The work presented in this thesis can be utilised by electrical steel manufacturers and transformer designers to design energy efficient transformer cores possessing lower life cycle costs, thereby increasing financial gains. A novel computer based algorithm to predict losses of 3-phase, 3-limb transformer cores built with high permeability grain oriented steel (HGO) and conventional grain oriented steel (CGO) is presented. The algorithm utilises parameters like transformer geometry, global flux distribution, localised loss data and material properties thus enhancing the accuracy of the predicted results which were 1% of the measured values. This algorithm has contributed to new knowledge in the no-load loss prediction approach. Six, geometrically identical, 350 kVA stacked five packet 3-phase, 3-limb transformer cores assembled with HGO, CGO and four mixed combinations of HGO and CGO laminations in multi step lap (MSL) joint configuration were tested for the global flux density distribution and no-load loss. The investigation results are novel and suggest that the bolt hole diameter (slot width) and lamination width ratio affects the packet to packet variation of . This is a new contribution to the flux distribution regime in transformer cores. The no-load loss experimental results are novel and suggest that the variation of no-load losses with CGO content in mixed cores was non-linear because of the packet to packet variation of . This is a new contribution of knowledge in the field of mixed core loss behaviour. Novel data sets of localised specific loss increase and localised relative permeability decrease around different sizes of holes and slots for HGO and CGO were processed from data obtained by two dimensional finite element (FE) analysis. The data sets are a new contribution in the field of predicting localised magnetic properties around holes and slots.

In this thesis, an investigation of the use of the poly-harmonic distortion model and related techniques is conducted, and applied to model fundamental and harmonic load-pull. Contained within the thesis is a detailed review of the development of the poly-harmonic distortion model and related methods. This thesis shows that although the poly-harmonic distortion model improves on the prediction of fundamental load-pull, over Hot-S-Parameters it still has a limited range of application. To address this observation, higher order models have been investigated along with Fourier methods allowing rapid extraction of the behavioral models. These methods allow conclusions to be drawn on the accuracy of the extracted models, by the direct observation of the magnitudes of the model coefficients. The thesis is concluded with the presentation of the results from third party, using a model extracted using the methods discussed in this thesis. The Model is of a 0.5W GaAs pHEMT at 9 GHz and is used within the design of a Class-J MMIC amplifier.

A load control scheme that used frequency measurements of a smart meter for controlling domestic appliances to provide primary response was investigated. An experimental rig was developed to test and demonstrate the scheme. The amount of loads to be controlled to limit the frequency drop of the Great Britain (GB) power system to a set of minimum allowable frequencies was found. Operating speeds and the limitations of the components of the load controller in providing primary response are discussed. It is shown that if smart meters are to play any role in primary response then the speed at which the system frequency is measured must be increased very considerably. Load profile of fridges/freezers, washers/dryers and hobs/ovens in the GB power system were constructed. Then the percentages of appliances required to be in the load control scheme to shed the estimated amount of controllable loads, were calculated. It is found that the total controllable load requirement can be provided using fridges and freezers alone. Since many washers/dryers and hobs/ovens do not operate at night, they can not then provide a significant amount of controllable loads. However, using these appliances in the day time, the amount of fridges and freezers in the load control scheme can be reduced significantly. The ability of the proposed smart metering system in the UK to report available demand response from the appliances to the network operator was investigated. It was found that the communication network would not support reporting demand response in near real-time. Using load profiles of appliances for 40,000 houses, it was shown that by installing aggregation devices at distribution transformers and substations, the demand response can be reported to the network operator every minute. By aggregating and sending changes only, the impact of reporting demand response in near real-time on communication network reduces significantly. The ability of a state estimator to estimate distribution network voltages using smart meter measurements obtained on the previous day was evaluated. The improvement 10 of the accuracy of the estimated voltages with the number of nodes providing near real-time measurements obtained from distributed generators was also investigated. It was found that when the voltages are estimated using the previous day’s measurements without using any near real-time measurement, the voltage error at all nodes were high. By using near-real time measurements obtained from distributed generators, the error can be reduced significantly.

