherently limits the amount of infor- mation available about its structural health, as any damage which develops in unexpected circumstances or while the pipeline is not being inspected may remain undetected. Such lack of information hinders the reliability of any prognosis and of Risk-Based Inspection and Maintenance strategies, increases the risk of unexpected critical damage development and pipeline failure, and forces the use of costly time-based maintenance, following the safe-life design approach. Conversely, if sufficient information about pipeline integrity were avail- able to produce reliable prognoses, then it would become possible to dramatically reduce the risk of unexpected failures and to utilise cost-efficient condition-based maintenance, which prescribes the replacement of a pipeline only when it is about to suffer critical dam- age and has therefore reached the actual end of its operational life. In this way, pipeline networks would become safer and more reliable while at the same time more productive and less costly. This thesis introduces and demonstrates a Structural Health Monitoring ap- proach that has the potential to fill the integrity information gap and ultimately enable the use of condition-based pipeline maintenance. This approach, embodied by a practical au- tomated pipeline damage detection procedure, complements permanently installed guided wave sensors to create a complete pipeline health monitoring solution. Utilising experimen- tal data from a permanently installed guided wave sensor installed on a purpose-built NPS 8 Schedule 40 pipe loop facility at BP's Naperville Campus, it is shown that the procedure is very effective at detecting and quantifying actual damage, thereby achieving the intended aim of this thesis.

The objective of this work is the formulation, development and implementation of Artificial Neural Networks (ANNs) to turbulent combustion problems, for the representation of reduced chemical kinetics. Although ANNs are general and robust tools for simulating dynamical systems within reasonable computational times, their employment in combustion has been limited. In previous studies, ANNs were trained with data collected from either the test case of interest or from a similar problem. To overcome this training drawback, in this work, ANNs are trained with samples generated from an abstract problem; the laminar flamelet equation, allowing the simulation of a wide range of problems. To achieve this, the first step is to reduce a detailed chemical mechanism to a manageable number of variables. This task is performed by the Rate-Controlled Constrained Equilibrium (RCCE) reduction method. The training data sets consist of the composition of points with random mixture fraction, recorded from flamelets with random strain rates. The training, testing and simulation of the ANNs is carried out via the Self-Organising Map - Multilayer Perceptrons (SOM-MLPs) approach. The SOM-MLPs combination takes advantage of a reference map and splits the chemical space into domains of chemical similarity, allowing the employment of a separate MLP for each sub-domain. The RCCE-ANNs tabulation is used to replace conventional chemistry integration methods in RANS computations and LES of real turbulent flames. In the context of RANS the interaction of turbulence and combustion is described by using a PDF method utilising stochastic Lagrangian particles. In LES the sub-grid PDF is represented by an ensemble of Eulerian stochastic fields. Test cases include non-premixed and partially premixed turbulent flames in both non-piloted and piloted burner configurations. The comparison between RCCE-ANNs, real-time RCCE and experimental measurements shows good overall agreement in reproducing the overall flame structure and a significant speed-up of CPU time by the RCCE-ANN method.

