The lack of methods for rigorous dynamic analysis of foundations and structures for vibrating machines has resulted in below optimum performance and in some cases reduction of life of machines, structures and foundations. The costs and complexities of these machines make it necessary to conduct proper geotechnical site investigations. and dynamic analyses to obtain the response of the soil, foundation and structure as a system to excitation. In order to highlight the use of dynamic analyses, the response of the foundations and structures were compared to the "rule of thumb" which is based on mass ratio. Furthermore sensitivity analyses were carried out comprising the following variables: • Shear modulus of soil, G • Poisson's ratio of soil,Y • Type of structure (ie raft, table top and multi-storey) • Stiffness of structure • Stiffness of foundation The fundamentals of structural dynamics have not been dealt with in this dissertation. The dynamic analyses were carried out using a finite element analysis program called Strand 7. The results were typical of a finite element analysis, giving stresses, strains, deflections, amplitudes, frequencies and velocities of vibration. The traditional "elastic halfspace model" is deficient as it does not account for soil comprising various layers. The theory has been based on an isolated circular footing. Most foundations are located in soils with layered mediums, are rectangular and in some cases are affected by the interaction of foundations in close proximity. Furthermore there is a need to account for the non-linear effects and properties of soil. It is therefore becoming more attractive to adopt mathematical models of soils using finite elements, where the visco-elastoplastic properties of soils can be realized and modeled. Furthermore the finite element method overcomes limitations such as layering and shapes or foot-prints of foundations. The "rule of thumb" or mass ratio method of design procedure is as follows:- • firstly the requirements of stresses and serviceability must be satisfied. This is usual in a statically loaded system. the ratio of the machine mass to that of the foundation together with the structure should be greater than 3 in the case of a revolving machine and 5 in the case of a reciprocating machine. • in order to obtain uniform settlement of the foundation the distance of the combined centre of gravity of the machine and foundation from the centre of area in contact with the soil is limited to 5% of the corresponding dimension of the foundation. It is evident that the mass ratio lacks accuracy in that there are several parameters that are required to describe the satisfactory performance of a system such as amplitude, frequency and velocity of vibration. The finite element method allows for calibration of the model to account for the real behaviour of the system. Calibration is generally conducted using sensitive transducers called accelerometers. The accelerometers produce power spectral density (PSD) graphs from which deflections and stresses can be back calculated. The deflections and stresses are compared with calculated deflections and stresses. Descriptions of the methods of analysis followed by presentation of results, discussions and interpretations have been included. To motivate the use of dynamic analyses case histories have been presented and discussed. Finally the dissertation concludes with findings of the study together with recommendations for the way forward in terms of research. / Thesis (M.Sc.)-University of Natal, Durban, 2002.

Power utilities around the world are now in the practice of installing fiber optic cables on their high voltage transmission networks. These high-speed communication channels can, not only transmit data needed for utility operation, but the unused fiber capacity may also be rented to others for communication. All dielectric self-supporting (ADSS) fiber optic cable appears to be the fiber optic cable most frequently installed by power utilities as it is more economical, has a larger fiber capacity and may be installed on a transmission line without de-energization. When installed however, ADSS fiber optic cable does undergo some degree of degradation caused either by armor rod corona at the towers or dry-band arcing. A comprehensive literature survey regarding both phenomena is presented in this study, as well as current mitigation techniques. Different models that describe the process of dry-band arcing are discussed, including those where an equivalent circuit is used to represent a polluted fiber optic cable in a high voltage environment. An implementation of this model on a MATLAB® based computer program is used to evaluate parameters such as leakage current magnitude, which may be used to predict the possibility of dry-band arcing. This leakage current is also compared to simulated results that were generated using a power system analysis program called Alternate Transients Program (ATP). A finite element package, FEMLAB®, was used to model the experimental system, prior to construction. A single-phase transmission line with an accompanying fiber optic cable was constructed. The leakage current magnitude obtained from this experiment was subsequently compared to those obtained from the simulations. These leakage current comparisons are discussed and explained in view of limitations with the theoretical models and refinements in the experimental techniques employed. The results clearly indicate that physical parameters like pollution severity, system voltage, length of span and the point of attachment of the ADSS fiber optic cable in the tower play a significant role in the determination of leakage currents induced on the outer sheath of the cable. These induced currents result in the formation of 'dry bands', due to joule heating, and this could result in arcing activity that erodes the fiber optic cable. / Thesis (M.Sc.Eng)-University of Natal, Durban, 2003.

