This thesis presents a study on HVLF AC transmission systems for a long distance Offshore Wind Farm (OWF) grid connection. A particular scheme highlights the use of a high voltage cycloconverter as a frequency changer at the sending end of the transmission system, in which the voltage is stepped up at 60 Hz before inputting to the cycloconverter. This eliminates the need for a high voltage low frequency transformer on the offshore platform, and also it allows the use of standard 50/60 Hz generating equipment in the wind turbines. A modelling study has been undertaken to validate the operation of the system, including with the presence of transmission line faults. The study shows the effects of the pre-fault harmonics generated by the cycloconverter on the fault generated transients. Thus, the need to develop a new frequency based fault protection system for HVLF transmission is addressed. New transmission line fault detection and location algorithms for the HVLF system have been developed. Firstly, the frequency range of the post-fault generated transients were identified using Fast Fourier Transform (FFT) analysis. It was shown that these transient components, ranging from DC to high frequency, are distributed throughout the spectrum of the three-phase current signals. However, the FFT analysis provides the frequency information of these transients but without time information. In this thesis, the Wavelet Packets Transform (WPT) is introduced for the fault identification. The fault generated transients were detected by monitoring wavelet coefficients over a time window. The performance of the protection system under all possible fault scenarios of the HVLF transmission line are investigated. Finally, practical considerations, such as the impact of the fault inception angle and the switching and control of the cycloconverter on the fault detection and location algorithm’s accuracy were also investigated.

A model chemical plant was designed and constructed for the purpose of experiments in sophisticated control techniques using an on-line digital computer. It proved essential to develop a mathematical model of the major units of the plant which were two mixer-settler extraction tanks. The model was based on the physics and chemistry of the process but involved four indeterminate parameters required to represent the complex extraction mechanism. Qualitative comparisons were made between the responses of the plant and of the model to step changes in the input variables. Efforts were then directed towards the development of an adaptive model whose parameters were adjusted by a hill-climbing procedure. Tests were carried out using a second model with fixed parameters to represent the plant. These proved a failure both because of transient effects and the special features of the hill-climbing problem. New systems were developed to overcome these difficulties for a more restricted two parameter case. In addition an investigation was carried out into the use of a normalised error criterion as proposed by Nightingale and its application to adaptive modelling.

This work is concerned with the optimal conditions of flow control of countercurrent chemical processes. Such processes abound in chemical engineering, in a number of forms. A parallel treatment of both continuous and discrete systems is pursued, largely under the mantle of distributed control system theory. Examination of common features of counterflow processes yields performance criteria and control constraints which may be chosen as a measure of controller efficiency. Based on these criteria a standardised system is investigated according to several distinct approaches, each tested practically against a particular physical countercurrent system. The derivation of transfer functions relating exit boundary conditions to input flow and input boundary conditions is extended, in order to cover the case of spatially dependent flow dynamics, with reference to a solvent extraction process. The complexity and role of terms entering into the transfer function is assessed from a viewpoint of sub-optimal regulatory control. A more general approach surveys the theory of partial differential equations, in order to investigate approximate means of representation. Out of a number of variational approximation methods, it is shown that a discrete Fourier series provides the most suitable reduction to a lumped system. Using distributed gradient techniques, it is possible to obtain optimal flow profiles corresponding to regulatory and maximum extraction cost functions. Such a method is shown to extend to the case of spatially varying flow dynamics. Indirect optimal control methods are also applied to the distributed problem, when both dynamics and performance criteria are expressed in lumped system form. The multiplicative role of flow parameters is found to persist through both the direct and indirect methods of optimal control.

The stability of some particular classes of control systems described by ordinary nonlinear differential equations is considered. As a means of introduction to the problem, systems containing a single nonlinearity in an otherwise linear, time-invariant closed loop are examined. Stability criteria based on the frequency-response of the linear part of the system are established by constructing a Liapunov function of a 'quadratic plus integral of non-linearity' form. The problem is extended to cover those classes of control systems which contain several such nonlinear functions (i.e. multivariable control systems) and frequency domain stability criteria are established by constructing a Liapunov function akin to that described above. It is also asserted that stability criteria less restrictive than those obtained previously for these multivariable systems may be achieved by placing certain additional restrictions on the nonlinear functions. Some classes of systems containing nonlinear functions of a most general nature are considered in later chapters of this thesis. Frequency-domain stability criteria are established with the aid of quadratic forms of Liapunov functions. Again, if the complexities of these nonlinearities are reduced it is seen that less restrictive criteria than obtained previously may be established for these classes of systems. Emphasis is laid throughout upon the development of a unified approach to the problem of stability of the classes of systems considered. The criteria, once formulated, can be applied in practice without any further reference to the Liapunov function used.

