Large cyclo-converter drives using either synchronous or induction machines are found in a number of low speed high torque applications: these include steel mills, gearless cement mills, mine winders and ice breakers. The cyclo-converter output frequency is synthesised from portions of the high voltage supply system. At low frequencies this produces an output current waveform which is reasonably sinusoidal, as the frequency increases the waveform become more distorted. There is concern for the level of current harmonics produced by the cyclo-converter in both the motor windings and the supply system. The waveform distortion is mainly due to the switching action of the cyclo-converter, this is modified by the effects of interbridge changeover delay, commutation overlap, the supply and decoupling reactances. Industrial drives are required to conform to the Electricity Council's regulations on input current harmonics and voltage waveform distortion. In marine drives the main generators feed the ship propulsion system and the ship's auxiliary supplies. The level of interference produced by the converter input line current on the supply network is of particular concern. High levels of voltage distortion may cause malfunction of the auxiliary equipment and produce large torque pulsations in the supply alternator. The cyclo-converter generates unwanted harmonics in the output current waveforms, these are responsible for torque pulsations on the motor drive. With naval marine drives the torque pulsation is of particular interest, since this level must be minimised to increase the sensitivity of detecting instruments such as passive sonar. To date, mathematical models and computer based models have not detailed all the aspects which are responsible for the distortion to the converter input and output waveforms. This thesis seeks to address this and identify the effects on the converter performance of, the converter interbridqe delay, overlap and the supply parameters supply and decoupling reactance. Various motor models will also be examined. The author has used the mathematical work by Pelly as a basis, and has identified the percentage errors in the level of harmonic currents which are likely to be introduced if his results are used in designing new systems. Comparisons will be made with experimentally produced results obtained from a mine winder at the Wearmouth Colliery. Results from a lower powered cyclo-converter produced experimentally and computationally at Newcastle University will also be examined.

The ability that many blind people develop of detecting near objects by means of echo-location is well known. This investigation is concerned with the development of an aid that will enhance this ability by transmitting ultrasonic pulses into the environment and receiving the echoes returning from surrounding objects. The echoes are detected and heard as clicks by the user, who is able to determine the direction from which the sound is coming by using his natural localisation ability. Further, the clicks are "coded" by switching the receivers on and off, so that the rate of clicking indicates the distance of an object. A key component of the system is the ultrasonic transducer. Piezoelectric bimorph transducers have been selected for use as both transmitters and receivers, although the characteristics of the standard type have been readjusted to meet the requirements of the aid. In particular, the transducer bandwidth has been considerably extended to ensure a sharp sounding click. In order to provide accurate directional information it has been necessary to closely simulate naturally occurring interaural differences. Interaural time differences are readily provided by spacing the receiving transducers at the 'inter-ear distance apart, whereas the provision of accurate interaural intensity differences has necessitated the shaping of the directional response of the receiver to match that of the ear. A series of psychophysical tests has been performed to measure the ability of subjects using the device to localise objects in the horizontal plane. Considerable accuracy was shown with the best results being obtained using interaural time information alone or interaural intensity information alone, rather than a combination of both. Elementary outdoor trials with the device have shown the range and directional information to be effective with the indication of the shoreline being particularly clear. A progressive refinement of the aid should lead it to a stage where its usefulness to blind people can be evaluated.

The arrival of cheap and fast microcomputers has stimulated the development of machine condition monitoring systems. Such systems measure one or more machine performance parameters, with a view to detecting the early signs of failure and initiating some form of action. To help developing these systems, a thorough knowledge of the behaviour of the machine in its post-fault condition must be well known and well understood. With this view in mind, a simple, yet reasonably accurate method of analysing the induction machine magnetic circuit under normal and abnormal conditions at moderate cost has been developed. Numerical field analysis is applied to the calculation of induction motors. It is based on a very simple Magnetic Network Technique (MNT) solution of the magnetic field. The field is assumed to be two-dimensional. The three dimensionality of the machine is taken into account within the two dimensional model. The general time-dependence of the field and the motion of the rotor are modelled correctly in a step-by-step solution. The model uses relatively small amount of computation time when compared with the previous methods of analysis. This technique is used for the evaluation of the broken bars effects on the machines performance and magnetic field variation. A series of experimental tests have been carried out. The results obtained are directly compared with the computed ones and they showed a good correlation. Finally, suggestions for suitable methods for the detection of broken bars are given along with some suggestions for future work.

This thesis covers the investigation into the use of Space Vector Modulation for the control of a 4-leg matrix converter, which is capable of providing a 3-phase plus neutral supply from a standard balanced 3-phase source. Traditional 3x3 matrix converters have limited use in this application as they are only capable of supplying a balanced three-phase load. It would be desirable to be able to power unbalanced and non-linear loads, requiring that the converter provides a neutral connection. As with voltage source inverters, this goal can be achieved by extending the number of output legs in the matrix converter to four. In this thesis, a new Space Vector Modulation technique is proposed for this 4-leg, or 3x4, matrix converter. This technique is an extension of the method currently in use on 3x3 matrix converters, and so it allows the de-coupled control of both the input and output. The thesis then goes onto cover the build of a demonstration converter, looking at the different aspects which make up a converter, to finally go on to prove the theory, and a set of results are presented to validate this.