The main aim of this study was concerned with the design and commissioning of a corona cage, which could be used under Direct Current (DC) conditions. The cage was designed based on empirical formulas and equations as well as electric field simulations. The designed cage was then fabricated. The commissioning of the cage was undertaken in the High Voltage Direct Current (HVDC) laboratory at the University of Durban - Westville (UDW). Tests to determine the effects of a silicone coating as well as wind on the corona performance of conductors were undertaken. The tests were done in order to determine ways of improving the corona performance of conductors under HVDC potential. The tests were carried out using various conductor surface conditions. The wind tests were made possible by using a powerful fan. A silicone coating was also used to determine the effects that it would have in mitigation of corona activity on HVDC conductors. The conductors were tested without the coating, with half of their length coated and then fully coated. Results showed that the effect of wind on corona generation in a corona cage is minimal. The effect of the silicone coating was that it increased the corona currents measured in the corona cage. The conductors with no coating generated the lowest currents, the half coated conductors generated the second highest measured currents and the fully coated conductors generated the most corona. Analysis of the increased currents showed that the increase in corona currents due to the silicone coating could be attributed to three factors. Firstly the coating caused an increase in conductor to cage capacitance. Secondly, partial discharges could have occurred in the silicone due to microscopic air particles and lastly, the increase in corona currents could be ascribed to the effect of the boundary conditions on the boundary between the conductor and the coating. / Thesis (M.Sc.)-University of Durban-Westville, 2003.

This dissertation analyses the reliability of power transformers and its impact of failure on system performance. Eskom Distribution, Eastern Region is used as a practical case study, which has an installed transformer base of 6066MVA comprising of 428 transformers ranging from 1 MVA to 8OMVA with voltage levels of 6.6kV to I 32kV. The literature review illustrates the theory and principles of transformers, evolution and changes in design criteria, the function of cellulose and insulating oil, failure modes. operations and maintenance practices and factors affecting the distribution systems performance. This study included a conditional assessment and an oil analysis review of transformers at Eskom. A method to trend multiple oil samples was developed and illustrated. The research further investigates the reliability of series and parallel systems using actual component reliability values. A study was conducted to establish the degree of network firmness. Transformer failure data was analysed and were shown to be characteristic of a bathtub curve. Defects from on site inspections were analysed and identified oil leaks as a maintenance focus area. The Distribution Supply Loss Index was determined to be the major impact Key Performance Index due to transformer failures. Transformer failures using statistical methods, showed HV/LV winding to be the main component to fail. The cost of a transformer failure to Eskom and the customer was determined. International Benchmarking was investigated to establish the criteria for network reliability indices and to compare the network infrastructure and performance of international utilities and Eskom. The later part of the study involved the analysis of a risk ranking methodology to establish a risk ranking matrix. The transformers were ranked according to the matrix, identifying the high risk focus areas. Projects were raised within Eskom to replace the identified high risk transformers. This study has concluded that the reliability of transformers is impacted by the changes in transformer design, increased maintenance defects and inadequate transformer protection at substations. The reduced oil volume per kVA, increased hot spot and ambient temperature, and compact tank designs have resulted in the cellulose being overheated and fault gases being produced during normal operating conditions. The increase in load demand from the existing transformer fleet and a reduction in capital expenditure to maintain and build additional substations have also contributed to accelerated aging, since the transformers are forced to operate at 100 percent loading. There is an increase in transformer maintenance defects due to in sufficient operational staff, high staff turnover, reduced skills transfer. and insufficient network contingencies to allow for planned outages to clear the defects identified. The failure analysis showed that the main component to fail is HV/LV windings. The winding failures were traced to there being no or inadequate transformer protection at -20% of substations. / Thesis (M.Sc.)-University of KwaZulu-Natal, 2007.

