Low-density parity-check (LDPC) codes form a Shannon limit approaching class of linear block codes. With iterative decoding based on their Tanner graphs, they can achieve outstanding performance. Since their rediscovery in late 1990's, the design, construction, and decoding of LDPC codes as well as their generalization have become one of the focal research points. This thesis takes a few more steps in these directions. The first significant contribution of this thesis is the introduction of a new class of codes called Generalized Irregular Low-Density (GILD) parity-check codes, which are adapted from the previously known class of Generalized Low-Density (GLD) codes. GILD codes are generalization of irregular LDPC codes, and are shown to outperform GLD codes. In addition, GILD codes have a significant advantage over GLD codes in terms of encoding and decoding complexity. They are also able to match and even beat LDPC codes for small block lengths. The second significant contribution of this thesis is the proposition of several decoding algorithms. Two new decoding algolithms for LDPC codes are introduced. In principle and complexity these algorithms can be grouped with bit flipping algorithms. Two soft-input soft-output (SISO) decoding algorithms for linear block codes are also proposed. The first algorithm is based on Maximum a Posteriori Probability (MAP) decoding of low-weight subtrellis centered around a generated candidate codeword. The second algorithm modifies and utilizes the improved Kaneko's decoding algorithm for soft-input hard-output decoding. These hard outputs are converted to soft-decisions using reliability calculations. Simulation results indicate that the proposed algorithms provide a significant improvement in error performance over Chase-based algorithm and achieve practically optimal performance with a significant reduction in decoding complexity. An analytical expression for the union bound on the bit error probability of linear codes on the Gilbert-Elliott (GE) channel model is also derived. This analytical result is shown to be accurate in establishing the decoder performance in the range where obtaining sufficient data from simulation is impractical. / Thesis (Ph.D.)-University of KwaZulu-Natal, Durban, 2004.

Photovoltaic systems alongside energy storage systems are a recognized distributed generation (DG) technology deployed in stand-alone and grid connected system for urban and rural applications. DG system ranging in size from a few kilowatts up to 50 MW refers to a variety of small, modular power-generating technologies connected to the electric grid, and combined with energy management and storage systems to improve the operation of electricity delivery systems. DG provides solutions to two long standing problems of power system operation: non-availability at all times of sufficient power generation to meet peak demands and the lack of capacity of existing transmission lines to carry all the electricity needed by consumers. Installing DG at or near a customer load can eliminate the need to upgrade existing transmission/distribution networks to handle the extra power requirement. Since these distributed energy systems are inertia-less and possess large time constants (response times), there are significant interactions between these systems, the power converters and the distribution networks. This precipitates new dynamics and control limitations, which are unknown in the conventional electric power distribution networks. To perform effective load scheduling, high performance control and optimal operation of these energy systems require an understanding of the dynamic and steady state characteristics of the DG system. This thesis report presents a mathematical model for a Photovoltaic (PVG) generator system, including the energy-storage system. Laboratory test results for steady state performance analysis using various loads are presented and discussed. It concludes with a fundamental economic evaluation of system. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2006.

