The thesis is concerned with a study of the electrical engineering parameters of the lightning ground flash - i.e. the statistical distributions of peak current amplitudes, discharge current waveform characteristics, and flash striking distances - in the event of flashes to practical engineering structures. In view of its predominating frequency of occurrence in practical situations, the discharge of primary concern is the downward progressing and negatively charged ground flash. A central feature of this work is the establishment of a lightning research station (incorporating a 60 m instrumented mast) in the Transvaal highveld region of South Africa. The design of this station and the related measurement techniques are fully described. Preliminary results accumulated over a 6-year period of observation are presented, and include recordings obtained during direct strikes to the mast, as well as data from associated measurements of additional thunderstorm and lightning parameters. The latter studies include the use of closed circuit television video recordings, together with electrostatic field mills and lightning flash counters. Analysis of the resultant data serves to provide a comprehensive characterisation of the thunderstorm and lightning climatology in the region - on the basis of electrical activity. With only few exceptions, it is concluded that the characteristics of lightning observed in the. Transvaal region are generally consistent with the trends of data from other regions of the world. A unique aspect of the project is a study of lightning striking distances. An attempt to estimate these distances using bi-directional photography of flashes to the research mast is described, and several preliminary results are also presented - in conjunction with the associated measurements of discharge current amplitude. These results are compared with previously used relationships between striking distance and peak current. / Thesis (Ph.D.)-University of Natal, Durban, 1979.

Partial discharges (p.d.'s) are a major source of the progressive deterioration of insulation in / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, 2007.

Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, 2006.

Proper radio link design requires an accurate prediction of the effective earth radius factor (the kfactor) distribution, for the location where propagation is intended. Though a median value of k equals 4/3 is normally use for communication design purposes, in reality the true k-factor values differ, for different locations, globally. The effective earth radius factor distribution for Southern Africa was evaluated in the dissertation. The two Southern African countries chosen for the study are Botswana and Republic of South Africa. The dissertation reports in detail a study on the topic using three years radiosonde data obtain in Maun, Botswana and ten months radiosonde data collected in Durban, South Africa. An analytical model was proposed, which predicts the probability density function of the k-factor for the Southern Africa using the data from these two countries. Also a comparison of the data from the two countries was done in the analysis and reported in the write-up. The application ofthe work was also investigated and reported by simulating a radio link between Sherwood and Umlazi in Kwazulu-Natal Province of South Africa. The consequence of using inappropriate design value of k on link reliability was also investigated and reported. Recommendation for future work was given in the concluding chapter for future improvement on the study. Radio communication designers will find the results obtain in the report useful. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2005.

An Intelligent Network (IN) can be visualised as a central computer interconnected to the telecommunication network that allows the creation and deployment of services as consumer demand arises. Messages or information are transferred between the central processor and exchanges by a standardised form of signalling, known as Common Channel Signalling System Number Seven (CCS7). The thesis focuses on CCS7 performance, models and topologies. A requirement of the research entails the study of the CCS7 performance parameters: signalling delays and signalling network dependability. Signalling delays may be comprised of signalling message transfer delays within the message transfer part and queueing delays. Signalling network dependability usually include availability, reliability, maintainability and network robustness. For the purpose of modelling, the decomposition of the CCS7 message transfer part into subsystems is essential. A generic model for the message transfer part was used to implement certain functionalities of CCS7 in OPNET. OPNET (Optimised Network Engineering Tools) is a comprehensive engineering system capable of simulating large communication networks with detailed protocol modelling and performance analysis. The simulation software (OPNET) and performance parameters were used to analyse CCS7 networks resulting in the selection of a particular topology for a given region. The network structures analysed include the mesh topology, standard quad topology, hierarchical topology and Telstra's new topology. It was found that the mesh structure had the best simulated and availability/reliability results but was impractical for large networks due to the cost implications. This cost factor led to the selection of a hierarchical signalling system for Kwa-Zulu Natal. / Thesis (M.Sc.Eng.)-University of Durban-Westville, 1998.