The face of electricity generation, transmission and distribution is changing due to economic, technology and environmental incentives. Recently, interactive and intelligent electricity grid structures which consist of interconnected small/medium sized generators, power electronic technologies and energy storage elements have been developed to address the major shortcomings of the traditional electricity grid structure. Microgrids are key elements of these emerging grid structures. Although microgrids are accepted as possible solutions to power quality and power stability issues in ac power systems, the uncertainty in the ability of microgrids to cope with severe fluctuating load and fault conditions is a major concern in the operation of these new grid structures. This project was aimed at improving frequency control within a microgrid. Four objectives were identified and addressed to meet this aim. I. A weak microgrid network using an emulated internal combustion engine generator and associated loads was modelled. The emulation of a diesel generating set was achieved with a vector controlled induction motor driving a synchronous generator. The diesel engine emulation was achieved by incorporating a single delay into the speed control loop of the vector controlled induction motor. The modelled microgrid network is a very useful tool for the development of novel control schemes for frequency control within a microgrid. II. Simulation studies were carried out to investigate how a weak power system can be strengthened through the use of enhanced STATCOM and Supercapacitor energy storage. This assisted in understanding the limitations and performances of the novel algorithms proposed for frequency control improvement. III. Novel open and closed loop control algorithms for frequency control within a microgrid were proposed. The advantage of the open loop control scheme is its simplicity but the functionality of the control action is limited with the knowledge of the diesel engine transfer function and load current being important requirements. A closed loop control scheme was employed to address the limitations of the open loop control scheme. IV. A laboratory prototype of the microgrid network was developed and used in validating the novel control schemes proposed. The thesis describes the novel algorithms for frequency control using an intelligent STAtic COMpensator (STATCOM) and SuperCapacitor Energy Storage System (SCESS). The benefits and effectiveness of the proposed algorithms are given in the simulation and experimental results.

This thesis analyzes saliency-based sensorless control methods for AC surface mounted permanent magnet machines (PMSM), because PMSMs have features that make them attractive for use in industrial drives: small size, high efficiency, low maintenance, high dynamics, and high power density. The thesis focuses on four different HF injection sensorless methods, which utilize resistance and inductance based saliencies for position estimation: the measurement axis method, the eddy current resistance based saliency tracking method, the eddy current inductance based saliency tracking method, and the PWM switching frequency injection method. The emphasis is in the comparison of the four HF saliency tracking methods under various conditions such as steady state, load impact, speed reversal, and zero and low speed operation. The amplitude and frequency of the injection signals are also compared to choose the best HF injection signal for the four saliency tracking methods. The best sensorless control method using eddy current resistance based saliency is introduced and the experimental results confirm the expected advantages for this sensorless application. This thesis also describes the development and enhancement of current derivative measurement for saliency tracking methods, which uses the stator current transient response to the voltage vectors contained in the fundamental PWM sequence. Due to the HF switching oscillations caused by the switching of the IGBT and parasitic capacitance, the accuracy of the current measurement is reduced and requires a minimum vector time of approximately 6µs. A signal processing algorithm is proposed which uses current samples during the high frequency current oscillations, and can potentially reduce this minimum pulse time.

This thesis concerns the modified and improved, time-stepping, dynamic reluctance mesh (DRM) modelling technique for machines and its application to multiple machine systems with their control algorithms. Improvements are suggested which enable the stable solution of the resulting complex non-linear equations. The concept of finite element (FE) derived, overlap-curves has been introduced to facilitate the evaluation of the air-gap reluctances linking the teeth on the rotor to those on the stator providing good model accuracy and efficient computation. Motivated industrially, the aim of the work is to develop a fast and effective simulation tool principally for evaluating salient pole generator system designs including the generator, exciter and the automatic voltage regulator (AVR). The objective is to provide a modelling system capable of examining the detail of machine operation including saturation of main and leakage flux paths, slotting and space harmonics of the windings. Solutions are obtained in a sufficiently short computational time to facilitate efficient iterative design procedures in an industrial design office. The DRM modelling technique for electrical machines has been shown in this thesis to be a fast and efficient tool for electrical machine simulation. Predicted results for specific machine and system designs have been compared with FE solutions and with experimental results showing, that for engineering purposes, the technique yields excellent accuracy. The DRM method has a great advantage in multiple machine simulations. This is because magnetic field calculations are limited to evaluating only the most important information so saving computation time. A brushless generating system including the excitation system and control scheme has been modelled. Additionally a cascaded, doubly fed induction generator for wind generator applications has also been modelled. These different applications for the dynamic reluctance mesh method have proved that this approach yields an excellent machine and machine-system evaluation and design tool.

