The aim of this research is to investigate the structural response of PVC foam based sandwich structures, composite reinforced foam cores and fibre metal laminates (FMLs) subjected to quasi-static and dynamic loading conditions. It also includes the investigation of the mechanical properties and energy-absorbing characteristics of the novel hybrid materials and structures for their potential use in aerospace and a wide range of engineering applications. Firstly，a series of experimental tests have been undertaken to obtain the mechanical properties of all constituent materials and structural behavior of the composite structures, which are used to develop and validate numerical models. The material tests carried out include (1) tension properties of composite laminates and aluminium alloys, (2) compression of PVC foams, carbon and glass fibre rods and tubes, and fibre metal laminates in the edge wise and flat wise, (3) shear and bending of PVC foams, (4) Hopkinson Bar, (5) quasi-static and dynamic crushing of composite reinforced foams, and (6) projectile impact on fibre reinforced laminates, aluminium alloy panels, PVC foam based sandwich panels and fibre metal laminates. The corresponding failure modes are obtained to validate the numerical predictions. In addition, perforation energy and specific energy absorptions of various composite structures investigated are evaluated. Moreover, the rate-sensitivity of FMLs based on glass fibre reinforced epoxy and three aluminium alloys has been investigated though a series of quasi-static and impact perforation tests on multilayer configurations ranging from a simple 2/1 lay-up to a 5/4 stacking sequence. FMLs based on a combination of the composite and metal constituents exhibit a low degree of rate-sensitivity, with the impact perforation energy increasing slightly in passing from quasi-static to dynamic rates of loading. Then, finite element (FE) models are developed using the commercial code Abaqus/Explicit to simulate the impact response of PVC foam sandwich structures. The agreement between the numerical predictions and the experimental results is very good across the range of the structures and configurations investigated. The FE models have produced accurate predictions of the impact load-displacement responses, the perforation energies and the failure characteristics recorded. The analyses are used to estimate the energy absorbed by the skins and the core during the perforation process. The validated FE models are also used to investigate the effect of oblique loading and to study the impact response of sandwich panels on an aqueous environment and subjected to a pressure differential (equivalent to flying at an altitude of 10000 m). The modelling has been further undertaken on the low velocity impact response of the sandwich structures based on graded or composite reinforced PVC foam cores, with reasonably good correlation to the corresponding experimental results. Consequently, a series of finite element analyses have been conducted to investigate the influence of varying foam density, rod diameter, rod length and fibre type on the energy-absorbing characteristics of the reinforced foams. Perforation energies, impact resistance performance and unit cost of the structures have been evaluated. Furthermore, the low velocity impact response of fibre metal laminates has been studied numerically. Here, the composite layer in FMLs is modelled using the modified 3D Hashin’s failure criteria, which are implemented into the main programme through a user-defined subroutine, whilst aluminium alloys are modelled using Johnson-Cook plasticity and the corresponding damage criterion. A large number of simulations have been undertaken to cover FMLs with all stacking sequences and alloy types studied, which are compared with the experimental results in terms of the load-displacement trace and failure modes, with very good correlation. Similar modelling work has been carried out on the aluminium layer and composite layer individually. The energy to perforate the various FMLs is plotted and fitted on a single curve that can be used to predict the perforation energies of other configurations. The dynamic characteristics of the composite structures through a series experimental tests and numerical predictions investigated in this project can be used in the design of lightweight composite structures for energy-absorbing applications.

Emission legislation for passenger cars has become more stringent and the increasing demand for reduced fuel consumption has resulted in the introduction of complex new engine and after-treatment technologies involving significantly more control parameters. Vehicle manufacturers employ a time consuming engine parameter calibration process to optimise vehicle performance through the development of engine management system control maps. The traditional static calibration methods require an exponential increase in calibration time with additional calibration parameters and control objectives. To address this issue, this thesis develops and investigates a novel Inverse Optimal Behaviour Based Dynamic Calibration methodology and its application to diesel engines. This multi-stage methodology is based on dynamic black-box modelling and dynamic system optimisation. Firstly the engine behaviour is characterized by black-box models, based on data obtained in a rapid data collection process, for accurate dynamic representation of a subject engine. Then constrained dynamic optimisation is employed to find the optimal input-output behaviour. Finally the optimal input-output behaviour is used to identify feedforward dynamic controllers. The current study applies the methodology to an industrial state-of-the-art WAVERT model of a 1.5 litre Turbo EU6.1 Diesel engine acting as a virtual engine. The approach directly yields a feedforward controller in a nonlinear polynomial structure which can either be directly implemented in the engine-management system or converted to a dynamic or static look-up table format. The results indicate that the methodology is superior to the conventional static calibration approach in both computing efficiency and control performance. A low-cost Transient Testing Platform is presented in this work to carry out transient data collection experiments on a steady-state dynamometer with application to non-linear engine and emissions modelling using State Space Neural Networks. This modelling technique is shown to be superior to the polynomial models and achieves similar performance to non-linear autoregressive with exogenous input neural (NARMAX) network models. Numerical Dynamic Programming is investigated in a simplified engine calibration problem for a virtual engine to potentially improve the dynamic calibration optimisation stage. In a second study the novel dynamic calibration methodology is applied to the airpath control of a 3.0L Jaguar Land Rover (JLR) turbocharged Diesel engine utilizing a direct optimisation approach and State Space Neural Network models. A complete experimental application of the methodology is demonstrated in a vehicle where the vehicle-implemented calibration is obtained in a one-shot process solely from data obtained from the fast dynamic dynamometer testing. The results obtained demonstrate the potential of this methodology for the rapid development of efficient dynamic feedforward controllers based on limited data from the engine test bed.