Approximately 60% of all South Africans do not have access to electricity from the national grid and 80% of the dwellings in the rural areas are not electrified. This is due to the fact that many rural South Africans, similar to other rural markets in the developing world, live in sparsely populated, widely dispersed villages, which cannot be reached within the grid electrification program. HVDC technology provides a viable option to transmit electricity to small distant loads. The objective of the present study is to demonstrate the application of HVDC technology in a medium voltage distribution system, to provide electrical power to Kwa·Ximba, which is a small distant rural area, located in KwaZulu-Natal, South Africa. The proposed system generates electricity from a hydroelectric generation scheme namely Nagle Dam and transmits the excess power to Eskom's Caltoridge-Georgedale sub-transmission network for system enhancement purposes. Extensive technical and economical analyses of the proposed system has been conducted. An HVAC system was also considered for the same purposes in order to make technical and economical comparisons between the use of a HYDC and a HYAC system. In addition, grid extension from Eskom's Catoridge-Georgedale sub-transmission was considered to provide power to Kwa-Ximba without the use of a hydroelectric generation scheme. The proposed networks were therfeore (i) Network A:- Power supply to Kwa-Ximba. and the Catoridge-Georgedale sub-transmission network, from a hydroelectric generation scheme, using HYDC technology, (ii) Network B:- Power supply to Kwa-Ximba, and the Catoridge-Georgedale sub-transmission network, from a hydroelectric generation scheme, using HYAC technology and (iii) Network C:Power supply to Kwa-Ximba by extending Eskom's existing AC CatoridgeGeorgedale sub-transmission network with the hydroelectric generation scheme switched off. It is proposed that Nagle Dam, which is situated adjacent to Kwa-Ximba be used as a hydroelectric generation plant. In order to detennine the most efficient and cost effective use of generator sets, the flow rate, available hydraulic power and available electrical power from the year 2005 to the year 2032 were calculated. The increase in flow rate was based on an annual growth rate of l.5% in water demand. The increase in electrical power demand for Kwa-Ximba was calculated for the next 29 years based on an annual growth rate of 1.8 %. Load flow analyses was conducted on the various power line and busbars that constitute each of the networks, in order to determine the effectiveness of each network. In order to maintain flexibility in power generation, five sets of hydro electrical generators were chosen to give a combined power delivery of 20MW. The first three hydro electrical generators are rated at 5MW each, the fourth set rated at 3MW and the fifth set rated to deliver 2MW, (G I to GS operate 11 KV, 3 phase). The combination of generator sets in use (01 to 05) will vary depending on the electrical power demand in any given year. Analyses of the predicted load flow pattern revealed that in the year 2005, Kwa-Ximba will receive 10.5 MW of power while 8.64 MW of power will be used to enhance the Eskom's Catoridge-Georgedale sub-transmission network, with a 4% spinning reserve. By the year 2014 power supply to the subtransmission network will cease since Kwa-Ximba will be absorbing 12.2 MW of power with a 17.5% spinning reserve. By the year 2032, Kwa-Ximba will absorb 17MW of power with a spinning reserve of 14.63%. The converter stations required for the HYDC transmission network (Network A) will be equipped with YSC and PWM technology and have a true power rating of a 20MW. This wi ll be adequate to supply Kwa-Ximba's power demand right up until the year 2032 when the demand will be 17 MW. Converters will include 10BTs. Two 45 km long, 30 MW, 80 kV triple extruded polymetric HYDe cable will be buried 700mm below natural ground level. The Rectic Master software was used to select an appropriate overhead line for HVAC transmiss ion (Networks B and C). An aluminium, wolf conductor was selected to transmit 20MW of active power. Load flow analyses revealed that the HYDC link contributes positively to network stabil ity by absorbing more reactive power than the HYAC link. The HVDC system absorbed a combined (Kwa-Ximba, Catoridge-Georgdale sub-transmission network) reactive power of 22.04 MY AT, as opposed to the HVAC transmission system where a combined reactive power 1.89 MVAT was absorbed from the connected network. This demonstrated that the HVDC link had the ability to absorb more reactive power from the Catoridge-Georgedale sub-transmission network, therefore contributing positively to the enhancement and stabi lity of the sub-transmission network. Network A contributes more to system stability than Network B. It has also been shown that if Eskom's Catoridge-Georgedale sub-transmission network (Network C) is extended to supply electricity to Kwa-Ximba, this would result in system instability, in the long term. It is evident that Eskom would attain direct benefit from the installation of Newtork A, rather than Networks Band C. The technical and environmental differences noted in the present study, between the HVDC and HVAC systems does not, however, justify the economics to install a HVDe system in order to supply power to Kwa-Ximba. Economical analyses revealed that the implementation of Network A would cost 64% more to install and result in a 75% less annual net profit than Network B. Network B would yield the highest annual net profit for the developer. From the developer's perspective, Network B will be the most feasible network to implement. However, from Eskom's perspective, Network A will be the most bene ficial. Various recommendations have been made by the researcher that would benefit the community of Kwa-Ximba, Eskom and the developer in the long term. / Thesis (M.Sc.)-University of KwaZulu-Natal, 2005