The main aim of this project was to determine the extent to which the study of electric fields and ions in a laboratory can be used to study the electrical environment below High Voltage Direct Current (HVDC) transmission lines. The focus of the study was to set up small scale laboratory experiments and to compare these results to actual line measurements and to software simulations. The laboratory tests were undertaken at the HVDC Centre at the University of KwaZulu-Natal (Westville Campus). The software simulations that were conducted as part of this study were done using EPRI TL 3.0 and Microsoft Excel. Initially tests conducted were the measurement of the induced voltage and corona leakage current on a floating object. The next set of laboratory tests conducted was the measurement of ion current density and the electric field at ground level. The ion current density was measured with a Wilson Plate (lm2) and the electric field at ground level was measured using a JCI static monitor field meter (JCI 140) and a Monroe (257D) Portable Electrostatic Fieldmeter, with an elevated earth plane. Measurements of ion current density and electric field at ground level were also taken under an operating HVDC transmission line (Cahora Bassa to Apollo), in order to compare the laboratory measurements and simulations with real line measurements. The results have shown that the electrical parameters (i.e. ion current, induced voltages, corona currents, electric field, ion density, space charge) are higher under the negative pole as compared to the positive pole. The results of the laboratory measurements show that the ion currents under the negative polarity are almost double the ion currents that were measured under positive polarity, while the electric field under negative polarity was 20 percent higher than under positive polarity. Measurements of the electric field show that the total electric field below the line is greatly enhanced when corona generated space charge is present. The results of the EPRI TL Workstation simulations show good correlation with the EXCEL® simulations. However, there was poor correlation between EPRI simulations and test line measurements in the laboratory. The EPRI simulations show good correlation to the measured electric field values below the Cahora Bassa line. The comparison between the actual measurements on the test line and the Cahora Bassa line showed poor correlation and this was attributed to factors such as scaling, laboratory size constraints, ion concentration in laboratory, line loading and wind speeds. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2008.

The dynamic characteristic of transmission line conductors is very important in designing and constructing a new line or upgrading an existing one. This concept is an impediment to line design and construction because it normally determines the tension at which the line is strung and this in respect affects the tower height and the span length. Investigations into the phenomenon of mechanical oscillation of power line conductors have been extensively looked into by many researchers using concepts from mechanics and aerodynamics to try and predict the conductor dynamic behaviour. Findings have shown that precise prediction of conductor windinduced vibration is very difficult i.e. non-linearity. Over the years, various analytical models have been developed by researchers to try and predict the mechanical vibration of transmission line conductors. The first part of this dissertation considers the analysis of the model describing the transverse vibration of a conductor as a long, slender, simply supported beam, isotropic in nature and subjected to a concentrated force. The solution of this beam equation was used to obtain the conductor natural frequencies and mode shapes. Conductor self-damping was obtained by the introduction of both external and internal damping models into the equation of motion for the beam. Next, also using the same beam concept was the application of the finite element method (FEM) for the dynamic analysis of transmission line conductors. A finite element formulation was done to present a weak form of the problem; Galerkin‟s method was then applied to derive the governing equations for the finite element. Assembly of these finite element equations, the equation of motion for the transverse vibration of the conductor is obtained. A one dimensional finite element simulation was done using ABAQUS software to simulate its transverse displacement. The eigenvalues and natural frequencies for the conductors were calculated at three different tensions for two different conductors. The damping behaviour of the conductors was evaluated using the proportional damping (Rayleigh damping) model. The results obtained were then compared with the results from the analytical model and the comparison showed a very good agreement. An electrical equivalent for the conductor was developed based on the concept of mechanicalelectrical analogy, using the discrete simply supported beam model. The developed electrical equivalent circuit was then used to formulate the transfer function for the conductor. Matlab software was used to simulate the free response of the developed transfer function. Finally, the experimental study was conducted to validate both the analytical model and the FEM. Tests were done on a single span conductor using two testing methods i.e. free and force vibration. The test results are valid only for Aeolian vibration. From the test results the conductor‟s natural frequencies and damping were determined. The experimental results, as compared with the analytical results were used to validate the finite element simulation results obtained from the ABAQUS simulation. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2011.

