In recent years there has been an increased interest in developing fault current limiters for power distribution networks. This arises from the need to cope with the ever increasing short-circuit levels and to reduce the stress on system equipment, e.g. transformers, circuit-breakers and cables. It is also due to the increased concern about power quality, where fault current limiters are expected to play an important role in mitigating voltage sags during faults. Various devices to limit the fault current have been proposed, such as controlled fuses, tuned LC circuits, solid-state and superconducting fault current limiters. This research investigate the use of a novel technique to develop a solid-state Fault Current Limiting and Interrupting device (FCLID) suitable for low voltage distribution networks. The FCLID mainly consists of a high-speed bi-directional semiconductor switch, a varistor (non-linear resistor) and a snubber circuit; all connected in parallel. The semiconductor switch and the varistor share the fault current during the period of FCLID operation. To protect the semiconductor switch and the varistor from damage due to overheating, their temperatures are indirectly monitored in order to define the maximum operating time of the FCLID. A new method for estimating the junction temperature of the switching device and the varistor under transient condition has been developed and experimental tests were carried out to validate the proposed method. The energy handling capability of varistors and associated problems due to their non-linear characteristics have also been investigated. Experimental tests were carried out to measure the energy handling capability of the varistor using thermal imaging system. A new method for improving the current sharing between parallel varistors has been implemented. A computer model of the FCLID has been developed and implemented into a typical distribution network using MATLAB/ SIMULINK. The network performance under different conditions has been analysed. An experimental single-phase 230 V prototype FCLID was developed and tested under different operating conditions. Finally, the outcome of the theoretical, simulation and experimental phases of the research was used to establish the outline design specifications of a FCLID suitable for 11 kV distribution networks.

The Unified Power Flow Controller (UPFC) is the versatile FACTS controller that can control up to three transmission system parameters individually or simultaneously in appropriate combinations. The work presented in this thesis is concentrated on the modelling and control of the UPFC. The overall aim is to provide effective tools for optimising the impact of the UPFC in the reinforcement of a transmission system. Existing modelling techniques for the UPFC together with the associated control strategies have been systematically reviewed. An exact power injection model is proposed which is based on the polar representation of the UPFC parameters and includes the reactive power capability of the shunt inverter. In addition, a steady-state model based on an ideal controlled voltage source has been developed using MATLAB/SIMULINK which provides a useful tool to analyse and develop the UPFC control system. The UPFC internal limits have been identified and accordingly, the feasible operating area of a transmission system incorporating a UPFC has been determined based on the UPFC maximum limits. The influence of both the series and shunt inverters on this controlled area has been analysed. The impact of a change in the system short circuit level on the UPFC operation and the size of the feasible area has also been investigated. Three modern controllers have been designed and tested for controlling the UPFC in a power flow mode for the series part and a voltage control mode for the shunt part. These controllers are: a fuzzy knowledge based controller, an artificial neural network based controller and a neuro-fuzzy based controller. For the former, the fuzzy rules are deduced from the relationship between the controlled power system parameters and the UPFC control variables. The second is a simple RBFNN controller which is constructed from a single neuron and trained on-line by a gradient descent algorithm. The third controller is designed using the adaptive capabilities of neural networks to estimate and tune the fuzzy rules. Computer simulation and experimental implementation of a UPFC using DS 1103 data acquisition board have been used to verify the proposed control strategies. In the experimental lab model, two 6-pulse inverters implementing the SPWM technique have been used to realise the UPFC system.