Smouldering combustion is the slow, low-temperature, flameless burning of porous fuels and the most persistent type of combustion, different from flaming combustion. Smouldering is the dominant phenomena in fires of coal and natural deposits of peat which are the largest and longest burning fires on Earth. These megafires fires contribute considerably to annual greenhouse gas emissions roughly equivalent to 15% of the man-made emissions, and result in the widespread destruction of global ecosystems and regional haze events. Moreover, the atmospheric release of ancient carbon in soil and the sensitivity of peat ignition to higher temperatures and drier conditions create a positive feedback mechanism to climate change. Compared to flaming combustion, smouldering combustion can be initiated with a much weaker ignition source, and provide a hazard shortcut to flaming. Once ignited, the persistent smouldering fires can consume a huge amount of earth biomass, and burn for very long periods of time (days, years and centuries) despite extensive firefighting efforts or climate changes. For the past few decades, there have been some experimental studies on smouldering peat fires of different scales. However, very few computational work has been done to systematically study such emerging fire phenomena before the research undertaken in this thesis. This thesis is presented in a manuscript style: each chapter takes the form of an independent paper, which has been published or submitted to a journal publication. A final chapter summarizes the conclusions, and suggests potential areas of future research. Chapter 1 first proposes a comprehensive 5-step kinetic model based on thermogravimetric analysis (TGA) to describe the heterogeneous reactions in smouldering combustion of peat. The corresponding kinetic parameters are inversely modelled using genetic algorithm (GA). This 5-step (including drying) kinetic model successfully explains the TG data of four different peat soils from different geographical locations. The chemical validity of the scheme is also investigated by incorporating it into a one-dimensional (1-D) plug-flow model. The reaction and species distributions of two most common fire spread modes, lateral and in-depth spread, are successfully simulated. Chapter 2 presents a new comprehensive 1-D model of a reactive porous media to solve the conservation equations and the proposed 5-step heterogeneous chemical kinetics. This model is used to simulate several ignition experiments on bench-scale peat samples in the literature. The model first predicts the smouldering thresholds, relating to the critical moisture content (MC) and inert content (IC). The modelling results show a good agreement with experiments for a wide range of peat types and organic soils. The influences of the kinetic parameters, physical properties, and ignition protocol on initiating the peat fire are also investigated. Chapter 3 continues to optimize this 1-D model to investigate the vertical in-depth spread of smouldering fires into peat columns 20-30 cm deep with heterogeneous profiles of MC, IC and density. Modelling results reveal that smouldering combustion can spread over peat layers with a very high MC (~250%) if the layer is thin and located below a thick and drier layer. It is also found that the critical MC for extinction can be much higher than the previously reported critical MC for ignition. Furthermore, depths of burn (DOB) in peat fire is successfully predicted, and shows a good agreement with experiments on 18 field peat samples in the literature. Chapter 4 further looks into the kinetic schemes of different complexities to explain the TGA of two peat soils under various atmospheric oxygen concentration. Their best kinetic parameters are fast searched via Kissinger-genetic algorithm (K-GA) method, and the oxidation model is determined for the first time. Then, the kinetic model is applied into the 1-D model to simulate the peat experiment with fire propagation apparatus (FPA) in the literature. Try peat samples are used to minimize the influence of moisture, and ignited under both sub- and super-atmospheric oxygen concentration. Modelling results show a good agreement with experiment, and the stochastic sensitivity analysis is used to test the model sensitivity to multiple physicochemical properties. Chapter 5 investigates the interactions of atmospheric oxygen and fuel moisture in smouldering wildfires with the proposed 1-D model. Modelling results reveal a nonlinear correlation existing between the critical fuel moisture and atmospheric oxygen as MC increases, a greater increase in oxygen concentration is required for both ignition and fire spread. Smouldering fires on dry fuel can survive at a substantially lower oxygen concentration (~11%) than flaming fires, and fuel type and chemistry may play important roles especially in high MC. The predicted spread rate of smouldering peat fire is on the order of 1 mm/min, much slower than flaming fires. In addition, the rate of fire spread increases in an oxygen-richer atmosphere, while decreases over a wetter fuel. Chapter 6 presents an experimental study on smouldering fires spreading over bench-scale peat samples under various moisture and wind conditions. The periodic 'overhang' phenomenon is observed where the smouldering fire spreads beneath the top surface, and the overhang thickness is found to increase with peat MC and the wind speed. Experimental results show that the lateral spread rate decreases with MC, while increases with the wind speed. As peat MC increases, the fire spread behaviour becomes less sensitive to the wind condition and its depth. A simple heat transfer analysis is proposed to explain the influence of moisture and wind on the spread rate profile, and suggests that the overhang phenomena is caused by the spread rate difference between the top and the lower peat layers. Chapter 7 summarizes the research of this thesis, and discuss the possible areas for future research.