A series of six equilibrium turbulent boundary layer flows hais been established with values of H from 1.3 to 2.3, and measurements made in them of the effect on the boundary layer development of a single two- dimensional roughness element of mainly square cross-section mounted near the start of the equilibrium region. It is shown that the local increment of the momentum thickness caused by the element is well-predicted by the flat-plate correlation of Gaudet and Johnson, a correlation which is here shown to be universally valid and that, for all flows except for the most adverse pressure gradient, a satisfactory prediction of the subsequent boundary layer development can be made with the aid of relationships proposed by Professor Bradshaw, for the change in H at the roughness element. For the flow with the largest value of H the prediction method for the development fails even in the absence of the element which, in fact, hats little influence on the flow. The discrepancy between calculation and experiment is much larger than can be accounted for by normal stress terms and the reasons for this discrepancy are not entirely evident. However, the essential outcome of the experiment is clear that the incremental drag of a roughness element depends on wall variables. In consequence, the effect of an element which is of small height compared with the boundary layer thickness is negligible in flows with strongly adverse pressure gradients. It is also demonstrated that the length of the separation region behind small roughness elements decreases as the pressure gradient increases adversely.

A central problem in control system design is how to design a controller to guarantee that the closed-loop system is robustly stable and that performance requirements are satisfied despite the presence of model uncertainties and exogenous disturbance signals. The analysis problem, that is the assessment of control systems with respect to robust stability and robust performance, can be adequately solved using the structured singular value u as introduced by Doyle. The corresponding design problem (how to choose a controller K to minimize u) is still largely unsolved, but an approximate solution can be found using Doyle's D - K iteration. In this thesis we present an alternative algorithm, called u - K iteration, which works by flattening the structured singular value u over frequency. As a prelude to this a classical loop shaping approach to robust performance is presented for SISO systems, and is also based on flattening u. In u-synthesis it is often the case that real uncertainties are modelled as complex perturbations but the conservatism so introduced can be severe. On the other hand, if real uncertainties are modelled as real perturbations then D - K iteration is not relevant. It is shown that u - K iteration still works for real perturbations. In addition, a geometric approach for computing the structured singular value for a scalar problem with respect to real and/or complex uncertainty is described. This provides insight into the relationship between real u and complex u. A robust performance problem is considered for a 2-input 2-output high purity distillation column which is an ill-conditioned plant. Analysis reveals the potentially damaging effects on robustness of ill-conditioning. A design is carried out using u - K iteration and the "optimum" u compared with that obtained by Doyle and by Freudenberg for the same problem.

A robust output feedback pole assignment method, which seeks to achieve a robust solution in the sense that the assigned poles are as insensitive as possible to perturbations in the system parameters, is studied. In particular, this work is concerned with pole assignment in a specified region rather than assignment to exact positions, whereby the freedom to obtain a robust solution may be realized. The robust output feedback pole assignment problem is formulated as an optimization problem with a special structure in matrix form. Efficient optimization methods and numerical algorithms for solving such a problem are proposed by introducing a concept of the derivative of a matrix valued function. The homotopy method, which is known as a globally convergent method, is applied to solve the robust output feedback pole assignment problem to overcome possible difficulties with the choice of feasible starting point. A new algorithm based on the homotopy approach for solving the pole assignment problem is proposed. Numerical examples of the robust pole assignment problem demonstrate how the homotopy algorithm globally converges to optimal solutions regardless of initial starting points with an appropriately defined homotopy mapping. The proposed algorithms are illustrated using an aircraft case study. It is seen that the controllers obtained using robust pole assignment methods yield the robust flight control and maintain the closed-loop system properties closer to the nominal ones. They are shown to be more robust than those obtained by an alternative direct pole assignment method which is frequently used to develop aircraft control strategies without attempting to optimize any robustness criterion. Indeed, the robust output feedback pole assignment method proposed in this study is a method which can be applied for control system design to achieve one important design objective, robustness.

