No description available.

The research involves devising new control and compensation methods for DC power distribution systems containing power electronics loads with constant power characteristic. A compensation method called an active damping network is examined, which utilises capacitor and inductor energy storage elements and high frequency switching devices. This network is controlled to behave as a small-signal resistor so that any instabilities in the power distribution system due to the interaction between interconnected sub-systems will be damped and eliminated successfully. This method is studied and analysed in detail. The system was designed and simulated, and was verified by experimental results. The results show that stable operation is achieved of a small-scale power distribution system that contains power electronic loads with constant power characteristics such as DC-DC converters and electrical drives.

Frequency Response Analysis (FRA) has been developed to detect winding movement and deformation in transformers. In FRA diagnosis, the decision on a transformer is made based on the comparison between two FRA responses and any significant difference would potentially indicate mechanical problems with the winding. However, the ability to interpret the 'differences' when comparing the FRA responses is a great challenge. The main objective of this thesis is to develop an understanding of the FRA responses towards establishing an FRA interpretation guideline. The overall approach used in this research consists of four elements. Initially, the Transfer Function Estimation (TFE) technique is developed as an FRA comparative tool. However, this technique requires the correlation of the resulting key parameters with the deformation criteria and this demands an understanding of the FRA responses. Secondly, experimental studies on single windings are used to gain a firm understanding of the features of their FRA responses and these studies covered the interleaved, plain and intershielded disc type windings. The lumped ladder network representation of the winding is also employed to provide a mathematical back-up. Thirdly, verified FRA simulation models are used to investigate the FRA responses of three-phase transformers. Finally, an FRA interpretation guideline is developed and applied on the analysis of the FRA responses of a suspected faulty transformer. The TFE technique was specifically developed to represent the FRA response with key parameters. These were the complex conjugate pairs of poles and zeroes which specified the resonant frequencies of the FRA response. An auto-transformer with broken axial clamping fault would result in the changes of the key parameters which corresponded with frequency shifts of 4-5%. In terms of the structure of single windings, they can be categorised into windings with either high- or low- series capacitance in proportion to the shunt capacitance. Correspondingly, the FRA responses of windings of high series capacitance exhibited the increasing trend of magnitude in the frequency range between 20kHz and 500kHz while the windings of low series capacitance displayed the steady magnitude trend with the resonances and anti-resonances (camel humps) features in the frequency range between 20kHz and 2MHz. The FRA response of windings in the real transformer configuration would be affected by the core and the interaction between the windings. While the core only influenced the FRA response in the low frequencies of hundreds of Hz, the frequency range between 2kHz and 20kHz was predominantly affected by the interaction between the windings in the transformer. Double peak feature was found mostly in the FRA responses of the autotransformer while only single major resonance was observed from the two-winding transformer. The bulk winding capacitances and the mutual couplings between the windings in the same phase influenced these features significantly. In most occasions, the features of the winding structure remained in the frequency ranges specified previously. An FRA interpretation guideline was developed and the success of this guideline was demonstrated by its application onto a suspected faulty transformer. The analyses on the FRA responses of the series winding indicated winding movements of the tertiary and common windings. These were confirmed by the visual inspection on this transformer which clearly showed hoop buckling in the common winding and the collapsed tertiary winding.

System-generated overvoltages are increasingly becoming a critical factor in determining the insulation level of the system, particularly at the higher transmission voltages. The ability to predict these overvoltages for given system conditions is therefore of paramount importance during the planning stage of a power network. This thesis is concerned mainly with excessive transient voltages generated in composite transformer feeders, particularly due to resonance. The thesis commences with a comprehensive review of over-voltages produced in power systems (Chapter 1). Power frequency, sustained and transient overvoltages of internal and external origin are discussed, together with measures employed to protect system equipment against their effects. Factors which are critical to system insulation levels at the higher voltages are established. Chapter 2 is concerned more specifically with the modifying influence of a transformer integrally connected to the feeder on switching overvoltages. Particular consideration is given to conditions under which these overvoltages may be accentuated by resonance and ferroresonance effects. The specific aspects of these phenomena which form the basis of the research effort discussed in later Chapters, are identified. The relative merits of available analogue and digital computer methods which can be used in the study of transients are presented in Chapter 3. The lattice diagram technique adapted for digital computation and the transient network analyser, both of which are utilised in some of the investigations, are described in more detail. Chapter 4 presents the rudiments of an efficient digital computer method based on the compensation theorem. This method is capable of accurately simulating travelling-wave phenomena in transmission lines as well as the transient response of complex lumped and non-linear sub-networks to switching stimuli. Single-phase results of investigations utilising the methods described commence in Chapter 5. The effects of system constants and various circuit arrangements on linear resonant energisation overvoltages in transformer feeders are assessed in Chapters 5 and 6 respectively. Conditions potentially onerous in respect of overvoltage magnitudes are identified and simple analytical techniques for their predictions are derived. A comprehensive mathematical model of three-phase transformer feeder circuits is presented in Chapter 7. This model incorporates mutual effects between the phases of the line and both the electric and magnetic interactions within a transformer of a composite core geometry, including the non-linear saturation characteristics. The compensation method described earlier is extended to three phase circuits and applied to this model to study the effects of sequential pole closure, transformer saturation and winding connections in a resonant network (Chapter 8). The phenomena of resonance and ferroresonance in double circuit transformer feeders following isolation of one circuit from the rest of the system are dealt with in Chapters 9 and 10, respectively. Analytical methods for predicting the occurrence of linear resonance overvoltages in a reactively compensated circuit which is being dropped, and preventive measures, are suggested in Chapter 9. The fundamental aspects of ferroresonant oscillations, their effects on system equipment and methods for their (ii) suppression, are set out in Chapter 10. The compensation method is used to assess the effects of certain circuit constants and characteristics.