The aim of this study is to develop the fundamentals of materials modelling to enable effective process control of hot stamping for forming UHSS panels with graded properties for optimised functional performance. A selective heating and press hardening strategy is adopted to grade the microstructural distribution of a press hardened component through differential heat treatment of the blank prior to forming. Comprehensive material models, to enable prediction of austenite formation and deformation behaviours of boron steel under hot forming conditions, as well as the dynamic response of a press hardened part with tailored properties in collision situations, have been developed based on experimental investigations and mechanism studies. The research work is concerned with four aspects: feasibility of the selective heating and press hardening strategy, austenite formation in boron steel during selective heating, thermo-mechanical properties of boron steel under hot stamping, and mechanical properties of boron steel with various microstructures at room temperature. Feasibility studies for the selective heating and press hardening strategy were carried out through a designed experimental programme. A lab-scale demonstrator part was designed and relevant manufacturing and property-assessment processes were defined. A heating technique and selective-heating rigs were designed to enable certain microstructural distributions in blanks to be obtained. A hot stamping tool set was designed for forming and quenching the parts. Test pieces were formed under various heating conditions to obtain demonstrator parts having variously graded microstructures. Microstructural distributions in the as-formed parts were determined through hardness testing and microstructural observation. Ultimately, the structural performance of the parts was evaluated through bending tests. Heat treatment tests were performed to study the formation of austenite in boron steel during selective heating. Characterisation of the effects of heating rate and temperature on transformation behavior was conducted based on the test results. A unified austenite formation model, capable of predicting full or partial austenite formation under both isothermal and non-isothermal conditions, was developed, and determined from the heat treatment test results. Hot tensile tests were performed to study the thermo-mechanical properties of the austenite and initial phase (ferrite and pearlite) of boron steel. The viscoplastic deformation behaviours of the both phase states were analysed in terms of strain rate and temperature dependence based on the test results. A viscoplastic-damage constitutive model, capable of describing the thermo-mechanical response of boron steel in a state corresponding to hot stamping after selective heating, was proposed. Values of constants in the model for both the austenite and initial phase were calibrated from the hot tensile test results. Dynamic and quasi-static tensile testes combined with hardness testing and microstructural observation were carried out to study the mechanical properties of press hardened boron steel with various microstructures at room temperature. Based on the results, the strain rate sensitivity of the martensite and initial phase of boron steel was characterised; the relationships between mechanical properties (true ultimate tensile strength, 0.2% proof stress, elongation, and hardness) and phase composition (volume fraction of martensite), for boron steel with various microstructures, were rationalised. Finally, a viscoplastic-damage constitutive model, capable of predicting the mechanical response of a press hardened boron steel part with graded properties being subjected to crash situations in automobiles, were developed, and determined from the test results.

The work presented in this thesis discusses the development of an innovative method for manufacturing tailored gears, specifically bi-metallic lightweight gears through the use of the forging process. Utilising this method, gears can be constructed from multiple metals, where high strength, high density materials are located in regions of high stress concentration, such as the tooth flank, tooth root and regions in contact with shaft attachment mechanisms. On the other hand, lower strength, lower density materials can be located at less critical regions, such as the central region of the gear, hence reducing weight. A patent has been filed for the production process of forging multi-material lightweight gears. To investigate this process, bi-metallic gear construction was studied, where a high strength outer ring; and low strength cylindrical or annular core placed within the confines of the ring, allowed for the production of high strength teeth. Experimental and simulation work was conducted to better understand the material flow which occurs during the forging process, and hence its implications on the structural integrity of the gear. To allow for experimental trials, a tool set was designed and manufactured, and used in conjunction with a forming press to forge gears of a spur gear profile. Gears were produced under both cold and hot forging conditions using model materials (lead and copper) and engineering alloys (aluminium and mild steel) respectively. This construction was evaluated for a range of ring thicknesses. A simplified Finite Element (FE) model was established to analyse the material flow and ring thickness distribution during the cold forging operation. Data for the materials commercially pure lead, copper (C101), aluminium alloy (Al 6082), mild steel (230M07) and gear steel (16MnCr5) were obtained through compressive experiments undertaken on Instron and Gleeble testing machines. Constitutive equations were calibrated to a unified constitutive equation model incorporating the physical parameters of stress, plastic strain rate, isotropic hardening, and dislocation density to model the behaviour of aluminium alloy, mild steel and gear steel allowing for the creation of a FE model representing the hot forging process. Furthermore, three locking mechanisms between the two materials were examined: macro-mechanical locking, micro-mechanical locking and diffusion bonding; which when coupled together may prevent disengaging during operation. In addition, the root and contact stresses experienced by bi-metallic gears were also compared to a single material steel gear through an FE model to identify performance differences. Finally, recommendations and future research directions are presented.