Leaking, friction and wear of seals are concerns for machine designers and users everywhere. Although perfect sealing may be the general aim, in practice, considering apparently identical seals in the same application, some may seal while some may not. This is due, at least in part, to surface-related random phenomena. Therefore, the importance of considering the reliability of sealed joints cannot be overemphasized. Up to now, there is no paper in the published literature about the reliability analysis of static sealed joints. All of these facts provide the motivation for the current research work. A computer simulation model for the leakage analysis of static sealed joints has been developed based on the percolation theory. The features of the leakage simulation model can be concluded as follows: (1) It reveals the effect of random properties of rough surfaces on the sealing performance and makes it possible to apply the statistical concepts in discussing the sealing reliability of static sealed joints; (2) It provides much simpler and more economic tool for the statistical analysis of leakage by computer simulation than by experiments; (3) It makes it possible to describe the leakage phenomenon more accurately using the leakage path model instead of the clearance between surface centre-lines; (4) It eliminates the need for individual asperity model of rough surfaces, because the actual digitized surface is used directly. The relationship between the leakage probability and the applied load, which is of great general interest to the designers of static sealed joints, has been predicted by the leakage simulation model. The simulated results show that for a given leakage probability, the required load will increase as the value of RMS o height a increases or the value of correlation length decreases. It is confirmed that a certain value of contact ratio can be used as the criterion for identifying the reliability of static sealed joints with a certain confidence level. The contact ratio criterion provides a simple, inexpensive and useful tool to evaluate the effects of rough surfaces, material properties and applied load on the sealing reliability of static sealed joints. However, in order to be of practical use, experimental work is required to evaluate its validity.

The final properties of hot rolled products depend on the microstructure developed during the hot rolling process. The kinetics of thermomechanical events which determine the microstructure depend very much on the strain accumulated at each deformation. Aluminium and its alloys exhibit a high static recovery which reduces the strain in the interpass times, although this strain may be retrieved when applying a second deformation. This thesis is about determining the extent of static recovery in hot rolling in terms of a magnitude called 'Retained Strain', to be used in a computer model for hot rolling. A methodology to measure the extent of static recovery and retrieval of strain is proposed. This methodology is based on the development of empirical equations which were obtained after analysing experimental data from plane strain compression tests from two different sources. In the analysed data the extent of static recovery is related to the deformation conditions so that a relationship between static recovery and the Zener-Hollomon parameter is established. The retrieval of strain depends on the time interval between deformations, so that after very short intervals the strain retrieves immediately, even at very small second strains, but after very long intervals strain may not retrieve at all. The Sheffield Model for hot Rolling is modified to include simulation of static recovery and retrieval of strain. Industrial schedules for Al-l%Mg and c.p. A1 are studied. With the 'Recovery model' recrystallization is inhibited. With the 'Retrieved model' the effects of recovery are almost eliminated. Modelling recovery and retrieval of strain has little effect on industrial hot rolling schedules designed to produce a full recrystallized final microstructure, but it is expected to have a major influence on other hot rolling schedules where a final fine subgrain microstructure, characteristic of recovery, is required.

Airlines use the polar routes because for some flights the shortest route (the great circle) takes them over the polar regions. However, a lack of VHF ground stations over the oceans and within polar regions means that aircraft have to rely on long-range HF communications via the ionosphere. The ionosphere is affected by space weather more intensively in the high latitude regions. In severe space weather conditions, ionospheric HF communications become impossible, and aircraft have to be re-routed to lower latitudes. This increases the flight time, cost, passenger and flight loadings and impact on the environment. The research reported in this thesis designs a model of HF propagation that can provide nowcasting (and forecasting) to advice airlines of the likely quality of HF communications. The model consists of electron density profile, ray-tracing, and signal loss and antenna gain considerations. An electron density model partially comprises a background ionosphere, and the work presented in this thesis contributes by proposing an assimilative model based on the International Reference Ionosphere (IRI). This is then perturbed by ionospheric structures that occur within the high latitudes regions. Ionospheric absorption is then applied, and the gain at both antennas (transmitter and receiver) included. Finally, maps of signal coverage are generated for a specific date, time and frequency. Assessment of IRI model was discussed through comparisons of the prediction with the standard IRI model. Comparisons revealed improvements in foF2 using IRI. The predicted HF signal characteristics have been compared with the observations. The model shows an ability to yield results with good agreement with the observations. Features were reproduced through the model for two periods: one showing the influence of an absorption event, and the other experienced much off great circle propagation. A method of using the output of the nowcast model has been introduced with examples.