For many years, it was generally understood that a.c. transmission systems could not be controlled fast enough to handle dynamic system conditions. The dynamic system problems were usually handled by over-design, which resulted in under utilisation of the system. Flexible AC Transmission System (FACTS) devices play an important role in improving the dynamic performance of a power system and hence achieve better utilisation of the available system. They are principally employed to 'rapidly' control one or more of the three main parameters directly affecting a.c. power transmission, namely the system impedance, magnitude and phase angle of the system voltage. The Voltage Source Inverter (VSI) is the basic building block of most FACTS devices. The multi-level VSI topologies are becoming the favourite power circuits for the 2nd generation of FACTS shunt compensators. The research reported in this thesis is to investigate the reliability of Voltage Source Inverter topologies that are used in high power applications, mainly the Advanced Static VAr Compensator (ASVC). The inherent redundancy of the diode-clamped multi-level VSI topology, with respect to short-circuit and open- circuit device faults, is investigated using a space-vector nodal representation. The harmonic performance of the ASVC under normal and during 'device fault' operating conditions is also investigated. A new multi-level inverter topology is proposed to improve the robustness of the conventional diode-clamped VSI topology. Harmonic spectrum 'recovery' techniques to be utilised in the event of device failure are also investigated and discussed. An adaptive PWM controller is proposed to maintain an acceptable low order harmonic performance for the ASVC under normal and abnormal operating conditions. The results obtained show that the proposed system can maintain uninterrupted operational performance throughout certain device failure conditions. An experimental 3-level discharge path protection switch clamped (DPPSC) VSI system has been designed, constructed and analysed. To demonstrate the 3-level adaptive SHEM strategy, the adaptive DPPS controller was implemented on a TMS320F240 DSP evaluation module (EVM). The results were in good agreement with those predicted in the analytical and simulation parts of the work. The research carried out in this work showed that under loaded operating conditions, the low frequency harmonic components targeted by SHEM techniques are not fully eliminated from the output voltage spectrum. This investigation revealed that this is due to the harmonic interaction between the a.c. and d.c. sides of the multi-level inverter. A new 'Dynamic Selective Harmonic Elimination Modulation (DSHEM)' scheme is proposed to overcome this problem. The DSHEM dynamically adjusts the switching angles according to the system operating point. The proposed method is verified using simulation and the experimental model.

A practical Vertical Axis Wind Turbine (VAWTs) based on a Darrieus rotor has been designed and tested and found to be capable of self-starting at wind speeds above 4m/s. The self-start feature has been achieved by replacing the usual symmetrical aerofoil blade in the VAWT rotor and by using a concentric Savonius rotor or semi-cylinder turbine. A computer program was produced to compute the power coefficient versus tip speed ratio characteristics of a selected aerofoil profile employed in a VAWT. The program accounts for chord length, pitch angle, number of blades, and rotor radius at any wind speed. The published data from 40 aerofoil sections were assessed, taking into account the two main criteria — self-starting and efficiency. Computational fluid dynamics software (ANSYS, Flotran) has been used to investigate the lift and drag performance of a NACA 66-212 and NACA 4421 aerofoils in order to check the computer program predictions. Excellent agreement was obtained for the static aerofoil assessment, but only after special ICEM Computational Fluid Dynamics (CFD) meshing interface routines were utilised. However, agreement between the theoretical and published results was not good for the rotating aerofoils in a VAWT. Thus, further CFD work was not pursued and in preference, an experimental route was initiated. In the first series of wind tunnel tests involving three candidate profiles, good agreement was found between the experimental results and the mathematical models. The aerofoils chosen were the NACA 661-212, the 51223 and the Clark-Y standard aerofoils. A number of prototype VAWTs were fabricated and tested for the influence of the blade pitch angle, the chord length ratio, with 2 or 3 blades. The aerofoil surfaces were made from aluminium sheet with a standard surface finish. The prototype designs were tested in the Northumbria University low speed wind tunnel facility - the models were 0.4 m. high with a 0.4 m diameter. The torque versus wind speed characteristics were recorded and analysed. The S 1223 profile was found to be self-starting with high efficiency. This model generated a high power coefficient of about 0.3 at a tip speed ratio of 1.2. The second series of tests were carried out in field sites in the UK with a 2 m diameter straight—bladed Darrieus rotor prototype with 3 blades using the S1223 blade section. Three field trials were undertaken in the UK to produce realistic performance characteristics for wind conditions of 4-10 m/s. The maximum power coefficient of this machine was found to be 0.18 at a tip speed ratio of 1.2. In addition, an alternative semi-cylinder turbine combined with a Darrieus rotor was fabricated and tested in the UK. It demonstrated the advantage that it could self-start at lower wind speeds, that is 3m/s but delivers approximately 50% less power than that obtained from the first proposed design. A final phase of testing was carried out with an enlarged and modified 3 m diameter prototype installed at a shrimp farm in Thailand to demonstrate how the unit could be used to replace an equivalent 2 HP 2-stroke diesel engine and hence eliminate its inherent emission pollution problems. A Savonius rotor was fitted to the prototype to improve self-start capabilities at a wind speed of 4 m/s for a practical application which by its nature required a high starting torque. The designs are easy to fabricate, low cost, pollution free and have been demonstrated to be ideal for applications in developing countries where there are sufficient wind resources.

