The research work presented in this thesis is concerned with the modelling of this new generation of power electronics controllers with a view to conduct comprehensive power systems harmonic analyses. An issue of paramount importance in this research is the representation of the self-commutated valves used by the controllers addressed in this work. Such a representation is based on switching functions that enable the realization of flexible and comprehensive harmonic models. Modularity is another key issue of great importance in this research, and the model of the voltage source converter is used as the basic building block with which to assemble harmonic models of actual power systems controllers. In this research the complex Fourier series in the form of operational matrices was used to derive the harmonic models. Also, a novel methodology is presented in this thesis for conducting transient analysis of electric networks containing non-linearities and power electronic components. The methodology is termed the extended harmonic domain. This method is based on the use of time-dependent Fourier series, operational matrices, state-space representation and averaging methods. With this method, state-space equations for linear circuit, non-linear circuits, and power electronics controllers models are obtained. The state variables are the harmonic coefficients of x(t) instead of x(t) itself. The solution of the state-space equations gives the dynamic response of the harmonic coefficients of x(t). Moreover, a new harmonic power flow methodology, based on the instantaneous power flow balance concept, the harmonic domain, and Newton-Raphson method, is developed and explained in the thesis. This method is based on the instantaneous power balance as opposed to the active and reactive power balance, followed by traditional harmonic power flow methods. The power system and the power electronics controllers are modelled entirely in the harmonic domain.

No description available.

The aim of this PhD project was to develop the technology of dielectrophoresis on the sub-micrometre scale and to use DEP to manipulate sub-micrometre particles and measure their dielectric properties. Of particular interest was the application of DEP to viruses, the largest of which is approximately 250 nm in diameter. A system for virus characterisation, identification and separation based on DEP would be a major milestone in this field of research, as well as having beneficial medical and biotechnological uses. Particles with a diameter between 1nm and 1m are referred to as Colloidal particles and the dynamics of their movement are complicated by the effects of thermal energy and Brownian motion. High electric fields are required to dominate these effects but signals with high potentials and high frequencies are difficult to generate. Semiconductor manufacturing techniques can be used to fabricate micro-electrode structures which can produce high electric fields from relatively low potentials. Lithography based manufacturing techniques were developed to produce suitable electrodes for dielectrophoresis on a scale small enough to manipulate sub-micrometre particles. Detailed electric field patterns were numerically calculated for these electrodes, so that the dielectrophoretic force could be simulated, predicted and compared with experimental measurements of particle movement. The dielectric properties of latex spheres with diameters from 93 nm to 557 nm were determined through observation and measurement of the DEP movement; new theories were postulated to account for the results which did not conform to accepted theories. A rod shaped plant virus, Tobacco Mosaic Virus (TMV) was also studied and its dielectric properties determined from the experimental results. TMV is 300 nm long with a cylindrical radius of 9nm, a shape of particle which is very different from a sphere and one which has not been studied by this method previously. An expression for the frequency dependent dielectrophoretic force on such a particle was derived and values of the dielectrophoretic force on the particle were measured and compared with the theoretical model.

The objective of the work of this thesis is to design, fabricate, and characterise a GaAs MESFET based monolithic microwave integrated circuit (MMIC) downconverter which operates at Ka-band frequency (35GHz). In the course of the project active and passive elements required for the MMIC were designed, fabricated, characterised and their equivalent circuit models extracted. Fabrication processes for passives, actives and MMIC were realised using mainly electron beam lithography (EBL) techniques. The main findings of this thesis were: Devices - Influence of gate recess offset on MESFETs The MESFETs were patterned by EBL and gate recessing was accomplished by selective dry etching. The influence of the gate recess offset on the small signal AC equivalent circuit, DC device characteristics, overall high frequency device performance, and low frequency noise behaviour of 0.2 m gate length GaAs MESFETs implemented in the low noise amplifier (LNA) circuit design in this thesis was investigated. Numerical simulations of the AC small signal equivalent circuit dependence were carried out in order to help understand the effects observed. Good qualitative agreement between measured and simulated response was obtained. - Schottky diodes The performance of Schottky-contact diodes used in the MMIC mixer were studied as a function of their geometry and processing conditions. Passives - CPW losses Losses in coplanar interconnect topologies (coplanar waveguide and slotline) using different metallisation processes were investigated. - CPW to slotline transitions A range of coplanar waveguide to slotline transitions required for the MMIC mixer were studied. Broadband performance with insertion loss of < 0.5dB per transition was observed. Transmission line models of the structures have been implemented to enable circuit performance to be predicted and designed to suit the application frequency. The effect of parasitic modes on transitions performance was also investigated.