Numerical modelling of scour around offshore structures is still a challenging research topic for engineers and scientists due to the complex flow-structure-seabed interactions. In comparison to single-phase models and Eulerian models with Exner equation, a multiphase approach has advantages in interpreting the flow-particle and particle-particle interactions. In the present study, an Euler-Lagrange multiphase approach is adopted to develop a new scour model in order to simulate the air-water-sediment interplay simultaneously while being computationally efficient. The model is able to represent free-surface flow with a mobile bed, which is often critical for realistic scour modelling. Based on the open source computational fluid dynamics (CFD) software package OpenFOAM®, the model solves the Navier-Stokes equations on an Eulerian computational grid. The sediment particles are traced using the multiphase particle-in-cell (MP-PIC) method in a Lagrangian approach. The drag force from the fluid, body forces and inter-particle stresses as well as the interphase momentum transfer are all accounted for in the model. The model system is calibrated using several simple test cases, including a falling particle and steady flow passing isolated blocks, to identify optimal parameters for model operation. The model is then validated against available experimental data on a steady current around a vertical cylinder and sand suspension under oscillatory sheet flow, amongst other tests, with satisfactory agreement. Application of the model against laboratory experiments includes benchmark scour cases underneath a horizontal pipeline under currents and waves, respectively. The tunnel erosion and lee-wake erosion stages are captured well by the model. The scour prediction matches with the measurements. In addition, the onset of scour is reproduced vigorously without any additional numerical assumptions or approximations. The model's capability to resolve the scour process and reveal the mechanisms involved is presented well.

The simulation of sediment motion in a hydraulic model of an estuary or a tidal river is a complex problem, and one that has received considerable attention in recent years. As a result of this, there exists a large number of published theories, each of which might be used as a basis for the design of hydraulic models. This thesis seeks to understand and to relate several of these design methods, including those that have been adopted as standard practice in some of the leading hydraulic laboratories. Eight methods are examined in detail, and the similarities and the discrepancies noted. The conclusion is reached that there is still an insufficient amount of experimental data available with which to assess these conflicting proposals. A series of experiments is therefore undertaken rising a hydraulic model of the Tay estuary. Particular attention is given to the choice of the bed material and the sediment time scale. Five different bed materials are used in the investigation, and the development and the shape of the different bed formations are studied in detail. The experiments are repeated with a reduced vertical exaggeration. The results are correlated with each design method and the deviations evaluated systematically so that each method is ranked in order of agreement. Although most of the design methods are found to give a reasonable indication of the sediment time scale, no one method is found to be sufficiently comprehensive to include all aspects of sediment motion similarity. Certain conclusions are made in connection with this, and some proposals put forward concerning future work in this interesting field of study.

No description available.

No description available.

No description available.

No description available.

No description available.

In this work we present finite element-based simulations of magnetic nanostructures using the micromagnetic software packages Nmag and Finmag developed at the University of Southampton. As part of this work the package Finmag has been extended with the implementation of an eigenvalue-based method to compute resonant modes in magnetic nanosystems. The details of this implementation are discussed, including certain complications encountered in the context of a finite element discretisation scheme. The implementation is verified using results from an independently published study on eigenmodes of an elliptical nanodisc. We present two studies of domain walls in magnetic nanowires. The first one investigates field-driven domain wall motion in nanowires with edge roughness. A new roughness model is introduced which allows the systematic study of how edge roughness features influence the domain wall motion compared to the case of a smooth nanowire. While the large-scale behaviour, such as the asymptotic domain wall velocity, is largely unaffected by the roughness, it introduces marked local alterations to the domain wall trajectories and can lead to dynamic pinning, both below and above the Walker breakdown. It is shown that the effective pinning strength of the roughness features is strongest when their size is comparable to the size of the domain wall. The second domain wall study investigates different types of resonant modes (translational, breathing and twisting modes) of transverse domain walls pinned at notches in a magnetic nanowire. The different sensititivies of each mode type on the nanowire and notch geometry are investigated in detail. It is found that the translational and twisting mode respond relatively strongly to changes in the notch geometry, while the breathing mode is fairly robust to changes in the notches’ size, making it a promising candidate for applications. We finally present a study of resonant modes in an elliptical magnetic nanodisc representing the free layer of a spin-torque nano-oscillator. We demonstrate that the resonant frequencies and spatial mode profiles are altered in the presence of a magnetic nanoparticle. The dependence of the frequency shifts on the nanoparticle position and material parameters is studied systematically. It is shown that these frequency shifts exceed achievable linewidths in state-of-the-art spin-torque oscillators and that they can be maintained over large external field ranges (owing to to the fact that they are a direct response to the stray field of the nanoparticle and do not rely on changes to the magnetic ground state of the disc). This opens up promising applications for novel nano-sensing devices using frequency-based detection schemes.