There is an increasing interest in building millimetre-wave circuits on standard digital complementary metal oxide semiconductor (CMOS) technology for applications such as wireless local area networks (WLAN), automotive radar and remote sensing. This stems from the existing low cost, well-developed, high yield infrastructure for mass production. The overall aim of this thesis is to extend the operating range of 0.18um standard logic CMOS technology to millimetre-wave regime. To this end, microwave and millimetre-wave design, optimisation and modelling methodologies for active and passive devices and low noise circuit implementation are described. As part of the evaluation, new systematic and modular ways of making high performance passive and active devices such as spiral inductors, slow-wave coplanar waveguide (CPW) transmission lines, comb capacitors and NMOS transistors are proposed, designed, simulated, fabricated, modelled and analysed. Small-signal and noise de-embedding techniques are developed and verified up to 110 GHz, providing an increased accuracy in the device model, leading to a robust design at millimetre-wave frequencies. Reduced substrate losses resulting in increased quality factor are presented for optimised spiral inductor designs, featuring patterned floating shield (PFS), enabling improved matching network and a reduced chip area. Based on the proposed shielded slow-wave CPW, both the line attenuation and structure length are decreased, resulting in a more compact and simplified circuit design. An optimised transistor design, aimed at reducing the layout parasitic effects, was realised. The optimisation led to a significant improvement in the gain and noise performance of the transistor, extending its operation beyond the cut-off frequency (ft). By combining all the optimised components, low noise amplifiers (LNAs) operating at 25 GHz and 40 GHz were implemented and compared. These LNAs demonstrate state-of-the-art performance, with the 40 GHz LNA exhibiting the highest gain and lowest noise performance of any LNA reported using 0.18um CMOS technology. On the other hand, the 25 GHz LNA showed a comparable performance to other reported results in literature using several topologies implemented in CMOS technology. These findings will provide a framework for expansion to smaller CMOS technology nodes with the view of extending to sub millimetre-wave frequencies.

NAND flash memory has been widely used for data storage due to its high density, high throughput, and low power. However, as the flash memory scales to smaller process technologies and stores more bits per cell, its reliability is decreasing. The error correction coding can be used to significantly improve the data reliability; nevertheless, the advanced ECCs such as low-density parity-check (LDPC) codes generally demand soft decisions while NAND flash memory channel provides hard-decisions only. Extracting the soft information requires the accurate characterization of flash memory channel and the effective design of coding schemes. To this end, we have presented a novel LDPC-TCM coding scheme for the Multilevel Cell (MLC) flash memories. The a posteriori TCM decoding algorithm is used in the scheme to generate soft information, which is fed to the LDPC decoder for further correction of data bits. It has been demonstrated that the proposed scheme can achieve higher error correction performance than the traditional hard-decisions based flash coding algorithms, and is feasible in the design practice. Further with the LDPC-TCM, we believe it is important to characterize the flash memory channel and investigate a method to calculate the soft decision for each bit, with the available channel outputs. We studied the various noises and interferences occurring in the memory channel and mathematically formulated the probability density function of the overall noise distribution. Based on the results we derived the final distribution for the cell threshold voltages, which can be used to instruct the calculation of soft decisions. The discoveries on the theoretical level have been demonstrated to be consistent with the real channel behaviours. The channel characterization and model provided in this dissertation can enable more design of soft-decisions based ECCs for future NAND flash memories. The data pattern processing algorithm deals with the write patterns and targets to lower the proportion of patterns that would introduce data errors. On the other hand, the voltages applied to the memory cells charges the MOSFET capacitances frequently on programming these data patterns, leading to the power problem. The high energy consumption and current spikes also cause reliability issue to the data stored in the flash memory. This dissertation proposes a write pattern formatting algorithm (WPFA) attempting to solve the two problems together. We have designed and implemented the algorithm and evaluated its performance through both the software simulations and hardware synthesis.

Electric power systems have served well the consumers need for continuous, uninterrupted power supply of good quality and at the minimum possible cost. However, nowadays, the worldwide increasing demand on electric power, coupled with governmental policy changes towards “green” energy and emissions reduction have led to significant changes in the electric power generation. These changes have introduced many serious issues and problems to the electric power systems and although they have been efficiently addressed in the past years, now they need to be restudied and reanalysed taking into consideration all new developments. Distributed generation (DG), constitutes one of the most important developments in modern electric power systems and introduced many benefits as well as drawbacks. DG units are connected to the electric power system near load centres, thus, directly to the distribution network. DG units are larger in number than the more massive conventional power stations and are linked to the introduction of bidirectional power flow. As a result, the configuration of the traditional electric power systems and the networks’ operation have been prominently altered over the last years as soon as DG was introduced into the electric network. This progress has offered many challenges that need to be addressed such as those in terms of control and protection of electric power systems and particularly of distribution networks. The current PhD Thesis attempts to offer a contribution to the electricity distribution networks’ studies with particular reference to distributed generation and protection. In particular, the problems and the issues arising from the installation of DG units in distribution networks are studied. Research on the methods for improving voltage profiles and for reducing real and reactive power losses in distribution networks caused by DGs installation is conducted. Moreover, a decision making algorithm is developed and proposed for selecting the optimum size and location of DG in distribution networks. Furthermore, a new technique based on syntactic pattern recognition for the identification of power system signals used by protective relays is developed in an effort to contribute in the deterrence and reduction of faults. Finally, extensive studies in a distribution network have been conducted, with and without DGs, which aimed to identify the influence of several important parameters in the network’s lightning performance and with its main goal the limitation of lightning faults.