This dissertation is concerned with the application of tomography within the sugar industry; in particular non-accessible locations found in a sugar mill. In this study, the focus of research is that of a vacuum pan, and if better understood through tomographical techniques could significantly improve mill efficiency and throughput. The tomography system comprises unique mechanical interfaces, data acquisition modules and software algorithms in order to generate images which reflect the dynamics in the tomographical sensor zone. The distribution of gas (low conductivity) and liquid (high conductivity) within a tube is of main interest in order to understand the boiling dynamics and ultimately pan design. This is attained by determining the internal cross-sectional spatial distribution of conductivity of a number of tubes within the pan simultaneously. Thermal properties of the contents of the sugar pan (a syrup-like substance known as massecuite) at different boiling stages can be estimated based on the tomographical data. Data acquisition is achieved via an in-house designed electronic state machine. A neighbourhood back-projection reconstruction technique was developed in MATLAB in order to generate tomographical images. Results from the system have identified different boiling dynamics which improve the understanding and design of vacuum sugar pans. / Thesis (M.Sc.Eng)-University of KwaZulu-Natal, Durban, 2010.

The continuously growing demand for electric power requires transmitting larger amounts of power over long distances. An economically attractive solution to increase the power transfer through a long interconnection (up to a limit) without building new parallel circuits is to install series capacitor compensation on the transmission line. Large disturbances which constantly occur in power systems may disrupt the synchronous operation of the generators and lead to out-of-step conditions. Coordinated insertion and removal of the compensating capacitors in series with a transmission line is an approach that has been known for many years to be capable of enhancing the transient stability of power systems as well as providing additional damping to the power system oscillations. The relatively recent emergence of the thyristor controlled series capacitor (TCSC) has now made this method of transient stability enhancement practically feasible. This thesis compares a range of different strategies that have been proposed in the literature for control of series compensating reactance to enhance transient stability. Initially a simple swing-equation model of a single-generator power system, including an idealised controllable series compensator (CSC) is used to study the fundamental characteristics of the variable impedance control and its impact on transient stability. Subsequently, a detailed model of a small study system is developed, including a detailed representation of a TCSC, for more in-depth analysis. This detailed study system model is then used to compare three different transient stability control schemes for the TCSC, namely: generator speed-deviation based bang-bang control, discrete control based on an energy-function method, and nonlinear adaptive control. Time-domain results are presented to demonstrate the impact of the TCSC on first swing stability of the SMIB system with the above control schemes for various fault scenarios. The performance of each control scheme is also compared by evaluating the extent to which it extends the transient stability margin of the study system. For each of the three different TCSC control approaches considered, the results show that variable impedance control of the TCSC provides further improvement in the transient stability limits of the study system over and above the improvement that is obtained by having a fixed-impedance TCSC in the system. In the case of the bangbang and discrete control approaches, it is shown that a combination of a large steady state value of the TCSC compensation, together with a relative small range of variable TCSC reactance under transient conditions, offers. the best improvement in the transient stability limits for the studied system. The results also show that there is little difference in the extent to which the energy function method of TCSC control improves the transient stability limits over the improvement obtained using speed-deviation bang-bang control of the TCSC for the study system considered. / Thesis (M.Sc.)-University of KwaZulu-Natal, Durban, 2005.