Multistatic radar systems are of emerging interest as they can exploit spatial diversity, enabling improved performance and new applications. Their development is being fuelled by advances in enabling technolo gies in such fields as communications and Digital Signal Processing (DSP). Such systems differ from typical modern active radar systems through consisting of multiple spatially diverse transmitter and re ceiver sites. Due to this spatial diversity, these systems present challenges in managing their operation as well as in usefully combining the multiple sources of information to give an output to the radar operator. In this work, a novel digital Commercial Off-The-Shelf (COTS) based coherent multistatic radar system designed at University College London, named 'NetRad', has been developed to produce some of the first published experimental results, investigating the challenges of operating such a system, and determining what level of performance might be achievable. Full detail of the various stages involved in the combination of data from the component transmitter-receiver pairs within a multistatic system is investigated, and many of the practical issues inherent are discussed. Simulation and subsequent experimental verification of several centralised and decentralised detec tion algorithms in terms of localisation (resolution and parameter estimation) of targets was undertaken. The computational cost of the DSP involved in multistatic data fusion is also considered. This gave a clear demonstration of several of the benefits of multistatic radar. Resolution of multiple targets that would have been unresolvable in a conventional monostatic system was shown. Targets were also shown to be plotted as two-dimensional vector position and velocities from use of time delay and Doppler shift information only. A range of targets were used including some such as walking people which were particularly challenging due to the variability of Radar Cross Section (RCS). Performance improvements were found to be dependant on the type of multistatic radar, method of data fusion and target characteristics in question. It is likely that future work will look to further explore the optimisation of multistatic radar for the various measures of performance identified and discussed in this work.

Among the number of optical sources, vertical-cavity surface-emitting lasers (VCSELs) are relatively recent type of semiconductor laser devices, which are attractive for a number of applications particularly for free space optical (FSO) communication systems. In such systems reliable optical devices with lower power consumption and low cost are among the key requirements. VCSELs typically operate with unstable output polarization modes, and there is a need to improve their output power regarding to the polarization instability, particularly when introducing the optical feedback (OF). This thesis investigates a number of key properties of VCSEL including the polarization instability, hysteresis loop (HL), relative intensity noise (RIN) and how to control the polarization switching (PS). The investigations are based on the analytical studies and extensive experimental work. PS properties of VCSEL are investigated by introducing variable polarization optical feedback (VPOF) with the modulation frequency and modulation depth. The dependency conditions for the HL, RIN and PS are determined with VPOF. Under OF, the threshold current (Ith) of VCSEL is reduced by 11.5% and the PS, which is demonstrated theoretically and experimentally, is completely suppressed. The PS positions are depending on the polarization angle of OF, OF levels and the bias current. The PS disappeared with the modulation depth of 78.66%, whereas it is entirely vanished with the modulation frequency of 200 MHz. The hysteresis width of the VCSEL polarization modes is reduced by increasing the feedback level. The minimum RIN value of -156 dB/Hz is achieved at a zero degree of the polarization angle for the dominant polarization mode of VCSEL under VPOF. For the first time, a novel technique based on employing orthogonal polarization OF is proposed to supress the nonlinearity associated with the modulated VCSEL, where the second, third, and fourth harmonics are completely suppressed to the noise floor. Finally, optimal operating conditions for a high-quality polarization-resolved chaos synchronization of the polarization modes of VCSEL with VPOF are experimentally and theoretically studied. A perfect value of 99% of the correlation dynamics for the chaotic synchronization of the polarization modes of VCSEL is found with a zero time delay over a wide range of polarization angle. Finally, Simulink and Origin software version 6.1 are used in this work to simulate and plot the results. The simulation results are agreed with the experimental results, which show that the chaotic synchronization dynamic of the polarization modes can be achieved by VPOF.