This thesis presents novel formal verification techniques to verify the important property of inevitability of states in analog and mixed signal (AMS) circuits. Two techniques to verify the inevitability of phase locking in a Charge Pump Phase Lock Loop (PLL) circuit are presented: mixed deductivebounded and deductive-only verification approaches. The deductive-bounded approach uses Lyapunov-like certificates with bounded advection of sets to verify the inevitability of phase locking. The deductive-only technique uses a combination of Lyapunov and Escape certificates to verify the inevitability property. Both deductive-only and deductive-bounded verification approaches involve positivity/negativity checks of polynomials over semi-algebraic sets, which both belong to the NP-hard set of problems. The Sum of Squares (SOS) programming technique is used to transform the positivity tests of polynomials to the feasibility of semi-definite programs. The efficacy of the approach is demonstrated by verifying the inevitability of phase locking for a third and fourth order CP PLL. Similarly, the inevitability of oscillation in ring oscillators (ROs) is verified using a numeric-symbolic deductive approach. The global inevitability (of oscillation) property is specified as a conjunction of several sub-properties that are verified via different Lyapunov-like certificates in different subsets of the state space. The construction of these certificates is posed as the verification of First Order Formulas (FOFs) having Universal-Existential quantifiers. A tractable numeric-symbolic approach, based on SOS programming and Quantifier Elimination (QE), is used to verify these FOFs. The approach is applied to the verification of inevitability of oscillation in ROs with odd and even topologies. Furthermore, frequency domain properties specification and verification for analog oscillators is presented. The behaviour of an oscillator in the frequency domain is specified, while it operates in close proximity to the desired limit cycle, employing finite Fourier series representation of a periodic signal. To be sufficiently robust enough against parameter variations, robustness of parameters is introduced in these specifications. These frequency domain properties are verified using a mixed time-frequency domain technique based on Satisfiability Modulo Ordinary Differential Equation (SMODE). The efficacy of the technique is demonstrated for the benchmark voltage controlled and tunnel diode oscillators.

The reliability of electronic products is fast becoming of major importance with the demand for increased safety, especially in the automotive industry. Tracks, pads and vias on printed circuit boards can suffer a variety of problems if circuits are contaminated with electrical-conducting substances. Electrochemical migration, especially dendrite growth, has long been a concern in safety critical and durable systems, and current preventative methods tend to focus on various styles of printed circuit board protective coatings. These measures have a number of disadvantages, mainly process and material costs with extreme scepticism on their overall efficacy. Any design related developments that can minimise the impact of dendrite growth on reliability can lead to a more economic, durable and safer product. The work in this thesis provides a thorough literature search of the field of electrochemical migration on printed circuit boards. This study then develops a novel circuit-designorientated model, based on a multilevel full-factorial design, to study the effects of temperature, voltage and electrode gap on dendritic growth under saturated conditions. Preparation of several DC-biased copper-comb patterned printed circuit boards placed in temperature-controlled water-filled cuvettes enables the specific monitoring of dendrite activity, and detects a sharp current increase that accompanies a dendritic short circuit condition. A high R2 polynomial correlation-model is derived and it is noted that increased voltage and temperature and reduced track spacing increases the impact of dendritic growth on reliability. At voltages between 3 and 4V, gas bubble formation at the electrodes has the effect of increasing reliability by destroying the dendrite fuses. It is shown that dendrites may not grow below 1.25V, which coincides with the theoretical onset voltage for the decomposition of water. It was also demonstrated that electrically biased, watercontaminated printed circuit boards form extreme acid and alkaline regions close to the anode and cathode terminations, respectively, which can cause corrosion. The thesis proposes a novel approach, termed ‘design contingency’, for preventing dendritic growth through design optimisation.