This thesis is concerned with the interaction of an optical wave with a microwave in a waveguiding structure coupled by a second order nonlinearity. Emphasis is laid upon the generation of ultrashort electrical transients via optical rectification (OR) as well as cascading effects due to the interplay of OR and the linear electro-optic effect. A simple transmission line model is introduced to explain qualitatively the basic physical mechanisms of an externally induced polarisation in a travelling wave structure. For a quantitative description, evolution equations for the overall interaction between the microwave and the optical wave based on a coupled mode formalism are developed. The basic properties of the structure under consideration are discussion and techniques for their evaluation are introduced. A set of corresponding parameters for typical structures is estimated and used for calculations throughout the thesis. The generation of electrical signals from optical waves via OR is discussed in detail for the cases of single and mixed polarization optical modes in the structure. The self phase modulation due to cascading of OR and the electro-optic effect is elucidated. It is shown that continuous wave solutions of the conservative system are modulationally unstable in a large range of relevant system parameters. The possibility of formation of solitary waves due to the mutual interaction of optical wave and microwave is considered in the context of long wave short wave interaction. Basic properties of bright stationary solutions and their excitation are discussed. The possibility of formation of solitons due to microwave self-interaction is illuminated. The linear stability of bright solitary waves is investigated. The observed oscillations and radiation of perturbed propagated bound states are explained by the existence of discrete, quasi-bond internal modes of the stationary solutions. Collision scenarios are addressed.

A systematic investigation of the electrochemical behaviour of two sets of microelectrode arrays, fabricated by standard photolithographic and reactive-ion etching techniques, is presented. The first set of microelectrode arrays had a constant relative centre-centre spacing of 10r (where r is the electrode radius). As a value of r was decreased, the cyclic voltammograms recorded from the array became increasingly peak-shaped, due to merging of the diffusion fields of the individual electrodes. Furthermore, it was shown that the peak current densities obtained were largest for the arrays with the smallest individual electrodes, as were the signal-to-noise ratios (SNRs). Electroplating the individuals electrodes with platinum black was also shown to increase the peak currents and the SNRs, due to an increase in the effective surface area. Sigmoidal voltammograms, which are indicative of radial diffusion, were obtained for an individual electrode radius of 25 mm but not for arrays with smaller electrodes. To obtain radial diffusion for an array of 2.5 mm electrodes, it was shown (using a second set of microelectrode arrays) that a minimum relative centre-centre spacing of 40r is required. Further enhancement of the peak current densities were obtained by decreasing the size of the individual electrodes. A series of nanoelectrode arrays were fabricated using electron-beam lithography (EBL). The voltammograms obtained from these arrays exhibited a continual increase in the recorded peak current as the individual electrodes radius was decreased to a value of 110 nm. Since EBL is a slow and costly technique, nanoimprint lithography (NIL) was investigated as an alternative method of fabricating nanoelectrode arrays and comparable results were obtained from arrays produced by EBL and NIL. A dissolved oxygen and temperature sensor incorporating a working microelectrode array was also designed and fabricated. The sector comprised a densely packed array of 2.5 mm radius electrodes and a micro-reference electrode, both of which were covered with an agarose electrolyte gel enclosed in an SU8 chamber. A thermal resistor was included for temperature compensation of the dissolved oxygen measurements. The Ag|AgCl micro-reference electrode was found to be stable for approximately 80 hours in 0.1 M KCl, with 100 nA of current passing through it. Linear calibration curves were obtained from both temperature and dissolved oxygen measurements

Drift-Diffusion, Hydrodynamic and Monte Carlo simulations have been used in this work to simulate strained Si/SiGe devices for RF and CMOS applications. For numerical simulations of Si/SiGe devices, strain effects on the band structure of Si have been analyzed and analytical expressions are presented for parameters related to the bandgap and band alignment of Si/SiGe heterostructure. Optimization of n-type buried strained Si channel Si/SiGe MODFETs has been carried out in order to achieve high RF performance and high linearity. The impact of both lateral and vertical device geometries and different doping strategies has been investigated. The impact of the Ge content of the SiGe buffer on the performance of p-type surface channel strained Si/SiGe MOSFETs has been studied. Hydrodynamic device simulations have been used to assess the device performance of p-type strained Si/SiGe MOSFETs down to 35 nm gate lengths. Well-tempered strained Si MOSFETs with halo implants around the source/drain regions have been simulated and compared with those devices possessing only a single retrograde channel doping. The calibrations in respect of sub-100 nm Si and strained Si MOSFETs fabricated by IBM lead to a scaling study of those devices at 65 nm, 45 nm and 35 nm gate lengths. Using Drift-Diffusion simulations, ring oscillator circuit behaviour has been evaluated. Strained Si on insulator (SSOI) circuits have also been simulated and compared with strained Si circuits, Si circuits employing conventional surface channel MOSFETs along with SOI devices. Ensemble Monte Carlo simulations have been used to evaluate the device performance of n-type strained Si MOSFETs. A non-perturbative interface roughness scattering model has been used and validated by calibrating with respect to experimental mobility behaviour and device characteristics. The impact of interface roughness on the performance enhancement of strained Si MOSFETs has been investigated and evidence for reduced interface roughness scattering is presented, i.e., a smoother interface is suggested in strained Si MOSFETs. A 35 nm gate length Toshiba Si MOSFET has been simulated and the performance enhancement of 35 nm strained Si MOSFETs over the Toshiba Si device is predicted. Monte Carlo simulations are also employed to investigate the performance degradation due to soft-optical phonon scattering, which arises with the introduction of high-K gate dielectrics. Based on the device structures of the calibrated sub-100 nm n-type conventional and strained Si IBM MOSFETs, significant current degradation has been observed in devices with high-K gate dielectrics, HfO2 and Al2O3.