Dye-sensitised solar cells (DSSCs) based on metal oxide semiconductor photoanodes (i.e. TiO2, ZnO) offer a promising route for low-cost and transparent solar cells, especially suitable for indoor/outdoor applications, building and automotive integrated electricity generation. Apart from developing new dyes (or absorbers) for increasing absorption and stable fast-regenerated electrolytes, improving metal oxide photoanodes has also gained great research attention. In a conventional DSSC, the mesoporous TiO2 photoanodes assembled by TiO2 nanoparticles can support large surface areas for sufficient dye (absorber) loading, thus result in reasonably good solar cell performance. However, the poor pore-filling of large sized molecules (i.e. solid electrolyte) and inefficient electron transport lead to significant photo-generated charge recombination thus loss of photo-generated energy. Despite the reasonably good electron transport ability of the currently used particulate-based photoanodes, the requirements of high temperature processes for these photoanodes significantly limit the substrate and material choices. In this thesis, a low-temperature strategy was designed to synthesise crystallised metal oxide (ZnO and TiO2) nanostructures with controllable morphologies to be used as photoanodes with improved electron transport abilities for DSSCs. ZnO nanorods (NRs) with tailored nanostructure (i.e. growth direction, aspect ratio and surface distribution density) were synthesised on pre-seeded substrates using zinc salt based aqueous solutions at temperatures generally lower than 95 oC. The influences of reaction temperature, pH, concentration, reaction duration and additives were systematically investigated. The detailed studies of nanostructures, morphology, crystallinity and properties of the ZnO NRs led to an improved understanding of the synthesis process. Anatase TiO2 modification layer was achieved using a plasma ion assisted deposition (PIAD) without external heating or subsequent annealing. Effects of the deposition parameters (duration, gas flow rates and plasma energy) on TiO2 film properties (optical, structural and chemical activities) have been studied. By combining two lowtemperature processes (aqueous solution growth of ZnO NRs and PIAD of crystalline TiO2 nanostructures), nano-sculptured ZnO-TiO2 nanostructures were achieved. The ZnO NRs were covered with a layer of anatase TiO2 to form core-shell and foxtail-like nanostructures. These nanostructured photoanodes showed an improved electron transport as well as suppression of recombination capability in the DSSCs assembled by these photoanodes. A novel in-situ microfluidic control unit (MCU) was designed and applied in the aqueous solutions synthesis process, which provided an easy way to localize liquidphase reaction and realise selective synthesis and direct growth of nanostructures, all in a low-temperature and ambient pressure environment. The morphology of the nanostructures was controlled by varying the amount of additivities supplied by the MCU. This achieved a facile fabrication of Al-doped ZnO (AZO) nanoflakes vertically grown on flexible polymer substrates with enhanced dye loading and electron transporting capabilities. Flexible DSSCs with a significant enhancement (410% compared to ZnO NRs based devices) in the power conversion efficiency were obtained using the AZO nanoflakes photoanodes of 6 µm thickness, due to the enhancement in electron transport capability of the photoanodes and reduction in the recombination process.