The optimal capacitor placement problem is the determination of the optimal location of the shunt capacitors on the sub-transmission networks such that energy losses are minimised, the power factor and the network voltage profile are improved. During this period when Eskom is experiencing an unacceptably low generation reserve margin, it’s quite critical that the electrical Transmission and Distribution network losses be kept to a minimum to optimise on the scarce generation that is available to supply South Africa’s current and future power demand. One of the ways of minimising technical losses is through the optimal placement or installation of capacitor banks on the network. The placement of shunt capacitors on a bulk Transmission network is essentially to improve the voltage profile on the network, increase system security and reduce transmission losses. The optimal placement of shunt capacitors with the above objectives would assist in minimising the cost of the investment whilst maximising the return on investment to the utility. This research subject is treated as an optimization problem and hence optimization solutions were considered to address the “Optimal capacitor placement problem”. This optimisation problem is solved for all loading levels i.e. peak, standard and off peak periods and for different seasons in a given typical year. This thesis investigates the capability of Genetic Algorithms technique in solving this optimisation problem. Genetic algorithms utilize a guided search principle to develop a robust solution to this research problem. Given their capability to traverse the complicated search space with a multivariate objective function, Genetic Algorithm are versatile and robust to locate the global optimum of the objective function. These Genetic Algorithms (GA’s) were implemented on real sub transmission networks modelled on DigSilent/ Powerfactory. The modelled GA’s on DigSilent were then tested on different network types i.e. commercial, mining, residential and industrial load mixes. The solutions determined by the different GA’s are then compared in terms of time taken to locate the solution, reliability and robustness. The most reliable GA is then identified and recommended as the preferred optimisation approach. A methodology of using GA’s to solve the above mentioned problem is therefore proposed / Thesis (M.Sc.)-University of KwaZulu-Natal, Durban, 2009.

Also known as High Altitude Platform Stations (HAPS), these systems employ unmanned or manned, solar or fuel energy airships or aircraft carrying payloads with transponders and antennas. These remote airships or aircraft offer a much more cost effective solution for coverage of certain regions including: urban, suburban, rural and other environments with low population densities. The Stratospheric Communications Platform (SCP) network offers a better solution than existing Cellular Radio Systems for telecommunication and multimedia services, with greater speed of transmission than even optical modes. It would be virtually impossible to construct land lines and microwave networks in remote, one thousand square-kilometre rural areas. There are other drawbacks to wired deployment as well. The cost of copper wire is astronomical, the terrain harsh and inaccessible and the population scattered. The aim of this dissertation is to illustrate that the use of a platform of this nature is suitable to the rural environment of South Africa. This work includes a case study to ascertain the feasibility of a high altitude platform approach to telecommunication service provision for rural areas. Realising its feasibility has led to an intensive study of rain attenuation. The specific attenuation calculated for a South African rural area (Ulundi) is compared to lTV values. A performance evaluation of the SCP has been done via a link budget calculation with the calculated attenuation values used as input parameters. The advantages of SCPs due to lower path loss, mobile power consumption and system cost are documented. A cellular network architecture is proposed and future investigations into station-keeping techniques, payload power and platform placement are discussed. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, 2006

Synchronous turbo alternators form the basic building block of a generation scheme. Turbo alternators are highly susceptible to foreign material ingress. In industry this phenomenon is on the rise. We evaluated the effects of foreign materials on the generator and condition monitoring equipment. The sources of foreign materials and methods to reduce ingress were investigated. We further evaluated industry best-practices on foreign material exclusion. Information is scarce, as no classic research material is available on foreign material exclusion. We therefore used industry foreign material exclusion best-practices and expert opinions were gathered. Eleven global incidents of foreign material exclusion failure were investigated. The areas of interest were: types of foreign material, area of ingress, condition monitoring response, component damage, root cause and prevention. We point out that turbo alternators are vulnerable to foreign material ingress mainly due to weak foreign material exclusion practices. We categorised the foreign materials and their effects on the operational parts of the turbo generator. We found that the air gap was the most susceptible to ingress. We identified the Generator Core Monitor as a possible solution to minimising foreign material damage. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2009.