Computer simulations using accurate chemical kinetic models are increasingly being used to support the development of combustion technologies and fuels. This action is essential for the reduction of hazardous and green-house gas emissions as well as for efficiency improvement of combustion applications. Consequently, the need to incorporate detailed chemistry in the simulation of combustion processes resulted in an increased interest in developing effective tools for mechanism reduction, from both accuracy and efficiency point of view. In this work, the Rate-Controlled Constrained Equilibrium (RCCE) and the Computational Singular Perturbation (CSP) methods are combined in order to generate an automatic technique for chemical kinetics reduction. The former method identifies the steady-state species and fast reactions while the latter simplifies the kinetics of complex reacting systems. The non steady state species provided by CSP represent the constraints employed in the RCCE code which systematically reduces the detailed mechanism. The benefits of combining the two reduction methods are briefly assessed for H2-air and C2H2-air chemical mechanisms and other two laminar premixed flames are thoroughly investigated i.e. the CH4-air and C3H8-air flames. The detailed chemical kinetics mechanisms used for describing the later two gas mixtures consist of 53 species and 325 reactions and 118 species and $665$ reactions, respectively. The direct numerical solution of 1-D laminar premixed flames is computed using the premixed flame code, providing accurate data for the proposed methodology of detailed chemistry reduction for methane-air and propane-air mixtures. The computational work involves the investigation of several chemical reduced models for each of the above gas mixtures in order to test the potential of the synergy between the two chemical mechanism reduction methodologies. These models are obtained by gradually increasing the number of constraints used for RCCE (resulting in reduced schemes with 12, 16 and 20 constraints for the methane/air flame and 15, 25, 35 and 45 constraints for the propane/air flame) as well as varying the equivalence ratio in the range of (0.8-1.2). The comparison with the results obtained from direct numerical simulations shows that the reduced models containing 20 constraints for the methane case and 45 constraints for the propane case provide good predictions of the laminar flame structure, including steady-state minor species both at stoichiometric and rich/lean mixtures, as well as adequate values of the corresponding burning velocities. Very good agreement with the detailed kinetic model as well as a significant computational time gain are observed. The study also derived a reduced chemical model that can predict flames of various mixture composition at specific pressure value. This reduced chemical model uses the same set of constraints for various equivalence ratio cases and it is able to predict global variables and species concentration within acceptable accuracy limits. The computation time by using this RCCE-CSP scheme is reduced to a third of the simulation time required in the case of applying the direct integration method. Overall, the results suggest that the combined RCCE-CSP is potentially a very reliable time-scale separation method for deriving low-dimensional models by the use of a fully automatic reduction algorithm. Since the proposed technique is an approach to automatically generate reduced chemical models and requires minor computations, it is recommended for the simplification of large detailed chemical kinetic mechanisms as well as for applications to turbulent combustion due to its potential for tabulation.

No description available.

Articular cartilage is a layered porous fibril-reinforced biological composite which performs essential load bearing and shock absorption in mammalian joints; however it is unable to satisfactorily self-repair and is therefore of interest to tissue engineers. Engineering functional tissue requires full mechanical characterisation of the target tissue. ‘Elastography’ is an ultrasonic method which can be used to find depth-dependent strains in a material following the application of a global strain. In this thesis, ultrasonic elastography are applied to determine the importance of the specialised structure of articular cartilage. Existing ultrasonic elastography techniques are adapted in terms of experimental accuracy, protocol variation and signal-processing to allow experimental comparisons to be made with native articular cartilage, a native non-layered cartilage and engineered tissue constructs at different stages of tissue development. A one-dimensional axisymmetric phenomenological model is constructed, allowing extraction of further material properties from articular cartilage experimental data. The results of these novel comparisons demonstrate the importance of the structure of the tissue in determining the global and depth-dependent elastic properties of the tissue. The global elastic modulus of articular cartilage is dependent on the strain applied, which was not the case for the non-layered sample. Depth-dependent moduli of articular cartilage samples match those previously reported, whereas the cartilage with the non-layered structure has an approximately homogeneous modulus. The changes occurring in the engineered tissue with respect to culture time are small but quantifiable. The suitability of this elastography technique for application to engineered tissue is examined and discussed. The implications of the results demonstrating the importance of the depth-dependent structure for tissue engineering are also addressed and further work suggested.