No description available.

The Lagrangian perturbation method for the NLS is revisited in the form of an equivalent direct problem. The analogy can be extended to arbitrarily perturbed systems. It is then possible to provide first order perturbation expansions for the fundamental soliton. The case of the damped NLS is considered and shown to fully comply with IST predictions. Subsequently the problem of NLS initial condition not corresponding to an exact soliton is examined. There are two issues that need to be considered: the location of the soliton solution and the modelling of the continuum. The location of the soliton solution is handled by considering the integrals of motion of the NLS. The improvement arises by the inclusion of the contributions due to the continuum. The results are compared with numerical calculations and are proved to be satisfactory provided that the initial pulse shape does not depart greatly from the Asech(z) functional form. The propagation problem is handled by considering the evolution of the soliton and the continuum separately and recombining them at the required time. Two cases are considered: the far field pattern and the position where the peak of the soliton lies. For the former the recombination of continuum with the soliton is achieved with the help of the inverse part of the IST. For the peak position a Bäcklund transform is considered. Results from both regimes are compared with numerical results and shown to agree satisfactorily.

Electricity usage has grown steadily ever since the first commercial generator came into operation more than one century ago. Power transmission networks too, have grown in size and in operational complexity to be able to handle the large blocks of electricity that travel from generator to consumers round-the-clock and with huge variations. At various stages of the development, state-of-the-art equipment, methods and techniques have been incorporated in the vast array of tools that power systems engineers have at their disposal to keep up with the demands imposed by the planning, management, operation and control of modern power systems. Transient stability has always been an issue of paramount importance in power system planning and operation. Arguably, most of the ideas and concepts associated with power system stability analysis were conceived many years ago. Nonetheless, continuous expansion of the network and the emergence of a new generation of fast acting, multi-purpose power system controllers have called for renewed research efforts in this all-important application area of power systems. In particular, there is growing concern that the power network is becoming more unbalanced, owing to higher operating voltages and a relentless drive for interconnection, and that unbalances may impair the effectiveness of power electronic-based loads and controllers. These are issues that may be difficult to address satisfactorily with conventional transient stability modelling approaches since they are based on the premise that the transmission network observes a perfect balance, even under faulted operating regimes. The study of a limited range of asymmetrical transient stability problems using conventional methods can be achieved, but only with great difficulty, which involves transforming the network into fictitious components (i. e. symmetrical components). This is significant since asymmetrical short-circuit faults constitute the largest percentage of faults that occur in the power network, and network designs based solely on the three-phase short-circuit-to-ground faults result in underengineered networks. Equally important issues are the widespread commissioning of modern power electronics controllers and the lack of suitable models and methods for assessing the impact of such controllers in network-wide operation with particular reference to transient stability and unbalanced operation. The research reported in this thesis addresses these issues and develops a direct time phasedomain model for conducting multimachine transient stability analysis where asymmetrical operating conditions and the impact of modem power electronics controllers are represented. In this simulation environment, AC synchronous and asynchronous generators are represented together with asynchronous motors. The set of non-linear equations describing the machines are solved using discretisation and the trapezoidal rule of integration. The proposed model is compared against an industry standard power system package for cases of symmetrical operation. The generality and versatility of the model is demonstrated when applied to the analysis of symmetrical and asymmetrical power system operations. An important aspect of this research is a drive towards the solution of transient stability in real-time, where the results produced are in actual world time. This is achieved by embedding the model into a commercially available multi-purpose real-time station. To this end, coherency-based synchronous generators equivalent has been developed to enable the solution of multimachine systems in real-time. The equivalent unit is obtained based on the aggregation of the coherent generators using phase-domain techniques. Dynamic loads in the form of asynchronous motors are implemented within the multimachine network. The adverse influences of motor operation on voltage problems in the network under symmetrical and asymmetrical conditions are analysed. Transient analysis of dispersed generation is also considered where the asynchronous machine is operated as a generator alongside synchronous generators. The behaviours of the two type of generators under various networks and operating conditions are presented. Models of power electronics controllers in the direct time phase-domain are also described in this thesis. The generalised models of the Static Var Compensator (SVC), Static Synchronous Compensator (STATCOM), Dynamic Voltage Restorer (DVR) and High Voltage Direct Current-Voltage Source Converter (HVDC-VSC) station are proposed. The SVC comprised of a fix capacitor and a thyristor controlled reactor (TCR) is developed. Here, switching functions are used to represent the operation of the thyristor. Models of STATCOM, DVR and HVDC-VSC station are developed based on the self-commutated voltage source converter (VSC) technology. The VSC is represented by the switching functions of its pulse width modulation (PWM) control, hence, providing a flexible model within the direct time phase-domain approach. The model of the VSC is implemented into the respective power electronics controllers enabling a convenient modular approach to be adopted. The power electronics controllers are incorporated into the multimachine environment for the analysis of transient and power quality related issues.