This thesis investigates areas of neural networks and their application to aspects of image processing. Three neural network models, namely the backpropagation network, the Hopfield network and the competitive network, are studied. First, the learning algorithms for backpropagation networks and competitive neural networks are studied and new algorithms are developed. The applications of Hopfield neural networks to image coding, and of backpropagation networks to image data compression, are then investigated, and new techniques are presented. The learning algorithms for feed-forward neural networks are studied. A modified backpropagation algorithm is developed for accelerating the learning of backpropagation networks. Simulation results for four benchmark tasks are presented which show that the new method provides improved acceleration of learning, and global convergence characteristics. The learning algorithms for competitive neural networks are investigated. By incorporating principles of statistical mechanics into competitive learning, a new simulated annealing procedure for training competitive neural networks, the Stochastic Competitive Learning Algorithm (SCLA), is developed. Simulation results are presented which show that the SCLA is insensitive to the initial values of weight vectors, and can achieve a lower cost function value than other established competitive learning algorithms, and is therefore a valuable tool for data clustering. The computational power of Hopfield neural networks is employed for image coding applications. A new Hopfield neural network based block truncation coding (HNNBTC) technique is developed. The new HNNBTC is shown to provide improved performance over established block truncation coding techniques both in terms of mean square error performance and visual quality of the coded images. Two variations of the HNNBTC technique are also presented which are shown to provide increased compression ratios without sacrificing much of the visual quality of the coded images. The application of backpropagation networks to image data compression are investigated, and two new techniques are developed. The ceniralised MLP (CMLP) image data compression scheme is developed, which is aimed at improving the networks' generalisation capabilities, thereby enabling them to effectively compress a wide range of novel images. The learning edge patterns using MLPs (LEPMLPs) scheme, for image data compression is developed, which is aimed at improving the reconstruction of edges, thus improving the visual quality of the reconstructed images. Tabulated experimental results and example reconstructed novel images, for the original method and the new techniques, are presented, which demonstrate the improved image compression perfonnance gained using these new techniques.

Synthetic aperture radar (SAP) is gaining increasing importance in the field of remote sensing. The main feature of image data transmitted from orbital satellite synthetic aperture radar is all-weather performance, which explains its importance in many applications. Unfortunately SAP image data is susceptible to "speckle noise", which means that considerable processing is required before full use can be made of the data. This thesis describes research performed into the enhancement of linear features, corresponding to ships' wakes, in noisy SAP satellite images. A summary of noise reduction techniques is described, and the suitability of each to the problem of enhancing linear features in SAR images is discussed. A survey of available hardware is also included, with an assessment of its suitability for the selected technique, namely the Radon transform. A transputer network is described in detail, together with OCCAM software to perform the Radon transform within a few seconds. The Radon transform, performed via the two-dimensional frequency domain, uses a technique based on the Fourier slice theorem, and in this thesis a previously reported ambiguity is confirmed and analysed. The development of a new Radon transform is described, which enables transformation into Radon space (for contrast enhancement), and back into image space, all via the frequency domain. This new transform was also implemented in OCCAM on the transputer network, and it is shown that the new transform has advantages over a previously reported line enhancement techniq'ie in that it eliminates the ambiguity found in the standard Radon transform, and reduces computation considerably when the inverse transform is also required.

The aims of the project were to investigate DC Thyristor Circuit Breaker (TCB) current interruption and appraise the current interruption performance in simulated systems. A generalized analysis was developed which enabled succinct assessment of the current interruption performance of different thyristor circuit breaker configurations. A 250V, 60A DC Thyristor Circuit Breaker was analysed and a study of TCB-system interaction was made. A novel thyristor circuit breaker switching overvoltage control technique was also investigated. The project findings on TCB-system interaction demonstrated that practical thyristor circuit breaker selection and design must be done in conjunction with the system configuration and its parameters. The normalized and analytical approaches were found to be useful for determining TCB design criteria but computer simulation tools, such as SPICE, are necessary for predicting the full range of TCB current interruption.