Radars and airborne electronic support measures (ESMs) systems are locked in a tactical battle to detect each other whilst remaining undetected. Traditionally, the ESM system has a range advantage. Low probability of intercept (LPI) waveform designers are, however, more heavily exploiting the matched filter radar advantage and hence degrading the range advantage. There have been literature and internal, SELEX Galileo proposals to regain some ESM processing gain of low probability of intercept (LPI) waveforms. This study, however, has sought digital signal processing (DSP) solutions which are: (1) computationally simple; (2) backward-compatible with existing SELEX Galileo digital receivers (DRxs) and (3) have low resource requirements. The two contributions are complementary and result in a detector which is suitable for detection of most radar waveforms. The first contribution is the application of spatially variant apodization (SVA) in a detection role. Compared to conventional window functions, SVA was found to be beneficial for the detection of sinusoidal radar waveforms as it surpassed the fixed window function detectors in all scenarios tested. The second contribution shows by simulation that simple spectral smoothing techniques improved DRx LPI detection capability to a level similar to more complicated non-parametric spectral estimators and far in excess of the conventional (modified) periodogram. The DSP algorithms were implemented using model-based design (MBD). The implication is that a detector with improved conventional and LPI waveform detection capability can be created from the intellectual property (IP). Estimates of the improvement in SELEX Galileo DRx system detection range are provided in the conclusion.

This research project investigates the development and validation of alternative simulation models for voltage source inverter fed permanent magnet synchronous machine drive systems which can rapidly and accurately analyse and evaluate the performance of PM machine drives and associated control system designs. Traditionally simulations have been conducted using switching models and state space average value methods. The simulation of switching models is time consuming and that of state space averaging involves complex mathematical transformation to d-q axis, with additional circuitry and this limits their application in a time critical design process. Even if the complex calculations of state space are overcome, the proposed model can still achieve better results. This thesis presents the development of fast multi system simulation models for permanent magnet synchronous motor and generator drive systems. The fast simulation model: Average Voltage Estimation Model (AVEM) was developed for two-level, three phase VSI-fed PMSM drive systems and two-level three phase full-scale back-back VSI incorporated in a PMSG wind energy conversion system. The method uses the principle of control strategy and switching function to derive the average phase voltage in one switching period and then uses the average voltages to drive piecewise-linear voltage sources across the terminals of the permanent magnet synchronous machine and three phase system. A voltage source inverter loss model was also developed and incorporated into the AVEM to simulate the drive system power flow and its performance evaluated. The average voltage estimation model is also used to estimate and simulate the energy output of the variable speed PMSG wind energy conversion system. Practical implementation of this technique is achieved using a DSP based controller and validation made through comparison of the DSP AVEM energy estimation method with calculated energy. The study also presents the development of detailed VSI switching models for a variable speed PMSM and a PMSG wind energy conversion system which serve as benchmarks for the proposed AVEM models. A detailed description of both models will be presented. Since models require a control strategy: these control strategies were also developed using the carrier-based sinusoidal (SPWM) and implemented with PI regulators. In the permanent magnet synchronous generator wind energy conversion system application, the SPWM is applied to control the speed of the generator side converter to track maximum power as wind speed varies using the developed passive MPPT control technique and controls the AC load side converter to maintained constant DC link voltage. The sinusoidal PWM control provides a simplified control suitable for the variable speed PMSM drive system and the PMSG wind energy conversion system. Lastly, this thesis presents a detailed development of an experimental test rig. The test rig is developed to provide flexibility for the validation and comparison of the results of both simulation models against real practical implementations for PMSM drive system and PMSG wind energy conversions system. Several simulation case studies were performed using the PORTUNUS simulation package to validate and analyse the steady state accuracy of the proposed average voltage estimation model and control system against the switching model. Experiments were also carried out to validate the results of the simulation models. The simulation models results are presented and compared with experimental results. Suitable steady state performance analysis of two-level, three phase voltage source inverter fed permanent magnet synchronous motor and two-level three phase full scale back-back voltage source inverter with permanent magnet synchronous generator drive simulation and experimental performance are also carried out. The results show good agreement of the proposed average voltage estimation model with the switching model and experimental data, and where necessary the reasons for differences are discussed. The simulation of the AVEM is approximately 50 times faster than the switching model. The limitation of the proposed model is also discussed; mainly it cannot be used for the study and analysis of the internal dynamics of the voltage source inverter. The results from the proposed modelling method utilising the average voltage estimation confirm that this method can be used as an alternative to the detailed switching model for fast simulation and steady state analysis of PM machine drive systems given the advantages of speed, simplicity and ease of implementation. Note that the proposed model is only used for steady state performance analysis; however, in future its application can be extended to transient analysis. In addition, the model is not about maximium power point tracking techniques but it can accommodate maximium power point tracking techniques. It should also be highlighted that exactly the same digital control block is used in both the switching and AVEM models thus allowing a true comparison of controller behaviour. The model developed in this research project has application beyond PMSM drive system and PMSG wind energy conversion system. It can be applied to modelling, simulation and control of other electrical machine drives such as induction machines, switched reluctance machines and three-phase VSI-fed systems.