Non-coherent detection is a simple form of signal detection and demodulation for digital communications. The main drawback of this detection method is the performance penalty incurred, since the channel state information is not known at the receiver. Multiple symbol detection (MSD) is a technique employed to close the gap between coherent and non-coherent detection schemes. Differentially encoded JW-ary phase shift keying (DM-PSK) is the classic modulation technique that is favourable for non-coherent detection. The main drawback for standard differential detection (SDD) has been the error floor incurred for frequency flat fading channels. Recently a decision feedback differential detection (DFDD) scheme, which uses the concept of MSD was proposed and offered significant performance gain over the SDD in the mobile flat fading channel, almost eliminating the error floor. This dissertation investigates multiple symbol decision feedback detection schemes, and proposes alternate adaptive strategies for non-coherent detection. An adaptive algorithm utilizing the numerically stable QR decomposition that does not require training symbols is proposed, named QR-DFDD. The QR-DFDD is modified to use a simpler QR decomposition method which incorporates sliding windows: QRSW-DFDD. This structure offers good tracking performance in flat fading conditions, while achieving near optimal DFDD performance. A bit interleaved coded decision feedback differential demodulation (DFDM) scheme, which takes advantage of the decision feedback concept and iterative decoding, was introduced by Lampe in 2001. This low complexity iterative demodulator relied on accurate channel statistics for optimal performance. In this dissertation an alternate adaptive DFDM is introduced using the recursive least squares (RLS) algorithm. The alternate iterative decoding procedure makes use of the convergence properties of the RLS algorithm that is more stable and achieves superior performance compared to the DFDM. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2006.

A possible lower cost alternative to medium voltage line arresters and parallel-connected surge arresters is the use of parallel-connected spark gaps across a metal oxide surge arrester. The function of the spark gap is to protect the transformer when the surge arrester fails. Clearly the breakdown voltage characteristics of such a gap need to be carefully co-ordinated with the transformer insulation and those of the arrester. Eskom (Electricity Supply Commission of South Africa) is the national electrical utility that provides the generation, transmission and distribution of electricity in South Africa. The majority of Eskom's electricity reticulation is done with either 11 kV or 22 kV electrical overhead networks. An unacceptable number of Eskom's pole mounted power transformers on these networks have failed over the past few years. The high failure rate of Distribution transformers in Eskom, South Africa has previously been highlighted and investigated in an MSc thesis, the most recent being the thesis completed at the University of KwaZulu-Natal by Chatterton [6]. The thesis proposed possible solutions to the problem but experienced high implementation costs and particular technical issues before widespread implementation could prove viable for the Distribution System. The average transformer failure rate for the Distribution Eastern Region for the twelve month period taken as a moving average was calculated to be 5.19 % per annum at the end of November 2005 and 3.84 % at the end of November 2006. (Eskom Eastern Region Plant report, November 2006). International norms seem to indicate that a transformer failure rate of between 0.5% and 1.0% per annum is acceptable, Chatterton [6]. The reason for the increased failure rates during 2005 was attributed to incorrect Ground Lead Disconnect (GLD) specifications by one of the major surge arrester manufacturers. The incorrect GLD specifications have resulted in premature and nuisance operations. These were triggered by low intensity lightning storms as a result of the lower threshold trigger values. These premature operations have left numerous transformers vulnerable for periods as long as six months and have resulted in the transformer failures increasing from 2.4% quoted by Chatterton [6] in 2002 to 5.19% in 2005 and a reduction to 3.84 % in November 2006 once the problem was identified and the GLDs corrected. See annexure A, Figure A1 of the Plant report for November 2006 for performance details. Hence, this manufacturing flaw and the GLD's sensitivity to specification necessitate further investigation into the spark gap as back-up protection. The spark gap therefore becomes more viable than line or double surge arresters due to its cost effectiveness and robustness. This thesis was based on an idea proposed by Eskom's (Industrial Association Resource Centre) IARC. The aim was to investigate the technical feasibility of using a spark gap to grade with a distribution class surge arrester whilst the surge arrester was operational. The purpose of the spark gap was to act as back-up protection when the arrester fails. Experimentation was undertaken via simulation using the FEMLAB software to model the most suitable gap and geometry for a given rod diameter. The breakdown characteristic of the rod was well understood and verified. Thus, the results obtained from the simulation were compared against the laboratory experiments for the same rod diameters and tip shapes used in the simulation. The results have been analyzed to determine whether the spark gap is a feasible solution for use with surge arresters to protect the transformer from induced strikes following arrester failure. / A possible lower cost alternative to medium voltage line arresters and parallel-connected / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, 2007. / Thesis (M.Sc.)-University of KwaZulu-Natal, 2007.