The technology of direct AC/AC power conversion (Matrix Converters) is gaining increasing interest in the scientific community, particularly for aerospace applications. The aim of this research project is to investigate the use of direct AC/AC three phase four-leg Matrix Converter as ground power unit to supply aircraft with power during stopover or maintenance in airports. The converter fourth leg is used to provide a path for the zero sequence components when feeding unbalanced or non-linear loads. A high bandwidth controller is required to regulate the output voltage of Matrix Converter with a 400Hz output frequency. However, the controller bandwidth is limited due to the reduced ratio between the converter switching frequency and the fundamental frequency. In this case undesirable, periodic errors and distortion will exist in the output voltage above all in the presence of a non-linear or unbalanced load. Digital repetitive control system is proposed to regulate the output voltage of a four-leg Matrix Converter in an ABC reference frame. The proposed control structure introduces a high gain at the fundamental and its integer multiple frequencies. Using the proposed repetitive controller will reduce the tracking error between the output and the reference voltage, as well as increasing the stability of the converter under balanced and unbalanced load conditions. Simulation studies using SABER and MATLAB software packages show that the proposed controller is able to regulate the output voltage during balanced and unbalanced load conditions and during the presence of non-linear load. In order to validate the effectiveness of the proposed controller, an experimental prototype of a 7.5KW has been implemented in PEMC laboratory using DSP/FPGA platform to control the converter prototype. The steady state and the dynamic performance of the proposed control strategy are investigated in details, and extensive experimental tests have showed that the proposed controller was able to offer high tracking accuracy, fast transient response and able to regulate the output voltage during balanced, unbalanced and non-linear loading.

In this thesis, the dye molecule cis-bis(isothiocyanato)bis(2,2'-bipyridyl-4,4'-dicarboxylato)-ruthenium(II) (N3) is studied on the rutile TiO2(110) and Au(111) surfaces. The molecules were deposited onto the surfaces using an ultra-high vacuum (UHV) electrospray deposition system. Thermally labile molecules such as N3 cannot be deposited using the typical method of thermal sublimation, so development of this deposition technique was a necessary step for entirely in situ experiments. The geometric and electronic structure of the samples are characterised using core-level and valence band photoemission spectroscopy, x-ray absorption fine structure spectroscopy, density functional theory, resonant x-ray emission spectroscopy and scanning tunnelling microscopy. These reveal that N3 bonds to TiO2(110) by deprotonation of the carboxyl groups of one bi-isonicotinic acid ligand so that its oxygen atoms bond to titanium atoms of the substrate, and one of the thiocyanate groups bonds via a sulphur atom to an oxygen atom of the substrate. N3 bonds to Au(111) via sulphur atoms with no deprotonation of the carboxylic groups, and at low coverages decorates the Au(111) herringbone reconstruction. For N3 on TiO2, a consideration of the energetics in relation to optical absorption is used to identify the main photoexcitation channel between occupied and unoccupied molecular orbitals in this system, and also to quantify the relative binding energies of core and valence excitons. For N3 on Au(111), the energetics show that the highest occupied molecular orbital overlaps with the Au conduction band. The transfer of charge between the N3 molecule and the TiO2(110) and Au(111) surfaces was studied using resonant photoemission spectroscopy and resonant x-ray emission spectroscopy. These techniques, combined with knowledge gained about the geometric and electronic structure, are used to determine the locations and electronic levels of N3 from which charge is readily transferred to the substrate. The core-hole clock implementation of resonant photoemission spectroscopy is used to reveal that electron delocalisation from N3 to TiO2(110) occurs within 16 femtoseconds.

This work considers the sensorless control of AC machines for the low and zero speed operation range. Two novel techniques have been developed that use the inherit nature of the inverter PWM to estimate the rotor position of the machine. The inherent back EMF and the saliency of AC machines can be utilized to identify the rotor/flux position. The zero vector current derivative (ZVCD) technique for permanent magnet synchronous machines (PMSM) utilizes both of these effects. No additional test signals are injected into the machine and the difficulty in sensing the machine terminal voltage at low speed is eased. Only three standard current transducers are used in the drive system. For the position/ speed estimator only the machine current derivative during the relatively long (at low speed) zero voltage vectors is used for obtaining the rotor position. Practical results show the operation of the drive at several torque and speed conditions including stand still. A further method has been developed for the sensorless control of induction machines. The high frequency harmonics present in a PWM inverter drive system can be used to detect an equivalent impendence saliency that shows modulation due to rotor/ flux position saliency. The proposed method focuses particularly on the extraction of spatial saliency modulation due to rotor bar slotting effect, which can be used to determine the mechanical rotor position. No additional signal injection is required; the method simply employs some of the inherent PWM carrier harmonics.