Operating faraway from maximum power point decreases the generated power from photovoltaic (PV) system. For optimum operation, it is necessary to continually track the maximum power point of the PV solar array. However with huge changes in external influences and the nonlinear relationship of electrical characteristics of PV panels it is a difficult problem to identify the maximum power point as a function of these influences. Many tracking control strategies have been proposed to track maximum power point such as perturb and observe, incremental conductance, parasitic capacitance, and neural networks. These proposed methods have some disadvantages such as high cost, difficulty, complexity and nonstability. This thesis presents a novel approach based on Adaptive NeuroFuzzy Inference System (ANFIS) to predict the maximum power point utilising the actual field data, which is performed in different environmental conditions. The short circuit current and open circuit voltage are used as inputs to PV panels instead of solar irradiation and cell junction temperature. The predicted $V_{max}$from ANFIS model is used as a reference voltage for fuzzy logic controller (FLC). The FLC is used to adjust the duty cycle of the electronic switch of two types of DC-DC converter. These DC-DC converters are used to interface between the load voltage and PV panels. The duty cycle of the electronic switch of the DC-DC converter is adjusted until the input voltage of the converter tracks the predicted $V_{max}$of the PV system. FLC rules and membership functions are designed to achieve the most promising performance at different environmental conditions, different load types and different rate of changes in the duty cycle of Buck-Boost and Buck converters. The membership functions and fuzzy rules of FLC are designed to balance between different required features such as quick tracking under different environmental conditions, high accuracy, stability and high efficiency.

Following world wide deregulation of the electrical power industry, the requirements for high quality and highly reliable power supplies are increasing. Comprehensive and systematic computer based methodologies and algorithms for fault locators in power systems are needed to support reliable, independent and comprehensive fault diagnosis or accurate fault location, fast fault detection and correct fault classification. This thesis describes work to develop novel fault diagnosis techniques and accurate fault location algorithms for fault locators using system-wide information of modern power systems. Fast and efficient modern communication techniques make it possible to apply the novel fault locator system in practice. The fault diagnosis techniques use fault messages obtained from current and voltage sensors installed in radial and meshed networks respectively. A faulty section in the meshed network can be detected by analysing the network's topologic structure and the measurements of the fundamental frequency voltages obtained from voltage sensors; while for the radial network, current sensors are used to trace the path of a fault. The fault diagnosis techniques can work independently from protection devices, and circuit breakers, and can be implemented economically in the low- and medium-voltage power systems, because current and voltage sensors can be made cheaply. In order to achieve very accurate fault location for the EHV transmission systems with long transmission lines and complex network topology, algorithms based on the wavelet transform, travelling wave concept and GPS synchronisation have been developed. When a fault occurs, fault generated transient waves propagate from fault point to line terminals and then to the other nodes in the whole network. Fault transient detectors are installed at nodes in the meshed network, at substations and customer terminals in the radial system to capture the time of fault transient signal arrival. From the time recorded and topologic network structure, very accurate fault location can be achieved. The thesis also describes and compares applications of digital Fourier transform, least squares method and Kalman filters for fast measurements in current and voltage sensors, for fault classification and fault detection, and the wavelet transform used in fault transient detectors. Methodologies and algorithms developed have been validated by ATP/EMTP simulation on different networks, such as, traditional twosource system and IEEE 14-bus system. The maIn advantage of the developed methodologies and algorithms for fault locators in power systems is the use of system-wide fault messages in the electrical networks. The fault locators will efficiently support the operators in the control centres and relevant substations with clear fault information allowing them to take suitable actions for fault emergency and restoration of power supplies. This will be beneficial in reducing the outage time required for inspection, damage repair and restoration of the power supply. Reliability, security and power quality will be improved for the customers.

The problem__of switching surges in recent years has been a great -- concern to transmissi. on line engineers since, with the raising of transmission voltages to 30P-40P kV-and above, the stage has been reaphed where switching surges become of importance equal-to, or greater-than,. those caused by lightning for the selection of insulation and protective levels. This theBiS seeks to increase the knowledge on switching suges, particularly, those due to line energisation and re-energisation during which the highest overvoltage magnitudes are generated. The mathematical model used in the anlysis-has been the theory of nitural, modes as applied to symmetric transmission lines in conjunction with the Fourier transform'technique which tpLkes into account the frequency dependence of the elemlents. Sequential energisation and-high speed re-energisation of three-phaia lines as modified by the system configuration on both sides. --of the circuit breaker has been analysed. Surge magnitude reduction techniques by speciaAy designed circuit-breakers such-as resistor and synchronous ewitchit'99 and the protective characteristics of lightning arresters are presented* In insulation coordinationt statistical methods are being introduced to design for the most economical arrangement of the insulation structures taki, 39 I into account that the hig1jeqt overvoltages--have a low frequency of appearancs' In order to assess the sto-. -hastic behaviour of switching overvoltages, a fast and accurate method has been developedp and the distribution curves of some of the switching surge problems are presented.

No description available.

No description available.

No description available.