The requirement for new and existing outdoor air insulated substations to support larger blocks of power in restricted spaces requires an investigation into new bus conductor systems. This study has considered the use of supported, as opposed to suspended, round tubular conductors that have high current capacity necessary for this purpose. The advantages are associated with the fact that smaller clearances are attainable as a result of the restriction of horizontal deflection under fault conditions, and vertical deflection under own weight (sag) due to the rigidity of the tubular conductors as long as the bus tube is correctly sized for the application. Since the tubes are sized for mechanical strength, they are generally oversized electrically, allowing greater flexibility in busbar configuration. Forces due to gravity, wind vibration, fault current and tubular conductor thermal expansion, as well as the restrictions on electromagnetic fields have been considered. The design criteria for conductor and insulator strength calculations also formed part of the study. The study was carried out in a step-by-step analysis of the above considerations. An Excel based programme was developed to analyse the sensitivity of conductor and insulator strength calculations, due to errors in the estimation of various parameters that are required for these calculations. The tolerances in manufacturing and deviations in experimental data were determined and used to evaluate the results obtained and provide a level of confidence that any errors that may arise could be mitigated against by choosing the correct components. The result of the study is a design guide that allows substation designers to develop tubular busbar systems that will operate successfully in the conditions they were designed for. The guide provides an integrated design approach with methods for calculating the forces to which rigid bus structures are subjected. / Thesis (M.Sc.)-University of KwaZulu-Natal, 2008.

High-temperature superconducting materials have found considerable technological application and still have a largely unrealised potential. The key to unlocking this potential depends on a better understanding of their properties; in particular, the maximum 'critical current density' which these materials, in the form of wires, tapes, thin-films and bulk monolithic forms, are able to support for high-current applications. The 'critical current density' in a polycrystalline high-temperature oxide superconductor system is determined by a percolation process of the super current through a three dimensional grain-boundary network. Grain-boundaries in these systems behave as Josephson junction 'weak-links' and they severely limit the critical current density in the presence of even moderate self or applied magnetic fields. In the present work, isothermal quasi-static magnetisation measurements on the polycrystalline YIBa2Cu307-x system are presented and analysed. An effective granular penetration depth in conjunction with a critical state model, which includes an approximate treatment for the percolation process, is used to obtain many of the salient physical parameters of the grain-boundary Josephson junctions and of the three-dimensional grain boundary junction network. Determination of the temperature and magnetic field dependence of several of these parameters, in particular a magnetic field-independent critical current which depends on the micro structure of the grain-boundary junction network, allows for testing and verification of models of the weak-link and network behaviour. This treatment has been carried out specifically on various silver doped polycrystalline Y1Ba2Cu30 7-x specimens in order to determine and quantify the effects of silver doping. An improvement in the critical current density with silver doping is explained in terms of silver scavenging and ridding grain-boundaries of impurities, and a proximity effect where trace amounts of silver residing in the grain-boundaries decreases the normal resistance of the grain-boundary Josephson junction. The insight gained from silver doping experimentation led to a macroscopic investigation into the joining of large single-domain YIBa2Cu307-x specimens for large-scale applications. / Thesis (Ph.D.)-University of KwaZulu-Natal, Durban, 2006.