No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

The aim of this research was to design and evaluate various Gallium Arsenide circuit elements such as logic gates, adders and multipliers suitable for high speed VLSI circuits. The issues addressed are the logic gate design and optimisation, evaluation of various buffering schemes and the impact of the algorithm on adder and multiplier performance for digital signal processing applications. This has led to the development of a design approach to produce high speed and low power dissipation Gallium Arsenide VLSI circuits. This is achieved by : Evaluating the well established Direct Coupled Logic (DCFL) gates and proposing an alternative gate, namely the Source Follower DCFL (SDCFL), to improve the noise margin and speed. Suggesting various buffering schemes to maintain high speed in areas where the fanout loading is high (eg. clock drivers). Comparing various adder types in terms of delay-power and delay-area products to arrive at a suitable architecture for Gallium Arsenide implementation and to determine the influence of the algorithm and layout approach on circuit performance. To investigate this further, a multiplier was also designed to assess the performance at higher levels of integration. Applying a new layout approach, called the 'ring notation*, to the adder and multiplier circuits in order to improve their delay-area product. Finally, the critical factors influencing the performance of the circuits are reviewed and a number of suggestions are given to maintain reliable operation at high speed.

This thesis develops solutions that accommodate the introduction of state estimation in High Voltage (HV) power distribution networks, and proposes methodologies that further enhance the value of state estimation in distribution network operation. Furthermore, it investigates the impact of Distributed Generation (DG) and Active Management (AM) on the infrastructure development of typical European distribution networks. In HV distribution networks, measurements are very limited and normally available at the main substation only. Thus, it is essential to introduce appropriately modelled pseudo measurements. This is necessary not only for the state estimation mathematical models to be established but also for state estimation to generate estimates of sufficient quality. Two approaches, one based on correlation coefficients and regression analysis and one based on Artificial Neural Networks (ANNs), are proposed. Distribution networks are not static. Faults, maintenance and emergencies constantly change their topology; sudden changes in major power injections significantly change their power flows and voltages. For the Distribution Management System (DMS) to be reliable, it is important that changes significantly changing the state of the network are immediately identified and taken into consideration before control actions are issued. A methodology for detection of network changes using state estimation and the Bayes theorem of conditional probability is introduced. Finally, the future infrastructure development of typical distribution networks of Germany, the Netherlands and Poland is examined. The technical, economic and environmental aspects of Passive Management (PM), AM and DG are assessed and quantified while the technical and economic efficiency of different AM strategies is evaluated.

No description available.

The objective of this thesis is to create a model, which provides a detailed description of the electrical load on a low voltage distribution network in the context of a typical UK urban feeder from a primary transformer. The aim of the work, when used in association with a power flow analysis package, is to help to provide a risk assessmenfto r over-voltagee vents and over-heatingi n the network when different levels of embedded generation are applied. A fine-grained domestic load model has been constructed with a three layered approach-to provide per-consumer, 1-minute averaged loads on an end-use basis. Datasets from the Load Research Group have been used as the starting point and form the basis for layer 1 which represents group-averaged demands on a halfhourly basis. Layer 2 of the model introduces diversity in terms of number of occupants, living space, ownership and soci?- economic factors. Layer -3 uses appliance duty cycles to create wider variations by random triggering to derive 1- minute loads from assigned half-hourly values. The domestic model has been adaptedf or use with smaller( sub MOW) non-domesticc onsumers. The research question for this study is whether or not the models provide an adequate representation of the electricity demand for a typical urban LV network, judged in terms of a variety of parameters. The output from the domestic model comparesw ell with measuredd ata giving realisticd emandc haracteristicsin terms of mean, peak, load factor and distribution. Compared against diversified peak demands currently in use within the industry, the model estimates values within 10% for groups fewer than 25 and 5% for groups of 100 or more. When used together with a power flow analysis package, the predicted voltage variation agrees with measured results in terms of mean value and distribution. The investigation of time and group averaging of demand, power factor surveys and, with a matching model for PV and solar thermal output, studies into electrical demand reduction within mixed communities are all possible additional applications for the model.