Constrained electrical power systems and the long lead times ne eded for new capacity necessitate interim demand side management measures such as t ime-of-use (TOU) pricing. This form of electricity pricing has the potential to reduce system peak demand and thus improve the efficiency of power systems. Such time differenti ated pricing mechanisms have been used successfully in the industrial and commercial secto rs to shift demand out of the peak periods but have yet to be implemented in the residential sector in South Africa (SA). TOU schemes are based on the cost of supply and reflect, in part , the changes in short-run marginal costs. In contrast the conventional residential tariffs in SA are based on flat rate structures and recover long-run costs only. The analysis of the impact of such schemes, for both the utility as well as the customers, is gaining importance once more, particularly when most utilities are contemplating the implementation of smart s ystems and advanced metering infrastructures and the costs associated with this. A recent TOU pilot project, HomeFlex, is analysed from an ec onometric point of view. Panel data sets for both treatment groups and the control group ar e obtained from the pilot project database for each customer in two separate experiments in two separate geographic areas. The Caves and Christensen approach is used and the const ant elasticity of substitution functional form is chosen. Conditioning variables such as daily consumpti on per customer as well as climate effects are included in the ordinary lea st squares regression in order to establish the relationship between peak and off-peak consumption and the extent of the substitutability of these two commodities. The elasticity of substitution estimates obtained for stage 1 of the analysis range from 0.339 to 0.384. The conditioning variables enter the analysis as modifie rs to the estimates but their effect is insignificant. The stage 2 estimates range from 0.457 to 0.518. The effect of the conditioning variables is also statistically insignificant at this stage. The effect of the daily and weekly price ratio is therefore the primary factor in det ermining the response of customers to TOU pricing in the HomeFlex project. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2009.

The following work documents the design of an HF band direct sequence point-to-point link as used in a rural environment. The dissertation begins with a description of the overall document layout. An introduction into the problems associated with providing rural access is then given, with special emphasis on wireless technologies. It is argued that the attributes of HF band radio make it a good candidate for providing wireless communications links for under serviced rural areas in South Africa. The pitfalls and disadvantages of using an ionospheric-based medium in which to propagate an electromagnetic wave are discussed and several solutions are put forward. One of these solutions is Direct Sequence Code Division Multiple Access (DS-CDMA). A thorough analysis of the principles of DS-CDMA is given with emphasis on its abilities to combat the negating effects of the HF channel. A CDMA HF system is then proposed, outlining the various practical and theoretical aspects. Next, an HF channel model is designed and simulated, first with no spreading or coding, then with just spreading and finally with spreading and coding. It is found that although the extra diversity of the spreading and coding aid in reducing the bit error probability, more detailed local measurements and refinement in the design of the channel model and simulated system are needed to increase performance. Various aspects of a practical system that was built are then discussed, highlighting issues such as hardware interfacing and the software design of a man-machine-interface. Field measurements are also given with estimates on an upper limit on expected performance of a practical system. Finally conclusions are given, detailing the achievements and shortcomings of the research. / Thesis (M.Sc.Eng.)-University of Durban-Westville, 2000.

This thesis consists of three main parts. An introduction to diode-pumped solid-state lasers, thermal modelling of solid-state lasers and rate-equation modelling of solid-state lasers. The first part explains the basic components and operation principles of a typical diode-end-pumped solid-state laser. The stimulated emission process, solid-state laser gain media, various pump geometries and a basic end-pumped laser resonator configuration are among the topics that are explained. Since thermal effects are one of the main limiting factors in the power-scaling of diode-pumped solid-state lasers, the second part of this thesis describes numerical and analytical thermal models that determine the thermal lens and thermally induced stresses in a laser crystal. As a first step, a time-independent numerical thermal model which calculates the three-dimensional temperature distribution in the laser crystal is implemented. In order to calculate the time dependent thermally induced stresses in a laser crystal, a coupled thermal-stress finite element analysis model was implemented. Even though some steady-state analytical solutions for simple crystal geometries do exist, the finite element analysis approach was taken so that the time dependent thermally induced stresses could be calculated for birefringent crystals of various geometries. In order to validate the numerical results, they are compared to experimental data and analytical solutions where possible. In the last part, the population dynamics inside the laser gain medium are described and modelled with a quasi-three-level rate-equation model. A comprehensive spatially resolved rate-equation model is developed and discussed. In order to simplify the implementation of the rate-equation model as a computer simulation, the spatial dependence of the laser parameters is ignored so that the model reduces to a singleelement plane-wave model. The simplified rate-equation model is implemented and solved numerically. The model is applied to a four-level CW and Q-switched Nd:YLF laser as well as a quasi-three-level QCW Tm:GdV04 laser. The models' predictions are thoroughly verified with experimental results and also with analytical solutions where possible. / Thesis (M.Sc.)-University of KwaZulu-Natal, 2008.