Series compensating capacitors were initially introduced in transmission networks mainly to increase the / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2009.

The power quality performance of South African utilities has been regulated through the application of NRS 048-2 standard. The earliest edition of the power quality standard (NRS 048- 2:1996) defines compatibility levels for voltage dips in the form of annual dip limits for each voltage dip type. Actual measured utility dip performance has consistently resulted in higher dip numbers than the limits imposed in the standard. On the other hand, the dip limits were considered to be less restrictive by industrial customers. The revised power quality standard addresses the difficulties in managing voltage dip performance based on dip limits as specified in the first edition of NRS 048-2. This new philosophy does not define dip limits; instead, utilities are required to develop specific strategies to manage dip performance according to customer requirements. This research work develops an alternative approach to the management of dip performance as opposed to the application of dip limits. The study analyses measured voltage dip records for a steel-processing plant and a pulp and paper plant. The supply network for each plant is modelled to define dip influence zones as a function of fault locations. The principal results of this study are critical circuits, causes of dips, dip influence zones and the key elements of the proposed approach in communicating dip performance. The optimised approach was presented to and adopted by the customers involved. / Thesis (M.Sc.)-University of KwaZulu-Natal, Durban, 2009.

Electrical power distribution networks are required to provide power to customers at nearconstant / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2010.

The South African economy is an emerging market and as such there is a continued and growing need for the efficient supply of cost effective electricity. The capital investment involved in the design, construction, installation and commissioning of overhead transmission line networks are high and so too are the subsequent maintenance and operation costs, incurred over their life cycle periods. The need to improve the electrical operating efficiency of existing and future electrical transmission networks, through the reduction of electrical losses, focused and motivated the research in this particular area. The results and findings produced by this research study show that the magnetic induction produced by the steel core in ACSR (Aluminium conductor, steel reinforced) conductors cause in increase in the ac power losses, associated ac-dc resistance ratio and the effective ac resistance of the conductor, whilst the conductor is energised during normal operation. More specifically, the key parameters that cause this increase in the effective ac resistance of the conductor, as a result of the magnetic induction produced by the steel core, are those of hysterisis and eddy current power losses in the steel core and an added power loss caused by the non-uniform redistribution of current in the layers of aluminum wires, due to the ‘transformer effect’. Therefore the addition of the conductor dc resistance value to the component resistances produced by the current redistribution and magnetic hysterisis & eddy current power losses, form the total effective ac conductor resistance. This is contrary to standard practice where assumption is made that the conductor ac and dc resistance values are equal. The factors which influence the magnetic induction, include amongst others; the ferromagnetic properties of the steel core, the physical construction of the conductor, the conductor operating/core temperature and the load current. In order to calculate the effective ac-resistance of multi-layer ACSR conductors a computer simulation program was developed, which was largely based on determining the impact of varying these key factors, by evaluating its effect on the ac resistance of the conductor. It was found through manipulation of these factors that the total effective ac resistance of the conductor could be reduced and significantly so with higher load currents. The conductor sample used in this research study is commonly known as TERN ACSR conductor in the South African market and it was shown that with practical changes in lay ratios or lay lengths, one is able to reduce the total effective ac resistance of the conductor and associated power losses. Several software simulation exercises were performed using the developed software simulation program, to ultimately produce a set of optimised lay-lengths (lay-ratios) for the TERN ACSR conductor, with the intention that these simulated parameters would be employed in the production of actual conductor samples. The intention going forward after the planned production trial runs would be to test these conductor samples to verify compliance, in meeting both electrical and mechanical performance requirements. It should be noted that the planned production trials and relevant conductor-testing processes did not form part of the scope of this research report but are processes that have been planned for in the near future. Although testing to IEC 61089 are post processes that are planned for outside of this research scope, the specification requirements of IEC61089 were incorporated into the various computer simulation exercises. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2009.